From c02e8807df12e9678d13a15ba6680fd1bc1b5be6 Mon Sep 17 00:00:00 2001 From: Sandeep Kumar Date: Wed, 14 Dec 2016 15:49:40 +0530 Subject: [PATCH 001/108] Fixes in Makefile of couple of samples - modified Makefile for hipblas_saxpy to replaced hcblas.so with hipblas.so as part of HCSWAP-100 - Resolved missing separator issue in peer2peer cookbook Makefile Change-Id: I678fea267eee1481f02da09379339ed78d3f95f2 --- samples/2_Cookbook/8_peer2peer/Makefile | 13 +++++++------ samples/7_Advanced/hipblas_saxpy/Makefile | 2 +- 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/samples/2_Cookbook/8_peer2peer/Makefile b/samples/2_Cookbook/8_peer2peer/Makefile index 5cb7473921..0bf9e6f93e 100644 --- a/samples/2_Cookbook/8_peer2peer/Makefile +++ b/samples/2_Cookbook/8_peer2peer/Makefile @@ -1,6 +1,6 @@ HIP_PATH?= $(wildcard /opt/rocm/hip) ifeq (,$(HIP_PATH)) - HIP_PATH=../../.. + HIP_PATH=../../.. endif HIPCC=$(HIP_PATH)/bin/hipcc @@ -22,14 +22,15 @@ CXX=$(HIPCC) $(EXECUTABLE): $(OBJECTS) - $(HIPCC) $(OBJECTS) -o $@ + $(HIPCC) $(OBJECTS) -o $@ + test: $(EXECUTABLE) - $(EXECUTABLE) + $(EXECUTABLE) clean: -rm -f $(EXECUTABLE) -rm -f $(OBJECTS) -rm -f $(HIP_PATH)/src/*.o + rm -f $(EXECUTABLE) + rm -f $(OBJECTS) + rm -f $(HIP_PATH)/src/*.o diff --git a/samples/7_Advanced/hipblas_saxpy/Makefile b/samples/7_Advanced/hipblas_saxpy/Makefile index ed88be2dd0..8586e75d25 100644 --- a/samples/7_Advanced/hipblas_saxpy/Makefile +++ b/samples/7_Advanced/hipblas_saxpy/Makefile @@ -12,7 +12,7 @@ endif ifeq (${HIP_PLATFORM}, hcc) HCBLAS_ROOT?= $(wildcard /opt/rocm/hcblas) HIPCC_FLAGS += -stdlib=libc++ -I$(HCBLAS_ROOT)/include - LIBS = -L$(HCBLAS_ROOT)/lib -lhcblas + LIBS = -L$(HCBLAS_ROOT)/lib -lhipblas -rpath $(HIP_PATH)/lib endif From f052f43b3b3d48ac79f2111c7da74c24f1ad29b2 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Mon, 19 Dec 2016 04:04:06 +0000 Subject: [PATCH 002/108] Updated doxygen documentation Change-Id: If04d1155173fba8d3e050f3259da8b3edc60e076 --- docs/RuntimeAPI/html/Synchonization.html | 117 ++ docs/RuntimeAPI/html/annotated.html | 145 ++ docs/RuntimeAPI/html/bc_s.png | Bin 0 -> 676 bytes docs/RuntimeAPI/html/bdwn.png | Bin 0 -> 147 bytes docs/RuntimeAPI/html/bug.html | 97 ++ .../html/classFakeMutex-members.html | 104 ++ docs/RuntimeAPI/html/classFakeMutex.html | 117 ++ .../html/classLockedAccessor-members.html | 105 ++ docs/RuntimeAPI/html/classLockedAccessor.html | 117 ++ .../html/classShortTid-members.html | 105 ++ docs/RuntimeAPI/html/classShortTid.html | 118 ++ docs/RuntimeAPI/html/classes.html | 130 ++ .../classihipCtxCriticalBase__t-members.html | 118 ++ .../html/classihipCtxCriticalBase__t.html | 164 ++ .../html/classihipCtxCriticalBase__t.png | Bin 0 -> 824 bytes .../html/classihipCtx__t-members.html | 115 ++ docs/RuntimeAPI/html/classihipCtx__t.html | 152 ++ .../html/classihipDevice__t-members.html | 111 ++ docs/RuntimeAPI/html/classihipDevice__t.html | 141 ++ .../html/classihipException-members.html | 103 ++ docs/RuntimeAPI/html/classihipException.html | 126 ++ docs/RuntimeAPI/html/classihipException.png | Bin 0 -> 435 bytes .../html/classihipFunction__t-members.html | 106 ++ .../RuntimeAPI/html/classihipFunction__t.html | 124 ++ .../classihipIpcMemHandle__t-members.html | 104 ++ .../html/classihipIpcMemHandle__t.html | 118 ++ .../html/classihipModule__t-members.html | 109 ++ docs/RuntimeAPI/html/classihipModule__t.html | 130 ++ ...lassihipStreamCriticalBase__t-members.html | 109 ++ .../html/classihipStreamCriticalBase__t.html | 145 ++ .../html/classihipStreamCriticalBase__t.png | Bin 0 -> 854 bytes .../html/classihipStream__t-members.html | 124 ++ docs/RuntimeAPI/html/classihipStream__t.html | 216 +++ docs/RuntimeAPI/html/closed.png | Bin 0 -> 132 bytes docs/RuntimeAPI/html/deprecated.html | 98 ++ .../html/device__util_8h_source.html | 138 ++ .../dir_68267d1309a1af8e8297ef4c3efbcdba.html | 130 ++ .../dir_82c4e5e99cd964a390bd8b007f79d8a6.html | 121 ++ .../dir_ce4f9d092cd8b82b7c46cef8ba457558.html | 121 ++ .../dir_d44c64559bbebec7f509842c48db8b23.html | 100 ++ docs/RuntimeAPI/html/doxygen.css | 1366 +++++++++++++++++ docs/RuntimeAPI/html/doxygen.png | Bin 0 -> 3779 bytes docs/RuntimeAPI/html/dynsections.js | 97 ++ docs/RuntimeAPI/html/files.html | 127 ++ docs/RuntimeAPI/html/ftv2blank.png | Bin 0 -> 86 bytes docs/RuntimeAPI/html/ftv2cl.png | Bin 0 -> 453 bytes docs/RuntimeAPI/html/ftv2doc.png | Bin 0 -> 746 bytes docs/RuntimeAPI/html/ftv2folderclosed.png | Bin 0 -> 616 bytes docs/RuntimeAPI/html/ftv2folderopen.png | Bin 0 -> 597 bytes docs/RuntimeAPI/html/ftv2lastnode.png | Bin 0 -> 86 bytes docs/RuntimeAPI/html/ftv2link.png | Bin 0 -> 746 bytes docs/RuntimeAPI/html/ftv2mlastnode.png | Bin 0 -> 246 bytes docs/RuntimeAPI/html/ftv2mnode.png | Bin 0 -> 246 bytes docs/RuntimeAPI/html/ftv2mo.png | Bin 0 -> 403 bytes docs/RuntimeAPI/html/ftv2node.png | Bin 0 -> 86 bytes docs/RuntimeAPI/html/ftv2ns.png | Bin 0 -> 388 bytes docs/RuntimeAPI/html/ftv2plastnode.png | Bin 0 -> 229 bytes docs/RuntimeAPI/html/ftv2pnode.png | Bin 0 -> 229 bytes docs/RuntimeAPI/html/ftv2splitbar.png | Bin 0 -> 314 bytes docs/RuntimeAPI/html/ftv2vertline.png | Bin 0 -> 86 bytes docs/RuntimeAPI/html/functions.html | 332 ++++ docs/RuntimeAPI/html/functions_rela.html | 106 ++ docs/RuntimeAPI/html/functions_vars.html | 329 ++++ docs/RuntimeAPI/html/globals.html | 567 +++++++ docs/RuntimeAPI/html/globals_defs.html | 152 ++ docs/RuntimeAPI/html/globals_enum.html | 119 ++ docs/RuntimeAPI/html/globals_eval.html | 146 ++ docs/RuntimeAPI/html/globals_func.html | 451 ++++++ docs/RuntimeAPI/html/globals_type.html | 119 ++ docs/RuntimeAPI/html/group__API.html | 245 +++ docs/RuntimeAPI/html/group__Context.html | 607 ++++++++ docs/RuntimeAPI/html/group__Device.html | 584 +++++++ docs/RuntimeAPI/html/group__Driver.html | 750 +++++++++ docs/RuntimeAPI/html/group__Error.html | 206 +++ docs/RuntimeAPI/html/group__Event.html | 356 +++++ docs/RuntimeAPI/html/group__Fence.html | 211 +++ docs/RuntimeAPI/html/group__GlobalDefs.html | 700 +++++++++ .../RuntimeAPI/html/group__HCC__Specific.html | 96 ++ docs/RuntimeAPI/html/group__HIP-ENV.html | 90 ++ docs/RuntimeAPI/html/group__Memory.html | 1319 ++++++++++++++++ docs/RuntimeAPI/html/group__PeerToPeer.html | 404 +++++ docs/RuntimeAPI/html/group__Profiler.html | 142 ++ docs/RuntimeAPI/html/group__Stream.html | 437 ++++++ docs/RuntimeAPI/html/group__Texture.html | 214 +++ docs/RuntimeAPI/html/hcc_8h_source.html | 130 ++ docs/RuntimeAPI/html/hcc__acc_8h_source.html | 138 ++ .../hcc__detail_2hip__complex_8h_source.html | 276 ++++ .../hcc__detail_2hip__fp16_8h_source.html | 278 ++++ .../html/hcc__detail_2hip__runtime_8h.html | 1191 ++++++++++++++ .../hcc__detail_2hip__runtime_8h_source.html | 915 +++++++++++ .../hcc__detail_2hip__runtime__api_8h.html | 589 +++++++ ...__detail_2hip__runtime__api_8h_source.html | 718 +++++++++ .../html/hcc__detail_2hip__texture_8h.html | 560 +++++++ .../hcc__detail_2hip__texture_8h_source.html | 333 ++++ .../hcc__detail_2hip__vector__types_8h.html | 431 ++++++ ..._detail_2hip__vector__types_8h_source.html | 547 +++++++ docs/RuntimeAPI/html/hierarchy.html | 150 ++ .../html/hip__common_8h_source.html | 186 +++ .../html/hip__complex_8h_source.html | 134 ++ docs/RuntimeAPI/html/hip__fp16_8h_source.html | 133 ++ docs/RuntimeAPI/html/hip__hcc_8cpp.html | 486 ++++++ docs/RuntimeAPI/html/hip__hcc_8h_source.html | 922 +++++++++++ docs/RuntimeAPI/html/hip__ldg_8h_source.html | 181 +++ .../html/hip__profile_8h_source.html | 138 ++ .../html/hip__runtime_8h_source.html | 155 ++ .../html/hip__runtime__api_8h_source.html | 431 ++++++ .../html/hip__texture_8h_source.html | 138 ++ docs/RuntimeAPI/html/hip__util_8h_source.html | 138 ++ .../html/hip__vector__types_8h_source.html | 138 ++ docs/RuntimeAPI/html/host__defines_8h.html | 147 ++ .../html/host__defines_8h_source.html | 160 ++ docs/RuntimeAPI/html/index.html | 99 ++ docs/RuntimeAPI/html/jquery.js | 31 + docs/RuntimeAPI/html/modules.html | 109 ++ docs/RuntimeAPI/html/nav_f.png | Bin 0 -> 153 bytes docs/RuntimeAPI/html/nav_g.png | Bin 0 -> 95 bytes docs/RuntimeAPI/html/nav_h.png | Bin 0 -> 98 bytes docs/RuntimeAPI/html/open.png | Bin 0 -> 123 bytes docs/RuntimeAPI/html/pages.html | 97 ++ docs/RuntimeAPI/html/search/all_0.html | 26 + docs/RuntimeAPI/html/search/all_0.js | 9 + docs/RuntimeAPI/html/search/all_1.html | 26 + docs/RuntimeAPI/html/search/all_1.js | 4 + docs/RuntimeAPI/html/search/all_10.html | 26 + docs/RuntimeAPI/html/search/all_10.js | 7 + docs/RuntimeAPI/html/search/all_11.html | 26 + docs/RuntimeAPI/html/search/all_11.js | 9 + docs/RuntimeAPI/html/search/all_12.html | 26 + docs/RuntimeAPI/html/search/all_12.js | 4 + docs/RuntimeAPI/html/search/all_13.html | 26 + docs/RuntimeAPI/html/search/all_13.js | 4 + docs/RuntimeAPI/html/search/all_14.html | 26 + docs/RuntimeAPI/html/search/all_14.js | 4 + docs/RuntimeAPI/html/search/all_15.html | 26 + docs/RuntimeAPI/html/search/all_15.js | 4 + docs/RuntimeAPI/html/search/all_2.html | 26 + docs/RuntimeAPI/html/search/all_2.js | 4 + docs/RuntimeAPI/html/search/all_3.html | 26 + docs/RuntimeAPI/html/search/all_3.js | 11 + docs/RuntimeAPI/html/search/all_4.html | 26 + docs/RuntimeAPI/html/search/all_4.js | 10 + docs/RuntimeAPI/html/search/all_5.html | 26 + docs/RuntimeAPI/html/search/all_5.js | 5 + docs/RuntimeAPI/html/search/all_6.html | 26 + docs/RuntimeAPI/html/search/all_6.js | 7 + docs/RuntimeAPI/html/search/all_7.html | 26 + docs/RuntimeAPI/html/search/all_7.js | 4 + docs/RuntimeAPI/html/search/all_8.html | 26 + docs/RuntimeAPI/html/search/all_8.js | 230 +++ docs/RuntimeAPI/html/search/all_9.html | 26 + docs/RuntimeAPI/html/search/all_9.js | 19 + docs/RuntimeAPI/html/search/all_a.html | 26 + docs/RuntimeAPI/html/search/all_a.js | 10 + docs/RuntimeAPI/html/search/all_b.html | 26 + docs/RuntimeAPI/html/search/all_b.js | 15 + docs/RuntimeAPI/html/search/all_c.html | 26 + docs/RuntimeAPI/html/search/all_c.js | 4 + docs/RuntimeAPI/html/search/all_d.html | 26 + docs/RuntimeAPI/html/search/all_d.js | 6 + docs/RuntimeAPI/html/search/all_e.html | 26 + docs/RuntimeAPI/html/search/all_e.js | 4 + docs/RuntimeAPI/html/search/all_f.html | 26 + docs/RuntimeAPI/html/search/all_f.js | 8 + docs/RuntimeAPI/html/search/classes_0.html | 26 + docs/RuntimeAPI/html/search/classes_0.js | 4 + docs/RuntimeAPI/html/search/classes_1.html | 26 + docs/RuntimeAPI/html/search/classes_1.js | 5 + docs/RuntimeAPI/html/search/classes_2.html | 26 + docs/RuntimeAPI/html/search/classes_2.js | 7 + docs/RuntimeAPI/html/search/classes_3.html | 26 + docs/RuntimeAPI/html/search/classes_3.js | 6 + docs/RuntimeAPI/html/search/classes_4.html | 26 + docs/RuntimeAPI/html/search/classes_4.js | 10 + docs/RuntimeAPI/html/search/classes_5.html | 26 + docs/RuntimeAPI/html/search/classes_5.js | 16 + docs/RuntimeAPI/html/search/classes_6.html | 26 + docs/RuntimeAPI/html/search/classes_6.js | 9 + docs/RuntimeAPI/html/search/classes_7.html | 26 + docs/RuntimeAPI/html/search/classes_7.js | 4 + docs/RuntimeAPI/html/search/classes_8.html | 26 + docs/RuntimeAPI/html/search/classes_8.js | 6 + docs/RuntimeAPI/html/search/classes_9.html | 26 + docs/RuntimeAPI/html/search/classes_9.js | 4 + docs/RuntimeAPI/html/search/classes_a.html | 26 + docs/RuntimeAPI/html/search/classes_a.js | 9 + docs/RuntimeAPI/html/search/close.png | Bin 0 -> 273 bytes docs/RuntimeAPI/html/search/defines_0.html | 26 + docs/RuntimeAPI/html/search/defines_0.js | 4 + docs/RuntimeAPI/html/search/defines_1.html | 26 + docs/RuntimeAPI/html/search/defines_1.js | 4 + docs/RuntimeAPI/html/search/enums_0.html | 26 + docs/RuntimeAPI/html/search/enums_0.js | 10 + docs/RuntimeAPI/html/search/enumvalues_0.html | 26 + docs/RuntimeAPI/html/search/enumvalues_0.js | 65 + docs/RuntimeAPI/html/search/files_0.html | 26 + docs/RuntimeAPI/html/search/files_0.js | 9 + docs/RuntimeAPI/html/search/functions_0.html | 26 + docs/RuntimeAPI/html/search/functions_0.js | 6 + docs/RuntimeAPI/html/search/functions_1.html | 26 + docs/RuntimeAPI/html/search/functions_1.js | 111 ++ docs/RuntimeAPI/html/search/groups_0.html | 26 + docs/RuntimeAPI/html/search/groups_0.js | 4 + docs/RuntimeAPI/html/search/groups_1.html | 26 + docs/RuntimeAPI/html/search/groups_1.js | 5 + docs/RuntimeAPI/html/search/groups_2.html | 26 + docs/RuntimeAPI/html/search/groups_2.js | 5 + docs/RuntimeAPI/html/search/groups_3.html | 26 + docs/RuntimeAPI/html/search/groups_3.js | 4 + docs/RuntimeAPI/html/search/groups_4.html | 26 + docs/RuntimeAPI/html/search/groups_4.js | 4 + docs/RuntimeAPI/html/search/groups_5.html | 26 + docs/RuntimeAPI/html/search/groups_5.js | 6 + docs/RuntimeAPI/html/search/groups_6.html | 26 + docs/RuntimeAPI/html/search/groups_6.js | 4 + docs/RuntimeAPI/html/search/groups_7.html | 26 + docs/RuntimeAPI/html/search/groups_7.js | 5 + docs/RuntimeAPI/html/search/groups_8.html | 26 + docs/RuntimeAPI/html/search/groups_8.js | 4 + docs/RuntimeAPI/html/search/groups_9.html | 26 + docs/RuntimeAPI/html/search/groups_9.js | 4 + docs/RuntimeAPI/html/search/mag_sel.png | Bin 0 -> 563 bytes docs/RuntimeAPI/html/search/nomatches.html | 12 + docs/RuntimeAPI/html/search/pages_0.html | 26 + docs/RuntimeAPI/html/search/pages_0.js | 4 + docs/RuntimeAPI/html/search/pages_1.html | 26 + docs/RuntimeAPI/html/search/pages_1.js | 4 + docs/RuntimeAPI/html/search/pages_2.html | 26 + docs/RuntimeAPI/html/search/pages_2.js | 4 + docs/RuntimeAPI/html/search/related_0.html | 26 + docs/RuntimeAPI/html/search/related_0.js | 4 + docs/RuntimeAPI/html/search/search.css | 271 ++++ docs/RuntimeAPI/html/search/search.js | 815 ++++++++++ docs/RuntimeAPI/html/search/search_l.png | Bin 0 -> 604 bytes docs/RuntimeAPI/html/search/search_m.png | Bin 0 -> 158 bytes docs/RuntimeAPI/html/search/search_r.png | Bin 0 -> 612 bytes docs/RuntimeAPI/html/search/typedefs_0.html | 26 + docs/RuntimeAPI/html/search/typedefs_0.js | 4 + docs/RuntimeAPI/html/search/typedefs_1.html | 26 + docs/RuntimeAPI/html/search/typedefs_1.js | 7 + docs/RuntimeAPI/html/search/variables_0.html | 26 + docs/RuntimeAPI/html/search/variables_0.js | 4 + docs/RuntimeAPI/html/search/variables_1.html | 26 + docs/RuntimeAPI/html/search/variables_1.js | 4 + docs/RuntimeAPI/html/search/variables_2.html | 26 + docs/RuntimeAPI/html/search/variables_2.js | 8 + docs/RuntimeAPI/html/search/variables_3.html | 26 + docs/RuntimeAPI/html/search/variables_3.js | 20 + docs/RuntimeAPI/html/search/variables_4.html | 26 + docs/RuntimeAPI/html/search/variables_4.js | 5 + docs/RuntimeAPI/html/search/variables_5.html | 26 + docs/RuntimeAPI/html/search/variables_5.js | 4 + docs/RuntimeAPI/html/search/variables_6.html | 26 + docs/RuntimeAPI/html/search/variables_6.js | 13 + docs/RuntimeAPI/html/search/variables_7.html | 26 + docs/RuntimeAPI/html/search/variables_7.js | 4 + docs/RuntimeAPI/html/search/variables_8.html | 26 + docs/RuntimeAPI/html/search/variables_8.js | 5 + docs/RuntimeAPI/html/search/variables_9.html | 26 + docs/RuntimeAPI/html/search/variables_9.js | 4 + docs/RuntimeAPI/html/search/variables_a.html | 26 + docs/RuntimeAPI/html/search/variables_a.js | 4 + docs/RuntimeAPI/html/search/variables_b.html | 26 + docs/RuntimeAPI/html/search/variables_b.js | 5 + docs/RuntimeAPI/html/search/variables_c.html | 26 + docs/RuntimeAPI/html/search/variables_c.js | 4 + docs/RuntimeAPI/html/search/variables_d.html | 26 + docs/RuntimeAPI/html/search/variables_d.js | 4 + docs/RuntimeAPI/html/search/variables_e.html | 26 + docs/RuntimeAPI/html/search/variables_e.js | 4 + docs/RuntimeAPI/html/search/variables_f.html | 26 + docs/RuntimeAPI/html/search/variables_f.js | 4 + .../RuntimeAPI/html/structDbName-members.html | 103 ++ docs/RuntimeAPI/html/structDbName.html | 114 ++ .../html/structLockedBase-members.html | 104 ++ docs/RuntimeAPI/html/structLockedBase.html | 131 ++ docs/RuntimeAPI/html/structLockedBase.png | Bin 0 -> 1146 bytes .../html/structProfTrigger-members.html | 106 ++ docs/RuntimeAPI/html/structProfTrigger.html | 127 ++ .../html/struct____half-members.html | 102 ++ docs/RuntimeAPI/html/struct____half.html | 111 ++ docs/RuntimeAPI/html/structchar3-members.html | 104 ++ docs/RuntimeAPI/html/structchar3.html | 117 ++ docs/RuntimeAPI/html/structchar4-members.html | 106 ++ docs/RuntimeAPI/html/structchar4.html | 127 ++ docs/RuntimeAPI/html/structdim3-members.html | 104 ++ docs/RuntimeAPI/html/structdim3.html | 124 ++ .../html/structdouble1-members.html | 102 ++ docs/RuntimeAPI/html/structdouble1.html | 111 ++ .../html/structdouble3-members.html | 104 ++ docs/RuntimeAPI/html/structdouble3.html | 117 ++ .../RuntimeAPI/html/structfloat1-members.html | 102 ++ docs/RuntimeAPI/html/structfloat1.html | 111 ++ .../RuntimeAPI/html/structfloat3-members.html | 104 ++ docs/RuntimeAPI/html/structfloat3.html | 117 ++ .../html/structhipArray-members.html | 105 ++ docs/RuntimeAPI/html/structhipArray.html | 120 ++ .../structhipChannelFormatDesc-members.html | 106 ++ .../html/structhipChannelFormatDesc.html | 123 ++ .../html/structhipDeviceArch__t-members.html | 118 ++ .../html/structhipDeviceArch__t.html | 176 +++ .../html/structhipDeviceProp__t-members.html | 127 ++ .../html/structhipDeviceProp__t.html | 216 +++ .../html/structhipDoubleComplex-members.html | 103 ++ .../html/structhipDoubleComplex.html | 114 ++ .../html/structhipFloatComplex-members.html | 103 ++ .../html/structhipFloatComplex.html | 114 ++ .../structhipPointerAttribute__t-members.html | 107 ++ .../html/structhipPointerAttribute__t.html | 130 ++ .../html/structihipEvent__t-members.html | 106 ++ docs/RuntimeAPI/html/structihipEvent__t.html | 123 ++ docs/RuntimeAPI/html/structint3-members.html | 104 ++ docs/RuntimeAPI/html/structint3.html | 117 ++ docs/RuntimeAPI/html/structlong3-members.html | 104 ++ docs/RuntimeAPI/html/structlong3.html | 117 ++ .../html/structlonglong3-members.html | 104 ++ docs/RuntimeAPI/html/structlonglong3.html | 117 ++ .../RuntimeAPI/html/structshort3-members.html | 104 ++ docs/RuntimeAPI/html/structshort3.html | 117 ++ .../html/structstruct__float-members.html | 103 ++ docs/RuntimeAPI/html/structstruct__float.html | 118 ++ .../html/structtextureReference-members.html | 104 ++ .../html/structtextureReference.html | 117 ++ .../RuntimeAPI/html/structuchar3-members.html | 104 ++ docs/RuntimeAPI/html/structuchar3.html | 117 ++ .../RuntimeAPI/html/structuchar4-members.html | 106 ++ docs/RuntimeAPI/html/structuchar4.html | 127 ++ docs/RuntimeAPI/html/structuint3-members.html | 104 ++ docs/RuntimeAPI/html/structuint3.html | 117 ++ .../RuntimeAPI/html/structulong3-members.html | 104 ++ docs/RuntimeAPI/html/structulong3.html | 117 ++ .../html/structulonglong3-members.html | 104 ++ docs/RuntimeAPI/html/structulonglong3.html | 117 ++ .../html/structushort3-members.html | 104 ++ docs/RuntimeAPI/html/structushort3.html | 117 ++ docs/RuntimeAPI/html/sync_off.png | Bin 0 -> 853 bytes docs/RuntimeAPI/html/sync_on.png | Bin 0 -> 845 bytes docs/RuntimeAPI/html/tab_a.png | Bin 0 -> 142 bytes docs/RuntimeAPI/html/tab_b.png | Bin 0 -> 169 bytes docs/RuntimeAPI/html/tab_h.png | Bin 0 -> 177 bytes docs/RuntimeAPI/html/tab_s.png | Bin 0 -> 184 bytes docs/RuntimeAPI/html/tabs.css | 60 + .../html/trace__helper_8h_source.html | 238 +++ 342 files changed, 35739 insertions(+) create mode 100644 docs/RuntimeAPI/html/Synchonization.html create mode 100644 docs/RuntimeAPI/html/annotated.html create mode 100644 docs/RuntimeAPI/html/bc_s.png create mode 100644 docs/RuntimeAPI/html/bdwn.png create mode 100644 docs/RuntimeAPI/html/bug.html create mode 100644 docs/RuntimeAPI/html/classFakeMutex-members.html create mode 100644 docs/RuntimeAPI/html/classFakeMutex.html create mode 100644 docs/RuntimeAPI/html/classLockedAccessor-members.html create mode 100644 docs/RuntimeAPI/html/classLockedAccessor.html create mode 100644 docs/RuntimeAPI/html/classShortTid-members.html create mode 100644 docs/RuntimeAPI/html/classShortTid.html create mode 100644 docs/RuntimeAPI/html/classes.html create mode 100644 docs/RuntimeAPI/html/classihipCtxCriticalBase__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipCtxCriticalBase__t.html create mode 100644 docs/RuntimeAPI/html/classihipCtxCriticalBase__t.png create mode 100644 docs/RuntimeAPI/html/classihipCtx__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipCtx__t.html create mode 100644 docs/RuntimeAPI/html/classihipDevice__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipDevice__t.html create mode 100644 docs/RuntimeAPI/html/classihipException-members.html create mode 100644 docs/RuntimeAPI/html/classihipException.html create mode 100644 docs/RuntimeAPI/html/classihipException.png create mode 100644 docs/RuntimeAPI/html/classihipFunction__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipFunction__t.html create mode 100644 docs/RuntimeAPI/html/classihipIpcMemHandle__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipIpcMemHandle__t.html create mode 100644 docs/RuntimeAPI/html/classihipModule__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipModule__t.html create mode 100644 docs/RuntimeAPI/html/classihipStreamCriticalBase__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipStreamCriticalBase__t.html create mode 100644 docs/RuntimeAPI/html/classihipStreamCriticalBase__t.png create mode 100644 docs/RuntimeAPI/html/classihipStream__t-members.html create mode 100644 docs/RuntimeAPI/html/classihipStream__t.html create mode 100644 docs/RuntimeAPI/html/closed.png create mode 100644 docs/RuntimeAPI/html/deprecated.html create mode 100644 docs/RuntimeAPI/html/device__util_8h_source.html create mode 100644 docs/RuntimeAPI/html/dir_68267d1309a1af8e8297ef4c3efbcdba.html create mode 100644 docs/RuntimeAPI/html/dir_82c4e5e99cd964a390bd8b007f79d8a6.html create mode 100644 docs/RuntimeAPI/html/dir_ce4f9d092cd8b82b7c46cef8ba457558.html create mode 100644 docs/RuntimeAPI/html/dir_d44c64559bbebec7f509842c48db8b23.html create mode 100644 docs/RuntimeAPI/html/doxygen.css create mode 100644 docs/RuntimeAPI/html/doxygen.png create mode 100644 docs/RuntimeAPI/html/dynsections.js create mode 100644 docs/RuntimeAPI/html/files.html create mode 100644 docs/RuntimeAPI/html/ftv2blank.png create mode 100644 docs/RuntimeAPI/html/ftv2cl.png create mode 100644 docs/RuntimeAPI/html/ftv2doc.png create mode 100644 docs/RuntimeAPI/html/ftv2folderclosed.png create mode 100644 docs/RuntimeAPI/html/ftv2folderopen.png create mode 100644 docs/RuntimeAPI/html/ftv2lastnode.png create mode 100644 docs/RuntimeAPI/html/ftv2link.png create mode 100644 docs/RuntimeAPI/html/ftv2mlastnode.png create mode 100644 docs/RuntimeAPI/html/ftv2mnode.png create mode 100644 docs/RuntimeAPI/html/ftv2mo.png create mode 100644 docs/RuntimeAPI/html/ftv2node.png create mode 100644 docs/RuntimeAPI/html/ftv2ns.png create mode 100644 docs/RuntimeAPI/html/ftv2plastnode.png create mode 100644 docs/RuntimeAPI/html/ftv2pnode.png create mode 100644 docs/RuntimeAPI/html/ftv2splitbar.png create mode 100644 docs/RuntimeAPI/html/ftv2vertline.png create mode 100644 docs/RuntimeAPI/html/functions.html create mode 100644 docs/RuntimeAPI/html/functions_rela.html create mode 100644 docs/RuntimeAPI/html/functions_vars.html create mode 100644 docs/RuntimeAPI/html/globals.html create mode 100644 docs/RuntimeAPI/html/globals_defs.html create mode 100644 docs/RuntimeAPI/html/globals_enum.html create mode 100644 docs/RuntimeAPI/html/globals_eval.html create mode 100644 docs/RuntimeAPI/html/globals_func.html create mode 100644 docs/RuntimeAPI/html/globals_type.html create mode 100644 docs/RuntimeAPI/html/group__API.html create mode 100644 docs/RuntimeAPI/html/group__Context.html create mode 100644 docs/RuntimeAPI/html/group__Device.html create mode 100644 docs/RuntimeAPI/html/group__Driver.html create mode 100644 docs/RuntimeAPI/html/group__Error.html create mode 100644 docs/RuntimeAPI/html/group__Event.html create mode 100644 docs/RuntimeAPI/html/group__Fence.html create mode 100644 docs/RuntimeAPI/html/group__GlobalDefs.html create mode 100644 docs/RuntimeAPI/html/group__HCC__Specific.html create mode 100644 docs/RuntimeAPI/html/group__HIP-ENV.html create mode 100644 docs/RuntimeAPI/html/group__Memory.html create mode 100644 docs/RuntimeAPI/html/group__PeerToPeer.html create mode 100644 docs/RuntimeAPI/html/group__Profiler.html create mode 100644 docs/RuntimeAPI/html/group__Stream.html create mode 100644 docs/RuntimeAPI/html/group__Texture.html create mode 100644 docs/RuntimeAPI/html/hcc_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__acc_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__complex_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__fp16_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h_source.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h.html create mode 100644 docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h_source.html create mode 100644 docs/RuntimeAPI/html/hierarchy.html create mode 100644 docs/RuntimeAPI/html/hip__common_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__complex_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__fp16_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__hcc_8cpp.html create mode 100644 docs/RuntimeAPI/html/hip__hcc_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__ldg_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__profile_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__runtime_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__runtime__api_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__texture_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__util_8h_source.html create mode 100644 docs/RuntimeAPI/html/hip__vector__types_8h_source.html create mode 100644 docs/RuntimeAPI/html/host__defines_8h.html create mode 100644 docs/RuntimeAPI/html/host__defines_8h_source.html create mode 100644 docs/RuntimeAPI/html/index.html create mode 100644 docs/RuntimeAPI/html/jquery.js create mode 100644 docs/RuntimeAPI/html/modules.html create mode 100644 docs/RuntimeAPI/html/nav_f.png create mode 100644 docs/RuntimeAPI/html/nav_g.png create mode 100644 docs/RuntimeAPI/html/nav_h.png create mode 100644 docs/RuntimeAPI/html/open.png create mode 100644 docs/RuntimeAPI/html/pages.html create mode 100644 docs/RuntimeAPI/html/search/all_0.html create mode 100644 docs/RuntimeAPI/html/search/all_0.js create mode 100644 docs/RuntimeAPI/html/search/all_1.html create mode 100644 docs/RuntimeAPI/html/search/all_1.js create mode 100644 docs/RuntimeAPI/html/search/all_10.html create mode 100644 docs/RuntimeAPI/html/search/all_10.js create mode 100644 docs/RuntimeAPI/html/search/all_11.html create mode 100644 docs/RuntimeAPI/html/search/all_11.js create mode 100644 docs/RuntimeAPI/html/search/all_12.html create mode 100644 docs/RuntimeAPI/html/search/all_12.js create mode 100644 docs/RuntimeAPI/html/search/all_13.html create mode 100644 docs/RuntimeAPI/html/search/all_13.js create mode 100644 docs/RuntimeAPI/html/search/all_14.html create mode 100644 docs/RuntimeAPI/html/search/all_14.js create mode 100644 docs/RuntimeAPI/html/search/all_15.html create mode 100644 docs/RuntimeAPI/html/search/all_15.js create mode 100644 docs/RuntimeAPI/html/search/all_2.html create mode 100644 docs/RuntimeAPI/html/search/all_2.js create mode 100644 docs/RuntimeAPI/html/search/all_3.html create mode 100644 docs/RuntimeAPI/html/search/all_3.js create mode 100644 docs/RuntimeAPI/html/search/all_4.html create mode 100644 docs/RuntimeAPI/html/search/all_4.js create mode 100644 docs/RuntimeAPI/html/search/all_5.html create mode 100644 docs/RuntimeAPI/html/search/all_5.js create mode 100644 docs/RuntimeAPI/html/search/all_6.html create mode 100644 docs/RuntimeAPI/html/search/all_6.js create mode 100644 docs/RuntimeAPI/html/search/all_7.html create mode 100644 docs/RuntimeAPI/html/search/all_7.js create mode 100644 docs/RuntimeAPI/html/search/all_8.html create mode 100644 docs/RuntimeAPI/html/search/all_8.js create mode 100644 docs/RuntimeAPI/html/search/all_9.html create mode 100644 docs/RuntimeAPI/html/search/all_9.js create mode 100644 docs/RuntimeAPI/html/search/all_a.html create mode 100644 docs/RuntimeAPI/html/search/all_a.js create mode 100644 docs/RuntimeAPI/html/search/all_b.html create mode 100644 docs/RuntimeAPI/html/search/all_b.js create mode 100644 docs/RuntimeAPI/html/search/all_c.html create mode 100644 docs/RuntimeAPI/html/search/all_c.js create mode 100644 docs/RuntimeAPI/html/search/all_d.html create mode 100644 docs/RuntimeAPI/html/search/all_d.js create mode 100644 docs/RuntimeAPI/html/search/all_e.html create mode 100644 docs/RuntimeAPI/html/search/all_e.js create mode 100644 docs/RuntimeAPI/html/search/all_f.html create mode 100644 docs/RuntimeAPI/html/search/all_f.js create mode 100644 docs/RuntimeAPI/html/search/classes_0.html create mode 100644 docs/RuntimeAPI/html/search/classes_0.js create mode 100644 docs/RuntimeAPI/html/search/classes_1.html create mode 100644 docs/RuntimeAPI/html/search/classes_1.js create mode 100644 docs/RuntimeAPI/html/search/classes_2.html create mode 100644 docs/RuntimeAPI/html/search/classes_2.js create mode 100644 docs/RuntimeAPI/html/search/classes_3.html create mode 100644 docs/RuntimeAPI/html/search/classes_3.js create mode 100644 docs/RuntimeAPI/html/search/classes_4.html create mode 100644 docs/RuntimeAPI/html/search/classes_4.js create mode 100644 docs/RuntimeAPI/html/search/classes_5.html create mode 100644 docs/RuntimeAPI/html/search/classes_5.js create mode 100644 docs/RuntimeAPI/html/search/classes_6.html create mode 100644 docs/RuntimeAPI/html/search/classes_6.js create mode 100644 docs/RuntimeAPI/html/search/classes_7.html create mode 100644 docs/RuntimeAPI/html/search/classes_7.js create mode 100644 docs/RuntimeAPI/html/search/classes_8.html create mode 100644 docs/RuntimeAPI/html/search/classes_8.js create mode 100644 docs/RuntimeAPI/html/search/classes_9.html create mode 100644 docs/RuntimeAPI/html/search/classes_9.js create mode 100644 docs/RuntimeAPI/html/search/classes_a.html create mode 100644 docs/RuntimeAPI/html/search/classes_a.js create mode 100644 docs/RuntimeAPI/html/search/close.png create mode 100644 docs/RuntimeAPI/html/search/defines_0.html create mode 100644 docs/RuntimeAPI/html/search/defines_0.js create mode 100644 docs/RuntimeAPI/html/search/defines_1.html create mode 100644 docs/RuntimeAPI/html/search/defines_1.js create mode 100644 docs/RuntimeAPI/html/search/enums_0.html create mode 100644 docs/RuntimeAPI/html/search/enums_0.js create mode 100644 docs/RuntimeAPI/html/search/enumvalues_0.html create mode 100644 docs/RuntimeAPI/html/search/enumvalues_0.js create mode 100644 docs/RuntimeAPI/html/search/files_0.html create mode 100644 docs/RuntimeAPI/html/search/files_0.js create mode 100644 docs/RuntimeAPI/html/search/functions_0.html create mode 100644 docs/RuntimeAPI/html/search/functions_0.js create mode 100644 docs/RuntimeAPI/html/search/functions_1.html create mode 100644 docs/RuntimeAPI/html/search/functions_1.js create mode 100644 docs/RuntimeAPI/html/search/groups_0.html create mode 100644 docs/RuntimeAPI/html/search/groups_0.js create mode 100644 docs/RuntimeAPI/html/search/groups_1.html create mode 100644 docs/RuntimeAPI/html/search/groups_1.js create mode 100644 docs/RuntimeAPI/html/search/groups_2.html create mode 100644 docs/RuntimeAPI/html/search/groups_2.js create mode 100644 docs/RuntimeAPI/html/search/groups_3.html create mode 100644 docs/RuntimeAPI/html/search/groups_3.js create mode 100644 docs/RuntimeAPI/html/search/groups_4.html create mode 100644 docs/RuntimeAPI/html/search/groups_4.js create mode 100644 docs/RuntimeAPI/html/search/groups_5.html create mode 100644 docs/RuntimeAPI/html/search/groups_5.js create mode 100644 docs/RuntimeAPI/html/search/groups_6.html create mode 100644 docs/RuntimeAPI/html/search/groups_6.js create mode 100644 docs/RuntimeAPI/html/search/groups_7.html create mode 100644 docs/RuntimeAPI/html/search/groups_7.js create mode 100644 docs/RuntimeAPI/html/search/groups_8.html create mode 100644 docs/RuntimeAPI/html/search/groups_8.js create mode 100644 docs/RuntimeAPI/html/search/groups_9.html create mode 100644 docs/RuntimeAPI/html/search/groups_9.js create mode 100644 docs/RuntimeAPI/html/search/mag_sel.png create mode 100644 docs/RuntimeAPI/html/search/nomatches.html create mode 100644 docs/RuntimeAPI/html/search/pages_0.html create mode 100644 docs/RuntimeAPI/html/search/pages_0.js create mode 100644 docs/RuntimeAPI/html/search/pages_1.html create mode 100644 docs/RuntimeAPI/html/search/pages_1.js create mode 100644 docs/RuntimeAPI/html/search/pages_2.html create mode 100644 docs/RuntimeAPI/html/search/pages_2.js create mode 100644 docs/RuntimeAPI/html/search/related_0.html create mode 100644 docs/RuntimeAPI/html/search/related_0.js create mode 100644 docs/RuntimeAPI/html/search/search.css create mode 100644 docs/RuntimeAPI/html/search/search.js create mode 100644 docs/RuntimeAPI/html/search/search_l.png create mode 100644 docs/RuntimeAPI/html/search/search_m.png create mode 100644 docs/RuntimeAPI/html/search/search_r.png create mode 100644 docs/RuntimeAPI/html/search/typedefs_0.html create mode 100644 docs/RuntimeAPI/html/search/typedefs_0.js create mode 100644 docs/RuntimeAPI/html/search/typedefs_1.html create mode 100644 docs/RuntimeAPI/html/search/typedefs_1.js create mode 100644 docs/RuntimeAPI/html/search/variables_0.html create mode 100644 docs/RuntimeAPI/html/search/variables_0.js create mode 100644 docs/RuntimeAPI/html/search/variables_1.html create mode 100644 docs/RuntimeAPI/html/search/variables_1.js create mode 100644 docs/RuntimeAPI/html/search/variables_2.html create mode 100644 docs/RuntimeAPI/html/search/variables_2.js create mode 100644 docs/RuntimeAPI/html/search/variables_3.html create mode 100644 docs/RuntimeAPI/html/search/variables_3.js create mode 100644 docs/RuntimeAPI/html/search/variables_4.html create mode 100644 docs/RuntimeAPI/html/search/variables_4.js create mode 100644 docs/RuntimeAPI/html/search/variables_5.html create mode 100644 docs/RuntimeAPI/html/search/variables_5.js create mode 100644 docs/RuntimeAPI/html/search/variables_6.html create mode 100644 docs/RuntimeAPI/html/search/variables_6.js create mode 100644 docs/RuntimeAPI/html/search/variables_7.html create mode 100644 docs/RuntimeAPI/html/search/variables_7.js create mode 100644 docs/RuntimeAPI/html/search/variables_8.html create mode 100644 docs/RuntimeAPI/html/search/variables_8.js create mode 100644 docs/RuntimeAPI/html/search/variables_9.html create mode 100644 docs/RuntimeAPI/html/search/variables_9.js create mode 100644 docs/RuntimeAPI/html/search/variables_a.html create mode 100644 docs/RuntimeAPI/html/search/variables_a.js create mode 100644 docs/RuntimeAPI/html/search/variables_b.html create mode 100644 docs/RuntimeAPI/html/search/variables_b.js create mode 100644 docs/RuntimeAPI/html/search/variables_c.html create mode 100644 docs/RuntimeAPI/html/search/variables_c.js create mode 100644 docs/RuntimeAPI/html/search/variables_d.html create mode 100644 docs/RuntimeAPI/html/search/variables_d.js create mode 100644 docs/RuntimeAPI/html/search/variables_e.html create mode 100644 docs/RuntimeAPI/html/search/variables_e.js create mode 100644 docs/RuntimeAPI/html/search/variables_f.html create mode 100644 docs/RuntimeAPI/html/search/variables_f.js create mode 100644 docs/RuntimeAPI/html/structDbName-members.html create mode 100644 docs/RuntimeAPI/html/structDbName.html create mode 100644 docs/RuntimeAPI/html/structLockedBase-members.html create mode 100644 docs/RuntimeAPI/html/structLockedBase.html create mode 100644 docs/RuntimeAPI/html/structLockedBase.png create mode 100644 docs/RuntimeAPI/html/structProfTrigger-members.html create mode 100644 docs/RuntimeAPI/html/structProfTrigger.html create mode 100644 docs/RuntimeAPI/html/struct____half-members.html create mode 100644 docs/RuntimeAPI/html/struct____half.html create mode 100644 docs/RuntimeAPI/html/structchar3-members.html create mode 100644 docs/RuntimeAPI/html/structchar3.html create mode 100644 docs/RuntimeAPI/html/structchar4-members.html create mode 100644 docs/RuntimeAPI/html/structchar4.html create mode 100644 docs/RuntimeAPI/html/structdim3-members.html create mode 100644 docs/RuntimeAPI/html/structdim3.html create mode 100644 docs/RuntimeAPI/html/structdouble1-members.html create mode 100644 docs/RuntimeAPI/html/structdouble1.html create mode 100644 docs/RuntimeAPI/html/structdouble3-members.html create mode 100644 docs/RuntimeAPI/html/structdouble3.html create mode 100644 docs/RuntimeAPI/html/structfloat1-members.html create mode 100644 docs/RuntimeAPI/html/structfloat1.html create mode 100644 docs/RuntimeAPI/html/structfloat3-members.html create mode 100644 docs/RuntimeAPI/html/structfloat3.html create mode 100644 docs/RuntimeAPI/html/structhipArray-members.html create mode 100644 docs/RuntimeAPI/html/structhipArray.html create mode 100644 docs/RuntimeAPI/html/structhipChannelFormatDesc-members.html create mode 100644 docs/RuntimeAPI/html/structhipChannelFormatDesc.html create mode 100644 docs/RuntimeAPI/html/structhipDeviceArch__t-members.html create mode 100644 docs/RuntimeAPI/html/structhipDeviceArch__t.html create mode 100644 docs/RuntimeAPI/html/structhipDeviceProp__t-members.html create mode 100644 docs/RuntimeAPI/html/structhipDeviceProp__t.html create mode 100644 docs/RuntimeAPI/html/structhipDoubleComplex-members.html create mode 100644 docs/RuntimeAPI/html/structhipDoubleComplex.html create mode 100644 docs/RuntimeAPI/html/structhipFloatComplex-members.html create mode 100644 docs/RuntimeAPI/html/structhipFloatComplex.html create mode 100644 docs/RuntimeAPI/html/structhipPointerAttribute__t-members.html create mode 100644 docs/RuntimeAPI/html/structhipPointerAttribute__t.html create mode 100644 docs/RuntimeAPI/html/structihipEvent__t-members.html create mode 100644 docs/RuntimeAPI/html/structihipEvent__t.html create mode 100644 docs/RuntimeAPI/html/structint3-members.html create mode 100644 docs/RuntimeAPI/html/structint3.html create mode 100644 docs/RuntimeAPI/html/structlong3-members.html create mode 100644 docs/RuntimeAPI/html/structlong3.html create mode 100644 docs/RuntimeAPI/html/structlonglong3-members.html create mode 100644 docs/RuntimeAPI/html/structlonglong3.html create mode 100644 docs/RuntimeAPI/html/structshort3-members.html create mode 100644 docs/RuntimeAPI/html/structshort3.html create mode 100644 docs/RuntimeAPI/html/structstruct__float-members.html create mode 100644 docs/RuntimeAPI/html/structstruct__float.html create mode 100644 docs/RuntimeAPI/html/structtextureReference-members.html create mode 100644 docs/RuntimeAPI/html/structtextureReference.html create mode 100644 docs/RuntimeAPI/html/structuchar3-members.html create mode 100644 docs/RuntimeAPI/html/structuchar3.html create mode 100644 docs/RuntimeAPI/html/structuchar4-members.html create mode 100644 docs/RuntimeAPI/html/structuchar4.html create mode 100644 docs/RuntimeAPI/html/structuint3-members.html create mode 100644 docs/RuntimeAPI/html/structuint3.html create mode 100644 docs/RuntimeAPI/html/structulong3-members.html create mode 100644 docs/RuntimeAPI/html/structulong3.html create mode 100644 docs/RuntimeAPI/html/structulonglong3-members.html create mode 100644 docs/RuntimeAPI/html/structulonglong3.html create mode 100644 docs/RuntimeAPI/html/structushort3-members.html create mode 100644 docs/RuntimeAPI/html/structushort3.html create mode 100644 docs/RuntimeAPI/html/sync_off.png create mode 100644 docs/RuntimeAPI/html/sync_on.png create mode 100644 docs/RuntimeAPI/html/tab_a.png create mode 100644 docs/RuntimeAPI/html/tab_b.png create mode 100644 docs/RuntimeAPI/html/tab_h.png create mode 100644 docs/RuntimeAPI/html/tab_s.png create mode 100644 docs/RuntimeAPI/html/tabs.css create mode 100644 docs/RuntimeAPI/html/trace__helper_8h_source.html diff --git a/docs/RuntimeAPI/html/Synchonization.html b/docs/RuntimeAPI/html/Synchonization.html new file mode 100644 index 0000000000..fa615f44d1 --- /dev/null +++ b/docs/RuntimeAPI/html/Synchonization.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: $title + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ +
+
+

Host-synchronous behavior:

+

The following commands are "host-asynchronous" - meaning they do not wait for any preceding commands to complete, and may return control to the host thread before the requested operation completes:

+
    +
  • Kernel launches (hipLaunchKernel() )
    • +
  • Asynchronous memory copies - any memory copy API which contains "Async", such as hipMemcpyAsync())
    • +
  • Any memory set (for example, hipMemset());
    • +
  • TODO
    • +
+

"Host-synchronous" commands have the following properties:

+
    +
  • wait for all previous commands to complete.
    • +
  • will not return control back to host until the command completes.
    • +
+

The following commands are "host-synchronous".

+
    +
  • hipMemcpy waits for preceding work in the same stream to complete.
    • +
+

Stream synchronization

+

Blocking

+

The term "blocking" has two meanings in HIP.

+

The first refers to synchronization commands (ie hipStreamSynchronize, hipEventSynchronize) that cause the host CPU to wait for GPU activity to complete. These can either use an active where the host CPU spin-waits on the synchronization variable, or can use an interrupt-based scheme where the core is interrupted when the wait completes. The second technique is referred to as "blocking" (ie hipDeviceBlockingSync, hipEventBlockingSync) while the first is referred to as "active". Active can be appropriate for short tasks where latency is critical, but comes at the expense of a CPU core dedicated to monitoring the event.

+

HIP_LAUNCH_BLOCKING (also can use CUDA_LAUNCH_BLOCKING)

+
    +
  • The following commands become host-synchronous and will not return until the requested command has completed:
      +
    • Kernel launches (hipKernelLaunch).
      • +
    • Memory set commands (hipMemset, hipMemsetAsync).
      • +
    • Memory copy commands (hipMemcpy, hipMemsetAsync).
      • +
    +
    • +
+

Note CUDA_LAUNCH_BLOCKING does add any pre-serialization to the commands and does not affect the concurrent stream behavior. For example, even when CUDA_LAUNCH_BLOCKING is set, kernels or data copy commands launched to separate streams can execute concurrently. Use the NULL stream if additional stream synchronization is desired.

+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/annotated.html b/docs/RuntimeAPI/html/annotated.html new file mode 100644 index 0000000000..eeeaa5f91f --- /dev/null +++ b/docs/RuntimeAPI/html/annotated.html @@ -0,0 +1,145 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Class List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + +
+ + + + +
+ +
+ +
+
+
Class List
+
+
+
Here are the classes, structs, unions and interfaces with brief descriptions:
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
oC__half
oCchar3
oCchar4
oCDbName
oCdim3
oCdouble1
oCdouble3
oCFakeMutex
oCfloat1
oCfloat3
oChipArray
oChipChannelFormatDesc
oChipDeviceArch_t
oChipDeviceProp_t
oChipDoubleComplex
oChipFloatComplex
oChipPointerAttribute_t
oCihipCtx_t
oCihipCtxCriticalBase_t
oCihipDevice_t
oCihipEvent_t
oCihipException
oCihipFunction_t
oCihipIpcMemHandle_t
oCihipModule_t
oCihipStream_t
oCihipStreamCriticalBase_t
oCint3
oCLockedAccessor
oCLockedBase
oClong3
oClonglong3
oCProfTrigger
oCshort3
oCShortTid
oCstruct_float
oCtextureReference
oCuchar3
oCuchar4
oCuint3
oCulong3
oCulonglong3
\Cushort3
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/bc_s.png b/docs/RuntimeAPI/html/bc_s.png new file mode 100644 index 0000000000000000000000000000000000000000..224b29aa9847d5a4b3902efd602b7ddf7d33e6c2 GIT binary patch literal 676 zcmV;V0$crwP)y__>=_9%My z{n931IS})GlGUF8K#6VIbs%684A^L3@%PlP2>_sk`UWPq@f;rU*V%rPy_ekbhXT&s z(GN{DxFv}*vZp`F>S!r||M`I*nOwwKX+BC~3P5N3-)Y{65c;ywYiAh-1*hZcToLHK ztpl1xomJ+Yb}K(cfbJr2=GNOnT!UFA7Vy~fBz8?J>XHsbZoDad^8PxfSa0GDgENZS zuLCEqzb*xWX2CG*b&5IiO#NzrW*;`VC9455M`o1NBh+(k8~`XCEEoC1Ybwf;vr4K3 zg|EB<07?SOqHp9DhLpS&bzgo70I+ghB_#)K7H%AMU3v}xuyQq9&Bm~++VYhF09a+U zl7>n7Jjm$K#b*FONz~fj;I->Bf;ule1prFN9FovcDGBkpg>)O*-}eLnC{6oZHZ$o% zXKW$;0_{8hxHQ>l;_*HATI(`7t#^{$(zLe}h*mqwOc*nRY9=?Sx4OOeVIfI|0V(V2 zBrW#G7Ss9wvzr@>H*`r>zE z+e8bOBgqIgldUJlG(YUDviMB`9+DH8n-s9SXRLyJHO1!=wY^79WYZMTa(wiZ!zP66 zA~!21vmF3H2{ngD;+`6j#~6j;$*f*G_2ZD1E;9(yaw7d-QnSCpK(cR1zU3qU0000< KMNUMnLSTYoA~SLT literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/bdwn.png b/docs/RuntimeAPI/html/bdwn.png new file mode 100644 index 0000000000000000000000000000000000000000..940a0b950443a0bb1b216ac03c45b8a16c955452 GIT binary patch literal 147 zcmeAS@N?(olHy`uVBq!ia0vp^>_E)H!3HEvS)PKZC{Gv1kP61Pb5HX&C2wk~_T + + + + + +HIP: Heterogenous-computing Interface for Portability: Bug List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ +
+
+
+
Bug List
+
+
+
+
Member hipGetDeviceProperties (hipDeviceProp_t *prop, int deviceId)
+

HCC always returns 0 for maxThreadsPerMultiProcessor

+

HCC always returns 0 for regsPerBlock

+

HCC always returns 0 for l2CacheSize

+
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classFakeMutex-members.html b/docs/RuntimeAPI/html/classFakeMutex-members.html new file mode 100644 index 0000000000..37207bf255 --- /dev/null +++ b/docs/RuntimeAPI/html/classFakeMutex-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
FakeMutex Member List
+
+
+ +

This is the complete list of members for FakeMutex, including all inherited members.

+ + + + +
lock() (defined in FakeMutex)FakeMutexinline
try_lock() (defined in FakeMutex)FakeMutexinline
unlock() (defined in FakeMutex)FakeMutexinline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classFakeMutex.html b/docs/RuntimeAPI/html/classFakeMutex.html new file mode 100644 index 0000000000..dd73a2063c --- /dev/null +++ b/docs/RuntimeAPI/html/classFakeMutex.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: FakeMutex Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +List of all members
+
+
FakeMutex Class Reference
+
+
+ + + + + + + + +

+Public Member Functions

+void lock ()
 
+bool try_lock ()
 
+void unlock ()
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classLockedAccessor-members.html b/docs/RuntimeAPI/html/classLockedAccessor-members.html new file mode 100644 index 0000000000..97ee053446 --- /dev/null +++ b/docs/RuntimeAPI/html/classLockedAccessor-members.html @@ -0,0 +1,105 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
LockedAccessor< T > Member List
+
+
+ +

This is the complete list of members for LockedAccessor< T >, including all inherited members.

+ + + + + +
LockedAccessor(T &criticalData, bool autoUnlock=true) (defined in LockedAccessor< T >)LockedAccessor< T >inline
operator->() (defined in LockedAccessor< T >)LockedAccessor< T >inline
unlock() (defined in LockedAccessor< T >)LockedAccessor< T >inline
~LockedAccessor() (defined in LockedAccessor< T >)LockedAccessor< T >inline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classLockedAccessor.html b/docs/RuntimeAPI/html/classLockedAccessor.html new file mode 100644 index 0000000000..1ca2f522a5 --- /dev/null +++ b/docs/RuntimeAPI/html/classLockedAccessor.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: LockedAccessor< T > Class Template Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +List of all members
+
+
LockedAccessor< T > Class Template Reference
+
+
+ + + + + + + + +

+Public Member Functions

+ LockedAccessor (T &criticalData, bool autoUnlock=true)
 
+void unlock ()
 
+T * operator-> ()
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classShortTid-members.html b/docs/RuntimeAPI/html/classShortTid-members.html new file mode 100644 index 0000000000..65b4c7a8ab --- /dev/null +++ b/docs/RuntimeAPI/html/classShortTid-members.html @@ -0,0 +1,105 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ShortTid Member List
+
+
+ +

This is the complete list of members for ShortTid, including all inherited members.

+ + + + + +
apiSeqNum() const (defined in ShortTid)ShortTidinline
incApiSeqNum() (defined in ShortTid)ShortTidinline
ShortTid() (defined in ShortTid)ShortTid
tid() const (defined in ShortTid)ShortTidinline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classShortTid.html b/docs/RuntimeAPI/html/classShortTid.html new file mode 100644 index 0000000000..8a98c9d95f --- /dev/null +++ b/docs/RuntimeAPI/html/classShortTid.html @@ -0,0 +1,118 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ShortTid Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +List of all members
+
+
ShortTid Class Reference
+
+
+ + + + + + + + +

+Public Member Functions

+int tid () const
 
+uint64_t incApiSeqNum ()
 
+uint64_t apiSeqNum () const
 
+
The documentation for this class was generated from the following files: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classes.html b/docs/RuntimeAPI/html/classes.html new file mode 100644 index 0000000000..fbbc40e7a7 --- /dev/null +++ b/docs/RuntimeAPI/html/classes.html @@ -0,0 +1,130 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Class Index + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + +
+ + + + +
+ +
+ +
+
+
Class Index
+
+
+
C | D | F | H | I | L | P | S | T | U | _
+ + + + + + + + + + + + + + + +
  D  
+
  S  
+		double1   hipFloatComplex   ihipStreamCriticalBase_t   
double3   hipPointerAttribute_t   int3   
DbName   ShortTid   
  f  
+
  i  
+
  l  
+
  F  
+
  _  
+
float1   ihipCtx_t   long3   
FakeMutex   __half   float3   ihipCtxCriticalBase_t   longlong3   
  L  
+
  c  
+
  h  
+		ihipDevice_t   
  s  
+
ihipEvent_t   
LockedAccessor   char3   hipArray   ihipException   short3   
LockedBase   char4   hipChannelFormatDesc   ihipFunction_t   struct_float   
  P  
+
  d  
+		hipDeviceArch_t   ihipIpcMemHandle_t   
  t  
+
hipDeviceProp_t   ihipModule_t   
ProfTrigger   dim3   hipDoubleComplex   ihipStream_t   textureReference   
+
C | D | F | H | I | L | P | S | T | U | _
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipCtxCriticalBase__t-members.html b/docs/RuntimeAPI/html/classihipCtxCriticalBase__t-members.html new file mode 100644 index 0000000000..574362a0ec --- /dev/null +++ b/docs/RuntimeAPI/html/classihipCtxCriticalBase__t-members.html @@ -0,0 +1,118 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipCtxCriticalBase_t< MUTEX_TYPE > Member List
+
+
+ +

This is the complete list of members for ihipCtxCriticalBase_t< MUTEX_TYPE >, including all inherited members.

+ + + + + + + + + + + + + + + + + + +
_mutex (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >private
_peers (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
addPeerWatcher(const ihipCtx_t *thisCtx, ihipCtx_t *peer) (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
addStream(ihipStream_t *stream) (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
const_streams() const (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >inline
ihipCtxCriticalBase_t(unsigned deviceCnt) (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >inline
isPeerWatcher(const ihipCtx_t *peer) (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
lock() (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >inlineprivate
LockedAccessor< ihipCtxCriticalBase_t > (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >friend
peerAgents() const (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >inline
peerCnt() const (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >inline
printPeerWatchers(FILE *f) const (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
removePeerWatcher(const ihipCtx_t *thisCtx, ihipCtx_t *peer) (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
resetPeerWatchers(ihipCtx_t *thisDevice) (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >
streams() (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >inline
unlock() (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >inlineprivate
~ihipCtxCriticalBase_t() (defined in ihipCtxCriticalBase_t< MUTEX_TYPE >)ihipCtxCriticalBase_t< MUTEX_TYPE >inline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipCtxCriticalBase__t.html b/docs/RuntimeAPI/html/classihipCtxCriticalBase__t.html new file mode 100644 index 0000000000..e18eee4e66 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipCtxCriticalBase__t.html @@ -0,0 +1,164 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipCtxCriticalBase_t< MUTEX_TYPE > Class Template Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +Friends | +List of all members
+
+
ihipCtxCriticalBase_t< MUTEX_TYPE > Class Template Reference
+
+
+
+Inheritance diagram for ihipCtxCriticalBase_t< MUTEX_TYPE >:
+
+
+ + +LockedBase< MUTEX_TYPE > + +
+ + + + + + + + + + + + + + + + + + + + + + + + +

+Public Member Functions

+ ihipCtxCriticalBase_t (unsigned deviceCnt)
 
+void addStream (ihipStream_t *stream)
 
+std::list< ihipStream_t * > & streams ()
 
+const std::list< ihipStream_t * > & const_streams () const
 
+bool isPeerWatcher (const ihipCtx_t *peer)
 
+bool addPeerWatcher (const ihipCtx_t *thisCtx, ihipCtx_t *peer)
 
+bool removePeerWatcher (const ihipCtx_t *thisCtx, ihipCtx_t *peer)
 
+void resetPeerWatchers (ihipCtx_t *thisDevice)
 
+void printPeerWatchers (FILE *f) const
 
+uint32_t peerCnt () const
 
+hsa_agent_t * peerAgents () const
 
+ + + +

+Public Attributes

+std::list< ihipCtx_t * > _peers
 
+ + + +

+Friends

+class LockedAccessor< ihipCtxCriticalBase_t >
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipCtxCriticalBase__t.png b/docs/RuntimeAPI/html/classihipCtxCriticalBase__t.png new file mode 100644 index 0000000000000000000000000000000000000000..7fe1e9a2dd6c507923ac73c813e1029bf5965ecd GIT binary patch literal 824 zcmeAS@N?(olHy`uVBq!ia0vp^uYfp!gBeIFNK57eDTx4|5ZC|z{{xvX-h3_XKQsZz z0^#Ar*{o=SC+jHsEpXcm1~inc_iy z&tr^2w`Oj=3kt3;QB)2T&ujRtD4W9$b2V##p0RJtIo+avRf_EZiYOX z(Q*rRs9mE?`i!Fx18KHyE1my?SiO z^V(vC_DhX5%nO-#7QfVV$e*>2>C3X0S|l*`KhEM%>``Poz%gqD!`{qAK>a3+wxT&Foh_U_aRUwz?pR_PLJ>j~di?MSyKmyf-oD z{}3k>x=wD9S + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipCtx_t Member List
+
+
+ +

This is the complete list of members for ihipCtx_t, including all inherited members.

+ + + + + + + + + + + + + + + +
_ctxFlags (defined in ihipCtx_t)ihipCtx_t
_defaultStream (defined in ihipCtx_t)ihipCtx_t
criticalData() (defined in ihipCtx_t)ihipCtx_tinline
getDevice() const (defined in ihipCtx_t)ihipCtx_tinline
getDeviceNum() const (defined in ihipCtx_t)ihipCtx_tinline
getWriteableDevice() const (defined in ihipCtx_t)ihipCtx_tinline
ihipCtx_t(ihipDevice_t *device, unsigned deviceCnt, unsigned flags) (defined in ihipCtx_t)ihipCtx_t
locked_addStream(ihipStream_t *s) (defined in ihipCtx_t)ihipCtx_t
locked_removeStream(ihipStream_t *s) (defined in ihipCtx_t)ihipCtx_t
locked_reset() (defined in ihipCtx_t)ihipCtx_t
locked_syncDefaultStream(bool waitOnSelf) (defined in ihipCtx_t)ihipCtx_t
locked_waitAllStreams() (defined in ihipCtx_t)ihipCtx_t
toString() const (defined in ihipCtx_t)ihipCtx_t
~ihipCtx_t() (defined in ihipCtx_t)ihipCtx_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipCtx__t.html b/docs/RuntimeAPI/html/classihipCtx__t.html new file mode 100644 index 0000000000..42115ba7d7 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipCtx__t.html @@ -0,0 +1,152 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipCtx_t Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
ihipCtx_t Class Reference
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + +

+Public Member Functions

+ ihipCtx_t (ihipDevice_t *device, unsigned deviceCnt, unsigned flags)
 
+void locked_addStream (ihipStream_t *s)
 
+void locked_removeStream (ihipStream_t *s)
 
+void locked_reset ()
 
+void locked_waitAllStreams ()
 
+void locked_syncDefaultStream (bool waitOnSelf)
 
+ihipCtxCritical_t & criticalData ()
 
+const ihipDevice_t * getDevice () const
 
+int getDeviceNum () const
 
+ihipDevice_t * getWriteableDevice () const
 
+std::string toString () const
 
+ + + + + +

+Public Attributes

+ihipStream_t * _defaultStream
 
+unsigned _ctxFlags
 
+
The documentation for this class was generated from the following files: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipDevice__t-members.html b/docs/RuntimeAPI/html/classihipDevice__t-members.html new file mode 100644 index 0000000000..bcff2ac95d --- /dev/null +++ b/docs/RuntimeAPI/html/classihipDevice__t-members.html @@ -0,0 +1,111 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipDevice_t Member List
+
+
+ +

This is the complete list of members for ihipDevice_t, including all inherited members.

+ + + + + + + + + + + +
_acc (defined in ihipDevice_t)ihipDevice_t
_computeUnitsihipDevice_t
_deviceId (defined in ihipDevice_t)ihipDevice_t
_hsaAgent (defined in ihipDevice_t)ihipDevice_t
_isLargeBar (defined in ihipDevice_t)ihipDevice_t
_primaryCtx (defined in ihipDevice_t)ihipDevice_t
_props (defined in ihipDevice_t)ihipDevice_t
getPrimaryCtx() const (defined in ihipDevice_t)ihipDevice_tinline
ihipDevice_t(unsigned deviceId, unsigned deviceCnt, hc::accelerator &acc) (defined in ihipDevice_t)ihipDevice_t
~ihipDevice_t() (defined in ihipDevice_t)ihipDevice_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipDevice__t.html b/docs/RuntimeAPI/html/classihipDevice__t.html new file mode 100644 index 0000000000..53dbfb48ce --- /dev/null +++ b/docs/RuntimeAPI/html/classihipDevice__t.html @@ -0,0 +1,141 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipDevice_t Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
ihipDevice_t Class Reference
+
+
+ + + + + + +

+Public Member Functions

+ ihipDevice_t (unsigned deviceId, unsigned deviceCnt, hc::accelerator &acc)
 
+ihipCtx_t * getPrimaryCtx () const
 
+ + + + + + + + + + + + + + + + +

+Public Attributes

+unsigned _deviceId
 
+hc::accelerator _acc
 
+hsa_agent_t _hsaAgent
 
+unsigned _computeUnits
 Number of compute units supported by the device:
 
+hipDeviceProp_t _props
 
+int _isLargeBar
 
+ihipCtx_t * _primaryCtx
 
+
The documentation for this class was generated from the following files: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipException-members.html b/docs/RuntimeAPI/html/classihipException-members.html new file mode 100644 index 0000000000..a95b0cec3d --- /dev/null +++ b/docs/RuntimeAPI/html/classihipException-members.html @@ -0,0 +1,103 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipException Member List
+
+
+ +

This is the complete list of members for ihipException, including all inherited members.

+ + + +
_code (defined in ihipException)ihipException
ihipException(hipError_t e) (defined in ihipException)ihipExceptioninline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipException.html b/docs/RuntimeAPI/html/classihipException.html new file mode 100644 index 0000000000..09965dc766 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipException.html @@ -0,0 +1,126 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipException Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
ihipException Class Reference
+
+
+
+Inheritance diagram for ihipException:
+
+
+ + + +
+ + + + +

+Public Member Functions

+ ihipException (hipError_t e)
 
+ + + +

+Public Attributes

+hipError_t _code
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipException.png b/docs/RuntimeAPI/html/classihipException.png new file mode 100644 index 0000000000000000000000000000000000000000..36add951e6d298079acb4b2f570e4ac87eab64cf GIT binary patch literal 435 zcmV;k0ZjghP)vTJkN^MxkN^Mxkifve1&Q1r00008bW%=J0RR90|NsC0)yh;d0003(NklPw?q0nLoN!SwatebnRKwNV%#_cmm(m|jk-LByNLhiccY}I$U zl60gIckRBiHjU}!HEa~kE_4nNpZlC=>xeR^*8 zUE|tIH~AiyNRqYPI+S~xsP8QIv>(#L_QNb!lE#{k{o7ngnp1Gyxg>eN0AN-N0655r zyDUCnC+>=M0ef*jt{;LMTmaRA8{FUmK&#*e_up_y(m@~q+^*pM5O*EwtZ5Yh*i~Hy zz?jW<6EJa=|uST`LEjJ&knVG}B_w0jv->AM_T&ZuLeIMNV z&bjph&5FF&9o(iWo3y{e-IB`v9M|_e-%{54&ou6b^sxP~Yp?$DhUpfNKh_ dJC`Ky*8`7!Gj#ISOkMy0002ovPDHLkV1npj(69gi literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/classihipFunction__t-members.html b/docs/RuntimeAPI/html/classihipFunction__t-members.html new file mode 100644 index 0000000000..b2146dd003 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipFunction__t-members.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipFunction_t Member List
+
+
+ +

This is the complete list of members for ihipFunction_t, including all inherited members.

+ + + + + + +
_kernel (defined in ihipFunction_t)ihipFunction_t
_kernelName (defined in ihipFunction_t)ihipFunction_t
_kernelSymbol (defined in ihipFunction_t)ihipFunction_t
ihipFunction_t(const char *name) (defined in ihipFunction_t)ihipFunction_tinline
~ihipFunction_t() (defined in ihipFunction_t)ihipFunction_tinline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipFunction__t.html b/docs/RuntimeAPI/html/classihipFunction__t.html new file mode 100644 index 0000000000..2b852c1910 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipFunction__t.html @@ -0,0 +1,124 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipFunction_t Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
ihipFunction_t Class Reference
+
+
+ + + + +

+Public Member Functions

+ ihipFunction_t (const char *name)
 
+ + + + + + + +

+Public Attributes

+const char * _kernelName
 
+hsa_executable_symbol_t _kernelSymbol
 
+uint64_t _kernel
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipIpcMemHandle__t-members.html b/docs/RuntimeAPI/html/classihipIpcMemHandle__t-members.html new file mode 100644 index 0000000000..87eee6dc9a --- /dev/null +++ b/docs/RuntimeAPI/html/classihipIpcMemHandle__t-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipIpcMemHandle_t Member List
+
+
+ +

This is the complete list of members for ihipIpcMemHandle_t, including all inherited members.

+ + + + +
ipc_handleihipIpcMemHandle_t
psize (defined in ihipIpcMemHandle_t)ihipIpcMemHandle_t
reserved (defined in ihipIpcMemHandle_t)ihipIpcMemHandle_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipIpcMemHandle__t.html b/docs/RuntimeAPI/html/classihipIpcMemHandle__t.html new file mode 100644 index 0000000000..b057cf392b --- /dev/null +++ b/docs/RuntimeAPI/html/classihipIpcMemHandle__t.html @@ -0,0 +1,118 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipIpcMemHandle_t Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
ihipIpcMemHandle_t Class Reference
+
+
+ + + + + + + + + +

+Public Attributes

+hsa_amd_ipc_memory_t ipc_handle
 ipc memory handle on ROCr
 
+char reserved [HIP_IPC_HANDLE_SIZE]
 
+size_t psize
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipModule__t-members.html b/docs/RuntimeAPI/html/classihipModule__t-members.html new file mode 100644 index 0000000000..0ee668b4a5 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipModule__t-members.html @@ -0,0 +1,109 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipModule_t Member List
+
+
+ +

This is the complete list of members for ihipModule_t, including all inherited members.

+ + + + + + + + + +
executable (defined in ihipModule_t)ihipModule_t
fileName (defined in ihipModule_t)ihipModule_t
ihipModule_t() (defined in ihipModule_t)ihipModule_tinline
object (defined in ihipModule_t)ihipModule_t
ptr (defined in ihipModule_t)ihipModule_t
registerFunction(ihipFunction_t *func) (defined in ihipModule_t)ihipModule_tinline
size (defined in ihipModule_t)ihipModule_t
~ihipModule_t() (defined in ihipModule_t)ihipModule_tinline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipModule__t.html b/docs/RuntimeAPI/html/classihipModule__t.html new file mode 100644 index 0000000000..0726617630 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipModule__t.html @@ -0,0 +1,130 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipModule_t Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
ihipModule_t Class Reference
+
+
+ + + + +

+Public Member Functions

+void registerFunction (ihipFunction_t *func)
 
+ + + + + + + + + + + +

+Public Attributes

+hsa_executable_t executable
 
+hsa_code_object_t object
 
+std::string fileName
 
+void * ptr
 
+size_t size
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipStreamCriticalBase__t-members.html b/docs/RuntimeAPI/html/classihipStreamCriticalBase__t-members.html new file mode 100644 index 0000000000..6cd21dc31f --- /dev/null +++ b/docs/RuntimeAPI/html/classihipStreamCriticalBase__t-members.html @@ -0,0 +1,109 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipStreamCriticalBase_t< MUTEX_TYPE > Member List
+
+
+ +

This is the complete list of members for ihipStreamCriticalBase_t< MUTEX_TYPE >, including all inherited members.

+ + + + + + + + + +
_av (defined in ihipStreamCriticalBase_t< MUTEX_TYPE >)ihipStreamCriticalBase_t< MUTEX_TYPE >
_kernelCnt (defined in ihipStreamCriticalBase_t< MUTEX_TYPE >)ihipStreamCriticalBase_t< MUTEX_TYPE >
_mutex (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >
ihipStreamCriticalBase_t(hc::accelerator_view av) (defined in ihipStreamCriticalBase_t< MUTEX_TYPE >)ihipStreamCriticalBase_t< MUTEX_TYPE >inline
lock() (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >inline
mlock() (defined in ihipStreamCriticalBase_t< MUTEX_TYPE >)ihipStreamCriticalBase_t< MUTEX_TYPE >inline
unlock() (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >inline
~ihipStreamCriticalBase_t() (defined in ihipStreamCriticalBase_t< MUTEX_TYPE >)ihipStreamCriticalBase_t< MUTEX_TYPE >inline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipStreamCriticalBase__t.html b/docs/RuntimeAPI/html/classihipStreamCriticalBase__t.html new file mode 100644 index 0000000000..c06d346cc3 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipStreamCriticalBase__t.html @@ -0,0 +1,145 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipStreamCriticalBase_t< MUTEX_TYPE > Class Template Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
ihipStreamCriticalBase_t< MUTEX_TYPE > Class Template Reference
+
+
+
+Inheritance diagram for ihipStreamCriticalBase_t< MUTEX_TYPE >:
+
+
+ + +LockedBase< MUTEX_TYPE > + +
+ + + + + + + + + + + +

+Public Member Functions

+ ihipStreamCriticalBase_t (hc::accelerator_view av)
 
+ihipStreamCriticalBase_t
+< StreamMutex > * 		mlock ()
 
- Public Member Functions inherited from LockedBase< MUTEX_TYPE >
+void lock ()
 
+void unlock ()
 
+ + + + + + + + +

+Public Attributes

+uint32_t _kernelCnt
 
+hc::accelerator_view _av
 
- Public Attributes inherited from LockedBase< MUTEX_TYPE >
+MUTEX_TYPE _mutex
 
+
The documentation for this class was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipStreamCriticalBase__t.png b/docs/RuntimeAPI/html/classihipStreamCriticalBase__t.png new file mode 100644 index 0000000000000000000000000000000000000000..4d8990f4ca2ca48ac6b812c10391032bb5bb60eb GIT binary patch literal 854 zcmeAS@N?(olHy`uVBq!ia0vp^|A9DwgBeIZoD+N)NJ#|vgt-3y4-$Xz=4)yHp$R}1 z7#}!rfVK0EJdn##666=m08|75S5Ji)F)%Pa^>lFzsbG9N_w}?}20U!XPg(ByuWa6+ zKS_TAL-ehw+kSH0*z{(i&Qtx{+j9Gkswq7RJ1GB0NhwbBPy5fw9Ukiw3*zk6&^DoMOyt4%Z{p|`_-~Q8%?@>2hJ8f^vKdt({ zlD-LsA-LKx&yXy6>%@=)JHs?hBeaYi`B_>SypI-~T z{c&~b^}COG{xAAB|HJ>af4(2B-SzLd_GiX_@~i)c;)VEU>C!Q?zna$YqYT4=Yj|%K_U*5WT^X{VDyji<$ZN9<$?BdV)b1V6nUo5|& zD=W#;FtzI6nMAQ@bLYA%TcWnzy7p{;%I@0Jj%V8}Tb_oc?6+##UT?`bx70dvZpP)T z<^CV8KMel!_0=@q)Bj%aym~CA`Z&5$V$IvgD(&hYW$%3#r2lbdw?435`QFv~`W03D zJtZIc^R9hZU2w0rJsY7v(GPm z>8)9I_Hn?!-{&m8Z8Q4cuz&Z4U#t@y?D>C8|6U(9!};I*Px`O^bN=7_=RCj7>WtWw z%QL^fsy}g5{RjV*`nP9yeSg`P_c=%{Dt%65_~tw2?f>%g>%}XA-!GkaTx`RG?WNn+ z%=^kM^WghS`G@-B?H|qZ2Ua)PFP$Runt%T%)0^VA8q|TAfWgz% K&t;ucLK6V?!m + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipStream_t Member List
+
+
+ +

This is the complete list of members for ihipStream_t, including all inherited members.

+ + + + + + + + + + + + + + + + + + + + + + + + +
_flags (defined in ihipStream_t)ihipStream_t
_id (defined in ihipStream_t)ihipStream_t
Auto enum value (defined in ihipStream_t)ihipStream_t
getCtx() const (defined in ihipStream_t)ihipStream_t
getDevice() const (defined in ihipStream_t)ihipStream_t
hipStreamQuery(hipStream_t)ihipStream_tfriend
ihipStream_t(ihipCtx_t *ctx, hc::accelerator_view av, unsigned int flags) (defined in ihipStream_t)ihipStream_t
launchModuleKernel(hc::accelerator_view av, hsa_signal_t signal, uint32_t blockDimX, uint32_t blockDimY, uint32_t blockDimZ, uint32_t gridDimX, uint32_t gridDimY, uint32_t gridDimZ, uint32_t groupSegmentSize, uint32_t sharedMemBytes, void *kernarg, size_t kernSize, uint64_t kernel) (defined in ihipStream_t)ihipStream_t
lockclose_postKernelCommand(const char *kernelName, hc::accelerator_view *av) (defined in ihipStream_t)ihipStream_t
locked_copyAsync(void *dst, const void *src, size_t sizeBytes, unsigned kind) (defined in ihipStream_t)ihipStream_t
locked_copySync(void *dst, const void *src, size_t sizeBytes, unsigned kind, bool resolveOn=true) (defined in ihipStream_t)ihipStream_t
locked_getAv() (defined in ihipStream_t)ihipStream_tinline
locked_recordEvent(hipEvent_t event) (defined in ihipStream_t)ihipStream_t
locked_wait(bool assertQueueEmpty=false) (defined in ihipStream_t)ihipStream_t
locked_waitEvent(hipEvent_t event) (defined in ihipStream_t)ihipStream_t
lockopen_preKernelCommand() (defined in ihipStream_t)ihipStream_t
operator<< (defined in ihipStream_t)ihipStream_tfriend
ScheduleMode enum name (defined in ihipStream_t)ihipStream_t
SeqNum_t typedef (defined in ihipStream_t)ihipStream_t
Spin enum value (defined in ihipStream_t)ihipStream_t
wait(LockedAccessor_StreamCrit_t &crit, bool assertQueueEmpty=false) (defined in ihipStream_t)ihipStream_t
Yield enum value (defined in ihipStream_t)ihipStream_t
~ihipStream_t() (defined in ihipStream_t)ihipStream_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/classihipStream__t.html b/docs/RuntimeAPI/html/classihipStream__t.html new file mode 100644 index 0000000000..9b06ac3bc4 --- /dev/null +++ b/docs/RuntimeAPI/html/classihipStream__t.html @@ -0,0 +1,216 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipStream_t Class Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Types | +Public Member Functions | +Public Attributes | +Friends | +List of all members
+
+
ihipStream_t Class Reference
+
+
+ + + + + + +

+Public Types

enum  ScheduleMode { Auto, +Spin, +Yield + }
 
+typedef uint64_t SeqNum_t
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Public Member Functions

+ ihipStream_t (ihipCtx_t *ctx, hc::accelerator_view av, unsigned int flags)
 
+void locked_copySync (void *dst, const void *src, size_t sizeBytes, unsigned kind, bool resolveOn=true)
 
+void locked_copyAsync (void *dst, const void *src, size_t sizeBytes, unsigned kind)
 
+LockedAccessor_StreamCrit_t lockopen_preKernelCommand ()
 
+void lockclose_postKernelCommand (const char *kernelName, hc::accelerator_view *av)
 
+void locked_wait (bool assertQueueEmpty=false)
 
+hc::accelerator_view * locked_getAv ()
 
+void locked_waitEvent (hipEvent_t event)
 
+void locked_recordEvent (hipEvent_t event)
 
+void wait (LockedAccessor_StreamCrit_t &crit, bool assertQueueEmpty=false)
 
+void launchModuleKernel (hc::accelerator_view av, hsa_signal_t signal, uint32_t blockDimX, uint32_t blockDimY, uint32_t blockDimZ, uint32_t gridDimX, uint32_t gridDimY, uint32_t gridDimZ, uint32_t groupSegmentSize, uint32_t sharedMemBytes, void *kernarg, size_t kernSize, uint64_t kernel)
 
+const ihipDevice_t * getDevice () const
 
+ihipCtx_t * getCtx () const
 
+ + + + + +

+Public Attributes

+SeqNum_t _id
 
+unsigned _flags
 
+ + + + + + +

+Friends

+std::ostream & operator<< (std::ostream &os, const ihipStream_t &s)
 
hipError_t hipStreamQuery (hipStream_t)
 Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not. More...
 
+

Friends And Related Function Documentation

+ +
+
+ + + + + +
+ + + + + + + + +
hipError_t hipStreamQuery (hipStream_t stream)
+ 		+friend
+
+ +

Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not.

+
Parameters
+ + +
[in]streamstream to query
+
+
+
Returns
hipSuccess, hipErrorNotReady, hipErrorInvalidResourceHandle
+

This is thread-safe and returns a snapshot of the current state of the queue. However, if other host threads are sending work to the stream, the status may change immediately after the function is called. It is typically used for debug.

+
See Also
hipStreamCreate, hipStreamCreateWithFlags, hipStreamWaitEvent, hipStreamSynchronize, hipStreamDestroy
+ +
+
+
The documentation for this class was generated from the following files: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/closed.png b/docs/RuntimeAPI/html/closed.png new file mode 100644 index 0000000000000000000000000000000000000000..98cc2c909da37a6df914fbf67780eebd99c597f5 GIT binary patch literal 132 zcmeAS@N?(olHy`uVBq!ia0vp^oFL4>1|%O$WD@{V-kvUwAr*{o@8{^CZMh(5KoB^r_<4^zF@3)Cp&&t3hdujKf f*?bjBoY!V+E))@{xMcbjXe@)LtDnm{r-UW|*e5JT literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/deprecated.html b/docs/RuntimeAPI/html/deprecated.html new file mode 100644 index 0000000000..3511316256 --- /dev/null +++ b/docs/RuntimeAPI/html/deprecated.html @@ -0,0 +1,98 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Deprecated List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ +
+
+
+
Deprecated List
+
+
+
+
Member hipFreeHost (void *ptr) __attribute__((deprecated("use hipHostFree instead")))
+
use hipHostFree() instead
+
Member hipHostAlloc (void **ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead")))
+
use hipHostMalloc() instead
+
Member hipMallocHost (void **ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead")))
+
use hipHostMalloc() instead
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/device__util_8h_source.html b/docs/RuntimeAPI/html/device__util_8h_source.html new file mode 100644 index 0000000000..7b1dbfd91e --- /dev/null +++ b/docs/RuntimeAPI/html/device__util_8h_source.html @@ -0,0 +1,138 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/src/device_util.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
device_util.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef DEVICE_UTIL_H
+
24 #define DEVICE_UTIL_H
+
25 
+
26 /*
+
27  Heap size computation for malloc and free device functions.
+
28 */
+
29 
+
30 #define NUM_PAGES_PER_THREAD 16
+
31 #define SIZE_OF_PAGE 64
+
32 #define NUM_THREADS_PER_CU 64
+
33 #define NUM_CUS_PER_GPU 64
+
34 #define NUM_PAGES NUM_PAGES_PER_THREAD * NUM_THREADS_PER_CU * NUM_CUS_PER_GPU
+
35 #define SIZE_MALLOC NUM_PAGES * SIZE_OF_PAGE
+
36 #define SIZE_OF_HEAP SIZE_MALLOC
+
37 
+
38 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/dir_68267d1309a1af8e8297ef4c3efbcdba.html b/docs/RuntimeAPI/html/dir_68267d1309a1af8e8297ef4c3efbcdba.html new file mode 100644 index 0000000000..96d309a573 --- /dev/null +++ b/docs/RuntimeAPI/html/dir_68267d1309a1af8e8297ef4c3efbcdba.html @@ -0,0 +1,130 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/src Directory Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ + +
+
+
+
src Directory Reference
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Files

file  device_util.cpp
 
file  device_util.h [code]
 
file  hip_context.cpp
 
file  hip_device.cpp
 
file  hip_error.cpp
 
file  hip_event.cpp
 
file  hip_fp16.cpp
 
file  hip_hcc.cpp
 
file  hip_hcc.h [code]
 
file  hip_ldg.cpp
 
file  hip_memory.cpp
 
file  hip_module.cpp
 
file  hip_peer.cpp
 
file  hip_stream.cpp
 
file  hip_util.h [code]
 
file  trace_helper.h [code]
 
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/dir_82c4e5e99cd964a390bd8b007f79d8a6.html b/docs/RuntimeAPI/html/dir_82c4e5e99cd964a390bd8b007f79d8a6.html new file mode 100644 index 0000000000..15a657aa89 --- /dev/null +++ b/docs/RuntimeAPI/html/dir_82c4e5e99cd964a390bd8b007f79d8a6.html @@ -0,0 +1,121 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip Directory Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ + +
+
+
+
hip Directory Reference
+
+
+ + + + +

+Directories

directory  hcc_detail
 
+ + + + + + + + + + + + + + + + + + + +

+Files

file  hcc.h [code]
 
file  hip_common.h [code]
 
file  hip_complex.h [code]
 
file  hip_fp16.h [code]
 
file  hip_profile.h [code]
 
file  hip_runtime.h [code]
 
file  hip_runtime_api.h [code]
 
file  hip_texture.h [code]
 
file  hip_vector_types.h [code]
 
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/dir_ce4f9d092cd8b82b7c46cef8ba457558.html b/docs/RuntimeAPI/html/dir_ce4f9d092cd8b82b7c46cef8ba457558.html new file mode 100644 index 0000000000..64041d944b --- /dev/null +++ b/docs/RuntimeAPI/html/dir_ce4f9d092cd8b82b7c46cef8ba457558.html @@ -0,0 +1,121 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail Directory Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ + +
+
+
+
hcc_detail Directory Reference
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + +

+Files

file  hcc_acc.h [code]
 
file  hip_complex.h [code]
 
file  hip_fp16.h [code]
 
file  hip_ldg.h [code]
 
file  hip_runtime.h [code]
 Contains definitions of APIs for HIP runtime.
 
file  hip_runtime_api.h [code]
 Contains C function APIs for HIP runtime. This file does not use any HCC builtin or special language extensions (-hc mode) ; those functions in hip_runtime.h.
 
file  hip_texture.h [code]
 HIP C++ Texture API for hcc compiler.
 
file  hip_vector_types.h [code]
 Defines the different newt vector types for HIP runtime.
 
file  host_defines.h [code]
 TODO-doc.
 
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/dir_d44c64559bbebec7f509842c48db8b23.html b/docs/RuntimeAPI/html/dir_d44c64559bbebec7f509842c48db8b23.html new file mode 100644 index 0000000000..719a287cc7 --- /dev/null +++ b/docs/RuntimeAPI/html/dir_d44c64559bbebec7f509842c48db8b23.html @@ -0,0 +1,100 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include Directory Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + +
+ +
+ + +
+
+
+
include Directory Reference
+
+
+ + + + +

+Directories

directory  hip
 
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/doxygen.css b/docs/RuntimeAPI/html/doxygen.css new file mode 100644 index 0000000000..f0f36f89f5 --- /dev/null +++ b/docs/RuntimeAPI/html/doxygen.css @@ -0,0 +1,1366 @@ +/* The standard CSS for doxygen 1.8.6 */ + +body, table, div, p, dl { + font: 400 14px/22px Roboto,sans-serif; +} + +/* @group Heading Levels */ + +h1.groupheader { + font-size: 150%; +} + +.title { + font: 400 14px/28px Roboto,sans-serif; + font-size: 150%; + font-weight: bold; + margin: 10px 2px; +} + +h2.groupheader { + border-bottom: 1px solid #879ECB; + color: #354C7B; + font-size: 150%; + font-weight: normal; + margin-top: 1.75em; + padding-top: 8px; + padding-bottom: 4px; + width: 100%; +} + +h3.groupheader { + font-size: 100%; +} + +h1, h2, h3, h4, h5, h6 { + -webkit-transition: text-shadow 0.5s linear; + -moz-transition: text-shadow 0.5s linear; + -ms-transition: text-shadow 0.5s linear; + -o-transition: text-shadow 0.5s linear; + transition: text-shadow 0.5s linear; + margin-right: 15px; +} + +h1.glow, h2.glow, h3.glow, h4.glow, h5.glow, h6.glow { + text-shadow: 0 0 15px cyan; +} + +dt { + font-weight: bold; +} + +div.multicol { + -moz-column-gap: 1em; + -webkit-column-gap: 1em; + -moz-column-count: 3; + -webkit-column-count: 3; +} + +p.startli, p.startdd { + margin-top: 2px; +} + +p.starttd { + margin-top: 0px; +} + +p.endli { + margin-bottom: 0px; +} + +p.enddd { + margin-bottom: 4px; +} + +p.endtd { + margin-bottom: 2px; +} + +/* @end */ + +caption { + font-weight: bold; +} + +span.legend { + font-size: 70%; + text-align: center; +} + +h3.version { + font-size: 90%; + text-align: center; +} + +div.qindex, div.navtab{ + background-color: #EBEFF6; + border: 1px solid #A3B4D7; + text-align: center; +} + +div.qindex, div.navpath { + width: 100%; + line-height: 140%; +} + +div.navtab { + margin-right: 15px; +} + +/* @group Link Styling */ + +a { + color: #3D578C; + font-weight: normal; + text-decoration: none; +} + +.contents a:visited { + color: #4665A2; +} + +a:hover { + text-decoration: underline; +} + +a.qindex { + font-weight: bold; +} + +a.qindexHL { + font-weight: bold; + background-color: #9CAFD4; + color: #ffffff; + border: 1px double #869DCA; +} + +.contents a.qindexHL:visited { + color: #ffffff; +} + +a.el { + font-weight: bold; +} + +a.elRef { +} + +a.code, a.code:visited, a.line, a.line:visited { + color: #4665A2; +} + +a.codeRef, a.codeRef:visited, a.lineRef, a.lineRef:visited { + color: #4665A2; +} + +/* @end */ + +dl.el { + margin-left: -1cm; +} + +pre.fragment { + border: 1px solid #C4CFE5; + background-color: #FBFCFD; + padding: 4px 6px; + margin: 4px 8px 4px 2px; + overflow: auto; + word-wrap: break-word; + font-size: 9pt; + line-height: 125%; + font-family: monospace, fixed; + font-size: 105%; +} + +div.fragment { + padding: 4px 6px; + margin: 4px 8px 4px 2px; + background-color: #FBFCFD; + border: 1px solid #C4CFE5; +} + +div.line { + font-family: monospace, fixed; + font-size: 13px; + min-height: 13px; + line-height: 1.0; + text-wrap: unrestricted; + white-space: -moz-pre-wrap; /* Moz */ + white-space: -pre-wrap; /* Opera 4-6 */ + white-space: -o-pre-wrap; /* Opera 7 */ + white-space: pre-wrap; /* CSS3 */ + word-wrap: break-word; /* IE 5.5+ */ + text-indent: -53px; + padding-left: 53px; + padding-bottom: 0px; + margin: 0px; + -webkit-transition-property: background-color, box-shadow; + -webkit-transition-duration: 0.5s; + -moz-transition-property: background-color, box-shadow; + -moz-transition-duration: 0.5s; + -ms-transition-property: background-color, box-shadow; + -ms-transition-duration: 0.5s; + -o-transition-property: background-color, box-shadow; + -o-transition-duration: 0.5s; + transition-property: background-color, box-shadow; + transition-duration: 0.5s; +} + +div.line.glow { + background-color: cyan; + box-shadow: 0 0 10px cyan; +} + + +span.lineno { + padding-right: 4px; + text-align: right; + border-right: 2px solid #0F0; + background-color: #E8E8E8; + white-space: pre; +} +span.lineno a { + background-color: #D8D8D8; +} + +span.lineno a:hover { + background-color: #C8C8C8; +} + +div.ah { + background-color: black; + font-weight: bold; + color: #ffffff; + margin-bottom: 3px; + margin-top: 3px; + padding: 0.2em; + border: solid thin #333; + border-radius: 0.5em; + -webkit-border-radius: .5em; + -moz-border-radius: .5em; + box-shadow: 2px 2px 3px #999; + -webkit-box-shadow: 2px 2px 3px #999; + -moz-box-shadow: rgba(0, 0, 0, 0.15) 2px 2px 2px; + background-image: -webkit-gradient(linear, left top, left bottom, from(#eee), to(#000),color-stop(0.3, #444)); + background-image: -moz-linear-gradient(center top, #eee 0%, #444 40%, #000); +} + +div.groupHeader { + margin-left: 16px; + margin-top: 12px; + font-weight: bold; +} + +div.groupText { + margin-left: 16px; + font-style: italic; +} + +body { + background-color: white; + color: black; + margin: 0; +} + +div.contents { + margin-top: 10px; + margin-left: 12px; + margin-right: 8px; +} + +td.indexkey { + background-color: #EBEFF6; + font-weight: bold; + border: 1px solid #C4CFE5; + margin: 2px 0px 2px 0; + padding: 2px 10px; + white-space: nowrap; + vertical-align: top; +} + +td.indexvalue { + background-color: #EBEFF6; + border: 1px solid #C4CFE5; + padding: 2px 10px; + margin: 2px 0px; +} + +tr.memlist { + background-color: #EEF1F7; +} + +p.formulaDsp { + text-align: center; +} + +img.formulaDsp { + +} + +img.formulaInl { + vertical-align: middle; +} + +div.center { + text-align: center; + margin-top: 0px; + margin-bottom: 0px; + padding: 0px; +} + +div.center img { + border: 0px; +} + +address.footer { + text-align: right; + padding-right: 12px; +} + +img.footer { + border: 0px; + vertical-align: middle; +} + +/* @group Code Colorization */ + +span.keyword { + color: #008000 +} + +span.keywordtype { + color: #604020 +} + +span.keywordflow { + color: #e08000 +} + +span.comment { + color: #800000 +} + +span.preprocessor { + color: #806020 +} + +span.stringliteral { + color: #002080 +} + +span.charliteral { + color: #008080 +} + +span.vhdldigit { + color: #ff00ff +} + +span.vhdlchar { + color: #000000 +} + +span.vhdlkeyword { + color: #700070 +} + +span.vhdllogic { + color: #ff0000 +} + +blockquote { + background-color: #F7F8FB; + border-left: 2px solid #9CAFD4; + margin: 0 24px 0 4px; + padding: 0 12px 0 16px; +} + +/* @end */ + +/* +.search { + color: #003399; + font-weight: bold; +} + +form.search { + margin-bottom: 0px; + margin-top: 0px; +} + +input.search { + font-size: 75%; + color: #000080; + font-weight: normal; + background-color: #e8eef2; +} +*/ + +td.tiny { + font-size: 75%; +} + +.dirtab { + padding: 4px; + border-collapse: collapse; + border: 1px solid #A3B4D7; +} + +th.dirtab { + background: #EBEFF6; + font-weight: bold; +} + +hr { + height: 0px; + border: none; + border-top: 1px solid #4A6AAA; +} + +hr.footer { + height: 1px; +} + +/* @group Member Descriptions */ + +table.memberdecls { + border-spacing: 0px; + padding: 0px; +} + +.memberdecls td, .fieldtable tr { + -webkit-transition-property: background-color, box-shadow; + -webkit-transition-duration: 0.5s; + -moz-transition-property: background-color, box-shadow; + -moz-transition-duration: 0.5s; + -ms-transition-property: background-color, box-shadow; + -ms-transition-duration: 0.5s; + -o-transition-property: background-color, box-shadow; + -o-transition-duration: 0.5s; + transition-property: background-color, box-shadow; + transition-duration: 0.5s; +} + +.memberdecls td.glow, .fieldtable tr.glow { + background-color: cyan; + box-shadow: 0 0 15px cyan; +} + +.mdescLeft, .mdescRight, +.memItemLeft, .memItemRight, +.memTemplItemLeft, .memTemplItemRight, .memTemplParams { + background-color: #F9FAFC; + border: none; + margin: 4px; + padding: 1px 0 0 8px; +} + +.mdescLeft, .mdescRight { + padding: 0px 8px 4px 8px; + color: #555; +} + +.memSeparator { + border-bottom: 1px solid #DEE4F0; + line-height: 1px; + margin: 0px; + padding: 0px; +} + +.memItemLeft, .memTemplItemLeft { + white-space: nowrap; +} + +.memItemRight { + width: 100%; +} + +.memTemplParams { + color: #4665A2; + white-space: nowrap; + font-size: 80%; +} + +/* @end */ + +/* @group Member Details */ + +/* Styles for detailed member documentation */ + +.memtemplate { + font-size: 80%; + color: #4665A2; + font-weight: normal; + margin-left: 9px; +} + +.memnav { + background-color: #EBEFF6; + border: 1px solid #A3B4D7; + text-align: center; + margin: 2px; + margin-right: 15px; + padding: 2px; +} + +.mempage { + width: 100%; +} + +.memitem { + padding: 0; + margin-bottom: 10px; + margin-right: 5px; + -webkit-transition: box-shadow 0.5s linear; + -moz-transition: box-shadow 0.5s linear; + -ms-transition: box-shadow 0.5s linear; + -o-transition: box-shadow 0.5s linear; + transition: box-shadow 0.5s linear; + display: table !important; + width: 100%; +} + +.memitem.glow { + box-shadow: 0 0 15px cyan; +} + +.memname { + font-weight: bold; + margin-left: 6px; +} + +.memname td { + vertical-align: bottom; +} + +.memproto, dl.reflist dt { + border-top: 1px solid #A8B8D9; + border-left: 1px solid #A8B8D9; + border-right: 1px solid #A8B8D9; + padding: 6px 0px 6px 0px; + color: #253555; + font-weight: bold; + text-shadow: 0px 1px 1px rgba(255, 255, 255, 0.9); + background-image:url('nav_f.png'); + background-repeat:repeat-x; + background-color: #E2E8F2; + /* opera specific markup */ + box-shadow: 5px 5px 5px rgba(0, 0, 0, 0.15); + border-top-right-radius: 4px; + border-top-left-radius: 4px; + /* firefox specific markup */ + -moz-box-shadow: rgba(0, 0, 0, 0.15) 5px 5px 5px; + -moz-border-radius-topright: 4px; + -moz-border-radius-topleft: 4px; + /* webkit specific markup */ + -webkit-box-shadow: 5px 5px 5px rgba(0, 0, 0, 0.15); + -webkit-border-top-right-radius: 4px; + -webkit-border-top-left-radius: 4px; + +} + +.memdoc, dl.reflist dd { + border-bottom: 1px solid #A8B8D9; + border-left: 1px solid #A8B8D9; + border-right: 1px solid #A8B8D9; + padding: 6px 10px 2px 10px; + background-color: #FBFCFD; + border-top-width: 0; + background-image:url('nav_g.png'); + background-repeat:repeat-x; + background-color: #FFFFFF; + /* opera specific markup */ + border-bottom-left-radius: 4px; + border-bottom-right-radius: 4px; + box-shadow: 5px 5px 5px rgba(0, 0, 0, 0.15); + /* firefox specific markup */ + -moz-border-radius-bottomleft: 4px; + -moz-border-radius-bottomright: 4px; + -moz-box-shadow: rgba(0, 0, 0, 0.15) 5px 5px 5px; + /* webkit specific markup */ + -webkit-border-bottom-left-radius: 4px; + -webkit-border-bottom-right-radius: 4px; + -webkit-box-shadow: 5px 5px 5px rgba(0, 0, 0, 0.15); +} + +dl.reflist dt { + padding: 5px; +} + +dl.reflist dd { + margin: 0px 0px 10px 0px; + padding: 5px; +} + +.paramkey { + text-align: right; +} + +.paramtype { + white-space: nowrap; +} + +.paramname { + color: #602020; + white-space: nowrap; +} +.paramname em { + font-style: normal; +} +.paramname code { + line-height: 14px; +} + +.params, .retval, .exception, .tparams { + margin-left: 0px; + padding-left: 0px; +} + +.params .paramname, .retval .paramname { + font-weight: bold; + vertical-align: top; +} + +.params .paramtype { + font-style: italic; + vertical-align: top; +} + +.params .paramdir { + font-family: "courier new",courier,monospace; + vertical-align: top; +} + +table.mlabels { + border-spacing: 0px; +} + +td.mlabels-left { + width: 100%; + padding: 0px; +} + +td.mlabels-right { + vertical-align: bottom; + padding: 0px; + white-space: nowrap; +} + +span.mlabels { + margin-left: 8px; +} + +span.mlabel { + background-color: #728DC1; + border-top:1px solid #5373B4; + border-left:1px solid #5373B4; + border-right:1px solid #C4CFE5; + border-bottom:1px solid #C4CFE5; + text-shadow: none; + color: white; + margin-right: 4px; + padding: 2px 3px; + border-radius: 3px; + font-size: 7pt; + white-space: nowrap; + vertical-align: middle; +} + + + +/* @end */ + +/* these are for tree view when not used as main index */ + +div.directory { + margin: 10px 0px; + border-top: 1px solid #A8B8D9; + border-bottom: 1px solid #A8B8D9; + width: 100%; +} + +.directory table { + border-collapse:collapse; +} + +.directory td { + margin: 0px; + padding: 0px; + vertical-align: top; +} + +.directory td.entry { + white-space: nowrap; + padding-right: 6px; + padding-top: 3px; +} + +.directory td.entry a { + outline:none; +} + +.directory td.entry a img { + border: none; +} + +.directory td.desc { + width: 100%; + padding-left: 6px; + padding-right: 6px; + padding-top: 3px; + border-left: 1px solid rgba(0,0,0,0.05); +} + +.directory tr.even { + padding-left: 6px; + background-color: #F7F8FB; +} + +.directory img { + vertical-align: -30%; +} + +.directory .levels { + white-space: nowrap; + width: 100%; + text-align: right; + font-size: 9pt; +} + +.directory .levels span { + cursor: pointer; + padding-left: 2px; + padding-right: 2px; + color: #3D578C; +} + +div.dynheader { + margin-top: 8px; + -webkit-touch-callout: none; + -webkit-user-select: none; + -khtml-user-select: none; + -moz-user-select: none; + -ms-user-select: none; + user-select: none; +} + +address { + font-style: normal; + color: #2A3D61; +} + +table.doxtable { + border-collapse:collapse; + margin-top: 4px; + margin-bottom: 4px; +} + +table.doxtable td, table.doxtable th { + border: 1px solid #2D4068; + padding: 3px 7px 2px; +} + +table.doxtable th { + background-color: #374F7F; + color: #FFFFFF; + font-size: 110%; + padding-bottom: 4px; + padding-top: 5px; +} + +table.fieldtable { + /*width: 100%;*/ + margin-bottom: 10px; + border: 1px solid #A8B8D9; + border-spacing: 0px; + -moz-border-radius: 4px; + -webkit-border-radius: 4px; + border-radius: 4px; + -moz-box-shadow: rgba(0, 0, 0, 0.15) 2px 2px 2px; + -webkit-box-shadow: 2px 2px 2px rgba(0, 0, 0, 0.15); + box-shadow: 2px 2px 2px rgba(0, 0, 0, 0.15); +} + +.fieldtable td, .fieldtable th { + padding: 3px 7px 2px; +} + +.fieldtable td.fieldtype, .fieldtable td.fieldname { + white-space: nowrap; + border-right: 1px solid #A8B8D9; + border-bottom: 1px solid #A8B8D9; + vertical-align: top; +} + +.fieldtable td.fieldname { + padding-top: 3px; +} + +.fieldtable td.fielddoc { + border-bottom: 1px solid #A8B8D9; + /*width: 100%;*/ +} + +.fieldtable td.fielddoc p:first-child { + margin-top: 0px; +} + +.fieldtable td.fielddoc p:last-child { + margin-bottom: 2px; +} + +.fieldtable tr:last-child td { + border-bottom: none; +} + +.fieldtable th { + background-image:url('nav_f.png'); + background-repeat:repeat-x; + background-color: #E2E8F2; + font-size: 90%; + color: #253555; + padding-bottom: 4px; + padding-top: 5px; + text-align:left; + -moz-border-radius-topleft: 4px; + -moz-border-radius-topright: 4px; + -webkit-border-top-left-radius: 4px; + -webkit-border-top-right-radius: 4px; + border-top-left-radius: 4px; + border-top-right-radius: 4px; + border-bottom: 1px solid #A8B8D9; +} + + +.tabsearch { + top: 0px; + left: 10px; + height: 36px; + background-image: url('tab_b.png'); + z-index: 101; + overflow: hidden; + font-size: 13px; +} + +.navpath ul +{ + font-size: 11px; + background-image:url('tab_b.png'); + background-repeat:repeat-x; + background-position: 0 -5px; + height:30px; + line-height:30px; + color:#8AA0CC; + border:solid 1px #C2CDE4; + overflow:hidden; + margin:0px; + padding:0px; +} + +.navpath li +{ + list-style-type:none; + float:left; + padding-left:10px; + padding-right:15px; + background-image:url('bc_s.png'); + background-repeat:no-repeat; + background-position:right; + color:#364D7C; +} + +.navpath li.navelem a +{ + height:32px; + display:block; + text-decoration: none; + outline: none; + color: #283A5D; + font-family: 'Lucida Grande',Geneva,Helvetica,Arial,sans-serif; + text-shadow: 0px 1px 1px rgba(255, 255, 255, 0.9); + text-decoration: none; +} + +.navpath li.navelem a:hover +{ + color:#6884BD; +} + +.navpath li.footer +{ + list-style-type:none; + float:right; + padding-left:10px; + padding-right:15px; + background-image:none; + background-repeat:no-repeat; + background-position:right; + color:#364D7C; + font-size: 8pt; +} + + +div.summary +{ + float: right; + font-size: 8pt; + padding-right: 5px; + width: 50%; + text-align: right; +} + +div.summary a +{ + white-space: nowrap; +} + +div.ingroups +{ + font-size: 8pt; + width: 50%; + text-align: left; +} + +div.ingroups a +{ + white-space: nowrap; +} + +div.header +{ + background-image:url('nav_h.png'); + background-repeat:repeat-x; + background-color: #F9FAFC; + margin: 0px; + border-bottom: 1px solid #C4CFE5; +} + +div.headertitle +{ + padding: 5px 5px 5px 10px; +} + +dl +{ + padding: 0 0 0 10px; +} + +/* dl.note, dl.warning, dl.attention, dl.pre, dl.post, dl.invariant, dl.deprecated, dl.todo, dl.test, dl.bug */ +dl.section +{ + margin-left: 0px; + padding-left: 0px; +} + +dl.note +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #D0C000; +} + +dl.warning, dl.attention +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #FF0000; +} + +dl.pre, dl.post, dl.invariant +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #00D000; +} + +dl.deprecated +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #505050; +} + +dl.todo +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #00C0E0; +} + +dl.test +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #3030E0; +} + +dl.bug +{ + margin-left:-7px; + padding-left: 3px; + border-left:4px solid; + border-color: #C08050; +} + +dl.section dd { + margin-bottom: 6px; +} + + +#projectlogo +{ + text-align: center; + vertical-align: bottom; + border-collapse: separate; +} + +#projectlogo img +{ + border: 0px none; +} + +#projectname +{ + font: 300% Tahoma, Arial,sans-serif; + margin: 0px; + padding: 2px 0px; +} + +#projectbrief +{ + font: 120% Tahoma, Arial,sans-serif; + margin: 0px; + padding: 0px; +} + +#projectnumber +{ + font: 50% Tahoma, Arial,sans-serif; + margin: 0px; + padding: 0px; +} + +#titlearea +{ + padding: 0px; + margin: 0px; + width: 100%; + border-bottom: 1px solid #5373B4; +} + +.image +{ + text-align: center; +} + +.dotgraph +{ + text-align: center; +} + +.mscgraph +{ + text-align: center; +} + +.diagraph +{ + text-align: center; +} + +.caption +{ + font-weight: bold; +} + +div.zoom +{ + border: 1px solid #90A5CE; +} + +dl.citelist { + margin-bottom:50px; +} + +dl.citelist dt { + color:#334975; + float:left; + font-weight:bold; + margin-right:10px; + padding:5px; +} + +dl.citelist dd { + margin:2px 0; + padding:5px 0; +} + +div.toc { + padding: 14px 25px; + background-color: #F4F6FA; + border: 1px solid #D8DFEE; + border-radius: 7px 7px 7px 7px; + float: right; + height: auto; + margin: 0 20px 10px 10px; + width: 200px; +} + +div.toc li { + background: url("bdwn.png") no-repeat scroll 0 5px transparent; + font: 10px/1.2 Verdana,DejaVu Sans,Geneva,sans-serif; + margin-top: 5px; + padding-left: 10px; + padding-top: 2px; +} + +div.toc h3 { + font: bold 12px/1.2 Arial,FreeSans,sans-serif; + color: #4665A2; + border-bottom: 0 none; + margin: 0; +} + +div.toc ul { + list-style: none outside none; + border: medium none; + padding: 0px; +} + +div.toc li.level1 { + margin-left: 0px; +} + +div.toc li.level2 { + margin-left: 15px; +} + +div.toc li.level3 { + margin-left: 30px; +} + +div.toc li.level4 { + margin-left: 45px; +} + +.inherit_header { + font-weight: bold; + color: gray; + cursor: pointer; + -webkit-touch-callout: none; + -webkit-user-select: none; + -khtml-user-select: none; + -moz-user-select: none; + -ms-user-select: none; + user-select: none; +} + +.inherit_header td { + padding: 6px 0px 2px 5px; +} + +.inherit { + display: none; +} + +tr.heading h2 { + margin-top: 12px; + margin-bottom: 4px; +} + +/* tooltip related style info */ + +.ttc { + position: absolute; + display: none; +} + +#powerTip { + cursor: default; + white-space: nowrap; + background-color: white; + border: 1px solid gray; + border-radius: 4px 4px 4px 4px; + box-shadow: 1px 1px 7px gray; + display: none; + font-size: smaller; + max-width: 80%; + opacity: 0.9; + padding: 1ex 1em 1em; + position: absolute; + z-index: 2147483647; +} + +#powerTip div.ttdoc { + color: grey; + font-style: italic; +} + +#powerTip div.ttname a { + font-weight: bold; +} + +#powerTip div.ttname { + font-weight: bold; +} + +#powerTip div.ttdeci { + color: #006318; +} + +#powerTip div { + margin: 0px; + padding: 0px; + font: 12px/16px Roboto,sans-serif; +} + +#powerTip:before, #powerTip:after { + content: ""; + position: absolute; + margin: 0px; +} + +#powerTip.n:after, #powerTip.n:before, +#powerTip.s:after, #powerTip.s:before, +#powerTip.w:after, #powerTip.w:before, +#powerTip.e:after, #powerTip.e:before, +#powerTip.ne:after, #powerTip.ne:before, +#powerTip.se:after, #powerTip.se:before, +#powerTip.nw:after, #powerTip.nw:before, +#powerTip.sw:after, #powerTip.sw:before { + border: solid transparent; + content: " "; + height: 0; + width: 0; + position: absolute; +} + +#powerTip.n:after, #powerTip.s:after, +#powerTip.w:after, #powerTip.e:after, +#powerTip.nw:after, #powerTip.ne:after, +#powerTip.sw:after, #powerTip.se:after { + border-color: rgba(255, 255, 255, 0); +} + +#powerTip.n:before, #powerTip.s:before, +#powerTip.w:before, #powerTip.e:before, +#powerTip.nw:before, #powerTip.ne:before, +#powerTip.sw:before, #powerTip.se:before { + border-color: rgba(128, 128, 128, 0); +} + +#powerTip.n:after, #powerTip.n:before, +#powerTip.ne:after, #powerTip.ne:before, +#powerTip.nw:after, #powerTip.nw:before { + top: 100%; +} + +#powerTip.n:after, #powerTip.ne:after, #powerTip.nw:after { + border-top-color: #ffffff; + border-width: 10px; + margin: 0px -10px; +} +#powerTip.n:before { + border-top-color: #808080; + border-width: 11px; + margin: 0px -11px; +} +#powerTip.n:after, #powerTip.n:before { + left: 50%; +} + +#powerTip.nw:after, #powerTip.nw:before { + right: 14px; +} + +#powerTip.ne:after, #powerTip.ne:before { + left: 14px; +} + +#powerTip.s:after, #powerTip.s:before, +#powerTip.se:after, #powerTip.se:before, +#powerTip.sw:after, #powerTip.sw:before { + bottom: 100%; +} + +#powerTip.s:after, #powerTip.se:after, #powerTip.sw:after { + border-bottom-color: #ffffff; + border-width: 10px; + margin: 0px -10px; +} + +#powerTip.s:before, #powerTip.se:before, #powerTip.sw:before { + border-bottom-color: #808080; + border-width: 11px; + margin: 0px -11px; +} + +#powerTip.s:after, #powerTip.s:before { + left: 50%; +} + +#powerTip.sw:after, #powerTip.sw:before { + right: 14px; +} + +#powerTip.se:after, #powerTip.se:before { + left: 14px; +} + +#powerTip.e:after, #powerTip.e:before { + left: 100%; +} +#powerTip.e:after { + border-left-color: #ffffff; + border-width: 10px; + top: 50%; + margin-top: -10px; +} +#powerTip.e:before { + border-left-color: #808080; + border-width: 11px; + top: 50%; + margin-top: -11px; +} + +#powerTip.w:after, #powerTip.w:before { + right: 100%; +} +#powerTip.w:after { + border-right-color: #ffffff; + border-width: 10px; + top: 50%; + margin-top: -10px; +} +#powerTip.w:before { + border-right-color: #808080; + border-width: 11px; + top: 50%; + margin-top: -11px; +} + +@media print +{ + #top { display: none; } + #side-nav { display: none; } + #nav-path { display: none; } + body { overflow:visible; } + h1, h2, h3, h4, h5, h6 { page-break-after: avoid; } + .summary { display: none; } + .memitem { page-break-inside: avoid; } + #doc-content + { + margin-left:0 !important; + height:auto !important; + width:auto !important; + overflow:inherit; + display:inline; + } +} + diff --git a/docs/RuntimeAPI/html/doxygen.png b/docs/RuntimeAPI/html/doxygen.png new file mode 100644 index 0000000000000000000000000000000000000000..3ff17d807fd8aa003bed8bb2a69e8f0909592fd1 GIT binary patch literal 3779 zcmV;!4m|ORP)tMIv#Q0*~7*`IBSO7_x;@a8#Zk6_PeKR_s92J&)(m+);m9Iz3blw)z#Gi zP!9lj4$%+*>Hz@HCmM9L9|8c+0u=!H$O3?R0Kgx|#WP<6fKfC8fM-CQZT|_r@`>VO zX^Hgb|9cJqpdJA5$MCEK`F_2@2Y@s>^+;pF`~jdI0Pvr|vl4`=C)EH@1IFe7pdJ8F zH(qGi004~QnF)Ggga~8v08kGAs2hKTATxr7pwfNk|4#_AaT>w8P6TV+R2kbS$v==} zAjf`s0g#V8lB+b3)5oEI*q+{Yt$MZDruD2^;$+(_%Qn+%v0X-bJO=;@kiJ^ygLBnC z?1OVv_%aex1M@jKU|Z~$eI?PoF4Vj>fDzyo zAiLfpXY*a^Sj-S5D0S3@#V$sRW)g)_1e#$%8xdM>Jm7?!h zu0P2X=xoN>^!4DoPRgph2(2va07yfpXF+WH7EOg1GY%Zn z7~1A<(z7Q$ktEXhW_?GMpHp9l_UL18F3KOsxu81pqoBiNbFSGsof-W z6~eloMoz=4?OOnl2J268x5rOY`dCk0us(uS#Ud4yqOr@?=Q57a}tit|BhY>}~frH1sP`ScHS_d)oqH^lYy zZ%VP`#10MlE~P?cE(%(#(AUSv_T{+;t@$U}El}(1ig`vZo`Rm;+5&(AYzJ^Ae=h2X z@Re%vHwZU>|f0NI&%$*4eJweC5OROQrpPMA@*w|o z()A==l}(@bv^&>H1Ob3C=<^|hob?0+xJ?QQ3-ueQC}zy&JQNib!OqSO@-=>XzxlSF zAZ^U*1l6EEmg3r};_HY>&Jo_{dOPEFTWPmt=U&F#+0(O59^UIlHbNX+eF8UzyDR*T z(=5X$VF3!gm@RooS-&iiUYGG^`hMR(07zr_xP`d!^BH?uD>Phl8Rdifx3Af^Zr`Ku ztL+~HkVeL#bJ)7;`=>;{KNRvjmc}1}c58Sr#Treq=4{xo!ATy|c>iRSp4`dzMMVd@ zL8?uwXDY}Wqgh4mH`|$BTXpUIu6A1-cSq%hJw;@^Zr8TP=GMh*p(m(tN7@!^D~sl$ zz^tf4II4|};+irE$Fnm4NTc5%p{PRA`%}Zk`CE5?#h3|xcyQsS#iONZ z6H(@^i9td!$z~bZiJLTax$o>r(p}3o@< zyD7%(>ZYvy=6$U3e!F{Z`uSaYy`xQyl?b{}eg|G3&fz*`QH@mDUn)1%#5u`0m$%D} z?;tZ0u(mWeMV0QtzjgN!lT*pNRj;6510Wwx?Yi_=tYw|J#7@(Xe7ifDzXuK;JB;QO z#bg~K$cgm$@{QiL_3yr}y&~wuv=P=#O&Tj=Sr)aCUlYmZMcw?)T?c%0rUe1cS+o!qs_ zQ6Gp)-{)V!;=q}llyK3|^WeLKyjf%y;xHku;9(vM!j|~<7w1c*Mk-;P{T&yG) z@C-8E?QPynNQ<8f01D`2qexcVEIOU?y}MG)TAE6&VT5`rK8s(4PE;uQ92LTXUQ<>^ ztyQ@=@kRdh@ebUG^Z6NWWIL;_IGJ2ST>$t!$m$qvtj0Qmw8moN6GUV^!QKNK zHBXCtUH8)RY9++gH_TUV4^=-j$t}dD3qsN7GclJ^Zc&(j6&a_!$jCf}%c5ey`pm~1)@{yI3 zTdWyB+*X{JFw#z;PwRr5evb2!ueWF;v`B0HoUu4-(~aL=z;OXUUEtG`_$)Oxw6FKg zEzY`CyKaSBK3xt#8gA|r_|Kehn_HYVBMpEwbn9-fI*!u*eTA1ef8Mkl1=!jV4oYwWYM}i`A>_F4nhmlCIC6WLa zY%;4&@AlnaG11ejl61Jev21|r*m+?Kru3;1tFDl}#!OzUp6c>go4{C|^erwpG*&h6bspUPJag}oOkN2912Y3I?(eRc@U9>z#HPBHC?nps7H5!zP``90!Q1n80jo+B3TWXp!8Pe zwuKuLLI6l3Gv@+QH*Y}2wPLPQ1^EZhT#+Ed8q8Wo z1pTmIBxv14-{l&QVKxAyQF#8Q@NeJwWdKk>?cpiJLkJr+aZ!Me+Cfp!?FWSRf^j2k z73BRR{WSKaMkJ>1Nbx5dan5hg^_}O{Tj6u%iV%#QGz0Q@j{R^Ik)Z*+(YvY2ziBG)?AmJa|JV%4UT$k`hcOg5r9R?5>?o~JzK zJCrj&{i#hG>N7!B4kNX(%igb%kDj0fOQThC-8mtfap82PNRXr1D>lbgg)dYTQ(kbx z`Ee5kXG~Bh+BHQBf|kJEy6(ga%WfhvdQNDuOfQoe377l#ht&DrMGeIsI5C<&ai zWG$|hop2@@q5YDa)_-A?B02W;#fH!%k`daQLEItaJJ8Yf1L%8x;kg?)k)00P-lH+w z)5$QNV6r2$YtnV(4o=0^3{kmaXn*Dm0F*fU(@o)yVVjk|ln8ea6BMy%vZAhW9|wvA z8RoDkVoMEz1d>|5(k0Nw>22ZT){V<3$^C-cN+|~hKt2)){+l-?3m@-$c?-dlzQ)q- zZ)j%n^gerV{|+t}9m1_&&Ly!9$rtG4XX|WQ8`xYzGC~U@nYh~g(z9)bdAl#xH)xd5a=@|qql z|FzEil{P5(@gy!4ek05i$>`E^G~{;pnf6ftpLh$h#W?^#4UkPfa;;?bsIe&kz!+40 zI|6`F2n020)-r`pFaZ38F!S-lJM-o&inOw|66=GMeP@xQU5ghQH{~5Uh~TMTd;I9` z>YhVB`e^EVj*S7JF39ZgNf}A-0DwOcTT63ydN$I3b?yBQtUI*_fae~kPvzoD$zjX3 zoqBe#>12im4WzZ=f^4+u=!lA|#r%1`WB0-6*3BL#at`47#ebPpR|D1b)3BjT34nYY z%Ds%d?5$|{LgOIaRO{{oC&RK`O91$fqwM0(C_TALcozu*fWHb%%q&p-q{_8*2Zsi^ zh1ZCnr^UYa;4vQEtHk{~zi>wwMC5o{S=$P0X681y`SXwFH?Ewn{x-MOZynmc)JT5v zuHLwh;tLfxRrr%|k370}GofLl7thg>ACWWY&msqaVu&ry+`7+Ss>NL^%T1|z{IGMA zW-SKl=V-^{(f!Kf^#3(|T2W47d(%JVCI4JgRrT1pNz>+ietmFToNv^`gzC@&O-)+i zPQ~RwK8%C_vf%;%e>NyTp~dM5;!C|N0Q^6|CEb7Bw=Vz~$1#FA;Z*?mKSC)Hl-20s t8QyHj(g6VK0RYbl8UjE)0O0w=e*@m04r>stuEhWV002ovPDHLkV1hl;dM*F} literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/dynsections.js b/docs/RuntimeAPI/html/dynsections.js new file mode 100644 index 0000000000..ed092c7f63 --- /dev/null +++ b/docs/RuntimeAPI/html/dynsections.js @@ -0,0 +1,97 @@ +function toggleVisibility(linkObj) +{ + var base = $(linkObj).attr('id'); + var summary = $('#'+base+'-summary'); + var content = $('#'+base+'-content'); + var trigger = $('#'+base+'-trigger'); + var src=$(trigger).attr('src'); + if (content.is(':visible')===true) { + content.hide(); + summary.show(); + $(linkObj).addClass('closed').removeClass('opened'); + $(trigger).attr('src',src.substring(0,src.length-8)+'closed.png'); + } else { + content.show(); + summary.hide(); + $(linkObj).removeClass('closed').addClass('opened'); + $(trigger).attr('src',src.substring(0,src.length-10)+'open.png'); + } + return false; +} + +function updateStripes() +{ + $('table.directory tr'). + removeClass('even').filter(':visible:even').addClass('even'); +} +function toggleLevel(level) +{ + $('table.directory tr').each(function(){ + var l = this.id.split('_').length-1; + var i = $('#img'+this.id.substring(3)); + var a = $('#arr'+this.id.substring(3)); + if (l + + + + + +HIP: Heterogenous-computing Interface for Portability: File List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + +
+ + + + +
+ +
+ +
+
+
File List
+
+
+
Here is a list of all documented files with brief descriptions:
+
[detail level 1234]
+ + + + + + + + + + + + + + + + + + + + + + + + + + + +
o-include
|\-hip
| o-hcc_detail
| |o*hcc_acc.h
| |o*hip_complex.h
| |o*hip_fp16.h
| |o*hip_ldg.h
| |o*hip_runtime.hContains definitions of APIs for HIP runtime
| |o*hip_runtime_api.hContains C function APIs for HIP runtime. This file does not use any HCC builtin or special language extensions (-hc mode) ; those functions in hip_runtime.h
| |o*hip_texture.hHIP C++ Texture API for hcc compiler
| |o*hip_vector_types.hDefines the different newt vector types for HIP runtime
| |\*host_defines.hTODO-doc
| o*hcc.h
| o*hip_common.h
| o*hip_complex.h
| o*hip_fp16.h
| o*hip_profile.h
| o*hip_runtime.h
| o*hip_runtime_api.h
| o*hip_texture.h
| \*hip_vector_types.h
\-src
 o*device_util.h
 o*hip_hcc.cpp
 o*hip_hcc.h
 o*hip_util.h
 \*trace_helper.h
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/ftv2blank.png b/docs/RuntimeAPI/html/ftv2blank.png new file mode 100644 index 0000000000000000000000000000000000000000..63c605bb4c3d941c921a4b6cfa74951e946bcb48 GIT binary patch literal 86 zcmeAS@N?(olHy`uVBq!ia0vp^0zfRr!3HExu9B$%QnH>djv*C{Z|`mdau^P8_z}#X h?B8GEpdi4(BFDx$je&7RrDQEg&ePS;Wt~$(69Dh@6T1Ka literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2cl.png b/docs/RuntimeAPI/html/ftv2cl.png new file mode 100644 index 0000000000000000000000000000000000000000..132f6577bf7f085344904602815a260d29f55d9b GIT binary patch literal 453 zcmV;$0XqJPP)VBF;ev;toEj8_OB0EQg5eYilIj#JZG_m^33l3^k4mtzx!TVD?g)Y$ zrvwRDSqT!wLIM$dWCIa$vtxE|mzbTzu-y&$FvF6WA2a{Wr1g}`WdPT-0JzEZ0IxAv z-Z+ejZc&H;I5-pb_SUB}04j0^V)3t{`z<7asDl2Tw3w3sP%)0^8$bhEg)IOTBcRXv zFfq~3&gvJ$F-U7mpBW8z1GY~HK&7h4^YI~Orv~wLnC0PP_dAkv;nzX{9Q|8Gv=2ca z@v)c9T;D#h`TZ2X&&$ff2wedmot995de~-s3I)yauahg;7qn*?1n?F$e+PwP37}~; z1NKUk7reVK^7A;$QRW7qAx40HHUZ<|k3U%nz(Ec`#i+q9K!dgcROAlCS?`L= v>#=f?wF5ZND!1uAfQsk;KN^4&*8~0npJiJ%2dj9(00000NkvXXu0mjfWVFf_ literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2doc.png b/docs/RuntimeAPI/html/ftv2doc.png new file mode 100644 index 0000000000000000000000000000000000000000..17edabff95f7b8da13c9516a04efe05493c29501 GIT binary patch literal 746 zcmV7=@pnbNXRFEm&G8P!&WHG=d)>K?YZ1bzou)2{$)) zumDct!>4SyxL;zgaG>wy`^Hv*+}0kUfCrz~BCOViSb$_*&;{TGGn2^x9K*!Sf0=lV zpP=7O;GA0*Jm*tTYj$IoXvimpnV4S1Z5f$p*f$Db2iq2zrVGQUz~yq`ahn7ck(|CE z7Gz;%OP~J6)tEZWDzjhL9h2hdfoU2)Nd%T<5Kt;Y0XLt&<@6pQx!nw*5`@bq#?l*?3z{Hlzoc=Pr>oB5(9i6~_&-}A(4{Q$>c>%rV&E|a(r&;?i5cQB=} zYSDU5nXG)NS4HEs0it2AHe2>shCyr7`6@4*6{r@8fXRbTA?=IFVWAQJL&H5H{)DpM#{W(GL+Idzf^)uRV@oB8u$ z8v{MfJbTiiRg4bza<41NAzrl{=3fl_D+$t+^!xlQ8S}{UtY`e z;;&9UhyZqQRN%2pot{*Ei0*4~hSF_3AH2@fKU!$NSflS>{@tZpDT4`M2WRTTVH+D? z)GFlEGGHe?koB}i|1w45!BF}N_q&^HJ&-tyR{(afC6H7|aml|tBBbv}55C5DNP8p3 z)~jLEO4Z&2hZmP^i-e%(@d!(E|KRafiU8Q5u(wU((j8un3OR*Hvj+t literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2folderclosed.png b/docs/RuntimeAPI/html/ftv2folderclosed.png new file mode 100644 index 0000000000000000000000000000000000000000..bb8ab35edce8e97554e360005ee9fc5bffb36e66 GIT binary patch literal 616 zcmV-u0+;=XP)a9#ETzayK)T~Jw&MMH>OIr#&;dC}is*2Mqdf&akCc=O@`qC+4i z5Iu3w#1M@KqXCz8TIZd1wli&kkl2HVcAiZ8PUn5z_kG@-y;?yK06=cA0U%H0PH+kU zl6dp}OR(|r8-RG+YLu`zbI}5TlOU6ToR41{9=uz^?dGTNL;wIMf|V3`d1Wj3y!#6` zBLZ?xpKR~^2x}?~zA(_NUu3IaDB$tKma*XUdOZN~c=dLt_h_k!dbxm_*ibDM zlFX`g{k$X}yIe%$N)cn1LNu=q9_CS)*>A zsX_mM4L@`(cSNQKMFc$RtYbx{79#j-J7hk*>*+ZZhM4Hw?I?rsXCi#mRWJ=-0LGV5a-WR0Qgt<|Nqf)C-@80`5gIz45^_20000IqP)X=#(TiCT&PiIIVc55T}TU}EUh*{q$|`3@{d>{Tc9Bo>e= zfmF3!f>fbI9#GoEHh0f`i5)wkLpva0ztf%HpZneK?w-7AK@b4Itw{y|Zd3k!fH?q2 zlhckHd_V2M_X7+)U&_Xcfvtw60l;--DgZmLSw-Y?S>)zIqMyJ1#FwLU*%bl38ok+! zh78H87n`ZTS;uhzAR$M`zZ`bVhq=+%u9^$5jDplgxd44}9;IRqUH1YHH|@6oFe%z( zo4)_>E$F&^P-f(#)>(TrnbE>Pefs9~@iN=|)Rz|V`sGfHNrJ)0gJb8xx+SBmRf@1l zvuzt=vGfI)<-F9!o&3l?>9~0QbUDT(wFdnQPv%xdD)m*g%!20>Bc9iYmGAp<9YAa( z0QgYgTWqf1qN++Gqp z8@AYPTB3E|6s=WLG?xw0tm|U!o=&zd+H0oRYE;Dbx+Na9s^STqX|Gnq%H8s(nGDGJ j8vwW|`Ts`)fSK|Kx=IK@RG@g200000NkvXXu0mjfauFEA literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2lastnode.png b/docs/RuntimeAPI/html/ftv2lastnode.png new file mode 100644 index 0000000000000000000000000000000000000000..63c605bb4c3d941c921a4b6cfa74951e946bcb48 GIT binary patch literal 86 zcmeAS@N?(olHy`uVBq!ia0vp^0zfRr!3HExu9B$%QnH>djv*C{Z|`mdau^P8_z}#X h?B8GEpdi4(BFDx$je&7RrDQEg&ePS;Wt~$(69Dh@6T1Ka literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2link.png b/docs/RuntimeAPI/html/ftv2link.png new file mode 100644 index 0000000000000000000000000000000000000000..17edabff95f7b8da13c9516a04efe05493c29501 GIT binary patch literal 746 zcmV7=@pnbNXRFEm&G8P!&WHG=d)>K?YZ1bzou)2{$)) zumDct!>4SyxL;zgaG>wy`^Hv*+}0kUfCrz~BCOViSb$_*&;{TGGn2^x9K*!Sf0=lV zpP=7O;GA0*Jm*tTYj$IoXvimpnV4S1Z5f$p*f$Db2iq2zrVGQUz~yq`ahn7ck(|CE z7Gz;%OP~J6)tEZWDzjhL9h2hdfoU2)Nd%T<5Kt;Y0XLt&<@6pQx!nw*5`@bq#?l*?3z{Hlzoc=Pr>oB5(9i6~_&-}A(4{Q$>c>%rV&E|a(r&;?i5cQB=} zYSDU5nXG)NS4HEs0it2AHe2>shCyr7`6@4*6{r@8fXRbTA?=IFVWAQJL&H5H{)DpM#{W(GL+Idzf^)uRV@oB8u$ z8v{MfJbTiiRg4bza<41NAzrl{=3fl_D+$t+^!xlQ8S}{UtY`e z;;&9UhyZqQRN%2pot{*Ei0*4~hSF_3AH2@fKU!$NSflS>{@tZpDT4`M2WRTTVH+D? z)GFlEGGHe?koB}i|1w45!BF}N_q&^HJ&-tyR{(afC6H7|aml|tBBbv}55C5DNP8p3 z)~jLEO4Z&2hZmP^i-e%(@d!(E|KRafiU8Q5u(wU((j8un3OR*Hvj+t literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2mlastnode.png b/docs/RuntimeAPI/html/ftv2mlastnode.png new file mode 100644 index 0000000000000000000000000000000000000000..0b63f6d38c4b9ec907b820192ebe9724ed6eca22 GIT binary patch literal 246 zcmVkw!R34#Lv2LOS^S2tZA31X++9RY}n zChwn@Z)Wz*WWHH{)HDtJnq&A2hk$b-y(>?@z0iHr41EKCGp#T5?07*qoM6N<$f(V3Pvj6}9 literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2mnode.png b/docs/RuntimeAPI/html/ftv2mnode.png new file mode 100644 index 0000000000000000000000000000000000000000..0b63f6d38c4b9ec907b820192ebe9724ed6eca22 GIT binary patch literal 246 zcmVkw!R34#Lv2LOS^S2tZA31X++9RY}n zChwn@Z)Wz*WWHH{)HDtJnq&A2hk$b-y(>?@z0iHr41EKCGp#T5?07*qoM6N<$f(V3Pvj6}9 literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2mo.png b/docs/RuntimeAPI/html/ftv2mo.png new file mode 100644 index 0000000000000000000000000000000000000000..4bfb80f76e65815989a9350ad79d8ce45380e2b1 GIT binary patch literal 403 zcmV;E0c`$>P)${!fXv7NWJ%@%u4(KapRY>T6_x;E zxE7kt!}Tiw8@d9Sd`rTGum>z#Q14vIm`wm1#-byD1muMi02@YNO5LRF0o!Y{`a!Ya z{^&p0Su|s705&2QxmqdexG+-zNKL3f@8gTQSJrKByfo+oNJ^-{|Mn||Q5SDwjQVsS zr1}7o5-QMs>gYIMD>GRw@$lT`z4r-_m{5U#cR{urD_)TOeY)(UD|qZ^&y`IVijqk~ xs(9-kWFr7E^!lgi8GsFK5kOY_{Xbgf0^etEU%fLevs?fG002ovPDHLkV1nB&vX1}& literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2node.png b/docs/RuntimeAPI/html/ftv2node.png new file mode 100644 index 0000000000000000000000000000000000000000..63c605bb4c3d941c921a4b6cfa74951e946bcb48 GIT binary patch literal 86 zcmeAS@N?(olHy`uVBq!ia0vp^0zfRr!3HExu9B$%QnH>djv*C{Z|`mdau^P8_z}#X h?B8GEpdi4(BFDx$je&7RrDQEg&ePS;Wt~$(69Dh@6T1Ka literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2ns.png b/docs/RuntimeAPI/html/ftv2ns.png new file mode 100644 index 0000000000000000000000000000000000000000..72e3d71c2892d6f00e259facebc88b45f6db2e35 GIT binary patch literal 388 zcmV-~0ek+5P)f+++#cT|!CkD&4pnIkeMEUEM*>`*9>+Juji$!h-mW%M^8s9957{3nvbrz^&=u<~TAUrFROkmt%^F~Ez+-c53Lv%iH3d38!Rv?K zrb&MYAhp;Gf<}wS;9ZZq2@;!uYG;=Z>~GKE^{HD4keu}lnyqhc>kWX^tQn|warJ~h zT+rtMkdz6aHoN%z(o|&wpu@@OpJnF_z{PA)6(FHw02iHslz^(N{4*+K9)QJHR87wT iTyp>aXaF{u2lxRou|^4tux6eB0000^P)R?RzRoKvklcaQ%HF6%rK2&ZgO(-ihJ_C zzrKgp4jgO( fd_(yg|3PpEQb#9`a?Pz_00000NkvXXu0mjftR`5K literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2pnode.png b/docs/RuntimeAPI/html/ftv2pnode.png new file mode 100644 index 0000000000000000000000000000000000000000..c6ee22f937a07d1dbfc27c669d11f8ed13e2f152 GIT binary patch literal 229 zcmV^P)R?RzRoKvklcaQ%HF6%rK2&ZgO(-ihJ_C zzrKgp4jgO( fd_(yg|3PpEQb#9`a?Pz_00000NkvXXu0mjftR`5K literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2splitbar.png b/docs/RuntimeAPI/html/ftv2splitbar.png new file mode 100644 index 0000000000000000000000000000000000000000..fe895f2c58179b471a22d8320b39a4bd7312ec8e GIT binary patch literal 314 zcmeAS@N?(olHy`uVBq!ia0vp^Yzz!63>-{AmhX=Jf(#6djGiuzAr*{o?=JLmPLyc> z_*`QK&+BH@jWrYJ7>r6%keRM@)Qyv8R=enp0jiI>aWlGyB58O zFVR20d+y`K7vDw(hJF3;>dD*3-?v=<8M)@x|EEGLnJsniYK!2U1 Y!`|5biEc?d1`HDhPgg&ebxsLQ02F6;9RL6T literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/ftv2vertline.png b/docs/RuntimeAPI/html/ftv2vertline.png new file mode 100644 index 0000000000000000000000000000000000000000..63c605bb4c3d941c921a4b6cfa74951e946bcb48 GIT binary patch literal 86 zcmeAS@N?(olHy`uVBq!ia0vp^0zfRr!3HExu9B$%QnH>djv*C{Z|`mdau^P8_z}#X h?B8GEpdi4(BFDx$je&7RrDQEg&ePS;Wt~$(69Dh@6T1Ka literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/functions.html b/docs/RuntimeAPI/html/functions.html new file mode 100644 index 0000000000..ca8882856d --- /dev/null +++ b/docs/RuntimeAPI/html/functions.html @@ -0,0 +1,332 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Class Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + +
+ + + + +
+ +
+ +
+
Here is a list of all documented class members with links to the class documentation for each member:
+ +

- _ -

+ + +

- a -

+ + +

- c -

+ + +

- h -

+ + +

- i -

+ + +

- l -

+ + +

- m -

+ + +

- n -

+ + +

- p -

+ + +

- r -

+ + +

- s -

+ + +

- t -

+ + +

- w -

+ + +

- x -

+ + +

- y -

+ + +

- z -

+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/functions_rela.html b/docs/RuntimeAPI/html/functions_rela.html new file mode 100644 index 0000000000..9ddcfb49b4 --- /dev/null +++ b/docs/RuntimeAPI/html/functions_rela.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Class Members - Related Functions + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + +
+ + + + +
+ +
+ +
+  +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/functions_vars.html b/docs/RuntimeAPI/html/functions_vars.html new file mode 100644 index 0000000000..4569adf399 --- /dev/null +++ b/docs/RuntimeAPI/html/functions_vars.html @@ -0,0 +1,329 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Class Members - Variables + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + +
+ + + + +
+ +
+ +
+  + +

- _ -

+ + +

- a -

+ + +

- c -

+ + +

- h -

+ + +

- i -

+ + +

- l -

+ + +

- m -

+ + +

- n -

+ + +

- p -

+ + +

- r -

+ + +

- s -

+ + +

- t -

+ + +

- w -

+ + +

- x -

+ + +

- y -

+ + +

- z -

+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/globals.html b/docs/RuntimeAPI/html/globals.html new file mode 100644 index 0000000000..1bdcacb7ee --- /dev/null +++ b/docs/RuntimeAPI/html/globals.html @@ -0,0 +1,567 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: File Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + +
+ + + + +
+ +
+ +
+
Here is a list of all documented file members with links to the documentation:
+ +

- _ -

+ + +

- d -

+ + +

- h -

+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/globals_defs.html b/docs/RuntimeAPI/html/globals_defs.html new file mode 100644 index 0000000000..6f7178061f --- /dev/null +++ b/docs/RuntimeAPI/html/globals_defs.html @@ -0,0 +1,152 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: File Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + +
+ + + + +
+ +
+ +
+  +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/globals_enum.html b/docs/RuntimeAPI/html/globals_enum.html new file mode 100644 index 0000000000..f11ce794d3 --- /dev/null +++ b/docs/RuntimeAPI/html/globals_enum.html @@ -0,0 +1,119 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: File Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + +
+ + + + +
+ +
+ +
+  +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/globals_eval.html b/docs/RuntimeAPI/html/globals_eval.html new file mode 100644 index 0000000000..c2ff74f32f --- /dev/null +++ b/docs/RuntimeAPI/html/globals_eval.html @@ -0,0 +1,146 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: File Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + +
+ + + + +
+ +
+ +
+  +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/globals_func.html b/docs/RuntimeAPI/html/globals_func.html new file mode 100644 index 0000000000..f21de520bb --- /dev/null +++ b/docs/RuntimeAPI/html/globals_func.html @@ -0,0 +1,451 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: File Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + +
+ + + + +
+ +
+ +
+  + +

- _ -

+ + +

- h -

+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/globals_type.html b/docs/RuntimeAPI/html/globals_type.html new file mode 100644 index 0000000000..99b173c22b --- /dev/null +++ b/docs/RuntimeAPI/html/globals_type.html @@ -0,0 +1,119 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: File Members + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + +
+ + + + +
+ +
+ +
+  +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__API.html b/docs/RuntimeAPI/html/group__API.html new file mode 100644 index 0000000000..8a7c5d7c71 --- /dev/null +++ b/docs/RuntimeAPI/html/group__API.html @@ -0,0 +1,245 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: HIP API + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Modules | +Functions
+
+
HIP API
+
+
+ + + + + + + + + + + + + + + + + + + + + + +

+Modules

 Device Management
 
 Error Handling
 
 Stream Management
 
 Event Management
 
 Memory Management
 
 Device Memory Access
 
 Initialization and Version
 
 Control
 
 HCC-Specific Accessors
 
 Texture Reference Management
 
+ + + + + + + + + + +

+Functions

hipError_t hipIpcGetMemHandle (hipIpcMemHandle_t *handle, void *devPtr)
 Gets an interprocess memory handle for an existing device memory allocation. More...
 
hipError_t hipIpcOpenMemHandle (void **devPtr, hipIpcMemHandle_t handle, unsigned int flags)
 Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process. More...
 
hipError_t hipIpcCloseMemHandle (void *devPtr)
 Close memory mapped with hipIpcOpenMemHandle. More...
 
+

Detailed Description

+

Defines the HIP API. See the individual sections for more information.

+

Function Documentation

+ +
+
+ + + + + + + + +
hipError_t hipIpcCloseMemHandle (void * devPtr)
+
+ +

Close memory mapped with hipIpcOpenMemHandle.

+

Unmaps memory returnd by hipIpcOpenMemHandle. The original allocation in the exporting process as well as imported mappings in other processes will be unaffected.

+

Any resources used to enable peer access will be freed if this is the last mapping using them.

+
Parameters
+ + +
devPtr- Device pointer returned by hipIpcOpenMemHandle
+
+
+
Returns
hipSuccess, hipErrorMapBufferObjectFailed, hipErrorInvalidResourceHandle,
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipIpcGetMemHandle (hipIpcMemHandle_t * handle,
void * devPtr 
)
+
+ +

Gets an interprocess memory handle for an existing device memory allocation.

+

Takes a pointer to the base of an existing device memory allocation created with hipMalloc and exports it for use in another process. This is a lightweight operation and may be called multiple times on an allocation without adverse effects.

+

If a region of memory is freed with hipFree and a subsequent call to hipMalloc returns memory with the same device address, hipIpcGetMemHandle will return a unique handle for the new memory.

+
Parameters
+ + + +
handle- Pointer to user allocated hipIpcMemHandle to return the handle in.
devPtr- Base pointer to previously allocated device memory
+
+
+
Returns
hipSuccess, hipErrorInvalidResourceHandle, hipErrorMemoryAllocation, hipErrorMapBufferObjectFailed,
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipIpcOpenMemHandle (void ** devPtr,
hipIpcMemHandle_t handle,
unsigned int flags 
)
+
+ +

Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process.

+

Maps memory exported from another process with hipIpcGetMemHandle into the current device address space. For contexts on different devices hipIpcOpenMemHandle can attempt to enable peer access between the devices as if the user called hipDeviceEnablePeerAccess. This behavior is controlled by the hipIpcMemLazyEnablePeerAccess flag. hipDeviceCanAccessPeer can determine if a mapping is possible.

+

Contexts that may open hipIpcMemHandles are restricted in the following way. hipIpcMemHandles from each device in a given process may only be opened by one context per device per other process.

+

Memory returned from hipIpcOpenMemHandle must be freed with hipIpcCloseMemHandle.

+

Calling hipFree on an exported memory region before calling hipIpcCloseMemHandle in the importing context will result in undefined behavior.

+
Parameters
+ + + + +
devPtr- Returned device pointer
handle- hipIpcMemHandle to open
flags- Flags for this operation. Must be specified as hipIpcMemLazyEnablePeerAccess
+
+
+
Returns
hipSuccess, hipErrorMapBufferObjectFailed, hipErrorInvalidResourceHandle, hipErrorTooManyPeers
+
Note
No guarantees are made about the address returned in *devPtr. In particular, multiple processes may not receive the same address for the same handle.
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Context.html b/docs/RuntimeAPI/html/group__Context.html new file mode 100644 index 0000000000..7699238778 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Context.html @@ -0,0 +1,607 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Management + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Management
Initialization and Version
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipCtxCreate (hipCtx_t *ctx, unsigned int flags, hipDevice_t device)
 Create a context and set it as current/ default context. More...
 
hipError_t hipCtxDestroy (hipCtx_t ctx)
 Destroy a HIP context. More...
 
hipError_t hipCtxPopCurrent (hipCtx_t *ctx)
 Pop the current/default context and return the popped context. More...
 
hipError_t hipCtxPushCurrent (hipCtx_t ctx)
 Push the context to be set as current/ default context. More...
 
hipError_t hipCtxSetCurrent (hipCtx_t ctx)
 Set the passed context as current/default. More...
 
hipError_t hipCtxGetCurrent (hipCtx_t *ctx)
 Get the handle of the current/ default context. More...
 
hipError_t hipCtxGetDevice (hipDevice_t *device)
 Get the handle of the device associated with current/default context. More...
 
hipError_t hipCtxGetApiVersion (hipCtx_t ctx, int *apiVersion)
 Returns the approximate HIP api version. More...
 
hipError_t hipCtxGetCacheConfig (hipFuncCache_t *cacheConfig)
 Set Cache configuration for a specific function. More...
 
hipError_t hipCtxSetCacheConfig (hipFuncCache_t cacheConfig)
 Set L1/Shared cache partition. More...
 
hipError_t hipCtxSetSharedMemConfig (hipSharedMemConfig config)
 Set Shared memory bank configuration. More...
 
hipError_t hipCtxGetSharedMemConfig (hipSharedMemConfig *pConfig)
 Get Shared memory bank configuration. More...
 
hipError_t hipCtxSynchronize (void)
 Blocks until the default context has completed all preceding requested tasks. More...
 
hipError_t hipCtxGetFlags (unsigned int *flags)
 Return flags used for creating default context. More...
 
hipError_t hipCtxEnablePeerAccess (hipCtx_t peerCtx, unsigned int flags)
 Enables direct access to memory allocations in a peer context. More...
 
hipError_t hipCtxDisablePeerAccess (hipCtx_t peerCtx)
 Disable direct access from current context's virtual address space to memory allocations physically located on a peer context.Disables direct access to memory allocations in a peer context and unregisters any registered allocations. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

Function Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipCtxCreate (hipCtx_t * ctx,
unsigned int flags,
hipDevice_t device 
)
+
+ +

Create a context and set it as current/ default context.

+
Parameters
+ + + + +
[out]ctx
[in]flags
[in]associateddevice handle
+
+
+
Returns
hipSuccess
+
See Also
hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxDestroy (hipCtx_t ctx)
+
+ +

Destroy a HIP context.

+
Parameters
+ + +
[in]ctxContext to destroy
+
+
+
Returns
hipSuccess, hipErrorInvalidValue
+
See Also
hipCtxCreate, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent,hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize , hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxDisablePeerAccess (hipCtx_t peerCtx)
+
+ +

Disable direct access from current context's virtual address space to memory allocations physically located on a peer context.Disables direct access to memory allocations in a peer context and unregisters any registered allocations.

+

Returns hipErrorPeerAccessNotEnabled if direct access to memory on peerDevice has not yet been enabled from the current device.

+
Parameters
+ + +
[in]peerCtx
+
+
+
Returns
hipSuccess, hipErrorPeerAccessNotEnabled
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+
Warning
PeerToPeer support is experimental.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipCtxEnablePeerAccess (hipCtx_t peerCtx,
unsigned int flags 
)
+
+ +

Enables direct access to memory allocations in a peer context.

+

Memory which already allocated on peer device will be mapped into the address space of the current device. In addition, all future memory allocations on peerDeviceId will be mapped into the address space of the current device when the memory is allocated. The peer memory remains accessible from the current device until a call to hipDeviceDisablePeerAccess or hipDeviceReset.

+
Parameters
+ + + +
[in]peerCtx
[in]flags
+
+
+
Returns
hipSuccess, hipErrorInvalidDevice, hipErrorInvalidValue, hipErrorPeerAccessAlreadyEnabled
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+
Warning
PeerToPeer support is experimental.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipCtxGetApiVersion (hipCtx_t ctx,
int * apiVersion 
)
+
+ +

Returns the approximate HIP api version.

+
Parameters
+ + + +
[in]ctxContext to check
[out]apiVersion
+
+
+
Returns
hipSuccess
+
Warning
The HIP feature set does not correspond to an exact CUDA SDK api revision. This function always set *apiVersion to 4 as an approximation though HIP supports some features which were introduced in later CUDA SDK revisions. HIP apps code should not rely on the api revision number here and should use arch feature flags to test device capabilities or conditional compilation.
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetDevice, hipCtxGetFlags, hipCtxPopCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxGetCacheConfig (hipFuncCache_t * cacheConfig)
+
+ +

Set Cache configuration for a specific function.

+
Parameters
+ + +
[out]cacheConfiguration
+
+
+
Returns
hipSuccess
+
Warning
AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures.
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxGetCurrent (hipCtx_t * ctx)
+
+ +

Get the handle of the current/ default context.

+
Parameters
+ + +
[out]ctx
+
+
+
Returns
hipSuccess, hipErrorInvalidContext
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetDevice, hipCtxGetFlags, hipCtxPopCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxGetDevice (hipDevice_t * device)
+
+ +

Get the handle of the device associated with current/default context.

+
Parameters
+ + +
[out]device
+
+
+
Returns
hipSuccess, hipErrorInvalidContext
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxGetFlags (unsigned int * flags)
+
+ +

Return flags used for creating default context.

+
Parameters
+ + +
[out]flags
+
+
+
Returns
hipSuccess
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxGetSharedMemConfig (hipSharedMemConfig * pConfig)
+
+ +

Get Shared memory bank configuration.

+
Parameters
+ + +
[out]sharedMemoryConfiguration
+
+
+
Returns
hipSuccess
+
Warning
AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures.
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxPopCurrent (hipCtx_t * ctx)
+
+ +

Pop the current/default context and return the popped context.

+
Parameters
+ + +
[out]ctx
+
+
+
Returns
hipSuccess, hipErrorInvalidContext
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxSetCurrent, hipCtxGetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxPushCurrent (hipCtx_t ctx)
+
+ +

Push the context to be set as current/ default context.

+
Parameters
+ + +
[in]ctx
+
+
+
Returns
hipSuccess, hipErrorInvalidContext
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize , hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxSetCacheConfig (hipFuncCache_t cacheConfig)
+
+ +

Set L1/Shared cache partition.

+
Parameters
+ + +
[in]cacheConfiguration
+
+
+
Returns
hipSuccess
+
Warning
AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures.
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxSetCurrent (hipCtx_t ctx)
+
+ +

Set the passed context as current/default.

+
Parameters
+ + +
[in]ctx
+
+
+
Returns
hipSuccess, hipErrorInvalidContext
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize , hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxSetSharedMemConfig (hipSharedMemConfig config)
+
+ +

Set Shared memory bank configuration.

+
Parameters
+ + +
[in]sharedMemoryConfiguration
+
+
+
Returns
hipSuccess
+
Warning
AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures.
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipCtxSynchronize (void )
+
+ +

Blocks until the default context has completed all preceding requested tasks.

+
Returns
hipSuccess
+
Warning
This function waits for all streams on the default context to complete execution, and then returns.
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxGetDevice
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Device.html b/docs/RuntimeAPI/html/group__Device.html new file mode 100644 index 0000000000..e1dddbbed7 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Device.html @@ -0,0 +1,584 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Device Management + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Device Management
HIP API
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipDeviceSynchronize (void)
 Waits on all active streams on current device. More...
 
hipError_t hipDeviceReset (void)
 The state of current device is discarded and updated to a fresh state. More...
 
hipError_t hipSetDevice (int deviceId)
 Set default device to be used for subsequent hip API calls from this thread. More...
 
hipError_t hipGetDevice (int *deviceId)
 Return the default device id for the calling host thread. More...
 
hipError_t hipGetDeviceCount (int *count)
 Return number of compute-capable devices. More...
 
hipError_t hipDeviceGetAttribute (int *pi, hipDeviceAttribute_t attr, int deviceId)
 Query for a specific device attribute. More...
 
hipError_t hipGetDeviceProperties (hipDeviceProp_t *prop, int deviceId)
 Returns device properties. More...
 
hipError_t hipDeviceSetCacheConfig (hipFuncCache_t cacheConfig)
 Set L1/Shared cache partition. More...
 
hipError_t hipDeviceGetCacheConfig (hipFuncCache_t *cacheConfig)
 Set Cache configuration for a specific function. More...
 
hipError_t hipDeviceGetLimit (size_t *pValue, hipLimit_t limit)
 Get Resource limits of current device. More...
 
hipError_t hipFuncSetCacheConfig (hipFuncCache_t config)
 Set Cache configuration for a specific function. More...
 
hipError_t hipDeviceGetSharedMemConfig (hipSharedMemConfig *pConfig)
 Returns bank width of shared memory for current device. More...
 
hipError_t hipDeviceSetSharedMemConfig (hipSharedMemConfig config)
 The bank width of shared memory on current device is set. More...
 
hipError_t hipSetDeviceFlags (unsigned flags)
 The current device behavior is changed according the flags passed. More...
 
hipError_t hipChooseDevice (int *device, const hipDeviceProp_t *prop)
 Device which matches hipDeviceProp_t is returned. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

Function Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipChooseDevice (int * device,
const hipDeviceProp_t * prop 
)
+
+ +

Device which matches hipDeviceProp_t is returned.

+
Parameters
+ + + +
[out]deviceID
[in]deviceproperties pointer
+
+
+
Returns
hipSuccess, hipErrorInvalidValue
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceGetAttribute (int * pi,
hipDeviceAttribute_t attr,
int deviceId 
)
+
+ +

Query for a specific device attribute.

+
Parameters
+ + + + +
[out]pipointer to value to return
[in]attrattribute to query
[in]deviceIdwhich device to query for information
+
+
+
Returns
hipSuccess, hipErrorInvalidDevice, hipErrorInvalidValue
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceGetCacheConfig (hipFuncCache_t * cacheConfig)
+
+ +

Set Cache configuration for a specific function.

+
Parameters
+ + +
[in]cacheConfig
+
+
+
Returns
hipSuccess, hipErrorInitializationError Note: AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceGetLimit (size_t * pValue,
hipLimit_t limit 
)
+
+ +

Get Resource limits of current device.

+
Parameters
+ + + +
[out]pValue
[in]limit
+
+
+
Returns
hipSuccess, #hipErrorUnsupportedLimit, hipErrorInvalidValue Note: Currently, only hipLimitMallocHeapSize is available
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceGetSharedMemConfig (hipSharedMemConfig * pConfig)
+
+ +

Returns bank width of shared memory for current device.

+
Parameters
+ + +
[out]pConfig
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInitializationError
+

Note: AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures.

+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceReset (void )
+
+ +

The state of current device is discarded and updated to a fresh state.

+

Calling this function deletes all streams created, memory allocated, kernels running, events created. Make sure that no other thread is using the device or streams, memory, kernels, events associated with the current device.

+
Returns
hipSuccess
+
See Also
hipDeviceSynchronize
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceSetCacheConfig (hipFuncCache_t cacheConfig)
+
+ +

Set L1/Shared cache partition.

+
Parameters
+ + +
[in]cacheConfig
+
+
+
Returns
hipSuccess, hipErrorInitializationError Note: AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures.
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceSetSharedMemConfig (hipSharedMemConfig config)
+
+ +

The bank width of shared memory on current device is set.

+
Parameters
+ + +
[in]config
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInitializationError
+

Note: AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures.

+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceSynchronize (void )
+
+ +

Waits on all active streams on current device.

+

When this command is invoked, the host thread gets blocked until all the commands associated with streams associated with the device. HIP does not support multiple blocking modes (yet!).

+
Returns
hipSuccess
+
See Also
hipSetDevice, hipDeviceReset
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipFuncSetCacheConfig (hipFuncCache_t config)
+
+ +

Set Cache configuration for a specific function.

+
Parameters
+ + +
[in]config;
+
+
+
Returns
hipSuccess, hipErrorInitializationError Note: AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures.
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipGetDevice (int * deviceId)
+
+ +

Return the default device id for the calling host thread.

+
Parameters
+ + +
[out]device*device is written with the default device
+
+
+

HIP maintains an default device for each thread using thread-local-storage. This device is used implicitly for HIP runtime APIs called by this thread. hipGetDevice returns in * device the default device for the calling host thread.

+
Returns
hipSuccess
+
See Also
hipSetDevice, hipGetDevicesizeBytes
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipGetDeviceCount (int * count)
+
+ +

Return number of compute-capable devices.

+
Parameters
+ + +
[output]count Returns number of compute-capable devices.
+
+
+
Returns
hipSuccess, hipErrorNoDevice
+

Returns in *count the number of devices that have ability to run compute commands. If there are no such devices, then hipGetDeviceCount will return hipErrorNoDevice. If 1 or more devices can be found, then hipGetDeviceCount returns hipSuccess.

+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipGetDeviceProperties (hipDeviceProp_t * prop,
int deviceId 
)
+
+ +

Returns device properties.

+
Parameters
+ + + +
[out]propwritten with device properties
[in]deviceIdwhich device to query for information
+
+
+
Returns
hipSuccess, hipErrorInvalidDevice
+
Bug:

HCC always returns 0 for maxThreadsPerMultiProcessor

+

HCC always returns 0 for regsPerBlock

+

HCC always returns 0 for l2CacheSize

+
+

Populates hipGetDeviceProperties with information for the specified device.

+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipSetDevice (int deviceId)
+
+ +

Set default device to be used for subsequent hip API calls from this thread.

+
Parameters
+ + +
[in]deviceIdValid device in range 0...hipGetDeviceCount().
+
+
+

Sets device as the default device for the calling host thread. Valid device id's are 0... (hipGetDeviceCount()-1).

+

Many HIP APIs implicitly use the "default device" :

+
    +
  • Any device memory subsequently allocated from this host thread (using hipMalloc) will be allocated on device.
    • +
  • Any streams or events created from this host thread will be associated with device.
    • +
  • Any kernels launched from this host thread (using hipLaunchKernel) will be executed on device (unless a specific stream is specified, in which case the device associated with that stream will be used).
    • +
+

This function may be called from any host thread. Multiple host threads may use the same device. This function does no synchronization with the previous or new device, and has very little runtime overhead. Applications can use hipSetDevice to quickly switch the default device before making a HIP runtime call which uses the default device.

+

The default device is stored in thread-local-storage for each thread. Thread-pool implementations may inherit the default device of the previous thread. A good practice is to always call hipSetDevice at the start of HIP coding sequency to establish a known standard device.

+
Returns
hipSuccess, hipErrorInvalidDevice, #hipErrorDeviceAlreadyInUse
+
See Also
hipGetDevice, hipGetDeviceCount
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipSetDeviceFlags (unsigned flags)
+
+ +

The current device behavior is changed according the flags passed.

+
Parameters
+ + +
[in]flagsThe schedule flags impact how HIP waits for the completion of a command running on a device. hipDeviceScheduleSpin : HIP runtime will actively spin in the thread which submitted the work until the command completes. This offers the lowest latency, but will consume a CPU core and may increase power. hipDeviceScheduleYield : The HIP runtime will yield the CPU to system so that other tasks can use it. This may increase latency to detect the completion but will consume less power and is friendlier to other tasks in the system. hipDeviceScheduleBlockingSync : On ROCm platform, this is a synonym for hipDeviceScheduleYield. hipDeviceScheduleAuto : Use a hueristic to select between Spin and Yield modes. If the number of HIP contexts is greater than the number of logical processors in the system, use Spin scheduling. Else use Yield scheduling.
+
+
+

hipDeviceMapHost : Allow mapping host memory. On ROCM, this is always allowed and the flag is ignored. hipDeviceLmemResizeToMax :

+
Warning
ROCm silently ignores this flag.
+
Returns
hipSuccess, hipErrorInvalidDevice, #hipErrorSetOnActiveProcess
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Driver.html b/docs/RuntimeAPI/html/group__Driver.html new file mode 100644 index 0000000000..c045006c7e --- /dev/null +++ b/docs/RuntimeAPI/html/group__Driver.html @@ -0,0 +1,750 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Initialization and Version + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Modules | +Functions
+
+
Initialization and Version
HIP API
+
+
+ + + + +

+Modules

 Management
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipInit (unsigned int flags)
 Explicitly initializes the HIP runtime. More...
 
hipError_t hipDeviceGet (hipDevice_t *device, int ordinal)
 Returns a handle to a compute device. More...
 
hipError_t hipDeviceComputeCapability (int *major, int *minor, hipDevice_t device)
 Returns the compute capability of the device. More...
 
hipError_t hipDeviceGetName (char *name, int len, hipDevice_t device)
 Returns an identifer string for the device. More...
 
hipError_t hipDeviceGetPCIBusId (char *pciBusId, int len, int device)
 Returns a PCI Bus Id string for the device, overloaded to take int device ID. More...
 
hipError_t hipDeviceGetByPCIBusId (int *device, const int *pciBusId)
 Returns a handle to a compute device. More...
 
hipError_t hipDeviceTotalMem (size_t *bytes, hipDevice_t device)
 Returns the total amount of memory on the device. More...
 
hipError_t hipDriverGetVersion (int *driverVersion)
 Returns the approximate HIP driver version. More...
 
hipError_t hipRuntimeGetVersion (int *runtimeVersion)
 Returns the approximate HIP Runtime version. More...
 
hipError_t hipModuleLoad (hipModule_t *module, const char *fname)
 Loads code object from file into a hipModule_t. More...
 
hipError_t hipModuleUnload (hipModule_t module)
 Frees the module. More...
 
hipError_t hipModuleGetFunction (hipFunction_t *function, hipModule_t module, const char *kname)
 Function with kname will be extracted if present in module. More...
 
hipError_t hipModuleGetGlobal (hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name)
 returns device memory pointer and size of the kernel present in the module with symbol name More...
 
hipError_t hipModuleLoadData (hipModule_t *module, const void *image)
 builds module from code object which resides in host memory. Image is pointer to that location. More...
 
hipError_t hipModuleLaunchKernel (hipFunction_t f, unsigned int gridDimX, unsigned int gridDimY, unsigned int gridDimZ, unsigned int blockDimX, unsigned int blockDimY, unsigned int blockDimZ, unsigned int sharedMemBytes, hipStream_t stream, void **kernelParams, void **extra)
 launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

Function Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceComputeCapability (int * major,
int * minor,
hipDevice_t device 
)
+
+ +

Returns the compute capability of the device.

+
Parameters
+ + + + +
[out]major
[out]minor
[in]device
+
+
+
Returns
hipSuccess, #hipErrorInavlidDevice
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceGet (hipDevice_t * device,
int ordinal 
)
+
+ +

Returns a handle to a compute device.

+
Parameters
+ + + +
[out]device
[in]ordinal
+
+
+
Returns
hipSuccess, #hipErrorInavlidDevice
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceGetByPCIBusId (int * device,
const int * pciBusId 
)
+
+ +

Returns a handle to a compute device.

+
Parameters
+ + + +
[out]devicehandle
[in]PCIBus ID
+
+
+
Returns
hipSuccess, #hipErrorInavlidDevice, hipErrorInvalidValue
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceGetName (char * name,
int len,
hipDevice_t device 
)
+
+ +

Returns an identifer string for the device.

+
Parameters
+ + + + +
[out]name
[in]len
[in]device
+
+
+
Returns
hipSuccess, #hipErrorInavlidDevice
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceGetPCIBusId (char * pciBusId,
int len,
int device 
)
+
+ +

Returns a PCI Bus Id string for the device, overloaded to take int device ID.

+
Parameters
+ + + + +
[out]pciBusId
[in]len
[in]device
+
+
+
Returns
hipSuccess, #hipErrorInavlidDevice
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceTotalMem (size_t * bytes,
hipDevice_t device 
)
+
+ +

Returns the total amount of memory on the device.

+
Parameters
+ + + +
[out]bytes
[in]device
+
+
+
Returns
hipSuccess, #hipErrorInavlidDevice
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDriverGetVersion (int * driverVersion)
+
+ +

Returns the approximate HIP driver version.

+
Parameters
+ + +
[out]driverVersion
+
+
+
Returns
hipSuccess, #hipErrorInavlidValue
+
Warning
The HIP feature set does not correspond to an exact CUDA SDK driver revision. This function always set *driverVersion to 4 as an approximation though HIP supports some features which were introduced in later CUDA SDK revisions. HIP apps code should not rely on the driver revision number here and should use arch feature flags to test device capabilities or conditional compilation.
+
See Also
hipRuntimeGetVersion
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipInit (unsigned int flags)
+
+ +

Explicitly initializes the HIP runtime.

+

Most HIP APIs implicitly initialize the HIP runtime. This API provides control over the timing of the initialization.

+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipModuleGetFunction (hipFunction_t * function,
hipModule_t module,
const char * kname 
)
+
+ +

Function with kname will be extracted if present in module.

+
Parameters
+ + + + +
[in]module
[in]kname
[out]function
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInvalidContext, hipErrorNotInitialized, hipErrorNotFound,
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipModuleGetGlobal (hipDeviceptr_t * dptr,
size_t * bytes,
hipModule_t hmod,
const char * name 
)
+
+ +

returns device memory pointer and size of the kernel present in the module with symbol name

+
Parameters
+ + + + + +
[out]dptr
[out[bytes
[in]hmod
[in]name
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorNotInitialized
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipModuleLaunchKernel (hipFunction_t f,
unsigned int gridDimX,
unsigned int gridDimY,
unsigned int gridDimZ,
unsigned int blockDimX,
unsigned int blockDimY,
unsigned int blockDimZ,
unsigned int sharedMemBytes,
hipStream_t stream,
void ** kernelParams,
void ** extra 
)
+
+ +

launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra

+
Parameters
+ + + + + + + + + + + + +
[in[f
[in]gridDimX
[in]gridDimY
[in]gridDimZ
[in]blockDimX
[in]blockDimY
[in]blockDimZ
[in]sharedMemBytes
[in]stream
[in]kernelParams
[in]extraaThe function takes the above arguments and run the kernel in hipFunction_t f. with launch parameters specified in gridDimX, gridDimY, gridDimZ, blockDimX, blockDimY and blockDimmZ. The amount of shared memory is specificed and can be used with HIP_DYNAMIC_SHARED. The arguemt extra is used to pass in the arguments for the kernel.
+
+
+
Returns
hipSuccess, hipInvalidDevice, hipErrorNotInitialized, hipErrorInvalidValue
+
Warning
kernellParams argument is not yet implemented in HIP. Please use extra instead. Please refer to hip_porting_driver_api.md for sample usage.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipModuleLoad (hipModule_t * module,
const char * fname 
)
+
+ +

Loads code object from file into a hipModule_t.

+
Parameters
+ + + +
[in]fname
[out]module
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInvalidContext, hipErrorFileNotFound, hipErrorOutOfMemory, hipErrorSharedObjectInitFailed, hipErrorNotInitialized
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipModuleLoadData (hipModule_t * module,
const void * image 
)
+
+ +

builds module from code object which resides in host memory. Image is pointer to that location.

+
Parameters
+ + + +
[in]image
[out]module
+
+
+
Returns
hipSuccess, hipErrorNotInitialized, hipErrorOutOfMemory, hipErrorNotInitialized
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipModuleUnload (hipModule_t module)
+
+ +

Frees the module.

+
Parameters
+ + +
[in]module
+
+
+
Returns
hipSuccess, hipInvalidValue module is freed and the code objects associated with it are destroyed
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipRuntimeGetVersion (int * runtimeVersion)
+
+ +

Returns the approximate HIP Runtime version.

+
Parameters
+ + +
[out]runtimeVersion
+
+
+
Returns
hipSuccess, #hipErrorInavlidValue
+
Warning
On HIP/HCC path this function returns HIP runtime patch version however on HIP/NVCC path this function return CUDA runtime version.
+
See Also
hipDriverGetVersion
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Error.html b/docs/RuntimeAPI/html/group__Error.html new file mode 100644 index 0000000000..03d5f89643 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Error.html @@ -0,0 +1,206 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Error Handling + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Error Handling
HIP API
+
+
+ + + + + + + + + + + + + + +

+Functions

hipError_t hipGetLastError (void)
 Return last error returned by any HIP runtime API call and resets the stored error code to hipSuccess. More...
 
hipError_t hipPeekAtLastError (void)
 Return last error returned by any HIP runtime API call. More...
 
const char * hipGetErrorName (hipError_t hip_error)
 Return name of the specified error code in text form. More...
 
const char * hipGetErrorString (hipError_t hipError)
 Return handy text string message to explain the error which occurred. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

Function Documentation

+ +
+
+ + + + + + + + +
const char* hipGetErrorName (hipError_t hip_error)
+
+ +

Return name of the specified error code in text form.

+
Parameters
+ + +
hip_errorError code to convert to name.
+
+
+
Returns
const char pointer to the NULL-terminated error name
+
See Also
hipGetErrorString, hipGetLastError, hipPeakAtLastError, hipError_t
+ +
+
+ +
+
+ + + + + + + + +
const char* hipGetErrorString (hipError_t hipError)
+
+ +

Return handy text string message to explain the error which occurred.

+
Parameters
+ + +
hipErrorError code to convert to string.
+
+
+
Returns
const char pointer to the NULL-terminated error string
+
Warning
: on HCC, this function returns the name of the error (same as hipGetErrorName)
+
See Also
hipGetErrorName, hipGetLastError, hipPeakAtLastError, hipError_t
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipGetLastError (void )
+
+ +

Return last error returned by any HIP runtime API call and resets the stored error code to hipSuccess.

+
Returns
return code from last HIP called from the active host thread
+

Returns the last error that has been returned by any of the runtime calls in the same host thread, and then resets the saved error to hipSuccess.

+
See Also
hipGetErrorString, hipGetLastError, hipPeakAtLastError, hipError_t
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipPeekAtLastError (void )
+
+ +

Return last error returned by any HIP runtime API call.

+
Returns
hipSuccess
+

Returns the last error that has been returned by any of the runtime calls in the same host thread. Unlike hipGetLastError, this function does not reset the saved error code.

+
See Also
hipGetErrorString, hipGetLastError, hipPeakAtLastError, hipError_t
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Event.html b/docs/RuntimeAPI/html/group__Event.html new file mode 100644 index 0000000000..73095e0234 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Event.html @@ -0,0 +1,356 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Event Management + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Event Management
HIP API
+
+
+ + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipEventCreateWithFlags (hipEvent_t *event, unsigned flags)
 Create an event with the specified flags. More...
 
hipError_t hipEventCreate (hipEvent_t *event)
 
hipError_t hipEventRecord (hipEvent_t event, hipStream_t stream)
 Record an event in the specified stream. More...
 
hipError_t hipEventDestroy (hipEvent_t event)
 Destroy the specified event. More...
 
hipError_t hipEventSynchronize (hipEvent_t event)
 Wait for an event to complete. More...
 
hipError_t hipEventElapsedTime (float *ms, hipEvent_t start, hipEvent_t stop)
 Return the elapsed time between two events. More...
 
hipError_t hipEventQuery (hipEvent_t event)
 Query event status. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

Function Documentation

+ +
+
+ + + + + + + + +
hipError_t hipEventCreate (hipEvent_t * event)
+
+

Create an event

+
Parameters
+ + +
[in,out]eventReturns the newly created event.
+
+
+
Returns
hipSuccess, hipErrorInitializationError, hipErrorInvalidValue, hipErrorLaunchFailure, hipErrorMemoryAllocation
+
See Also
hipEventCreateWithFlags, hipEventRecord, hipEventQuery, hipEventSynchronize, hipEventDestroy, hipEventElapsedTime
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipEventCreateWithFlags (hipEvent_t * event,
unsigned flags 
)
+
+ +

Create an event with the specified flags.

+
Parameters
+ + + +
[in,out]eventReturns the newly created event.
[in]flagsFlags to control event behavior. Valid values are hipEventDefault, hipEventBlockingSync, hipEventDisableTiming, hipEventInterprocess
+
+
+

hipEventDefault : Default flag. The event will use active synchronization and will support timing. Blocking synchronization provides lowest possible latency at the expense of dedicating a CPU to poll on the eevent. hipEventBlockingSync : The event will use blocking synchronization : if hipEventSynchronize is called on this event, the thread will block until the event completes. This can increase latency for the synchroniation but can result in lower power and more resources for other CPU threads. hipEventDisableTiming : Disable recording of timing information. On ROCM platform, timing information is always recorded and this flag has no performance benefit.

+
Warning
On HCC platform, hipEventInterprocess support is under development. Use of this flag will return an error.
+
Returns
hipSuccess, hipErrorInitializationError, hipErrorInvalidValue, hipErrorLaunchFailure, hipErrorMemoryAllocation
+
See Also
hipEventCreate, hipEventSynchronize, hipEventDestroy, hipEventElapsedTime
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipEventDestroy (hipEvent_t event)
+
+ +

Destroy the specified event.

+
Parameters
+ + +
[in]eventEvent to destroy.
+
+
+
Returns
hipSuccess, hipErrorInitializationError, hipErrorInvalidValue, hipErrorLaunchFailure
+

Releases memory associated with the event. If the event is recording but has not completed recording when hipEventDestroy() is called, the function will return immediately and the completion_future resources will be released later, when the hipDevice is synchronized.

+
See Also
hipEventCreate, hipEventCreateWithFlags, hipEventQuery, hipEventSynchronize, hipEventRecord, hipEventElapsedTime
+
Returns
hipSuccess
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipEventElapsedTime (float * ms,
hipEvent_t start,
hipEvent_t stop 
)
+
+ +

Return the elapsed time between two events.

+
Parameters
+ + + + +
[out]ms: Return time between start and stop in ms.
[in]start: Start event.
[in]stop: Stop event.
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorNotReady, hipErrorInvalidResourceHandle, hipErrorInitializationError, hipErrorLaunchFailure
+

Computes the elapsed time between two events. Time is computed in ms, with a resolution of approximately 1 us.

+

Events which are recorded in a NULL stream will block until all commands on all other streams complete execution, and then record the timestamp.

+

Events which are recorded in a non-NULL stream will record their timestamp when they reach the head of the specified stream, after all previous commands in that stream have completed executing. Thus the time that the event recorded may be significantly after the host calls hipEventRecord().

+

If hipEventRecord() has not been called on either event, then hipErrorInvalidResourceHandle is returned. If hipEventRecord() has been called on both events, but the timestamp has not yet been recorded on one or both events (that is, hipEventQuery() would return hipErrorNotReady on at least one of the events), then hipErrorNotReady is returned.

+
See Also
hipEventCreate, hipEventCreateWithFlags, hipEventQuery, hipEventDestroy, hipEventRecord, hipEventSynchronize
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipEventQuery (hipEvent_t event)
+
+ +

Query event status.

+
Parameters
+ + +
[in]eventEvent to query.
+
+
+
Returns
hipSuccess, hipErrorNotReady, hipErrorInvalidResourceHandle, hipErrorInvalidValue, hipErrorInitializationError, hipErrorLaunchFailure
+

Query the status of the specified event. This function will return hipErrorNotReady if all commands in the appropriate stream (specified to hipEventRecord()) have completed. If that work has not completed, or if hipEventRecord() was not called on the event, then hipSuccess is returned.

+
See Also
hipEventCreate, hipEventCreateWithFlags, hipEventRecord, hipEventDestroy, hipEventSynchronize, hipEventElapsedTime
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipEventRecord (hipEvent_t event,
hipStream_t stream 
)
+
+ +

Record an event in the specified stream.

+
Parameters
+ + + +
[in]eventevent to record.
[in]streamstream in which to record event.
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInitializationError, hipErrorInvalidResourceHandle, hipErrorLaunchFailure
+

hipEventQuery() or hipEventSynchronize() must be used to determine when the event transitions from "recording" (after hipEventRecord() is called) to "recorded" (when timestamps are set, if requested).

+

Events which are recorded in a non-NULL stream will transition to from recording to "recorded" state when they reach the head of the specified stream, after all previous commands in that stream have completed executing.

+

If hipEventRecord() has been previously called aon event, then this call will overwrite any existing state in event.

+

If this function is called on a an event that is currently being recorded, results are undefined - either outstanding recording may save state into the event, and the order is not guaranteed. This shoul be avoided.

+
See Also
hipEventCreate, hipEventCreateWithFlags, hipEventQuery, hipEventSynchronize, hipEventDestroy, hipEventElapsedTime
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipEventSynchronize (hipEvent_t event)
+
+ +

Wait for an event to complete.

+

This function will block until the event is ready, waiting for all previous work in the stream specified when event was recorded with hipEventRecord().

+

If hipEventRecord() has not been called on event, this function returns immediately.

+

TODO-hcc - This function needs to support hipEventBlockingSync parameter.

+
Parameters
+ + +
[in]eventEvent on which to wait.
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInitializationError, hipErrorInvalidResourceHandle, hipErrorLaunchFailure
+
See Also
hipEventCreate, hipEventCreateWithFlags, hipEventQuery, hipEventDestroy, hipEventRecord, hipEventElapsedTime
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Fence.html b/docs/RuntimeAPI/html/group__Fence.html new file mode 100644 index 0000000000..bb73801cca --- /dev/null +++ b/docs/RuntimeAPI/html/group__Fence.html @@ -0,0 +1,211 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Fence Functions + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Fence Functions
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

__device__ void __threadfence_block (void)
 threadfence_block makes writes visible to threads running in same block. More...
 
__device__ void __threadfence (void)
 threadfence makes wirtes visible to other threads running on same GPU. More...
 
__device__ void __threadfence_system (void)
 threadfence_system makes writes to pinned system memory visible on host CPU. More...
 
+__device__ unsigned __hip_ds_bpermute (int index, unsigned src)
 
+__device__ float __hip_ds_bpermutef (int index, float src)
 
+__device__ unsigned __hip_ds_permute (int index, unsigned src)
 
+__device__ float __hip_ds_permutef (int index, float src)
 
+__device__ unsigned __hip_ds_swizzle (unsigned int src, int pattern)
 
+__device__ float __hip_ds_swizzlef (float src, int pattern)
 
+__device__ int __hip_move_dpp (int src, int dpp_ctrl, int row_mask, int bank_mask, bool bound_ctrl)
 
+

Detailed Description

+

CUDA 8 device function features

+

Kernel launching

+
+
Warning
The HIP memory fence functions are currently not supported yet. If any of those threadfence stubs are reached by the application, you should set "export HSA_DISABLE_CACHE=1" to disable L1 and L2 caches.
+

On AMD platforms, the threadfence* routines are currently empty stubs.

+

Function Documentation

+ +
+
+ + + + + + + + +
__device__ void __threadfence (void )
+
+ +

threadfence makes wirtes visible to other threads running on same GPU.

+

void

+
Parameters
+ + +
void
+
+
+
Warning
__threadfence is a stub and map to no-op, application should set "export HSA_DISABLE_CACHE=1" to disable both L1 and L2 caches.
+ +
+
+ +
+
+ + + + + + + + +
__device__ void __threadfence_block (void )
+
+ +

threadfence_block makes writes visible to threads running in same block.

+

void

+
Parameters
+ + +
void
+
+
+
Warning
__threadfence_block is a stub and map to no-op.
+ +
+
+ +
+
+ + + + + + + + +
__device__ void __threadfence_system (void )
+
+ +

threadfence_system makes writes to pinned system memory visible on host CPU.

+

void

+
Parameters
+ + +
void
+
+
+
Warning
__threadfence_system is a stub and map to no-op.
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__GlobalDefs.html b/docs/RuntimeAPI/html/group__GlobalDefs.html new file mode 100644 index 0000000000..4182019c94 --- /dev/null +++ b/docs/RuntimeAPI/html/group__GlobalDefs.html @@ -0,0 +1,700 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Global enum and defines + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Classes | +Macros | +Typedefs | +Enumerations
+
+
Global enum and defines
+
+
+ + + + +

+Classes

struct  dim3
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Macros

#define hipStreamDefault   0x00
 Flags that can be used with hipStreamCreateWithFlags. More...
 
+#define hipStreamNonBlocking   0x01
 Stream does not implicitly synchronize with null stream.
 
#define hipEventDefault   0x0
 Flags that can be used with hipEventCreateWithFlags: More...
 
+#define hipEventBlockingSync   0x1
 Waiting will yield CPU. Power-friendly and usage-friendly but may increase latency.
 
+#define hipEventDisableTiming   0x2
 Disable event's capability to record timing information. May improve performance.
 
#define hipEventInterprocess   0x4
 Event can support IPC. More...
 
+#define hipHostMallocDefault   0x0
 Flags that can be used with hipHostMalloc.
 
+#define hipHostMallocPortable   0x1
 
+#define hipHostMallocMapped   0x2
 
+#define hipHostMallocWriteCombined   0x4
 
#define hipHostRegisterDefault   0x0
 Flags that can be used with hipHostRegister. More...
 
+#define hipHostRegisterPortable   0x1
 Memory is considered registered by all contexts. HIP only supports one context so this is always assumed true.
 
+#define hipHostRegisterMapped   0x2
 Map the allocation into the address space for the current device. The device pointer can be obtained with hipHostGetDevicePointer.
 
+#define hipHostRegisterIoMemory   0x4
 Not supported.
 
+#define hipDeviceScheduleAuto   0x0
 Automatically select between Spin and Yield.
 
+#define hipDeviceScheduleSpin   0x1
 Dedicate a CPU core to spin-wait. Provides lowest latency, but burns a CPU core and may consume more power.
 
+#define hipDeviceScheduleYield   0x2
 Yield the CPU to the operating system when waiting. May increase latency, but lowers power and is friendlier to other threads in the system.
 
+#define hipDeviceScheduleBlockingSync   0x4
 
+#define hipDeviceScheduleMask   0x7
 
+#define hipDeviceMapHost   0x8
 
+#define hipDeviceLmemResizeToMax   0x16
 
+ + + + + + + + + + + + + +

+Typedefs

+typedef enum hipError_t hipError_t
 
+typedef enum hipDeviceAttribute_t hipDeviceAttribute_t
 
typedef enum hipFuncCache_t hipFuncCache_t
 
typedef enum hipSharedMemConfig hipSharedMemConfig
 
typedef struct dim3 dim3
 
typedef enum hipMemcpyKind hipMemcpyKind
 
+ + + + + + + + + + + +

+Enumerations

enum  hipError_t {
+  hipSuccess = 0, +hipErrorOutOfMemory = 2, +hipErrorNotInitialized = 3, +hipErrorDeinitialized = 4, +
+  hipErrorProfilerDisabled = 5, +hipErrorProfilerNotInitialized = 6, +hipErrorProfilerAlreadyStarted = 7, +hipErrorProfilerAlreadyStopped = 8, +
+  hipErrorInvalidImage = 200, +hipErrorInvalidContext = 201, +hipErrorContextAlreadyCurrent = 202, +hipErrorMapFailed = 205, +
+  hipErrorUnmapFailed = 206, +hipErrorArrayIsMapped = 207, +hipErrorAlreadyMapped = 208, +hipErrorNoBinaryForGpu = 209, +
+  hipErrorAlreadyAcquired = 210, +hipErrorNotMapped = 211, +hipErrorNotMappedAsArray = 212, +hipErrorNotMappedAsPointer = 213, +
+  hipErrorECCNotCorrectable = 214, +hipErrorUnsupportedLimit = 215, +hipErrorContextAlreadyInUse = 216, +hipErrorPeerAccessUnsupported = 217, +
+  hipErrorInvalidKernelFile = 218, +hipErrorInvalidGraphicsContext = 219, +hipErrorInvalidSource = 300, +hipErrorFileNotFound = 301, +
+  hipErrorSharedObjectSymbolNotFound = 302, +hipErrorSharedObjectInitFailed = 303, +hipErrorOperatingSystem = 304, +hipErrorInvalidHandle = 400, +
+  hipErrorNotFound = 500, +hipErrorIllegalAddress = 700, +hipErrorInvalidSymbol = 701, +hipErrorMissingConfiguration = 1001, +
+  hipErrorMemoryAllocation = 1002, +hipErrorInitializationError = 1003, +hipErrorLaunchFailure = 1004, +hipErrorPriorLaunchFailure = 1005, +
+  hipErrorLaunchTimeOut = 1006, +hipErrorLaunchOutOfResources = 1007, +hipErrorInvalidDeviceFunction = 1008, +hipErrorInvalidConfiguration = 1009, +
+  hipErrorInvalidDevice = 1010, +hipErrorInvalidValue = 1011, +hipErrorInvalidDevicePointer = 1017, +hipErrorInvalidMemcpyDirection = 1021, +
+  hipErrorUnknown = 1030, +hipErrorInvalidResourceHandle = 1033, +hipErrorNotReady = 1034, +hipErrorNoDevice = 1038, +
+  hipErrorPeerAccessAlreadyEnabled = 1050, +hipErrorPeerAccessNotEnabled = 1051, +hipErrorRuntimeMemory = 1052, +hipErrorRuntimeOther = 1053, +
+  hipErrorHostMemoryAlreadyRegistered = 1061, +hipErrorHostMemoryNotRegistered = 1062, +hipErrorMapBufferObjectFailed = 1071, +hipErrorTbd +
+ }
 
enum  hipDeviceAttribute_t {
+  hipDeviceAttributeMaxThreadsPerBlock, +hipDeviceAttributeMaxBlockDimX, +hipDeviceAttributeMaxBlockDimY, +hipDeviceAttributeMaxBlockDimZ, +
+  hipDeviceAttributeMaxGridDimX, +hipDeviceAttributeMaxGridDimY, +hipDeviceAttributeMaxGridDimZ, +hipDeviceAttributeMaxSharedMemoryPerBlock, +
+  hipDeviceAttributeTotalConstantMemory, +hipDeviceAttributeWarpSize, +hipDeviceAttributeMaxRegistersPerBlock, +hipDeviceAttributeClockRate, +
+  hipDeviceAttributeMemoryClockRate, +hipDeviceAttributeMemoryBusWidth, +hipDeviceAttributeMultiprocessorCount, +hipDeviceAttributeComputeMode, +
+  hipDeviceAttributeL2CacheSize, +hipDeviceAttributeMaxThreadsPerMultiProcessor, +hipDeviceAttributeComputeCapabilityMajor, +hipDeviceAttributeComputeCapabilityMinor, +
+  hipDeviceAttributeConcurrentKernels, +hipDeviceAttributePciBusId, +hipDeviceAttributePciDeviceId, +hipDeviceAttributeMaxSharedMemoryPerMultiprocessor, +
+  hipDeviceAttributeIsMultiGpuBoard +
+ }
 
enum  hipFuncCache_t { hipFuncCachePreferNone, +hipFuncCachePreferShared, +hipFuncCachePreferL1, +hipFuncCachePreferEqual + }
 
enum  hipSharedMemConfig { hipSharedMemBankSizeDefault, +hipSharedMemBankSizeFourByte, +hipSharedMemBankSizeEightByte + }
 
enum  hipMemcpyKind {
+  hipMemcpyHostToHost = 0, +hipMemcpyHostToDevice = 1, +hipMemcpyDeviceToHost = 2, +hipMemcpyDeviceToDevice =3, +
+  hipMemcpyDefault = 4 +
+ }
 
+

Detailed Description

+

Macro Definition Documentation

+ +
+
+ + + + +
#define hipEventDefault   0x0
+
+ +

Flags that can be used with hipEventCreateWithFlags:

+

Default flags

+ +
+
+ +
+
+ + + + +
#define hipEventInterprocess   0x4
+
+ +

Event can support IPC.

+
Warning
- not supported in HIP.
+ +
+
+ +
+
+ + + + +
#define hipHostRegisterDefault   0x0
+
+ +

Flags that can be used with hipHostRegister.

+

Memory is Mapped and Portable

+ +
+
+ +
+
+ + + + +
#define hipStreamDefault   0x00
+
+ +

Flags that can be used with hipStreamCreateWithFlags.

+

Default stream creation flags. These are used with hipStreamCreate().

+ +
+
+

Typedef Documentation

+ +
+
+ + + + +
typedef struct dim3 dim3
+
+

Struct for data in 3D

+ +
+
+ +
+
+ + + + +
typedef enum hipFuncCache_t hipFuncCache_t
+
+
Warning
On AMD devices and recent Nvidia devices, these hints and controls are ignored.
+ +
+
+ +
+
+ + + + +
typedef enum hipMemcpyKind hipMemcpyKind
+
+

Memory copy types

+ +
+
+ +
+
+ + + + +
typedef enum hipSharedMemConfig hipSharedMemConfig
+
+
Warning
On AMD devices and recent Nvidia devices, these hints and controls are ignored.
+ +
+
+

Enumeration Type Documentation

+ +
+
+ + + + +
enum hipDeviceAttribute_t
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + +
Enumerator
hipDeviceAttributeMaxThreadsPerBlock  +

Maximum number of threads per block.

+
hipDeviceAttributeMaxBlockDimX  +

Maximum x-dimension of a block.

+
hipDeviceAttributeMaxBlockDimY  +

Maximum y-dimension of a block.

+
hipDeviceAttributeMaxBlockDimZ  +

Maximum z-dimension of a block.

+
hipDeviceAttributeMaxGridDimX  +

Maximum x-dimension of a grid.

+
hipDeviceAttributeMaxGridDimY  +

Maximum y-dimension of a grid.

+
hipDeviceAttributeMaxGridDimZ  +

Maximum z-dimension of a grid.

+
hipDeviceAttributeMaxSharedMemoryPerBlock  +

Maximum shared memory available per block in bytes.

+
hipDeviceAttributeTotalConstantMemory  +

Constant memory size in bytes.

+
hipDeviceAttributeWarpSize  +

Warp size in threads.

+
hipDeviceAttributeMaxRegistersPerBlock  +

Maximum number of 32-bit registers available to a thread block. This number is shared by all thread blocks simultaneously resident on a multiprocessor.

+
hipDeviceAttributeClockRate  +

Peak clock frequency in kilohertz.

+
hipDeviceAttributeMemoryClockRate  +

Peak memory clock frequency in kilohertz.

+
hipDeviceAttributeMemoryBusWidth  +

Global memory bus width in bits.

+
hipDeviceAttributeMultiprocessorCount  +

Number of multiprocessors on the device.

+
hipDeviceAttributeComputeMode  +

Compute mode that device is currently in.

+
hipDeviceAttributeL2CacheSize  +

Size of L2 cache in bytes. 0 if the device doesn't have L2 cache.

+
hipDeviceAttributeMaxThreadsPerMultiProcessor  +

Maximum resident threads per multiprocessor.

+
hipDeviceAttributeComputeCapabilityMajor  +

Major compute capability version number.

+
hipDeviceAttributeComputeCapabilityMinor  +

Minor compute capability version number.

+
hipDeviceAttributeConcurrentKernels  +

Device can possibly execute multiple kernels concurrently.

+
hipDeviceAttributePciBusId  +

PCI Bus ID.

+
hipDeviceAttributePciDeviceId  +

PCI Device ID.

+
hipDeviceAttributeMaxSharedMemoryPerMultiprocessor  +

Maximum Shared Memory Per Multiprocessor.

+
hipDeviceAttributeIsMultiGpuBoard  +

Multiple GPU devices.

+
+ +
+
+ +
+
+ + + + +
enum hipError_t
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
Enumerator
hipSuccess  +

Successful completion.

+
hipErrorInvalidContext  +

Produced when input context is invalid.

+
hipErrorInvalidKernelFile  +

In CUDA DRV, it is CUDA_ERROR_INVALID_PTX.

+
hipErrorMemoryAllocation  +

Memory allocation error.

+
hipErrorInitializationError  +

TODO comment from hipErrorInitializationError.

+
hipErrorLaunchFailure  +

An exception occurred on the device while executing a kernel.

+
hipErrorLaunchOutOfResources  +

Out of resources error.

+
hipErrorInvalidDevice  +

DeviceID must be in range 0...#compute-devices.

+
hipErrorInvalidValue  +

One or more of the parameters passed to the API call is NULL or not in an acceptable range.

+
hipErrorInvalidDevicePointer  +

Invalid Device Pointer.

+
hipErrorInvalidMemcpyDirection  +

Invalid memory copy direction.

+
hipErrorUnknown  +

Unknown error.

+
hipErrorInvalidResourceHandle  +

Resource handle (hipEvent_t or hipStream_t) invalid.

+
hipErrorNotReady  +

Indicates that asynchronous operations enqueued earlier are not ready. This is not actually an error, but is used to distinguish from hipSuccess (which indicates completion). APIs that return this error include hipEventQuery and hipStreamQuery.

+
hipErrorNoDevice  +

Call to hipGetDeviceCount returned 0 devices.

+
hipErrorPeerAccessAlreadyEnabled  +

Peer access was already enabled from the current device.

+
hipErrorPeerAccessNotEnabled  +

Peer access was never enabled from the current device.

+
hipErrorRuntimeMemory  +

HSA runtime memory call returned error. Typically not seen in production systems.

+
hipErrorRuntimeOther  +

HSA runtime call other than memory returned error. Typically not seen in production systems.

+
hipErrorHostMemoryAlreadyRegistered  +

Produced when trying to lock a page-locked memory.

+
hipErrorHostMemoryNotRegistered  +

Produced when trying to unlock a non-page-locked memory.

+
hipErrorMapBufferObjectFailed  +

Produced when the IPC memory attach failed from ROCr.

+
hipErrorTbd  +

Marker that more error codes are needed.

+
+ +
+
+ +
+
+ + + + +
enum hipFuncCache_t
+
+
Warning
On AMD devices and recent Nvidia devices, these hints and controls are ignored.
+ + + + + +
Enumerator
hipFuncCachePreferNone  +

no preference for shared memory or L1 (default)

+
hipFuncCachePreferShared  +

prefer larger shared memory and smaller L1 cache

+
hipFuncCachePreferL1  +

prefer larger L1 cache and smaller shared memory

+
hipFuncCachePreferEqual  +

prefer equal size L1 cache and shared memory

+
+ +
+
+ +
+
+ + + + +
enum hipMemcpyKind
+
+

Memory copy types

+ + + + + + +
Enumerator
hipMemcpyHostToHost  +

Host-to-Host Copy.

+
hipMemcpyHostToDevice  +

Host-to-Device Copy.

+
hipMemcpyDeviceToHost  +

Device-to-Host Copy.

+
hipMemcpyDeviceToDevice  +

Device-to-Device Copy.

+
hipMemcpyDefault  +

Runtime will automatically determine copy-kind based on virtual addresses.

+
+ +
+
+ +
+
+ + + + +
enum hipSharedMemConfig
+
+
Warning
On AMD devices and recent Nvidia devices, these hints and controls are ignored.
+ + + + +
Enumerator
hipSharedMemBankSizeDefault  +

The compiler selects a device-specific value for the banking.

+
hipSharedMemBankSizeFourByte  +

Shared mem is banked at 4-bytes intervals and performs best when adjacent threads access data 4 bytes apart.

+
hipSharedMemBankSizeEightByte  +

Shared mem is banked at 8-byte intervals and performs best when adjacent threads access data 4 bytes apart.

+
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__HCC__Specific.html b/docs/RuntimeAPI/html/group__HCC__Specific.html new file mode 100644 index 0000000000..b9940e0a53 --- /dev/null +++ b/docs/RuntimeAPI/html/group__HCC__Specific.html @@ -0,0 +1,96 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: HCC-Specific Accessors + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+
HCC-Specific Accessors
HIP API
+
+
+

----------------------------------------------------------------------------------------------—

+

The following calls are only supported when compiler HIP with HCC. To produce portable code, use of these calls must be guarded #ifdef checks:

+
#ifdef __HCC__
+
hc::accelerator acc;
+
hipError_t err = hipHccGetAccelerator(deviceId, &acc)
+
#endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__HIP-ENV.html b/docs/RuntimeAPI/html/group__HIP-ENV.html new file mode 100644 index 0000000000..f38a717247 --- /dev/null +++ b/docs/RuntimeAPI/html/group__HIP-ENV.html @@ -0,0 +1,90 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: HIP Environment Variables + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+
HIP Environment Variables
+
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Memory.html b/docs/RuntimeAPI/html/group__Memory.html new file mode 100644 index 0000000000..338ce818a2 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Memory.html @@ -0,0 +1,1319 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Memory Management + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Memory Management
HIP API
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipPointerGetAttributes (hipPointerAttribute_t *attributes, void *ptr)
 Return attributes for the specified pointer. More...
 
hipError_t hipMalloc (void **ptr, size_t size)
 Allocate memory on the default accelerator. More...
 
hipError_t hipMallocHost (void **ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead")))
 Allocate pinned host memory [Deprecated]. More...
 
hipError_t hipHostMalloc (void **ptr, size_t size, unsigned int flags)
 Allocate device accessible page locked host memory. More...
 
hipError_t hipHostAlloc (void **ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead")))
 Allocate device accessible page locked host memory [Deprecated]. More...
 
hipError_t hipHostGetDevicePointer (void **devPtr, void *hstPtr, unsigned int flags)
 Get Device pointer from Host Pointer allocated through hipHostMalloc. More...
 
hipError_t hipHostGetFlags (unsigned int *flagsPtr, void *hostPtr)
 Return flags associated with host pointer. More...
 
hipError_t hipHostRegister (void *hostPtr, size_t sizeBytes, unsigned int flags)
 Register host memory so it can be accessed from the current device. More...
 
hipError_t hipHostUnregister (void *hostPtr)
 Un-register host pointer. More...
 
hipError_t hipMallocPitch (void **ptr, size_t *pitch, size_t width, size_t height)
 
hipError_t hipFree (void *ptr)
 Free memory allocated by the hcc hip memory allocation API. This API performs an implicit hipDeviceSynchronize() call. If pointer is NULL, the hip runtime is initialized and hipSuccess is returned. More...
 
hipError_t hipFreeHost (void *ptr) __attribute__((deprecated("use hipHostFree instead")))
 Free memory allocated by the hcc hip host memory allocation API. [Deprecated]. More...
 
hipError_t hipHostFree (void *ptr)
 Free memory allocated by the hcc hip host memory allocation API This API performs an implicit hipDeviceSynchronize() call. If pointer is NULL, the hip runtime is initialized and hipSuccess is returned. More...
 
hipError_t hipMemcpy (void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind)
 Copy data from src to dst. More...
 
hipError_t hipMemcpyHtoD (hipDeviceptr_t dst, void *src, size_t sizeBytes)
 Copy data from Host to Device. More...
 
hipError_t hipMemcpyDtoH (void *dst, hipDeviceptr_t src, size_t sizeBytes)
 Copy data from Device to Host. More...
 
hipError_t hipMemcpyDtoD (hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes)
 Copy data from Device to Device. More...
 
hipError_t hipMemcpyHtoDAsync (hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream)
 Copy data from Host to Device asynchronously. More...
 
hipError_t hipMemcpyDtoHAsync (void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
 Copy data from Device to Host asynchronously. More...
 
hipError_t hipMemcpyDtoDAsync (hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
 Copy data from Device to Device asynchronously. More...
 
hipError_t hipMemcpyToSymbol (const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind)
 Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset bytes from the start of symbol symbol. More...
 
hipError_t hipMemcpyToSymbolAsync (const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream)
 Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset bytes from the start of symbol symbol. More...
 
hipError_t hipMemcpyAsync (void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream)
 Copy data from src to dst asynchronously. More...
 
hipError_t hipMemset (void *dst, int value, size_t sizeBytes)
 Copy data from src to dst asynchronously. More...
 
hipError_t hipMemsetAsync (void *dst, int value, size_t sizeBytes, hipStream_t stream)
 Fills the first sizeBytes bytes of the memory area pointed to by dev with the constant byte value value. More...
 
hipError_t hipMemGetInfo (size_t *free, size_t *total)
 Query memory info. Return snapshot of free memory, and total allocatable memory on the device. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

The following CUDA APIs are not currently supported:

+
    +
  • cudaMalloc3D
    • +
  • cudaMalloc3DArray
    • +
  • TODO - more 2D, 3D, array APIs here.
    • +
+

Function Documentation

+ +
+
+ + + + + + + + +
hipError_t hipFree (void * ptr)
+
+ +

Free memory allocated by the hcc hip memory allocation API. This API performs an implicit hipDeviceSynchronize() call. If pointer is NULL, the hip runtime is initialized and hipSuccess is returned.

+
Parameters
+ + +
[in]ptrPointer to memory to be freed
+
+
+
Returns
hipSuccess
+
+hipErrorInvalidDevicePointer (if pointer is invalid, including host pointers allocated with hipHostMalloc)
+
See Also
hipMalloc, hipMallocPitch, hipMallocArray, hipFreeArray, hipHostFree, hipMalloc3D, hipMalloc3DArray, hipHostMalloc
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipFreeHost (void * ptr)
+
+ +

Free memory allocated by the hcc hip host memory allocation API. [Deprecated].

+
Parameters
+ + +
[in]ptrPointer to memory to be freed
+
+
+
Returns
hipSuccess, hipErrorInvalidValue (if pointer is invalid, including device pointers allocated with hipMalloc)
+
Deprecated:
use hipHostFree() instead
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipHostAlloc (void ** ptr,
size_t size,
unsigned int flags 
)
+
+ +

Allocate device accessible page locked host memory [Deprecated].

+
Parameters
+ + + + +
[out]ptrPointer to the allocated host pinned memory
[in]sizeRequested memory size
[in]flagsType of host memory allocation
+
+
+
Returns
hipSuccess, hipErrorMemoryAllocation
+
Deprecated:
use hipHostMalloc() instead
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipHostFree (void * ptr)
+
+ +

Free memory allocated by the hcc hip host memory allocation API This API performs an implicit hipDeviceSynchronize() call. If pointer is NULL, the hip runtime is initialized and hipSuccess is returned.

+
Parameters
+ + +
[in]ptrPointer to memory to be freed
+
+
+
Returns
hipSuccess, hipErrorInvalidValue (if pointer is invalid, including device pointers allocated with hipMalloc)
+
See Also
hipMalloc, hipMallocPitch, hipFree, hipMallocArray, hipFreeArray, hipMalloc3D, hipMalloc3DArray, hipHostMalloc
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipHostGetDevicePointer (void ** devPtr,
void * hstPtr,
unsigned int flags 
)
+
+ +

Get Device pointer from Host Pointer allocated through hipHostMalloc.

+
Parameters
+ + + + +
[out]dstPtrDevice Pointer mapped to passed host pointer
[in]hstPtrHost Pointer allocated through hipHostMalloc
[in]flagsFlags to be passed for extension
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorMemoryAllocation
+
See Also
hipSetDeviceFlags, hipHostMalloc
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipHostGetFlags (unsigned int * flagsPtr,
void * hostPtr 
)
+
+ +

Return flags associated with host pointer.

+
Parameters
+ + + +
[out]flagsPtrMemory location to store flags
[in]hostPtrHost Pointer allocated through hipHostMalloc
+
+
+
Returns
hipSuccess, hipErrorInvalidValue
+
See Also
hipHostMalloc
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipHostMalloc (void ** ptr,
size_t size,
unsigned int flags 
)
+
+ +

Allocate device accessible page locked host memory.

+
Parameters
+ + + + +
[out]ptrPointer to the allocated host pinned memory
[in]sizeRequested memory size
[in]flagsType of host memory allocation
+
+
+
Returns
hipSuccess, hipErrorMemoryAllocation
+
See Also
hipSetDeviceFlags, hipHostFree
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipHostRegister (void * hostPtr,
size_t sizeBytes,
unsigned int flags 
)
+
+ +

Register host memory so it can be accessed from the current device.

+
Parameters
+ + + + +
[out]hostPtrPointer to host memory to be registered.
[in]sizeBytessize of the host memory
[in]flags.See below.
+
+
+

Flags:

+ +

After registering the memory, use hipHostGetDevicePointer to obtain the mapped device pointer. On many systems, the mapped device pointer will have a different value than the mapped host pointer. Applications must use the device pointer in device code, and the host pointer in device code.

+

On some systems, registered memory is pinned. On some systems, registered memory may not be actually be pinned but uses OS or hardware facilities to all GPU access to the host memory.

+

Developers are strongly encouraged to register memory blocks which are aligned to the host cache-line size. (typically 64-bytes but can be obtains from the CPUID instruction).

+

If registering non-aligned pointers, the application must take care when register pointers from the same cache line on different devices. HIP's coarse-grained synchronization model does not guarantee correct results if different devices write to different parts of the same cache block - typically one of the writes will "win" and overwrite data from the other registered memory region.

+
Returns
hipSuccess, hipErrorMemoryAllocation
+
See Also
hipHostUnregister, hipHostGetFlags, hipHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipHostUnregister (void * hostPtr)
+
+ +

Un-register host pointer.

+
Parameters
+ + +
[in]hostPtrHost pointer previously registered with hipHostRegister
+
+
+
Returns
Error code
+
See Also
hipHostRegister
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipMalloc (void ** ptr,
size_t size 
)
+
+ +

Allocate memory on the default accelerator.

+
Parameters
+ + + +
[out]ptrPointer to the allocated memory
[in]sizeRequested memory size
+
+
+
Returns
hipSuccess
+
See Also
hipMallocPitch, hipFree, hipMallocArray, hipFreeArray, hipMalloc3D, hipMalloc3DArray, hipHostFree, hipHostMalloc
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipMallocHost (void ** ptr,
size_t size 
)
+
+ +

Allocate pinned host memory [Deprecated].

+
Parameters
+ + + +
[out]ptrPointer to the allocated host pinned memory
[in]sizeRequested memory size
+
+
+
Returns
hipSuccess, hipErrorMemoryAllocation
+
Deprecated:
use hipHostMalloc() instead
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMallocPitch (void ** ptr,
size_t * pitch,
size_t width,
size_t height 
)
+
+

Allocates at least width (in bytes) * height bytes of linear memory Padding may occur to ensure alighnment requirements are met for the given row The change in width size due to padding will be returned in *pitch. Currently the alignment is set to 128 bytes

+
Parameters
+ + + + + +
[out]ptrPointer to the allocated device memory
[out]pitchPitch for allocation (in bytes)
[in]widthRequested pitched allocation width (in bytes)
[in]heightRequested pitched allocation height
+
+
+
Returns
Error code
+
See Also
hipMalloc, hipFree, hipMallocArray, hipFreeArray, hipHostFree, hipMalloc3D, hipMalloc3DArray, hipHostMalloc
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpy (void * dst,
const void * src,
size_t sizeBytes,
hipMemcpyKind kind 
)
+
+ +

Copy data from src to dst.

+

It supports memory from host to device, device to host, device to device and host to host The src and dst must not overlap.

+

For hipMemcpy, the copy is always performed by the current device (set by hipSetDevice). For multi-gpu or peer-to-peer configurations, it is recommended to set the current device to the device where the src data is physically located. For optimal peer-to-peer copies, the copy device must be able to access the src and dst pointers (by calling hipDeviceEnablePeerAccess with copy agent as the current device and src/dest as the peerDevice argument. if this is not done, the hipMemcpy will still work, but will perform the copy using a staging buffer on the host.

+
Parameters
+ + + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
[in]copyTypeMemory copy type
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorMemoryFree, #hipErrorUnknowni
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyAsync (void * dst,
const void * src,
size_t sizeBytes,
hipMemcpyKind kind,
hipStream_t stream 
)
+
+ +

Copy data from src to dst asynchronously.

+
Warning
If host or dest are not pinned, the memory copy will be performed synchronously. For best performance, use hipHostMalloc to allocate host memory that is transferred asynchronously.
+
+on HCC hipMemcpyAsync does not support overlapped H2D and D2H copies. For hipMemcpy, the copy is always performed by the device associated with the specified stream.
+

For multi-gpu or peer-to-peer configurations, it is recommended to use a stream which is a attached to the device where the src data is physically located. For optimal peer-to-peer copies, the copy device must be able to access the src and dst pointers (by calling hipDeviceEnablePeerAccess with copy agent as the current device and src/dest as the peerDevice argument. if this is not done, the hipMemcpy will still work, but will perform the copy using a staging buffer on the host.

+
Parameters
+ + + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
[in]accelerator_viewAccelerator view which the copy is being enqueued
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorMemoryFree, hipErrorUnknown
+
See Also
hipMemcpy, hipMemcpy2D, hipMemcpyToArray, hipMemcpy2DToArray, hipMemcpyFromArray, hipMemcpy2DFromArray, hipMemcpyArrayToArray, hipMemcpy2DArrayToArray, hipMemcpyToSymbol, hipMemcpyFromSymbol, hipMemcpy2DAsync, hipMemcpyToArrayAsync, hipMemcpy2DToArrayAsync, hipMemcpyFromArrayAsync, hipMemcpy2DFromArrayAsync, hipMemcpyToSymbolAsync, hipMemcpyFromSymbolAsync
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyDtoD (hipDeviceptr_t dst,
hipDeviceptr_t src,
size_t sizeBytes 
)
+
+ +

Copy data from Device to Device.

+
Parameters
+ + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
+
+
+
Returns
hipSuccess, #hipErrorDeInitialized, #hipErrorNotInitialized, hipErrorInvalidContext, hipErrorInvalidValue
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyDtoDAsync (hipDeviceptr_t dst,
hipDeviceptr_t src,
size_t sizeBytes,
hipStream_t stream 
)
+
+ +

Copy data from Device to Device asynchronously.

+
Parameters
+ + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
+
+
+
Returns
hipSuccess, #hipErrorDeInitialized, #hipErrorNotInitialized, hipErrorInvalidContext, hipErrorInvalidValue
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyDtoH (void * dst,
hipDeviceptr_t src,
size_t sizeBytes 
)
+
+ +

Copy data from Device to Host.

+
Parameters
+ + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
+
+
+
Returns
hipSuccess, #hipErrorDeInitialized, #hipErrorNotInitialized, hipErrorInvalidContext, hipErrorInvalidValue
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyDtoHAsync (void * dst,
hipDeviceptr_t src,
size_t sizeBytes,
hipStream_t stream 
)
+
+ +

Copy data from Device to Host asynchronously.

+
Parameters
+ + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
+
+
+
Returns
hipSuccess, #hipErrorDeInitialized, #hipErrorNotInitialized, hipErrorInvalidContext, hipErrorInvalidValue
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyHtoD (hipDeviceptr_t dst,
void * src,
size_t sizeBytes 
)
+
+ +

Copy data from Host to Device.

+
Parameters
+ + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
+
+
+
Returns
hipSuccess, #hipErrorDeInitialized, #hipErrorNotInitialized, hipErrorInvalidContext, hipErrorInvalidValue
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyHtoDAsync (hipDeviceptr_t dst,
void * src,
size_t sizeBytes,
hipStream_t stream 
)
+
+ +

Copy data from Host to Device asynchronously.

+
Parameters
+ + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
+
+
+
Returns
hipSuccess, #hipErrorDeInitialized, #hipErrorNotInitialized, hipErrorInvalidContext, hipErrorInvalidValue
+
See Also
hipArrayCreate, hipArrayDestroy, hipArrayGetDescriptor, hipMemAlloc, hipMemAllocHost, hipMemAllocPitch, hipMemcpy2D, hipMemcpy2DAsync, hipMemcpy2DUnaligned, hipMemcpyAtoA, hipMemcpyAtoD, hipMemcpyAtoH, hipMemcpyAtoHAsync, hipMemcpyDtoA, hipMemcpyDtoD, hipMemcpyDtoDAsync, hipMemcpyDtoH, hipMemcpyDtoHAsync, hipMemcpyHtoA, hipMemcpyHtoAAsync, hipMemcpyHtoDAsync, hipMemFree, hipMemFreeHost, hipMemGetAddressRange, hipMemGetInfo, hipMemHostAlloc, hipMemHostGetDevicePointer
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyToSymbol (const char * symbolName,
const void * src,
size_t sizeBytes,
size_t offset,
hipMemcpyKind kind 
)
+
+ +

Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset bytes from the start of symbol symbol.

+

The memory areas may not overlap. Symbol can either be a variable that resides in global or constant memory space, or it can be a character string, naming a variable that resides in global or constant memory space. Kind can be either hipMemcpyHostToDevice or hipMemcpyDeviceToDevice TODO: cudaErrorInvalidSymbol and cudaErrorInvalidMemcpyDirection is not supported, use hipErrorUnknown for now.

+
Parameters
+ + + + + + +
[in]symbolName- Symbol destination on device
[in]src- Data being copy from
[in]sizeBytes- Data size in bytes
[in]offset- Offset from start of symbol in bytes
[in]kind- Type of transfer
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorMemoryFree, hipErrorUnknown
+
See Also
hipMemcpy, hipMemcpy2D, hipMemcpyToArray, hipMemcpy2DToArray, hipMemcpyFromArray, hipMemcpy2DFromArray, hipMemcpyArrayToArray, hipMemcpy2DArrayToArray, hipMemcpyFromSymbol, hipMemcpyAsync, hipMemcpy2DAsync, hipMemcpyToArrayAsync, hipMemcpy2DToArrayAsync, hipMemcpyFromArrayAsync, hipMemcpy2DFromArrayAsync, hipMemcpyToSymbolAsync, hipMemcpyFromSymbolAsync
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyToSymbolAsync (const char * symbolName,
const void * src,
size_t sizeBytes,
size_t offset,
hipMemcpyKind kind,
hipStream_t stream 
)
+
+ +

Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset bytes from the start of symbol symbol.

+

The memory areas may not overlap. Symbol can either be a variable that resides in global or constant memory space, or it can be a character string, naming a variable that resides in global or constant memory space. Kind can be either hipMemcpyHostToDevice or hipMemcpyDeviceToDevice hipMemcpyToSymbolAsync() is asynchronous with respect to the host, so the call may return before copy is complete. TODO: cudaErrorInvalidSymbol and cudaErrorInvalidMemcpyDirection is not supported, use hipErrorUnknown for now.

+
Parameters
+ + + + + + +
[in]symbolName- Symbol destination on device
[in]src- Data being copy from
[in]sizeBytes- Data size in bytes
[in]offset- Offset from start of symbol in bytes
[in]kind- Type of transfer
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorMemoryFree, hipErrorUnknown
+
See Also
hipMemcpy, hipMemcpy2D, hipMemcpyToArray, hipMemcpy2DToArray, hipMemcpyFromArray, hipMemcpy2DFromArray, hipMemcpyArrayToArray, hipMemcpy2DArrayToArray, hipMemcpyFromSymbol, hipMemcpyAsync, hipMemcpy2DAsync, hipMemcpyToArrayAsync, hipMemcpy2DToArrayAsync, hipMemcpyFromArrayAsync, hipMemcpy2DFromArrayAsync, hipMemcpyToSymbolAsync, hipMemcpyFromSymbolAsync
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipMemGetInfo (size_t * free,
size_t * total 
)
+
+ +

Query memory info. Return snapshot of free memory, and total allocatable memory on the device.

+

Returns in *free a snapshot of the current free memory.

+
Returns
hipSuccess, hipErrorInvalidDevice, hipErrorInvalidValue
+
Warning
On HCC, the free memory only accounts for memory allocated by this process and may be optimistic.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemset (void * dst,
int value,
size_t sizeBytes 
)
+
+ +

Copy data from src to dst asynchronously.

+

It supports memory from host to device, device to host, device to device and host to host.

+
Parameters
+ + + + + +
[out]dstData being copy to
[in]srcData being copy from
[in]sizeBytesData size in bytes
[in]accelerator_viewAccelerator view which the copy is being enqueued
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorMemoryFree
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemsetAsync (void * dst,
int value,
size_t sizeBytes,
hipStream_t stream 
)
+
+ +

Fills the first sizeBytes bytes of the memory area pointed to by dev with the constant byte value value.

+

hipMemsetAsync() is asynchronous with respect to the host, so the call may return before the memset is complete. The operation can optionally be associated to a stream by passing a non-zero stream argument. If stream is non-zero, the operation may overlap with operations in other streams.

+
Parameters
+ + + + + +
[out]dstPointer to device memory
[in]value- Value to set for each byte of specified memory
[in]sizeBytes- Size in bytes to set
[in]stream- Stream identifier
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorMemoryFree
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipPointerGetAttributes (hipPointerAttribute_t * attributes,
void * ptr 
)
+
+ +

Return attributes for the specified pointer.

+
Parameters
+ + + +
[out]attributesfor the specified pointer
[in]pointerto get attributes for
+
+
+
Returns
hipSuccess, hipErrorInvalidDevice, hipErrorInvalidValue
+
See Also
hipGetDeviceCount, hipGetDevice, hipSetDevice, hipChooseDevice
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__PeerToPeer.html b/docs/RuntimeAPI/html/group__PeerToPeer.html new file mode 100644 index 0000000000..6038b9a5e5 --- /dev/null +++ b/docs/RuntimeAPI/html/group__PeerToPeer.html @@ -0,0 +1,404 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Device Memory Access + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Macros | +Functions
+
+
Device Memory Access
HIP API
+
+
+ + + + +

+Macros

+#define USE_PEER_NON_UNIFIED   1
 
+ + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipDeviceCanAccessPeer (int *canAccessPeer, int deviceId, int peerDeviceId)
 Determine if a device can access a peer's memory. More...
 
hipError_t hipDeviceEnablePeerAccess (int peerDeviceId, unsigned int flags)
 Enable direct access from current device's virtual address space to memory allocations physically located on a peer device. More...
 
hipError_t hipDeviceDisablePeerAccess (int peerDeviceId)
 Disable direct access from current device's virtual address space to memory allocations physically located on a peer device. More...
 
hipError_t hipMemGetAddressRange (hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr)
 Get information on memory allocations. More...
 
hipError_t hipMemcpyPeer (void *dst, int dstDeviceId, const void *src, int srcDeviceId, size_t sizeBytes)
 Copies memory from one device to memory on another device. More...
 
hipError_t hipMemcpyPeerAsync (void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream)
 Copies memory from one device to memory on another device. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+
Warning
PeerToPeer support is experimental.
+

Function Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceCanAccessPeer (int * canAccessPeer,
int deviceId,
int peerDeviceId 
)
+
+ +

Determine if a device can access a peer's memory.

+
Parameters
+ + + + +
[out]canAccessPeerReturns the peer access capability (0 or 1)
[in]device- device from where memory may be accessed.
[in]peerDevice- device where memory is physically located
+
+
+

Returns "1" in canAccessPeer if the specified device is capable of directly accessing memory physically located on peerDevice , or "0" if not.

+

Returns "0" in canAccessPeer if deviceId == peerDeviceId, and both are valid devices : a device is not a peer of itself.

+
Returns
hipSuccess,
+
+hipErrorInvalidDevice if deviceId or peerDeviceId are not valid devices
+
Warning
PeerToPeer support is experimental.
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipDeviceDisablePeerAccess (int peerDeviceId)
+
+ +

Disable direct access from current device's virtual address space to memory allocations physically located on a peer device.

+

Returns hipErrorPeerAccessNotEnabled if direct access to memory on peerDevice has not yet been enabled from the current device.

+
Parameters
+ + +
[in]peerDeviceId
+
+
+
Returns
hipSuccess, hipErrorPeerAccessNotEnabled
+
Warning
PeerToPeer support is experimental.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipDeviceEnablePeerAccess (int peerDeviceId,
unsigned int flags 
)
+
+ +

Enable direct access from current device's virtual address space to memory allocations physically located on a peer device.

+

Memory which already allocated on peer device will be mapped into the address space of the current device. In addition, all future memory allocations on peerDeviceId will be mapped into the address space of the current device when the memory is allocated. The peer memory remains accessible from the current device until a call to hipDeviceDisablePeerAccess or hipDeviceReset.

+
Parameters
+ + + +
[in]peerDeviceId
[in]flagsReturns hipSuccess, hipErrorInvalidDevice, hipErrorInvalidValue,
+
+
+
Returns
hipErrorPeerAccessAlreadyEnabled if peer access is already enabled for this device.
+
Warning
PeerToPeer support is experimental.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyPeer (void * dst,
int dstDeviceId,
const void * src,
int srcDeviceId,
size_t sizeBytes 
)
+
+ +

Copies memory from one device to memory on another device.

+
Parameters
+ + + + + + +
[out]dst- Destination device pointer.
[in]dstDeviceId- Destination device
[in]src- Source device pointer
[in]srcDeviceId- Source device
[in]sizeBytes- Size of memory copy in bytes
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInvalidDevice
+
Warning
PeerToPeer support is experimental.
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyPeerAsync (void * dst,
int dstDevice,
const void * src,
int srcDevice,
size_t sizeBytes,
hipStream_t stream 
)
+
+ +

Copies memory from one device to memory on another device.

+
Parameters
+ + + + + + + +
[out]dst- Destination device pointer.
[in]dstDevice- Destination device
[in]src- Source device pointer
[in]srcDevice- Source device
[in]sizeBytes- Size of memory copy in bytes
[in]stream- Stream identifier
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInvalidDevice
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemGetAddressRange (hipDeviceptr_t * pbase,
size_t * psize,
hipDeviceptr_t dptr 
)
+
+ +

Get information on memory allocations.

+
Parameters
+ + + + +
[out]pbase- BAse pointer address
[out]psize- Size of allocation
[in]dptr-Device Pointer
+
+
+
Returns
hipSuccess, hipErrorInvalidDevicePointer
+
See Also
hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Profiler.html b/docs/RuntimeAPI/html/group__Profiler.html new file mode 100644 index 0000000000..ab54cc9f59 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Profiler.html @@ -0,0 +1,142 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Control + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Control
HIP API
+
+
+ + + + + + + + +

+Functions

hipError_t hipProfilerStart ()
 Start recording of profiling information When using this API, start the profiler with profiling disabled. (–startdisabled) More...
 
hipError_t hipProfilerStop ()
 Stop recording of profiling information. When using this API, start the profiler with profiling disabled. (–startdisabled) More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+
Warning
The cudaProfilerInitialize API format for "configFile" is not supported.
+

Function Documentation

+ +
+
+ + + + + + + +
hipError_t hipProfilerStart ()
+
+ +

Start recording of profiling information When using this API, start the profiler with profiling disabled. (–startdisabled)

+
Warning
: hipProfilerStart API is under development.
+ +
+
+ +
+
+ + + + + + + +
hipError_t hipProfilerStop ()
+
+ +

Stop recording of profiling information. When using this API, start the profiler with profiling disabled. (–startdisabled)

+
Warning
: hipProfilerStop API is under development.
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Stream.html b/docs/RuntimeAPI/html/group__Stream.html new file mode 100644 index 0000000000..4230158ab1 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Stream.html @@ -0,0 +1,437 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Stream Management + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Typedefs | +Functions
+
+
Stream Management
HIP API
+
+
+ + + + +

+Typedefs

typedef void(* hipStreamCallback_t )(hipStream_t stream, hipError_t status, void *userData)
 
+ + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipStreamCreate (hipStream_t *stream)
 Create an asynchronous stream. More...
 
hipError_t hipStreamCreateWithFlags (hipStream_t *stream, unsigned int flags)
 Create an asynchronous stream. More...
 
hipError_t hipStreamDestroy (hipStream_t stream)
 Destroys the specified stream. More...
 
hipError_t hipStreamQuery (hipStream_t stream)
 Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not. More...
 
hipError_t hipStreamSynchronize (hipStream_t stream)
 Wait for all commands in stream to complete. More...
 
hipError_t hipStreamWaitEvent (hipStream_t stream, hipEvent_t event, unsigned int flags)
 Make the specified compute stream wait for an event. More...
 
hipError_t hipStreamGetFlags (hipStream_t stream, unsigned int *flags)
 Return flags associated with this stream. More...
 
hipError_t hipStreamAddCallback (hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags)
 Adds a callback to be called on the host after all currently enqueued items in the stream have completed. For each cudaStreamAddCallback call, a callback will be executed exactly once. The callback will block later work in the stream until it is finished. More...
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+

The following Stream APIs are not (yet) supported in HIP:

+
    +
  • cudaStreamAddCallback
    • +
  • cudaStreamAttachMemAsync
    • +
  • cudaStreamCreateWithPriority
    • +
  • cudaStreamGetPriority
    • +
  • cudaStreamWaitEvent
    • +
+

Typedef Documentation

+ +
+
+ + + + +
typedef void(* hipStreamCallback_t)(hipStream_t stream, hipError_t status, void *userData)
+
+

Stream CallBack struct

+ +
+
+

Function Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipStreamAddCallback (hipStream_t stream,
hipStreamCallback_t callback,
void * userData,
unsigned int flags 
)
+
+ +

Adds a callback to be called on the host after all currently enqueued items in the stream have completed. For each cudaStreamAddCallback call, a callback will be executed exactly once. The callback will block later work in the stream until it is finished.

+
Parameters
+ + + + + +
[in]stream- Stream to add callback to
[in]callback- The function to call once preceding stream operations are complete
[in]userData- User specified data to be passed to the callback function
[in]flags- Reserved for future use, must be 0
+
+
+
Returns
hipSuccess, hipErrorInvalidResourceHandle, #hipErrorNotSupported
+
See Also
hipStreamCreate, hipStreamCreateWithFlags, hipStreamQuery, hipStreamSynchronize, hipStreamWaitEvent, hipStreamDestroy
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipStreamCreate (hipStream_t * stream)
+
+ +

Create an asynchronous stream.

+
Parameters
+ + +
[in,out]streamValid pointer to hipStream_t. This function writes the memory with the newly created stream.
+
+
+
Returns
hipSuccess, hipErrorInvalidValue
+

Create a new asynchronous stream. stream returns an opaque handle that can be used to reference the newly created stream in subsequent hipStream* commands. The stream is allocated on the heap and will remain allocated even if the handle goes out-of-scope. To release the memory used by the stream, applicaiton must call hipStreamDestroy.

+
Returns
hipSuccess, hipErrorInvalidValue
+
See Also
hipStreamCreateWithFlags, hipStreamSynchronize, hipStreamWaitEvent, hipStreamDestroy
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipStreamCreateWithFlags (hipStream_t * stream,
unsigned int flags 
)
+
+ +

Create an asynchronous stream.

+
Parameters
+ + + +
[in,out]streamPointer to new stream
[in]flagsto control stream creation.
+
+
+
Returns
hipSuccess, hipErrorInvalidValue
+

Create a new asynchronous stream. stream returns an opaque handle that can be used to reference the newly created stream in subsequent hipStream* commands. The stream is allocated on the heap and will remain allocated even if the handle goes out-of-scope. To release the memory used by the stream, applicaiton must call hipStreamDestroy. Flags controls behavior of the stream. See hipStreamDefault, hipStreamNonBlocking.

+
See Also
hipStreamCreate, hipStreamSynchronize, hipStreamWaitEvent, hipStreamDestroy
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipStreamDestroy (hipStream_t stream)
+
+ +

Destroys the specified stream.

+
Parameters
+ + +
[in,out]streamValid pointer to hipStream_t. This function writes the memory with the newly created stream.
+
+
+
Returns
hipSuccess hipErrorInvalidResourceHandle
+

Destroys the specified stream.

+

If commands are still executing on the specified stream, some may complete execution before the queue is deleted.

+

The queue may be destroyed while some commands are still inflight, or may wait for all commands queued to the stream before destroying it.

+
See Also
hipStreamCreate, hipStreamCreateWithFlags, hipStreamQuery, hipStreamWaitEvent, hipStreamSynchronize
+
Returns
hipSuccess, hipErrorInvalidResourceHandle
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + +
hipError_t hipStreamGetFlags (hipStream_t stream,
unsigned int * flags 
)
+
+ +

Return flags associated with this stream.

+
Parameters
+ + + +
[in]streamstream to be queried
[in,out]flagsPointer to an unsigned integer in which the stream's flags are returned
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInvalidResourceHandle
+
+hipSuccess hipErrorInvalidValue hipErrorInvalidResourceHandle
+

Return flags associated with this stream in *flags.

+
See Also
hipStreamCreateWithFlags
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipStreamQuery (hipStream_t stream)
+
+ +

Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not.

+
Parameters
+ + +
[in]streamstream to query
+
+
+
Returns
hipSuccess, hipErrorNotReady, hipErrorInvalidResourceHandle
+

This is thread-safe and returns a snapshot of the current state of the queue. However, if other host threads are sending work to the stream, the status may change immediately after the function is called. It is typically used for debug.

+
See Also
hipStreamCreate, hipStreamCreateWithFlags, hipStreamWaitEvent, hipStreamSynchronize, hipStreamDestroy
+ +
+
+ +
+
+ + + + + + + + +
hipError_t hipStreamSynchronize (hipStream_t stream)
+
+ +

Wait for all commands in stream to complete.

+
Parameters
+ + +
[in]streamstream identifier.
+
+
+
Returns
hipSuccess, hipErrorInvalidResourceHandle
+

If the null stream is specified, this command blocks until all This command honors the hipDeviceLaunchBlocking flag, which controls whether the wait is active or blocking. This command is host-synchronous : the host will block until the stream is empty.

+
See Also
hipStreamCreate, hipStreamCreateWithFlags, hipStreamWaitEvent, hipStreamDestroy
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipStreamWaitEvent (hipStream_t stream,
hipEvent_t event,
unsigned int flags 
)
+
+ +

Make the specified compute stream wait for an event.

+
Parameters
+ + + + +
[in]streamstream to make wait.
[in]eventevent to wait on
[in]flagscontrol operation [must be 0]
+
+
+
Returns
hipSuccess, hipErrorInvalidResourceHandle
+

This function inserts a wait operation into the specified stream. All future work submitted to stream will wait until event reports completion before beginning execution. This function is host-asynchronous and the function may return before the wait has completed.

+
See Also
hipStreamCreate, hipStreamCreateWithFlags, hipStreamSynchronize, hipStreamDestroy
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/group__Texture.html b/docs/RuntimeAPI/html/group__Texture.html new file mode 100644 index 0000000000..21560396c7 --- /dev/null +++ b/docs/RuntimeAPI/html/group__Texture.html @@ -0,0 +1,214 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Texture Reference Management + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+Functions
+
+
Texture Reference Management
HIP API
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipChannelFormatDesc hipCreateChannelDesc (int x, int y, int z, int w, hipChannelFormatKind f)
 Returns a channel descriptor using the specified format. More...
 
+template<typename T >
hipChannelFormatDesc hipCreateChannelDesc ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< int > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< unsigned int > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< long > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< unsigned long > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< float > ()
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipBindTexture (size_t *offset, struct texture< T, dim, readMode > &tex, const void *devPtr, const struct hipChannelFormatDesc *desc, size_t size=UINT_MAX)
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipBindTexture (size_t *offset, struct texture< T, dim, readMode > &tex, const void *devPtr, size_t size=UINT_MAX)
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipBindTextureToArray (struct texture< T, dim, readMode > &tex, hipArray *array)
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipUnbindTexture (struct texture< T, dim, readMode > &tex)
 
+

Detailed Description

+

----------------------------------------------------------------------------------------------—

+
Warning
The HIP texture API implements a small subset of full texture API. Known limitations include:
    +
  • Only point sampling is supported.
    • +
  • Only C++ APIs are provided.
    • +
  • Many APIs and modes are not implemented.
    • +
+
+

The HIP texture support is intended to allow use of texture cache on hardware where this is beneficial.

+

The following CUDA APIs are not currently supported:

+
    +
  • cudaBindTexture2D
    • +
  • cudaBindTextureToArray
    • +
  • cudaBindTextureToMipmappedArray
    • +
  • cudaGetChannelDesc
    • +
  • cudaGetTextureReference
    • +
+

Function Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipChannelFormatDesc hipCreateChannelDesc (int x,
int y,
int z,
int w,
hipChannelFormatKind f 
)
+
+ +

Returns a channel descriptor using the specified format.

+
Parameters
+ + + + + + +
[in]xX component
[in]yY component
[in]zZ component
[in]wW component
[in]fChannel format
+
+
+
Returns
Channel descriptor with format f
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc_8h_source.html b/docs/RuntimeAPI/html/hcc_8h_source.html new file mode 100644 index 0000000000..720e8ffa6e --- /dev/null +++ b/docs/RuntimeAPI/html/hcc_8h_source.html @@ -0,0 +1,130 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hcc.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HIP_HCC_H
+
24 #define HIP_HCC_H
+
25 
+
26 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
27 #include "hip/hcc_detail/hcc_acc.h"
+
28 #endif
+
29 
+
30 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__acc_8h_source.html b/docs/RuntimeAPI/html/hcc__acc_8h_source.html new file mode 100644 index 0000000000..06918d855c --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__acc_8h_source.html @@ -0,0 +1,138 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hcc_acc.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hcc_acc.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HCC_ACC_H
+
24 #define HCC_ACC_H
+
25 #include "hip/hip_runtime_api.h"
+
26 
+
27 #if __cplusplus
+
28 #ifdef __HCC__
+
29 #include <hc.hpp>
+
34 hipError_t hipHccGetAccelerator(int deviceId, hc::accelerator *acc);
+
35 
+
42 hipError_t hipHccGetAcceleratorView(hipStream_t stream, hc::accelerator_view **av);
+
43 #endif
+
44 #endif
+
45 
+
46 #endif
+
hipError_t
hipError_t
Definition: hip_runtime_api.h:152
+
ihipStream_t
Definition: hip_hcc.h:463
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__complex_8h_source.html b/docs/RuntimeAPI/html/hcc__detail_2hip__complex_8h_source.html new file mode 100644 index 0000000000..ba44cc5c16 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__complex_8h_source.html @@ -0,0 +1,276 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_complex.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_complex.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HIPCOMPLEX_H
+
24 #define HIPCOMPLEX_H
+
25 
+
26 typedef struct{
+
27  float x;
+
28  float y;
+
29 }hipFloatComplex;
+
30 
+
31 __device__ static inline float hipCrealf(hipFloatComplex z){
+
32  return z.x;
+
33 }
+
34 
+
35 __device__ static inline float hipCimagf(hipFloatComplex z){
+
36  return z.y;
+
37 }
+
38 
+
39 __device__ static inline hipFloatComplex make_hipFloatComplex(float a, float b){
+
40  hipFloatComplex z;
+
41  z.x = a;
+
42  z.y = b;
+
43  return z;
+
44 }
+
45 
+
46 __device__ static inline hipFloatComplex hipConjf(hipFloatComplex z){
+
47  hipFloatComplex ret;
+
48  ret.x = z.x;
+
49  ret.y = -z.y;
+
50  return ret;
+
51 }
+
52 
+
53 __device__ static inline float hipCsqabsf(hipFloatComplex z){
+
54  return z.x * z.x + z.y * z.y;
+
55 }
+
56 
+
57 __device__ static inline hipFloatComplex hipCaddf(hipFloatComplex p, hipFloatComplex q){
+
58  return make_hipFloatComplex(p.x + q.x, p.y + q.y);
+
59 }
+
60 
+
61 __device__ static inline hipFloatComplex hipCsubf(hipFloatComplex p, hipFloatComplex q){
+
62  return make_hipFloatComplex(p.x - q.x, p.y - q.y);
+
63 }
+
64 
+
65 __device__ static inline hipFloatComplex hipCmulf(hipFloatComplex p, hipFloatComplex q){
+
66  return make_hipFloatComplex(p.x * q.x - p.y * q.y, p.y * q.x + p.x * q.y);
+
67 }
+
68 
+
69 __device__ static inline hipFloatComplex hipCdivf(hipFloatComplex p, hipFloatComplex q){
+
70  float sqabs = hipCsqabsf(q);
+
71  hipFloatComplex ret;
+
72  ret.x = (p.x * q.x + p.y * q.y)/sqabs;
+
73  ret.y = (p.y * q.x - p.x * q.y)/sqabs;
+
74  return ret;
+
75 }
+
76 
+
77 __device__ static inline float hipCabsf(hipFloatComplex z){
+
78  return sqrtf(hipCsqabsf(z));
+
79 }
+
80 
+
81 
+
82 typedef struct{
+
83  double x;
+
84  double y;
+
85 }hipDoubleComplex;
+
86 
+
87 __device__ static inline double hipCreal(hipDoubleComplex z){
+
88  return z.x;
+
89 }
+
90 
+
91 __device__ static inline double hipCimag(hipDoubleComplex z){
+
92  return z.y;
+
93 }
+
94 
+
95 __device__ static inline hipDoubleComplex make_hipDoubleComplex(double a, double b){
+
96  hipDoubleComplex z;
+
97  z.x = a;
+
98  z.y = b;
+
99  return z;
+
100 }
+
101 
+
102 __device__ static inline hipDoubleComplex hipConj(hipDoubleComplex z){
+
103  hipDoubleComplex ret;
+
104  ret.x = z.x;
+
105  ret.y = z.y;
+
106  return ret;
+
107 }
+
108 
+
109 __device__ static inline double hipCsqabs(hipDoubleComplex z){
+
110  return z.x * z.x + z.y * z.y;
+
111 }
+
112 
+
113 __device__ static inline hipDoubleComplex hipCadd(hipDoubleComplex p, hipDoubleComplex q){
+
114  return make_hipDoubleComplex(p.x + q.x, p.y + q.y);
+
115 }
+
116 
+
117 __device__ static inline hipDoubleComplex hipCsub(hipDoubleComplex p, hipDoubleComplex q){
+
118  return make_hipDoubleComplex(p.x - q.x, p.y - q.y);
+
119 }
+
120 
+
121 __device__ static inline hipDoubleComplex hipCmul(hipDoubleComplex p, hipDoubleComplex q){
+
122  return make_hipDoubleComplex(p.x * q.x - p.y * q.y, p.y * q.x + p.x * q.y);
+
123 }
+
124 
+
125 __device__ static inline hipDoubleComplex hipCdiv(hipDoubleComplex p, hipDoubleComplex q){
+
126  double sqabs = hipCsqabs(q);
+
127  hipDoubleComplex ret;
+
128  ret.x = (p.x * q.x + p.y * q.y)/sqabs;
+
129  ret.y = (p.y * q.x - p.x * q.y)/sqabs;
+
130  return ret;
+
131 }
+
132 
+
133 __device__ static inline double hipCabs(hipDoubleComplex z){
+
134  return sqrtf(hipCsqabs(z));
+
135 }
+
136 
+
137 typedef hipFloatComplex hipComplex;
+
138 
+
139 __device__ static inline hipComplex make_hipComplex(float x,
+
140  float y){
+
141  return make_hipFloatComplex(x, y);
+
142 }
+
143 
+
144 __device__ static inline hipFloatComplex hipComplexDoubleToFloat
+
145 (hipDoubleComplex z){
+
146  return make_hipFloatComplex((float)z.x, (float)z.y);
+
147 }
+
148 
+
149 __device__ static inline hipDoubleComplex hipComplexFloatToDouble
+
150 (hipFloatComplex z){
+
151  return make_hipDoubleComplex((double)z.x, (double)z.y);
+
152 }
+
153 
+
154 __device__ static inline hipComplex hipCfmaf(hipComplex p, hipComplex q, hipComplex r){
+
155  float real = (p.x * q.x) + r.x;
+
156  float imag = (q.x * p.y) + r.y;
+
157 
+
158  real = -(p.y * q.y) + real;
+
159  imag = (p.x * q.y) + imag;
+
160 
+
161  return make_hipComplex(real, imag);
+
162 }
+
163 
+
164 __device__ static inline hipDoubleComplex hipCfma(hipDoubleComplex p, hipDoubleComplex q, hipDoubleComplex r){
+
165  float real = (p.x * q.x) + r.x;
+
166  float imag = (q.x * p.y) + r.y;
+
167 
+
168  real = -(p.y * q.y) + real;
+
169  imag = (p.x * q.y) + imag;
+
170 
+
171  return make_hipDoubleComplex(real, imag);
+
172 }
+
173 
+
174 #endif
+
hipFloatComplex
Definition: hip_complex.h:26
+
hipDoubleComplex
Definition: hip_complex.h:82
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__fp16_8h_source.html b/docs/RuntimeAPI/html/hcc__detail_2hip__fp16_8h_source.html new file mode 100644 index 0000000000..27e7ea4ddb --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__fp16_8h_source.html @@ -0,0 +1,278 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_fp16.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_fp16.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HIP_FP16_H
+
24 #define HIP_FP16_H
+
25 
+
26 #include "hip/hip_runtime.h"
+
27 
+
28 typedef struct{
+
29  unsigned x: 16;
+
30 } __half;
+
31 
+
32 
+
33 typedef struct __attribute__((aligned(4))){
+
34  __half p,q;
+
35 } __half2;
+
36 
+
37 typedef __half half;
+
38 typedef __half2 half2;
+
39 
+
40 /*
+
41 Arithmetic functions
+
42 */
+
43 
+
44 __device__ __half __hadd(const __half a, const __half b);
+
45 
+
46 __device__ __half __hadd_sat(const __half a, const __half b);
+
47 
+
48 __device__ __half __hfma(const __half a, const __half b, const __half c);
+
49 
+
50 __device__ __half __hfma_sat(const __half a, const __half b, const __half c);
+
51 
+
52 __device__ __half __hmul(const __half a, const __half b);
+
53 
+
54 __device__ __half __hmul_sat(const __half a, const __half b);
+
55 
+
56 __device__ __half __hneq(const __half a);
+
57 
+
58 __device__ __half __hsub(const __half a, const __half b);
+
59 
+
60 __device__ __half __hsub_sat(const __half a, const __half b);
+
61 
+
62 
+
63 
+
64 /*
+
65 Half2 Arithmetic Instructions
+
66 */
+
67 
+
68 __device__ __half2 __hadd2(const __half2 a, const __half2 b);
+
69 
+
70 __device__ __half2 __hadd2_sat(const __half2 a, const __half2 b);
+
71 
+
72 __device__ __half2 __hfma2(const __half2 a, const __half2 b, const __half2 c);
+
73 
+
74 __device__ __half2 __hfma2_sat(const __half2 a, const __half2 b, const __half2 c);
+
75 
+
76 __device__ __half2 __hmul2(const __half2 a, const __half2 b);
+
77 
+
78 __device__ __half2 __hmul2_sat(const __half2 a, const __half2 b);
+
79 
+
80 __device__ __half2 __hneq2(const __half2 a);
+
81 
+
82 __device__ __half2 __hsub2(const __half2 a, const __half2 b);
+
83 
+
84 __device__ __half2 __hsub2_sat(const __half2 a, const __half2 b);
+
85 
+
86 /*
+
87 Half Cmps
+
88 */
+
89 
+
90 __device__ bool __heq(const __half a, const __half b);
+
91 
+
92 __device__ bool __hge(const __half a, const __half b);
+
93 
+
94 __device__ bool __hgt(const __half a, const __half b);
+
95 
+
96 __device__ bool __hisinf(const __half a);
+
97 
+
98 __device__ bool __hisnan(const __half a);
+
99 
+
100 __device__ bool __hle(const __half a, const __half b);
+
101 
+
102 __device__ bool __hlt(const __half a, const __half b);
+
103 
+
104 __device__ bool __hne(const __half a, const __half b);
+
105 
+
106 /*
+
107 Half2 Cmps
+
108 */
+
109 
+
110 __device__ bool __hbeq2(const __half2 a, const __half2 b);
+
111 
+
112 __device__ bool __hbge2(const __half2 a, const __half2 b);
+
113 
+
114 __device__ bool __hbgt2(const __half2 a, const __half2 b);
+
115 
+
116 __device__ bool __hble2(const __half2 a, const __half2 b);
+
117 
+
118 __device__ bool __hblt2(const __half2 a, const __half2 b);
+
119 
+
120 __device__ bool __hbne2(const __half2 a, const __half2 b);
+
121 
+
122 __device__ __half2 __heq2(const __half2 a, const __half2 b);
+
123 
+
124 __device__ __half2 __hge2(const __half2 a, const __half2 b);
+
125 
+
126 __device__ __half2 __hgt2(const __half2 a, const __half2 b);
+
127 
+
128 __device__ __half2 __hisnan2(const __half2 a);
+
129 
+
130 __device__ __half2 __hle2(const __half2 a, const __half2 b);
+
131 
+
132 __device__ __half2 __hlt2(const __half2 a, const __half2 b);
+
133 
+
134 __device__ __half2 __hne2(const __half2 a, const __half2 b);
+
135 
+
136 
+
137 /*
+
138 Half Cnvs and Data Mvmnt
+
139 */
+
140 
+
141 __device__ __half2 __float22half2_rn(const float2 a);
+
142 
+
143 __device__ __half __float2half(const float a);
+
144 
+
145 __device__ __half2 __float2half2_rn(const float a);
+
146 
+
147 __device__ __half2 __floats2half2_rn(const float a, const float b);
+
148 
+
149 __device__ float2 __half22float2(const __half2 a);
+
150 
+
151 __device__ float __half2float(const __half a);
+
152 
+
153 __device__ __half2 __half2half2(const __half a);
+
154 
+
155 __device__ __half2 __halves2half2(const __half a, const __half b);
+
156 
+
157 __device__ float __high2float(const __half2 a);
+
158 
+
159 __device__ __half __high2half(const __half2 a);
+
160 
+
161 __device__ __half2 __high2half2(const __half2 a);
+
162 
+
163 __device__ __half2 __highs2half2(const __half2 a, const __half2 b);
+
164 
+
165 __device__ float __low2float(const __half2 a);
+
166 
+
167 __device__ __half __low2half(const __half2 a);
+
168 
+
169 __device__ __half2 __low2half2(const __half2 a);
+
170 
+
171 __device__ __half2 __lows2half2(const __half2 a, const __half2 b);
+
172 
+
173 __device__ __half2 __lowhigh2highlow(const __half2 a);
+
174 
+
175 __device__ __half2 __low2half2(const __half2 a, const __half2 b);
+
176 
+
177 #endif
+
__half
Definition: hip_fp16.h:28
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h.html b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h.html new file mode 100644 index 0000000000..9b7b4d8f90 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h.html @@ -0,0 +1,1191 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_runtime.h File Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+Macros | +Functions | +Variables
+
+
hip_runtime.h File Reference
+
+
+ +

Contains definitions of APIs for HIP runtime. +More...

+
#include <math.h>
+#include <string.h>
+#include <stddef.h>
+#include <hip/hip_runtime_api.h>
+#include <grid_launch.h>
+#include <hip/hcc_detail/host_defines.h>
+
+

Go to the source code of this file.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Macros

+#define NVCC_COMPAT
 
+#define CUDA_SUCCESS   hipSuccess
 
+#define hipLaunchParm   grid_launch_parm
 
+#define __launch_bounds__(requiredMaxThreadsPerBlock, minBlocksPerMultiprocessor)
 
+#define clock_t   long long int
 
+#define hipThreadIdx_x   (hc_get_workitem_id(0))
 
+#define hipThreadIdx_y   (hc_get_workitem_id(1))
 
+#define hipThreadIdx_z   (hc_get_workitem_id(2))
 
+#define hipBlockIdx_x   (hc_get_group_id(0))
 
+#define hipBlockIdx_y   (hc_get_group_id(1))
 
+#define hipBlockIdx_z   (hc_get_group_id(2))
 
+#define hipBlockDim_x   (hc_get_group_size(0))
 
+#define hipBlockDim_y   (hc_get_group_size(1))
 
+#define hipBlockDim_z   (hc_get_group_size(2))
 
+#define hipGridDim_x   (hc_get_num_groups(0))
 
+#define hipGridDim_y   (hc_get_num_groups(1))
 
+#define hipGridDim_z   (hc_get_num_groups(2))
 
+#define __syncthreads()   hc_barrier(CLK_LOCAL_MEM_FENCE)
 
+#define HIP_KERNEL_NAME(...)   __VA_ARGS__
 
+#define HIP_SYMBOL(X)   #X
 
#define HIP_DYNAMIC_SHARED(type, var)
 
+#define HIP_DYNAMIC_SHARED_ATTRIBUTE   __attribute__((address_space(3)))
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

+__device__ float acosf (float x)
 
+__device__ float acoshf (float x)
 
+__device__ float asinf (float x)
 
+__device__ float asinhf (float x)
 
+__device__ float atan2f (float y, float x)
 
+__device__ float atanf (float x)
 
+__device__ float atanhf (float x)
 
+__device__ float cbrtf (float x)
 
+__device__ float ceilf (float x)
 
+__device__ float copysignf (float x, float y)
 
+__device__ float coshf (float x)
 
+__device__ float cyl_bessel_i0f (float x)
 
+__device__ float cyl_bessel_i1f (float x)
 
+__device__ float erfcf (float x)
 
+__device__ float erfcinvf (float y)
 
+__device__ float erfcxf (float x)
 
+__device__ float erff (float x)
 
+__device__ float erfinvf (float y)
 
+__device__ float exp2f (float x)
 
+__device__ float expm1f (float x)
 
+__device__ float fabsf (float x)
 
+__device__ float fdimf (float x, float y)
 
+__device__ __host__ float fdividef (float x, float y)
 
+__device__ float floorf (float x)
 
+__device__ float fmaf (float x, float y, float z)
 
+__device__ float fmaxf (float x, float y)
 
+__device__ float fminf (float x, float y)
 
+__device__ float fmodf (float x, float y)
 
+__device__ float frexpf (float x, float y)
 
+__device__ float hypotf (float x, float y)
 
+__device__ float ilogbf (float x)
 
+__host__ __device__ unsigned isfinite (float a)
 
+__device__ unsigned isinf (float a)
 
+__device__ unsigned isnan (float a)
 
+__device__ float j0f (float x)
 
+__device__ float j1f (float x)
 
+__device__ float jnf (int n, float x)
 
+__device__ float ldexpf (float x, int exp)
 
+__device__ float lgammaf (float x)
 
+__device__ long long int llrintf (float x)
 
+__device__ long long int llroundf (float x)
 
+__device__ float log1pf (float x)
 
+__device__ float logbf (float x)
 
+__device__ long int lrintf (float x)
 
+__device__ long int lroundf (float x)
 
+__device__ float modff (float x, float *iptr)
 
+__device__ float nanf (const char *tagp)
 
+__device__ float nearbyintf (float x)
 
+__device__ float nextafterf (float x, float y)
 
+__device__ float norm3df (float a, float b, float c)
 
+__device__ float norm4df (float a, float b, float c, float d)
 
+__device__ float normcdff (float y)
 
+__device__ float normcdfinvf (float y)
 
+__device__ float normf (int dim, const float *a)
 
+__device__ float rcbrtf (float x)
 
+__device__ float remainderf (float x, float y)
 
+__device__ float remquof (float x, float y, int *quo)
 
+__device__ float rhypotf (float x, float y)
 
+__device__ float rintf (float x)
 
+__device__ float rnorm3df (float a, float b, float c)
 
+__device__ float rnorm4df (float a, float b, float c, float d)
 
+__device__ float rnormf (int dim, const float *a)
 
+__device__ float roundf (float x)
 
+__device__ float rsqrtf (float x)
 
+__device__ float scalblnf (float x, long int n)
 
+__device__ float scalbnf (float x, int n)
 
+__host__ __device__ unsigned signbit (float a)
 
+__device__ void sincospif (float x, float *sptr, float *cptr)
 
+__device__ float sinhf (float x)
 
+__device__ float sinpif (float x)
 
+__device__ float sqrtf (float x)
 
+__device__ float tanhf (float x)
 
+__device__ float tgammaf (float x)
 
+__device__ float truncf (float x)
 
+__device__ float y0f (float x)
 
+__device__ float y1f (float x)
 
+__device__ float ynf (int n, float x)
 
+__host__ __device__ float cospif (float x)
 
+__device__ double acos (double x)
 
+__device__ double acosh (double x)
 
+__device__ double asin (double x)
 
+__device__ double asinh (double x)
 
+__device__ double atan (double x)
 
+__device__ double atan2 (double y, double x)
 
+__device__ double atanh (double x)
 
+__device__ double cbrt (double x)
 
+__device__ double ceil (double x)
 
+__device__ double copysign (double x, double y)
 
+__device__ double cos (double x)
 
+__device__ double cosh (double x)
 
+__host__ __device__ double cospi (double x)
 
+__device__ double cyl_bessel_i0 (double x)
 
+__device__ double cyl_bessel_i1 (double x)
 
+__device__ double erf (double x)
 
+__device__ double erfc (double x)
 
+__device__ double erfcinv (double y)
 
+__device__ double erfcx (double x)
 
+__device__ double erfinv (double x)
 
+__device__ double exp (double x)
 
+__device__ double exp10 (double x)
 
+__device__ double exp2 (double x)
 
+__device__ double expm1 (double x)
 
+__device__ double fabs (double x)
 
+__device__ double fdim (double x, double y)
 
+__device__ double fdivide (double x, double y)
 
+__device__ double floor (double x)
 
+__device__ double fma (double x, double y, double z)
 
+__device__ double fmax (double x, double y)
 
+__device__ double fmin (double x, double y)
 
+__device__ double fmod (double x, double y)
 
+__device__ double frexp (double x, int *nptr)
 
+__device__ double hypot (double x, double y)
 
+__device__ double ilogb (double x)
 
+__host__ __device__ unsigned isfinite (double x)
 
+__device__ unsigned isinf (double x)
 
+__device__ unsigned isnan (double x)
 
+__device__ double j0 (double x)
 
+__device__ double j1 (double x)
 
+__device__ double jn (int n, double x)
 
+__device__ double ldexp (double x, int exp)
 
+__device__ double lgamma (double x)
 
+__device__ long long llrint (double x)
 
+__device__ long long llround (double x)
 
+__device__ double log (double x)
 
+__device__ double log10 (double x)
 
+__device__ double log1p (double x)
 
+__device__ double log2 (double x)
 
+__device__ double logb (double x)
 
+__device__ long int lrint (double x)
 
+__device__ long int lround (double x)
 
+__device__ double modf (double x, double *iptr)
 
+__device__ double nan (const char *tagp)
 
+__device__ double nearbyint (double x)
 
+__device__ double nextafter (double x, double y)
 
+__device__ double norm (int dim, const double *t)
 
+__device__ double norm3d (double a, double b, double c)
 
+__device__ double norm4d (double a, double b, double c, double d)
 
+__device__ double normcdf (double y)
 
+__device__ double normcdfinv (double y)
 
+__device__ double pow (double x, double y)
 
+__device__ double rcbrt (double x)
 
+__device__ double remainder (double x, double y)
 
+__device__ double remquo (double x, double y, int *quo)
 
+__device__ double rhypot (double x, double y)
 
+__device__ double rint (double x)
 
+__device__ double rnorm (int dim, const double *t)
 
+__device__ double rnorm3d (double a, double b, double c)
 
+__device__ double rnorm4d (double a, double b, double c, double d)
 
+__device__ double round (double x)
 
+__host__ __device__ double rsqrt (double x)
 
+__device__ double scalbln (double x, long int n)
 
+__device__ double scalbn (double x, int n)
 
+__host__ __device__ unsigned signbit (double a)
 
+__device__ double sin (double a)
 
+__device__ void sincos (double x, double *sptr, double *cptr)
 
+__device__ void sincospi (double x, double *sptr, double *cptr)
 
+__device__ double sinh (double x)
 
+__host__ __device__ double sinpi (double x)
 
+__device__ double sqrt (double x)
 
+__device__ double tan (double x)
 
+__device__ double tanh (double x)
 
+__device__ double tgamma (double x)
 
+__device__ double trunc (double x)
 
+__device__ double y0 (double x)
 
+__device__ double y1 (double y)
 
+__device__ double yn (int n, double x)
 
+__device__ long long int clock64 ()
 
+__device__ clock_t clock ()
 
+__device__ int atomicAdd (int *address, int val)
 
+__device__ unsigned int atomicAdd (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicAdd (unsigned long long int *address, unsigned long long int val)
 
+__device__ float atomicAdd (float *address, float val)
 
+__device__ int atomicSub (int *address, int val)
 
+__device__ unsigned int atomicSub (unsigned int *address, unsigned int val)
 
+__device__ int atomicExch (int *address, int val)
 
+__device__ unsigned int atomicExch (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicExch (unsigned long long int *address, unsigned long long int val)
 
+__device__ float atomicExch (float *address, float val)
 
+__device__ int atomicMin (int *address, int val)
 
+__device__ unsigned int atomicMin (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicMin (unsigned long long int *address, unsigned long long int val)
 
+__device__ int atomicMax (int *address, int val)
 
+__device__ unsigned int atomicMax (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicMax (unsigned long long int *address, unsigned long long int val)
 
+__device__ int atomicCAS (int *address, int compare, int val)
 
+__device__ unsigned int atomicCAS (unsigned int *address, unsigned int compare, unsigned int val)
 
+__device__ unsigned long long int atomicCAS (unsigned long long int *address, unsigned long long int compare, unsigned long long int val)
 
+__device__ int atomicAnd (int *address, int val)
 
+__device__ unsigned int atomicAnd (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicAnd (unsigned long long int *address, unsigned long long int val)
 
+__device__ int atomicOr (int *address, int val)
 
+__device__ unsigned int atomicOr (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicOr (unsigned long long int *address, unsigned long long int val)
 
+__device__ int atomicXor (int *address, int val)
 
+__device__ unsigned int atomicXor (unsigned int *address, unsigned int val)
 
+__device__ unsigned long long int atomicXor (unsigned long long int *address, unsigned long long int val)
 
+__device__ unsigned int atomicInc (unsigned int *address, unsigned int val)
 
+__device__ unsigned int atomicDec (unsigned int *address, unsigned int val)
 
+__device__ int __mul24 (int arg1, int arg2)
 
+__device__ unsigned int __umul24 (unsigned int arg1, unsigned int arg2)
 
+__device__ unsigned int __popc (unsigned int input)
 
+__device__ unsigned int __popcll (unsigned long long int input)
 
+__device__ unsigned int __clz (unsigned int input)
 
+__device__ unsigned int __clzll (unsigned long long int input)
 
+__device__ unsigned int __clz (int input)
 
+__device__ unsigned int __clzll (long long int input)
 
+__device__ unsigned int __ffs (unsigned int input)
 
+__device__ unsigned int __ffsll (unsigned long long int input)
 
+__device__ unsigned int __ffs (int input)
 
+__device__ unsigned int __ffsll (long long int input)
 
+__device__ unsigned int __brev (unsigned int input)
 
+__device__ unsigned long long int __brevll (unsigned long long int input)
 
+__device__ int __all (int input)
 
+__device__ int __any (int input)
 
+__device__ unsigned long long int __ballot (int input)
 
+__device__ int __shfl (int input, int lane, int width)
 
+__device__ int __shfl_up (int input, unsigned int lane_delta, int width)
 
+__device__ int __shfl_down (int input, unsigned int lane_delta, int width)
 
+__device__ int __shfl_xor (int input, int lane_mask, int width)
 
+__device__ float __shfl (float input, int lane, int width)
 
+__device__ float __shfl_up (float input, unsigned int lane_delta, int width)
 
+__device__ float __shfl_down (float input, unsigned int lane_delta, int width)
 
+__device__ float __shfl_xor (float input, int lane_mask, int width)
 
+__host__ __device__ int min (int arg1, int arg2)
 
+__host__ __device__ int max (int arg1, int arg2)
 
+__device__ __attribute__ ((address_space(3))) void *__get_dynamicgroupbaseptr()
 
+__device__ float __hip_precise_cosf (float)
 
+__device__ float __hip_precise_exp10f (float)
 
+__device__ float __hip_precise_expf (float)
 
+__device__ float __hip_precise_frsqrt_rn (float)
 
+__device__ float __hip_precise_fsqrt_rd (float)
 
+__device__ float __hip_precise_fsqrt_rn (float)
 
+__device__ float __hip_precise_fsqrt_ru (float)
 
+__device__ float __hip_precise_fsqrt_rz (float)
 
+__device__ float __hip_precise_log10f (float)
 
+__device__ float __hip_precise_log2f (float)
 
+__device__ float __hip_precise_logf (float)
 
+__device__ float __hip_precise_powf (float, float)
 
+__device__ void __hip_precise_sincosf (float, float *, float *)
 
+__device__ float __hip_precise_sinf (float)
 
+__device__ float __hip_precise_tanf (float)
 
+__device__ double __hip_precise_dsqrt_rd (double)
 
+__device__ double __hip_precise_dsqrt_rn (double)
 
+__device__ double __hip_precise_dsqrt_ru (double)
 
+__device__ double __hip_precise_dsqrt_rz (double)
 
+ __attribute__ ((const)) float __hip_fast_cosf(float) __asm("llvm.cos.f32")
 
+__device__ float __hip_fast_exp10f (float)
 
+__device__ float __hip_fast_expf (float)
 
+__device__ float __hip_fast_frsqrt_rn (float)
 
+__device__ float __hip_fast_fsqrt_rn (float)
 
+__device__ float __hip_fast_fsqrt_ru (float)
 
+__device__ float __hip_fast_fsqrt_rz (float)
 
+__device__ float __hip_fast_log10f (float)
 
+__device__ float __hip_fast_logf (float)
 
+__device__ float __hip_fast_powf (float, float)
 
+__device__ void __hip_fast_sincosf (float, float *, float *)
 
+__device__ float __hip_fast_tanf (float)
 
+float __asm ("llvm.fma.f32")
 
+double __asm ("llvm.fma.f64")
 
+__device__ float sinf (float)
 
+__device__ float cosf (float)
 
+__device__ float tanf (float)
 
+__device__ void sincosf (float, float *, float *)
 
+__device__ float logf (float)
 
+__device__ float log2f (float)
 
+__device__ float log10f (float)
 
+__device__ float expf (float)
 
+__device__ float exp10f (float)
 
+__device__ float powf (float, float)
 
+__device__ float __cosf (float x)
 
+__device__ float __exp10f (float x)
 
+__device__ float __expf (float x)
 
+__device__ float __frsqrt_rn (float x)
 
+__device__ float __fsqrt_rd (float x)
 
+__device__ float __fsqrt_rn (float x)
 
+__device__ float __fsqrt_ru (float x)
 
+__device__ float __fsqrt_rz (float x)
 
+__device__ float __log10f (float x)
 
+__device__ float __log2f (float x)
 
+__device__ float __logf (float x)
 
+__device__ float __powf (float base, float exponent)
 
+__device__ void __sincosf (float x, float *s, float *c)
 
+__device__ float __sinf (float x)
 
+__device__ float __tanf (float x)
 
+__device__ float __fmaf_rd (float x, float y, float z)
 
+__device__ float __fmaf_rn (float x, float y, float z)
 
+__device__ float __fmaf_ru (float x, float y, float z)
 
+__device__ float __fmaf_rz (float x, float y, float z)
 
+__device__ float __frcp_rd (float x)
 
+__device__ float __frcp_rn (float x)
 
+__device__ float __frcp_ru (float x)
 
+__device__ float __frcp_rz (float x)
 
+__device__ double __dsqrt_rd (double x)
 
+__device__ double __dsqrt_rn (double x)
 
+__device__ double __dsqrt_ru (double x)
 
+__device__ double __dsqrt_rz (double x)
 
+__device__ double __fma_rd (double x, double y, double z)
 
+__device__ double __fma_rn (double x, double y, double z)
 
+__device__ double __fma_ru (double x, double y, double z)
 
+__device__ double __fma_rz (double x, double y, double z)
 
+__device__ double __drcp_rd (double x)
 
+__device__ double __drcp_rn (double x)
 
+__device__ double __drcp_ru (double x)
 
+__device__ double __drcp_rz (double x)
 
__device__ void __threadfence_block (void)
 threadfence_block makes writes visible to threads running in same block. More...
 
__device__ void __threadfence (void)
 threadfence makes wirtes visible to other threads running on same GPU. More...
 
__device__ void __threadfence_system (void)
 threadfence_system makes writes to pinned system memory visible on host CPU. More...
 
+__device__ unsigned __hip_ds_bpermute (int index, unsigned src)
 
+__device__ float __hip_ds_bpermutef (int index, float src)
 
+__device__ unsigned __hip_ds_permute (int index, unsigned src)
 
+__device__ float __hip_ds_permutef (int index, float src)
 
+__device__ unsigned __hip_ds_swizzle (unsigned int src, int pattern)
 
+__device__ float __hip_ds_swizzlef (float src, int pattern)
 
+__device__ int __hip_move_dpp (int src, int dpp_ctrl, int row_mask, int bank_mask, bool bound_ctrl)
 
+__device__ void * __hip_hc_malloc (size_t)
 
+__device__ void * __hip_hc_free (void *ptr)
 
+__device__ char4 __hip_hc_add8pk (char4, char4)
 
+__device__ char4 __hip_hc_sub8pk (char4, char4)
 
+__device__ char4 __hip_hc_mul8pk (char4, char4)
 
+ + + + + + + + + +

+Variables

+int HIP_TRACE_API
 
+const int warpSize
 
+ float
 
+ double
 
+

Detailed Description

+

Contains definitions of APIs for HIP runtime.

+

Macro Definition Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + +
#define HIP_DYNAMIC_SHARED( type,
 var 
)
+
+Value:
__attribute__((address_space(3))) type* var = \
+
(__attribute__((address_space(3))) type*)__get_dynamicgroupbaseptr(); \
+

extern shared

+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h_source.html b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h_source.html new file mode 100644 index 0000000000..7ad654c0d0 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime_8h_source.html @@ -0,0 +1,915 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_runtime.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_runtime.h
+
+
+Go to the documentation of this file.
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
28 //#pragma once
+
29 #ifndef HIP_RUNTIME_H
+
30 #define HIP_RUNTIME_H
+
31 
+
32 //---
+
33 // Top part of file can be compiled with any compiler
+
34 
+
35 
+
36 //#include <cstring>
+
37 #if __cplusplus
+
38 #include <cmath>
+
39 #else
+
40 #include <math.h>
+
41 #include <string.h>
+
42 #include <stddef.h>
+
43 #endif
+
44 // Define NVCC_COMPAT for CUDA compatibility
+
45 #define NVCC_COMPAT
+
46 #define CUDA_SUCCESS hipSuccess
+
47 
+
48 #include <hip/hip_runtime_api.h>
+
49 //#include "hip/hcc_detail/hip_hcc.h"
+
50 //---
+
51 // Remainder of this file only compiles with HCC
+
52 #ifdef __HCC__
+
53 #include <grid_launch.h>
+
54 
+
55 #if defined (GRID_LAUNCH_VERSION) and (GRID_LAUNCH_VERSION >= 20)
+
56 // Use field names for grid_launch 2.0 structure, if HCC supports GL 2.0.
+
57 #else
+
58 #error (HCC must support GRID_LAUNCH_20)
+
59 #endif
+
60 
+
61 extern int HIP_TRACE_API;
+
62 
+
63 //TODO-HCC-GL - change this to typedef.
+
64 //typedef grid_launch_parm hipLaunchParm ;
+
65 #define hipLaunchParm grid_launch_parm
+
66 #ifdef __cplusplus
+
67 //#include <hip/hcc_detail/hip_texture.h>
+
68 #include <hip/hcc_detail/hip_ldg.h>
+
69 #endif
+
70 #include <hip/hcc_detail/host_defines.h>
+
71 // TODO-HCC remove old definitions ; ~1602 hcc supports __HCC_ACCELERATOR__ define.
+
72 #if defined (__KALMAR_ACCELERATOR__) && !defined (__HCC_ACCELERATOR__)
+
73 #define __HCC_ACCELERATOR__ __KALMAR_ACCELERATOR__
+
74 #endif
+
75 
+
76 // Feature tests:
+
77 #if defined(__HCC_ACCELERATOR__) && (__HCC_ACCELERATOR__ != 0)
+
78 // Device compile and not host compile:
+
79 
+
80 //TODO-HCC enable __HIP_ARCH_HAS_ATOMICS__ when HCC supports these.
+
81  // 32-bit Atomics:
+
82 #define __HIP_ARCH_HAS_GLOBAL_INT32_ATOMICS__ (1)
+
83 #define __HIP_ARCH_HAS_GLOBAL_FLOAT_ATOMIC_EXCH__ (1)
+
84 #define __HIP_ARCH_HAS_SHARED_INT32_ATOMICS__ (1)
+
85 #define __HIP_ARCH_HAS_SHARED_FLOAT_ATOMIC_EXCH__ (1)
+
86 #define __HIP_ARCH_HAS_FLOAT_ATOMIC_ADD__ (0)
+
87 
+
88 // 64-bit Atomics:
+
89 #define __HIP_ARCH_HAS_GLOBAL_INT64_ATOMICS__ (1)
+
90 #define __HIP_ARCH_HAS_SHARED_INT64_ATOMICS__ (0)
+
91 
+
92 // Doubles
+
93 #define __HIP_ARCH_HAS_DOUBLES__ (1)
+
94 
+
95 //warp cross-lane operations:
+
96 #define __HIP_ARCH_HAS_WARP_VOTE__ (1)
+
97 #define __HIP_ARCH_HAS_WARP_BALLOT__ (1)
+
98 #define __HIP_ARCH_HAS_WARP_SHUFFLE__ (1)
+
99 #define __HIP_ARCH_HAS_WARP_FUNNEL_SHIFT__ (0)
+
100 
+
101 //sync
+
102 #define __HIP_ARCH_HAS_THREAD_FENCE_SYSTEM__ (0)
+
103 #define __HIP_ARCH_HAS_SYNC_THREAD_EXT__ (0)
+
104 
+
105 // misc
+
106 #define __HIP_ARCH_HAS_SURFACE_FUNCS__ (0)
+
107 #define __HIP_ARCH_HAS_3DGRID__ (1)
+
108 #define __HIP_ARCH_HAS_DYNAMIC_PARALLEL__ (0)
+
109 
+
110 #endif /* Device feature flags */
+
111 
+
112 
+
113 //TODO-HCC this is currently ignored by HCC target of HIP
+
114 #define __launch_bounds__(requiredMaxThreadsPerBlock, minBlocksPerMultiprocessor)
+
115 
+
116 // Detect if we are compiling C++ mode or C mode
+
117 #if defined(__cplusplus)
+
118 #define __HCC_CPP__
+
119 #elif defined(__STDC_VERSION__)
+
120 #define __HCC_C__
+
121 #endif
+
122 
+
123 __device__ float acosf(float x);
+
124 __device__ float acoshf(float x);
+
125 __device__ float asinf(float x);
+
126 __device__ float asinhf(float x);
+
127 __device__ float atan2f(float y, float x);
+
128 __device__ float atanf(float x);
+
129 __device__ float atanhf(float x);
+
130 __device__ float cbrtf(float x);
+
131 __device__ float ceilf(float x);
+
132 __device__ float copysignf(float x, float y);
+
133 __device__ float coshf(float x);
+
134 __device__ float cyl_bessel_i0f(float x);
+
135 __device__ float cyl_bessel_i1f(float x);
+
136 __device__ float erfcf(float x);
+
137 __device__ float erfcinvf(float y);
+
138 __host__ float erfcinvf(float y);
+
139 __device__ float erfcxf(float x);
+
140 __host__ float erfcxf(float x);
+
141 __device__ float erff(float x);
+
142 __device__ float erfinvf(float y);
+
143 __host__ float erfinvf(float y);
+
144 __device__ float exp2f(float x);
+
145 __device__ float expm1f(float x);
+
146 __device__ float fabsf(float x);
+
147 __device__ float fdimf(float x, float y);
+
148 __device__ __host__ float fdividef(float x, float y);
+
149 __device__ float floorf(float x);
+
150 __device__ float fmaf(float x, float y, float z);
+
151 __device__ float fmaxf(float x, float y);
+
152 __device__ float fminf(float x, float y);
+
153 __device__ float fmodf(float x, float y);
+
154 __device__ float frexpf(float x, float y);
+
155 __device__ float hypotf(float x, float y);
+
156 __device__ float ilogbf(float x);
+
157 __host__ __device__ unsigned isfinite(float a);
+
158 __device__ unsigned isinf(float a);
+
159 __device__ unsigned isnan(float a);
+
160 __device__ float j0f(float x);
+
161 __device__ float j1f(float x);
+
162 __device__ float jnf(int n, float x);
+
163 __device__ float ldexpf(float x, int exp);
+
164 __device__ float lgammaf(float x);
+
165 __device__ long long int llrintf(float x);
+
166 __device__ long long int llroundf(float x);
+
167 __device__ float log1pf(float x);
+
168 __device__ float logbf(float x);
+
169 __device__ long int lrintf(float x);
+
170 __device__ long int lroundf(float x);
+
171 __device__ float modff(float x, float *iptr);
+
172 __device__ float nanf(const char* tagp);
+
173 __device__ float nearbyintf(float x);
+
174 __device__ float nextafterf(float x, float y);
+
175 __device__ float norm3df(float a, float b, float c);
+
176 __host__ float norm3df(float a, float b, float c);
+
177 __device__ float norm4df(float a, float b, float c, float d);
+
178 __host__ float norm4df(float a, float b, float c, float d);
+
179 __device__ float normcdff(float y);
+
180 __host__ float normcdff(float y);
+
181 __device__ float normcdfinvf(float y);
+
182 __host__ float normcdfinvf(float y);
+
183 __device__ float normf(int dim, const float *a);
+
184 __device__ float rcbrtf(float x);
+
185 __host__ float rcbrtf(float x);
+
186 __device__ float remainderf(float x, float y);
+
187 __device__ float remquof(float x, float y, int *quo);
+
188 __device__ float rhypotf(float x, float y);
+
189 __host__ float rhypotf(float x, float y);
+
190 __device__ float rintf(float x);
+
191 __device__ float rnorm3df(float a, float b, float c);
+
192 __host__ float rnorm3df(float a, float b, float c);
+
193 __device__ float rnorm4df(float a, float b, float c, float d);
+
194 __host__ float rnorm4df(float a, float b, float c, float d);
+
195 __device__ float rnormf(int dim, const float* a);
+
196 __host__ float rnormf(int dim, const float* a);
+
197 __device__ float roundf(float x);
+
198 __device__ float rsqrtf(float x);
+
199 __device__ float scalblnf(float x, long int n);
+
200 __device__ float scalbnf(float x, int n);
+
201 __host__ __device__ unsigned signbit(float a);
+
202 __device__ void sincospif(float x, float *sptr, float *cptr);
+
203 __host__ void sincospif(float x, float *sptr, float *cptr);
+
204 __device__ float sinhf(float x);
+
205 __device__ float sinpif(float x);
+
206 __device__ float sqrtf(float x);
+
207 __device__ float tanhf(float x);
+
208 __device__ float tgammaf(float x);
+
209 __device__ float truncf(float x);
+
210 __device__ float y0f(float x);
+
211 __device__ float y1f(float x);
+
212 __device__ float ynf(int n, float x);
+
213 
+
214 __host__ __device__ float cospif(float x);
+
215 __host__ __device__ float sinpif(float x);
+
216 __device__ float sqrtf(float x);
+
217 __host__ __device__ float rsqrtf(float x);
+
218 
+
219 __device__ double acos(double x);
+
220 __device__ double acosh(double x);
+
221 __device__ double asin(double x);
+
222 __device__ double asinh(double x);
+
223 __device__ double atan(double x);
+
224 __device__ double atan2(double y, double x);
+
225 __device__ double atanh(double x);
+
226 __device__ double cbrt(double x);
+
227 __device__ double ceil(double x);
+
228 __device__ double copysign(double x, double y);
+
229 __device__ double cos(double x);
+
230 __device__ double cosh(double x);
+
231 __host__ __device__ double cospi(double x);
+
232 __device__ double cyl_bessel_i0(double x);
+
233 __device__ double cyl_bessel_i1(double x);
+
234 __device__ double erf(double x);
+
235 __device__ double erfc(double x);
+
236 __device__ double erfcinv(double y);
+
237 __device__ double erfcx(double x);
+
238 __device__ double erfinv(double x);
+
239 __device__ double exp(double x);
+
240 __device__ double exp10(double x);
+
241 __device__ double exp2(double x);
+
242 __device__ double expm1(double x);
+
243 __device__ double fabs(double x);
+
244 __device__ double fdim(double x, double y);
+
245 __device__ double fdivide(double x, double y);
+
246 __device__ double floor(double x);
+
247 __device__ double fma(double x, double y, double z);
+
248 __device__ double fmax(double x, double y);
+
249 __device__ double fmin(double x, double y);
+
250 __device__ double fmod(double x, double y);
+
251 __device__ double frexp(double x, int *nptr);
+
252 __device__ double hypot(double x, double y);
+
253 __device__ double ilogb(double x);
+
254 __host__ __device__ unsigned isfinite(double x);
+
255 __device__ unsigned isinf(double x);
+
256 __device__ unsigned isnan(double x);
+
257 __device__ double j0(double x);
+
258 __device__ double j1(double x);
+
259 __device__ double jn(int n, double x);
+
260 __device__ double ldexp(double x, int exp);
+
261 __device__ double lgamma(double x);
+
262 __device__ long long llrint(double x);
+
263 __device__ long long llround(double x);
+
264 __device__ double log(double x);
+
265 __device__ double log10(double x);
+
266 __device__ double log1p(double x);
+
267 __device__ double log2(double x);
+
268 __device__ double logb(double x);
+
269 __device__ long int lrint(double x);
+
270 __device__ long int lround(double x);
+
271 __device__ double modf(double x, double *iptr);
+
272 __device__ double nan(const char* tagp);
+
273 __device__ double nearbyint(double x);
+
274 __device__ double nextafter(double x, double y);
+
275 __device__ double norm(int dim, const double* t);
+
276 __device__ double norm3d(double a, double b, double c);
+
277 __host__ double norm3d(double a, double b, double c);
+
278 __device__ double norm4d(double a, double b, double c, double d);
+
279 __host__ double norm4d(double a, double b, double c, double d);
+
280 __device__ double normcdf(double y);
+
281 __host__ double normcdf(double y);
+
282 __device__ double normcdfinv(double y);
+
283 __host__ double normcdfinv(double y);
+
284 __device__ double pow(double x, double y);
+
285 __device__ double rcbrt(double x);
+
286 __host__ double rcbrt(double x);
+
287 __device__ double remainder(double x, double y);
+
288 __device__ double remquo(double x, double y, int *quo);
+
289 __device__ double rhypot(double x, double y);
+
290 __host__ double rhypot(double x, double y);
+
291 __device__ double rint(double x);
+
292 __device__ double rnorm(int dim, const double* t);
+
293 __host__ double rnorm(int dim, const double* t);
+
294 __device__ double rnorm3d(double a, double b, double c);
+
295 __host__ double rnorm3d(double a, double b, double c);
+
296 __device__ double rnorm4d(double a, double b, double c, double d);
+
297 __host__ double rnorm4d(double a, double b, double c, double d);
+
298 __device__ double round(double x);
+
299 __host__ __device__ double rsqrt(double x);
+
300 __device__ double scalbln(double x, long int n);
+
301 __device__ double scalbn(double x, int n);
+
302 __host__ __device__ unsigned signbit(double a);
+
303 __device__ double sin(double a);
+
304 __device__ void sincos(double x, double *sptr, double *cptr);
+
305 __device__ void sincospi(double x, double *sptr, double *cptr);
+
306 __host__ void sincospi(double x, double *sptr, double *cptr);
+
307 __device__ double sinh(double x);
+
308 __host__ __device__ double sinpi(double x);
+
309 __device__ double sqrt(double x);
+
310 __device__ double tan(double x);
+
311 __device__ double tanh(double x);
+
312 __device__ double tgamma(double x);
+
313 __device__ double trunc(double x);
+
314 __device__ double y0(double x);
+
315 __device__ double y1(double y);
+
316 __device__ double yn(int n, double x);
+
317 
+
318 __host__ double erfcinv(double y);
+
319 __host__ double erfcx(double x);
+
320 __host__ double erfinv(double y);
+
321 __host__ double fdivide(double x, double y);
+
322 
+
323 // TODO - hipify-clang - change to use the function call.
+
324 //#define warpSize hc::__wavesize()
+
325 extern const int warpSize;
+
326 
+
327 
+
328 #define clock_t long long int
+
329 __device__ long long int clock64();
+
330 __device__ clock_t clock();
+
331 
+
332 //atomicAdd()
+
333 __device__ int atomicAdd(int* address, int val);
+
334 __device__ unsigned int atomicAdd(unsigned int* address,
+
335  unsigned int val);
+
336 
+
337 __device__ unsigned long long int atomicAdd(unsigned long long int* address,
+
338  unsigned long long int val);
+
339 
+
340 __device__ float atomicAdd(float* address, float val);
+
341 
+
342 
+
343 //atomicSub()
+
344 __device__ int atomicSub(int* address, int val);
+
345 
+
346 __device__ unsigned int atomicSub(unsigned int* address,
+
347  unsigned int val);
+
348 
+
349 
+
350 //atomicExch()
+
351 __device__ int atomicExch(int* address, int val);
+
352 
+
353 __device__ unsigned int atomicExch(unsigned int* address,
+
354  unsigned int val);
+
355 
+
356 __device__ unsigned long long int atomicExch(unsigned long long int* address,
+
357  unsigned long long int val);
+
358 
+
359 __device__ float atomicExch(float* address, float val);
+
360 
+
361 
+
362 //atomicMin()
+
363 __device__ int atomicMin(int* address, int val);
+
364 __device__ unsigned int atomicMin(unsigned int* address,
+
365  unsigned int val);
+
366 __device__ unsigned long long int atomicMin(unsigned long long int* address,
+
367  unsigned long long int val);
+
368 
+
369 
+
370 //atomicMax()
+
371 __device__ int atomicMax(int* address, int val);
+
372 __device__ unsigned int atomicMax(unsigned int* address,
+
373  unsigned int val);
+
374 __device__ unsigned long long int atomicMax(unsigned long long int* address,
+
375  unsigned long long int val);
+
376 
+
377 
+
378 //atomicCAS()
+
379 __device__ int atomicCAS(int* address, int compare, int val);
+
380 __device__ unsigned int atomicCAS(unsigned int* address,
+
381  unsigned int compare,
+
382  unsigned int val);
+
383 __device__ unsigned long long int atomicCAS(unsigned long long int* address,
+
384  unsigned long long int compare,
+
385  unsigned long long int val);
+
386 
+
387 
+
388 //atomicAnd()
+
389 __device__ int atomicAnd(int* address, int val);
+
390 __device__ unsigned int atomicAnd(unsigned int* address,
+
391  unsigned int val);
+
392 __device__ unsigned long long int atomicAnd(unsigned long long int* address,
+
393  unsigned long long int val);
+
394 
+
395 
+
396 //atomicOr()
+
397 __device__ int atomicOr(int* address, int val);
+
398 __device__ unsigned int atomicOr(unsigned int* address,
+
399  unsigned int val);
+
400 __device__ unsigned long long int atomicOr(unsigned long long int* address,
+
401  unsigned long long int val);
+
402 
+
403 
+
404 //atomicXor()
+
405 __device__ int atomicXor(int* address, int val);
+
406 __device__ unsigned int atomicXor(unsigned int* address,
+
407  unsigned int val);
+
408 __device__ unsigned long long int atomicXor(unsigned long long int* address,
+
409  unsigned long long int val);
+
410 
+
411 //atomicInc()
+
412 __device__ unsigned int atomicInc(unsigned int* address,
+
413  unsigned int val);
+
414 
+
415 
+
416 //atomicDec()
+
417 __device__ unsigned int atomicDec(unsigned int* address,
+
418  unsigned int val);
+
419 
+
420 //__mul24 __umul24
+
421 __device__ int __mul24(int arg1, int arg2);
+
422 __device__ unsigned int __umul24(unsigned int arg1, unsigned int arg2);
+
423 
+
424 // integer intrinsic function __poc __clz __ffs __brev
+
425 __device__ unsigned int __popc( unsigned int input);
+
426 __device__ unsigned int __popcll( unsigned long long int input);
+
427 __device__ unsigned int __clz(unsigned int input);
+
428 __device__ unsigned int __clzll(unsigned long long int input);
+
429 __device__ unsigned int __clz(int input);
+
430 __device__ unsigned int __clzll(long long int input);
+
431 __device__ unsigned int __ffs(unsigned int input);
+
432 __device__ unsigned int __ffsll(unsigned long long int input);
+
433 __device__ unsigned int __ffs(int input);
+
434 __device__ unsigned int __ffsll(long long int input);
+
435 __device__ unsigned int __brev( unsigned int input);
+
436 __device__ unsigned long long int __brevll( unsigned long long int input);
+
437 
+
438 
+
439 // warp vote function __all __any __ballot
+
440 __device__ int __all( int input);
+
441 __device__ int __any( int input);
+
442 __device__ unsigned long long int __ballot( int input);
+
443 
+
444 // warp shuffle functions
+
445 #ifdef __cplusplus
+
446 __device__ int __shfl(int input, int lane, int width=warpSize);
+
447 __device__ int __shfl_up(int input, unsigned int lane_delta, int width=warpSize);
+
448 __device__ int __shfl_down(int input, unsigned int lane_delta, int width=warpSize);
+
449 __device__ int __shfl_xor(int input, int lane_mask, int width=warpSize);
+
450 __device__ float __shfl(float input, int lane, int width=warpSize);
+
451 __device__ float __shfl_up(float input, unsigned int lane_delta, int width=warpSize);
+
452 __device__ float __shfl_down(float input, unsigned int lane_delta, int width=warpSize);
+
453 __device__ float __shfl_xor(float input, int lane_mask, int width=warpSize);
+
454 #else
+
455 __device__ int __shfl(int input, int lane, int width);
+
456 __device__ int __shfl_up(int input, unsigned int lane_delta, int width);
+
457 __device__ int __shfl_down(int input, unsigned int lane_delta, int width);
+
458 __device__ int __shfl_xor(int input, int lane_mask, int width);
+
459 __device__ float __shfl(float input, int lane, int width);
+
460 __device__ float __shfl_up(float input, unsigned int lane_delta, int width);
+
461 __device__ float __shfl_down(float input, unsigned int lane_delta, int width);
+
462 __device__ float __shfl_xor(float input, int lane_mask, int width);
+
463 #endif
+
464 
+
465 __host__ __device__ int min(int arg1, int arg2);
+
466 __host__ __device__ int max(int arg1, int arg2);
+
467 
+
468 __device__ __attribute__((address_space(3))) void* __get_dynamicgroupbaseptr();
+
469 
+
470 //TODO - add a couple fast math operations here, the set here will grow :
+
471 
+
472 // Single Precision Precise Math
+
473 __device__ float __hip_precise_cosf(float);
+
474 __device__ float __hip_precise_exp10f(float);
+
475 __device__ float __hip_precise_expf(float);
+
476 __device__ float __hip_precise_frsqrt_rn(float);
+
477 __device__ float __hip_precise_fsqrt_rd(float);
+
478 __device__ float __hip_precise_fsqrt_rn(float);
+
479 __device__ float __hip_precise_fsqrt_ru(float);
+
480 __device__ float __hip_precise_fsqrt_rz(float);
+
481 __device__ float __hip_precise_log10f(float);
+
482 __device__ float __hip_precise_log2f(float);
+
483 __device__ float __hip_precise_logf(float);
+
484 __device__ float __hip_precise_powf(float, float);
+
485 __device__ void __hip_precise_sincosf(float,float*,float*);
+
486 __device__ float __hip_precise_sinf(float);
+
487 __device__ float __hip_precise_tanf(float);
+
488 
+
489 // Double Precision Precise Math
+
490 __device__ double __hip_precise_dsqrt_rd(double);
+
491 __device__ double __hip_precise_dsqrt_rn(double);
+
492 __device__ double __hip_precise_dsqrt_ru(double);
+
493 __device__ double __hip_precise_dsqrt_rz(double);
+
494 
+
495 // Single Precision Fast Math
+
496 extern __attribute__((const)) float __hip_fast_cosf(float) __asm("llvm.cos.f32");
+
497 extern __attribute__((const)) float __hip_fast_exp2f(float) __asm("llvm.exp2.f32");
+
498 __device__ float __hip_fast_exp10f(float);
+
499 __device__ float __hip_fast_expf(float);
+
500 __device__ float __hip_fast_frsqrt_rn(float);
+
501 extern __attribute__((const)) float __hip_fast_fsqrt_rd(float) __asm("llvm.sqrt.f32");
+
502 __device__ float __hip_fast_fsqrt_rn(float);
+
503 __device__ float __hip_fast_fsqrt_ru(float);
+
504 __device__ float __hip_fast_fsqrt_rz(float);
+
505 __device__ float __hip_fast_log10f(float);
+
506 extern __attribute__((const)) float __hip_fast_log2f(float) __asm("llvm.log2.f32");
+
507 __device__ float __hip_fast_logf(float);
+
508 __device__ float __hip_fast_powf(float, float);
+
509 __device__ void __hip_fast_sincosf(float,float*,float*);
+
510 extern __attribute__((const)) float __hip_fast_sinf(float) __asm("llvm.sin.f32");
+
511 __device__ float __hip_fast_tanf(float);
+
512 extern __attribute__((const)) float __hip_fast_fmaf(float,float,float) __asm("llvm.fma.f32");
+
513 extern __attribute__((const)) float __hip_fast_frcp(float) __asm("llvm.amdgcn.rcp.f32");
+
514 
+
515 extern __attribute__((const)) double __hip_fast_dsqrt(double) __asm("llvm.sqrt.f64");
+
516 extern __attribute__((const)) double __hip_fast_fma(double,double,double) __asm("llvm.fma.f64");
+
517 extern __attribute__((const)) double __hip_fast_drcp(double) __asm("llvm.amdgcn.rcp.f64");
+
518 
+
519 #ifdef HIP_FAST_MATH
+
520 // Single Precision Precise Math when enabled
+
521 
+
522 __device__ inline float cosf(float x) {
+
523  return __hip_fast_cosf(x);
+
524 }
+
525 
+
526 __device__ inline float exp10f(float x) {
+
527  return __hip_fast_exp10f(x);
+
528 }
+
529 
+
530 __device__ inline float expf(float x) {
+
531  return __hip_fast_expf(x);
+
532 }
+
533 
+
534 __device__ inline float log10f(float x) {
+
535  return __hip_fast_log10f(x);
+
536 }
+
537 
+
538 __device__ inline float log2f(float x) {
+
539  return __hip_fast_log2f(x);
+
540 }
+
541 
+
542 __device__ inline float logf(float x) {
+
543  return __hip_fast_logf(x);
+
544 }
+
545 
+
546 __device__ inline float powf(float base, float exponent) {
+
547  return __hip_fast_powf(base, exponent);
+
548 }
+
549 
+
550 __device__ inline void sincosf(float x, float *s, float *c) {
+
551  return __hip_fast_sincosf(x, s, c);
+
552 }
+
553 
+
554 __device__ inline float sinf(float x) {
+
555  return __hip_fast_sinf(x);
+
556 }
+
557 
+
558 __device__ inline float tanf(float x) {
+
559  return __hip_fast_tanf(x);
+
560 }
+
561 
+
562 #else
+
563 
+
564 __device__ float sinf(float);
+
565 __device__ float cosf(float);
+
566 __device__ float tanf(float);
+
567 __device__ void sincosf(float, float*, float*);
+
568 __device__ float logf(float);
+
569 __device__ float log2f(float);
+
570 __device__ float log10f(float);
+
571 __device__ float expf(float);
+
572 __device__ float exp10f(float);
+
573 __device__ float powf(float, float);
+
574 
+
575 #endif
+
576 // Single Precision Fast Math
+
577 __device__ inline float __cosf(float x) {
+
578  return __hip_fast_cosf(x);
+
579 }
+
580 
+
581 __device__ inline float __exp10f(float x) {
+
582  return __hip_fast_exp10f(x);
+
583 }
+
584 
+
585 __device__ inline float __expf(float x) {
+
586  return __hip_fast_expf(x);
+
587 }
+
588 
+
589 __device__ inline float __frsqrt_rn(float x) {
+
590  return __hip_fast_frsqrt_rn(x);
+
591 }
+
592 
+
593 __device__ inline float __fsqrt_rd(float x) {
+
594  return __hip_fast_fsqrt_rd(x);
+
595 }
+
596 
+
597 __device__ inline float __fsqrt_rn(float x) {
+
598  return __hip_fast_fsqrt_rn(x);
+
599 }
+
600 
+
601 __device__ inline float __fsqrt_ru(float x) {
+
602  return __hip_fast_fsqrt_ru(x);
+
603 }
+
604 
+
605 __device__ inline float __fsqrt_rz(float x) {
+
606  return __hip_fast_fsqrt_rz(x);
+
607 }
+
608 
+
609 __device__ inline float __log10f(float x) {
+
610  return __hip_fast_log10f(x);
+
611 }
+
612 
+
613 __device__ inline float __log2f(float x) {
+
614  return __hip_fast_log2f(x);
+
615 }
+
616 
+
617 __device__ inline float __logf(float x) {
+
618  return __hip_fast_logf(x);
+
619 }
+
620 
+
621 __device__ inline float __powf(float base, float exponent) {
+
622  return __hip_fast_powf(base, exponent);
+
623 }
+
624 
+
625 __device__ inline void __sincosf(float x, float *s, float *c) {
+
626  return __hip_fast_sincosf(x, s, c);
+
627 }
+
628 
+
629 __device__ inline float __sinf(float x) {
+
630  return __hip_fast_sinf(x);
+
631 }
+
632 
+
633 __device__ inline float __tanf(float x) {
+
634  return __hip_fast_tanf(x);
+
635 }
+
636 
+
637 __device__ inline float __fmaf_rd(float x, float y, float z) {
+
638  return __hip_fast_fmaf(x, y, z);
+
639 }
+
640 
+
641 __device__ inline float __fmaf_rn(float x, float y, float z) {
+
642  return __hip_fast_fmaf(x, y, z);
+
643 }
+
644 
+
645 __device__ inline float __fmaf_ru(float x, float y, float z) {
+
646  return __hip_fast_fmaf(x, y, z);
+
647 }
+
648 
+
649 __device__ inline float __fmaf_rz(float x, float y, float z) {
+
650  return __hip_fast_fmaf(x, y, z);
+
651 }
+
652 
+
653 __device__ inline float __frcp_rd(float x) {
+
654  return __hip_fast_frcp(x);
+
655 }
+
656 
+
657 __device__ inline float __frcp_rn(float x) {
+
658  return __hip_fast_frcp(x);
+
659 }
+
660 
+
661 __device__ inline float __frcp_ru(float x) {
+
662  return __hip_fast_frcp(x);
+
663 }
+
664 
+
665 __device__ inline float __frcp_rz(float x) {
+
666  return __hip_fast_frcp(x);
+
667 }
+
668 
+
669 __device__ inline double __dsqrt_rd(double x) {
+
670  return __hip_fast_dsqrt(x);
+
671 }
+
672 
+
673 __device__ inline double __dsqrt_rn(double x) {
+
674  return __hip_fast_dsqrt(x);
+
675 }
+
676 
+
677 __device__ inline double __dsqrt_ru(double x) {
+
678  return __hip_fast_dsqrt(x);
+
679 }
+
680 
+
681 __device__ inline double __dsqrt_rz(double x) {
+
682  return __hip_fast_dsqrt(x);
+
683 }
+
684 
+
685 __device__ inline double __fma_rd(double x, double y, double z) {
+
686  return __hip_fast_fma(x, y, z);
+
687 }
+
688 
+
689 __device__ inline double __fma_rn(double x, double y, double z) {
+
690  return __hip_fast_fma(x, y, z);
+
691 }
+
692 
+
693 __device__ inline double __fma_ru(double x, double y, double z) {
+
694  return __hip_fast_fma(x, y, z);
+
695 }
+
696 
+
697 __device__ inline double __fma_rz(double x, double y, double z) {
+
698  return __hip_fast_fma(x, y, z);
+
699 }
+
700 
+
701 __device__ inline double __drcp_rd(double x) {
+
702  return __hip_fast_drcp(x);
+
703 }
+
704 
+
705 __device__ inline double __drcp_rn(double x) {
+
706  return __hip_fast_drcp(x);
+
707 }
+
708 
+
709 __device__ inline double __drcp_ru(double x) {
+
710  return __hip_fast_drcp(x);
+
711 }
+
712 
+
713 __device__ inline double __drcp_rz(double x) {
+
714  return __hip_fast_drcp(x);
+
715 }
+
716 
+
750 // __device__ void __threadfence_block(void);
+
751 extern "C" __device__ void __threadfence_block(void);
+
752 
+
762 // __device__ void __threadfence(void) __attribute__((deprecated("Provided for compile-time compatibility, not yet functional")));
+
763 extern "C" __device__ void __threadfence(void);
+
764 
+
774 //__device__ void __threadfence_system(void) __attribute__((deprecated("Provided with workaround configuration, see hip_kernel_language.md for details")));
+
775 __device__ void __threadfence_system(void) ;
+
776 
+
777 __device__ unsigned __hip_ds_bpermute(int index, unsigned src);
+
778 __device__ float __hip_ds_bpermutef(int index, float src);
+
779 __device__ unsigned __hip_ds_permute(int index, unsigned src);
+
780 __device__ float __hip_ds_permutef(int index, float src);
+
781 
+
782 __device__ unsigned __hip_ds_swizzle(unsigned int src, int pattern);
+
783 __device__ float __hip_ds_swizzlef(float src, int pattern);
+
784 
+
785 __device__ int __hip_move_dpp(int src, int dpp_ctrl, int row_mask, int bank_mask, bool bound_ctrl);
+
786 
+
787 // doxygen end Fence Fence
+
793 #define hipThreadIdx_x (hc_get_workitem_id(0))
+
794 #define hipThreadIdx_y (hc_get_workitem_id(1))
+
795 #define hipThreadIdx_z (hc_get_workitem_id(2))
+
796 
+
797 #define hipBlockIdx_x (hc_get_group_id(0))
+
798 #define hipBlockIdx_y (hc_get_group_id(1))
+
799 #define hipBlockIdx_z (hc_get_group_id(2))
+
800 
+
801 #define hipBlockDim_x (hc_get_group_size(0))
+
802 #define hipBlockDim_y (hc_get_group_size(1))
+
803 #define hipBlockDim_z (hc_get_group_size(2))
+
804 
+
805 #define hipGridDim_x (hc_get_num_groups(0))
+
806 #define hipGridDim_y (hc_get_num_groups(1))
+
807 #define hipGridDim_z (hc_get_num_groups(2))
+
808 
+
809 //extern "C" __device__ void* memcpy(void* dst, void* src, size_t size);
+
810 //extern "C" __device__ void* memset(void* ptr, uint8_t val, size_t size);
+
811 
+
812 extern "C" __device__ void* __hip_hc_malloc(size_t);
+
813 extern "C" __device__ void* __hip_hc_free(void *ptr);
+
814 
+
815 //extern "C" __device__ void* malloc(size_t size);
+
816 //extern "C" __device__ void* free(void *ptr);
+
817 
+
818 extern "C" __device__ char4 __hip_hc_add8pk(char4, char4);
+
819 extern "C" __device__ char4 __hip_hc_sub8pk(char4, char4);
+
820 extern "C" __device__ char4 __hip_hc_mul8pk(char4, char4);
+
821 
+
822 #define __syncthreads() hc_barrier(CLK_LOCAL_MEM_FENCE)
+
823 
+
824 #define HIP_KERNEL_NAME(...) __VA_ARGS__
+
825 #define HIP_SYMBOL(X) #X
+
826 
+
827 #ifdef __HCC_CPP__
+
828 extern hipStream_t ihipPreLaunchKernel(hipStream_t stream, dim3 grid, dim3 block, grid_launch_parm *lp, const char *kernelNameStr);
+
829 extern hipStream_t ihipPreLaunchKernel(hipStream_t stream, dim3 grid, size_t block, grid_launch_parm *lp, const char *kernelNameStr);
+
830 extern hipStream_t ihipPreLaunchKernel(hipStream_t stream, size_t grid, dim3 block, grid_launch_parm *lp, const char *kernelNameStr);
+
831 extern hipStream_t ihipPreLaunchKernel(hipStream_t stream, size_t grid, size_t block, grid_launch_parm *lp, const char *kernelNameStr);
+
832 extern void ihipPostLaunchKernel(const char *kernelName, hipStream_t stream, grid_launch_parm &lp);
+
833 
+
834 
+
835 // Due to multiple overloaded versions of ihipPreLaunchKernel, the numBlocks3D and blockDim3D can be either size_t or dim3 types
+
836 #define hipLaunchKernel(_kernelName, _numBlocks3D, _blockDim3D, _groupMemBytes, _stream, ...) \
+
837 do {\
+
838  grid_launch_parm lp;\
+
839  lp.dynamic_group_mem_bytes = _groupMemBytes; \
+
840  hipStream_t trueStream = (ihipPreLaunchKernel(_stream, _numBlocks3D, _blockDim3D, &lp, #_kernelName)); \
+
841  _kernelName (lp, ##__VA_ARGS__);\
+
842  ihipPostLaunchKernel(#_kernelName, trueStream, lp);\
+
843 } while(0)
+
844 
+
845 
+
846 #elif defined (__HCC_C__)
+
847 
+
848 //TODO - develop C interface.
+
849 
+
850 #endif
+
851 
+
856 // Macro to replace extern __shared__ declarations
+
857 // to local variable definitions
+
858 #define HIP_DYNAMIC_SHARED(type, var) \
+
859  __attribute__((address_space(3))) type* var = \
+
860  (__attribute__((address_space(3))) type*)__get_dynamicgroupbaseptr(); \
+
861 
+
862 #define HIP_DYNAMIC_SHARED_ATTRIBUTE __attribute__((address_space(3)))
+
863 
+
864 #endif // __HCC__
+
865 
+
866 
+
871 //extern int HIP_PRINT_ENV ; ///< Print all HIP-related environment variables.
+
872 //extern int HIP_TRACE_API; ///< Trace HIP APIs.
+
873 //extern int HIP_LAUNCH_BLOCKING ; ///< Make all HIP APIs host-synchronous
+
874 
+
880 // End doxygen API:
+
888 #endif
+
__threadfence
__device__ void __threadfence(void)
threadfence makes wirtes visible to other threads running on same GPU.
+
host_defines.h
TODO-doc.
+
dim3
Definition: hip_runtime_api.h:151
+
__host__
#define __host__
Definition: host_defines.h:35
+
__threadfence_system
__device__ void __threadfence_system(void)
threadfence_system makes writes to pinned system memory visible on host CPU.
Definition: device_util.cpp:2600
+
char4
Definition: hip_vector_types.h:140
+
__threadfence_block
__device__ void __threadfence_block(void)
threadfence_block makes writes visible to threads running in same block.
+
ihipStream_t
Definition: hip_hcc.h:463
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h.html b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h.html new file mode 100644 index 0000000000..5de643b669 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h.html @@ -0,0 +1,589 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_runtime_api.h File Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+Classes | +Macros | +Typedefs | +Enumerations | +Functions
+
+
hip_runtime_api.h File Reference
+
+
+ +

Contains C function APIs for HIP runtime. This file does not use any HCC builtin or special language extensions (-hc mode) ; those functions in hip_runtime.h. +More...

+
#include <stdint.h>
+#include <stddef.h>
+#include <hip/hcc_detail/host_defines.h>
+#include <hip/hip_runtime_api.h>
+
+

Go to the source code of this file.

+ + + + +

+Classes

struct  dim3
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Macros

+#define HIP_LAUNCH_PARAM_BUFFER_POINTER   ((void*) 0x01)
 
+#define HIP_LAUNCH_PARAM_BUFFER_SIZE   ((void*) 0x02)
 
+#define HIP_LAUNCH_PARAM_END   ((void*) 0x03)
 
+#define hipIpcMemLazyEnablePeerAccess   0
 
#define hipStreamDefault   0x00
 Flags that can be used with hipStreamCreateWithFlags. More...
 
+#define hipStreamNonBlocking   0x01
 Stream does not implicitly synchronize with null stream.
 
#define hipEventDefault   0x0
 Flags that can be used with hipEventCreateWithFlags: More...
 
+#define hipEventBlockingSync   0x1
 Waiting will yield CPU. Power-friendly and usage-friendly but may increase latency.
 
+#define hipEventDisableTiming   0x2
 Disable event's capability to record timing information. May improve performance.
 
#define hipEventInterprocess   0x4
 Event can support IPC. More...
 
+#define hipHostMallocDefault   0x0
 Flags that can be used with hipHostMalloc.
 
+#define hipHostMallocPortable   0x1
 
+#define hipHostMallocMapped   0x2
 
+#define hipHostMallocWriteCombined   0x4
 
#define hipHostRegisterDefault   0x0
 Flags that can be used with hipHostRegister. More...
 
+#define hipHostRegisterPortable   0x1
 Memory is considered registered by all contexts. HIP only supports one context so this is always assumed true.
 
+#define hipHostRegisterMapped   0x2
 Map the allocation into the address space for the current device. The device pointer can be obtained with hipHostGetDevicePointer.
 
+#define hipHostRegisterIoMemory   0x4
 Not supported.
 
+#define hipDeviceScheduleAuto   0x0
 Automatically select between Spin and Yield.
 
+#define hipDeviceScheduleSpin   0x1
 Dedicate a CPU core to spin-wait. Provides lowest latency, but burns a CPU core and may consume more power.
 
+#define hipDeviceScheduleYield   0x2
 Yield the CPU to the operating system when waiting. May increase latency, but lowers power and is friendlier to other threads in the system.
 
+#define hipDeviceScheduleBlockingSync   0x4
 
+#define hipDeviceScheduleMask   0x7
 
+#define hipDeviceMapHost   0x8
 
+#define hipDeviceLmemResizeToMax   0x16
 
+#define USE_PEER_NON_UNIFIED   1
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Typedefs

+typedef struct ihipCtx_t * hipCtx_t
 
+typedef struct ihipDevice_t * hipDevice_t
 
+typedef struct ihipStream_t * hipStream_t
 
+typedef struct ihipIpcMemHandle_t * hipIpcMemHandle_t
 
+typedef struct
+ihipIpcEventHandle_t * 		hipIpcEventHandle_t
 
+typedef struct ihipModule_t * hipModule_t
 
+typedef struct ihipFunction_t * hipFunction_t
 
+typedef void * hipDeviceptr_t
 
+typedef struct ihipEvent_t * hipEvent_t
 
typedef enum hipFuncCache_t hipFuncCache_t
 
typedef enum hipSharedMemConfig hipSharedMemConfig
 
typedef struct dim3 dim3
 
typedef enum hipMemcpyKind hipMemcpyKind
 
typedef void(* hipStreamCallback_t )(hipStream_t stream, hipError_t status, void *userData)
 
+ + + + + + + + + +

+Enumerations

enum  hipLimit_t { hipLimitMallocHeapSize = 0x02 + }
 
enum  hipFuncCache_t { hipFuncCachePreferNone, +hipFuncCachePreferShared, +hipFuncCachePreferL1, +hipFuncCachePreferEqual + }
 
enum  hipSharedMemConfig { hipSharedMemBankSizeDefault, +hipSharedMemBankSizeFourByte, +hipSharedMemBankSizeEightByte + }
 
enum  hipMemcpyKind {
+  hipMemcpyHostToHost = 0, +hipMemcpyHostToDevice = 1, +hipMemcpyDeviceToHost = 2, +hipMemcpyDeviceToDevice =3, +
+  hipMemcpyDefault = 4 +
+ }
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipDeviceSynchronize (void)
 Waits on all active streams on current device. More...
 
hipError_t hipDeviceReset (void)
 The state of current device is discarded and updated to a fresh state. More...
 
hipError_t hipSetDevice (int deviceId)
 Set default device to be used for subsequent hip API calls from this thread. More...
 
hipError_t hipGetDevice (int *deviceId)
 Return the default device id for the calling host thread. More...
 
hipError_t hipGetDeviceCount (int *count)
 Return number of compute-capable devices. More...
 
hipError_t hipDeviceGetAttribute (int *pi, hipDeviceAttribute_t attr, int deviceId)
 Query for a specific device attribute. More...
 
hipError_t hipGetDeviceProperties (hipDeviceProp_t *prop, int deviceId)
 Returns device properties. More...
 
hipError_t hipDeviceSetCacheConfig (hipFuncCache_t cacheConfig)
 Set L1/Shared cache partition. More...
 
hipError_t hipDeviceGetCacheConfig (hipFuncCache_t *cacheConfig)
 Set Cache configuration for a specific function. More...
 
hipError_t hipDeviceGetLimit (size_t *pValue, hipLimit_t limit)
 Get Resource limits of current device. More...
 
hipError_t hipFuncSetCacheConfig (hipFuncCache_t config)
 Set Cache configuration for a specific function. More...
 
hipError_t hipDeviceGetSharedMemConfig (hipSharedMemConfig *pConfig)
 Returns bank width of shared memory for current device. More...
 
hipError_t hipDeviceSetSharedMemConfig (hipSharedMemConfig config)
 The bank width of shared memory on current device is set. More...
 
hipError_t hipSetDeviceFlags (unsigned flags)
 The current device behavior is changed according the flags passed. More...
 
hipError_t hipChooseDevice (int *device, const hipDeviceProp_t *prop)
 Device which matches hipDeviceProp_t is returned. More...
 
hipError_t hipGetLastError (void)
 Return last error returned by any HIP runtime API call and resets the stored error code to hipSuccess. More...
 
hipError_t hipPeekAtLastError (void)
 Return last error returned by any HIP runtime API call. More...
 
const char * hipGetErrorName (hipError_t hip_error)
 Return name of the specified error code in text form. More...
 
const char * hipGetErrorString (hipError_t hipError)
 Return handy text string message to explain the error which occurred. More...
 
hipError_t hipStreamCreate (hipStream_t *stream)
 Create an asynchronous stream. More...
 
hipError_t hipStreamCreateWithFlags (hipStream_t *stream, unsigned int flags)
 Create an asynchronous stream. More...
 
hipError_t hipStreamDestroy (hipStream_t stream)
 Destroys the specified stream. More...
 
hipError_t hipStreamQuery (hipStream_t stream)
 Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not. More...
 
hipError_t hipStreamSynchronize (hipStream_t stream)
 Wait for all commands in stream to complete. More...
 
hipError_t hipStreamWaitEvent (hipStream_t stream, hipEvent_t event, unsigned int flags)
 Make the specified compute stream wait for an event. More...
 
hipError_t hipStreamGetFlags (hipStream_t stream, unsigned int *flags)
 Return flags associated with this stream. More...
 
hipError_t hipStreamAddCallback (hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags)
 Adds a callback to be called on the host after all currently enqueued items in the stream have completed. For each cudaStreamAddCallback call, a callback will be executed exactly once. The callback will block later work in the stream until it is finished. More...
 
hipError_t hipEventCreateWithFlags (hipEvent_t *event, unsigned flags)
 Create an event with the specified flags. More...
 
hipError_t hipEventCreate (hipEvent_t *event)
 
hipError_t hipEventRecord (hipEvent_t event, hipStream_t stream)
 Record an event in the specified stream. More...
 
hipError_t hipEventDestroy (hipEvent_t event)
 Destroy the specified event. More...
 
hipError_t hipEventSynchronize (hipEvent_t event)
 Wait for an event to complete. More...
 
hipError_t hipEventElapsedTime (float *ms, hipEvent_t start, hipEvent_t stop)
 Return the elapsed time between two events. More...
 
hipError_t hipEventQuery (hipEvent_t event)
 Query event status. More...
 
hipError_t hipPointerGetAttributes (hipPointerAttribute_t *attributes, void *ptr)
 Return attributes for the specified pointer. More...
 
hipError_t hipMalloc (void **ptr, size_t size)
 Allocate memory on the default accelerator. More...
 
hipError_t hipMallocHost (void **ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead")))
 Allocate pinned host memory [Deprecated]. More...
 
hipError_t hipHostMalloc (void **ptr, size_t size, unsigned int flags)
 Allocate device accessible page locked host memory. More...
 
hipError_t hipHostAlloc (void **ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead")))
 Allocate device accessible page locked host memory [Deprecated]. More...
 
hipError_t hipHostGetDevicePointer (void **devPtr, void *hstPtr, unsigned int flags)
 Get Device pointer from Host Pointer allocated through hipHostMalloc. More...
 
hipError_t hipHostGetFlags (unsigned int *flagsPtr, void *hostPtr)
 Return flags associated with host pointer. More...
 
hipError_t hipHostRegister (void *hostPtr, size_t sizeBytes, unsigned int flags)
 Register host memory so it can be accessed from the current device. More...
 
hipError_t hipHostUnregister (void *hostPtr)
 Un-register host pointer. More...
 
hipError_t hipMallocPitch (void **ptr, size_t *pitch, size_t width, size_t height)
 
hipError_t hipFree (void *ptr)
 Free memory allocated by the hcc hip memory allocation API. This API performs an implicit hipDeviceSynchronize() call. If pointer is NULL, the hip runtime is initialized and hipSuccess is returned. More...
 
hipError_t hipFreeHost (void *ptr) __attribute__((deprecated("use hipHostFree instead")))
 Free memory allocated by the hcc hip host memory allocation API. [Deprecated]. More...
 
hipError_t hipHostFree (void *ptr)
 Free memory allocated by the hcc hip host memory allocation API This API performs an implicit hipDeviceSynchronize() call. If pointer is NULL, the hip runtime is initialized and hipSuccess is returned. More...
 
hipError_t hipMemcpy (void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind)
 Copy data from src to dst. More...
 
hipError_t hipMemcpyHtoD (hipDeviceptr_t dst, void *src, size_t sizeBytes)
 Copy data from Host to Device. More...
 
hipError_t hipMemcpyDtoH (void *dst, hipDeviceptr_t src, size_t sizeBytes)
 Copy data from Device to Host. More...
 
hipError_t hipMemcpyDtoD (hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes)
 Copy data from Device to Device. More...
 
hipError_t hipMemcpyHtoDAsync (hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream)
 Copy data from Host to Device asynchronously. More...
 
hipError_t hipMemcpyDtoHAsync (void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
 Copy data from Device to Host asynchronously. More...
 
hipError_t hipMemcpyDtoDAsync (hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
 Copy data from Device to Device asynchronously. More...
 
hipError_t hipMemcpyToSymbol (const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind)
 Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset bytes from the start of symbol symbol. More...
 
hipError_t hipMemcpyToSymbolAsync (const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream)
 Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset bytes from the start of symbol symbol. More...
 
hipError_t hipMemcpyAsync (void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream)
 Copy data from src to dst asynchronously. More...
 
hipError_t hipMemset (void *dst, int value, size_t sizeBytes)
 Copy data from src to dst asynchronously. More...
 
hipError_t hipMemsetAsync (void *dst, int value, size_t sizeBytes, hipStream_t stream)
 Fills the first sizeBytes bytes of the memory area pointed to by dev with the constant byte value value. More...
 
hipError_t hipMemGetInfo (size_t *free, size_t *total)
 Query memory info. Return snapshot of free memory, and total allocatable memory on the device. More...
 
hipError_t hipDeviceCanAccessPeer (int *canAccessPeer, int deviceId, int peerDeviceId)
 Determine if a device can access a peer's memory. More...
 
hipError_t hipDeviceEnablePeerAccess (int peerDeviceId, unsigned int flags)
 Enable direct access from current device's virtual address space to memory allocations physically located on a peer device. More...
 
hipError_t hipDeviceDisablePeerAccess (int peerDeviceId)
 Disable direct access from current device's virtual address space to memory allocations physically located on a peer device. More...
 
hipError_t hipMemGetAddressRange (hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr)
 Get information on memory allocations. More...
 
hipError_t hipMemcpyPeer (void *dst, int dstDeviceId, const void *src, int srcDeviceId, size_t sizeBytes)
 Copies memory from one device to memory on another device. More...
 
hipError_t hipMemcpyPeerAsync (void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream)
 Copies memory from one device to memory on another device. More...
 
hipError_t hipInit (unsigned int flags)
 Explicitly initializes the HIP runtime. More...
 
hipError_t hipCtxCreate (hipCtx_t *ctx, unsigned int flags, hipDevice_t device)
 Create a context and set it as current/ default context. More...
 
hipError_t hipCtxDestroy (hipCtx_t ctx)
 Destroy a HIP context. More...
 
hipError_t hipCtxPopCurrent (hipCtx_t *ctx)
 Pop the current/default context and return the popped context. More...
 
hipError_t hipCtxPushCurrent (hipCtx_t ctx)
 Push the context to be set as current/ default context. More...
 
hipError_t hipCtxSetCurrent (hipCtx_t ctx)
 Set the passed context as current/default. More...
 
hipError_t hipCtxGetCurrent (hipCtx_t *ctx)
 Get the handle of the current/ default context. More...
 
hipError_t hipCtxGetDevice (hipDevice_t *device)
 Get the handle of the device associated with current/default context. More...
 
hipError_t hipCtxGetApiVersion (hipCtx_t ctx, int *apiVersion)
 Returns the approximate HIP api version. More...
 
hipError_t hipCtxGetCacheConfig (hipFuncCache_t *cacheConfig)
 Set Cache configuration for a specific function. More...
 
hipError_t hipCtxSetCacheConfig (hipFuncCache_t cacheConfig)
 Set L1/Shared cache partition. More...
 
hipError_t hipCtxSetSharedMemConfig (hipSharedMemConfig config)
 Set Shared memory bank configuration. More...
 
hipError_t hipCtxGetSharedMemConfig (hipSharedMemConfig *pConfig)
 Get Shared memory bank configuration. More...
 
hipError_t hipCtxSynchronize (void)
 Blocks until the default context has completed all preceding requested tasks. More...
 
hipError_t hipCtxGetFlags (unsigned int *flags)
 Return flags used for creating default context. More...
 
hipError_t hipCtxEnablePeerAccess (hipCtx_t peerCtx, unsigned int flags)
 Enables direct access to memory allocations in a peer context. More...
 
hipError_t hipCtxDisablePeerAccess (hipCtx_t peerCtx)
 Disable direct access from current context's virtual address space to memory allocations physically located on a peer context.Disables direct access to memory allocations in a peer context and unregisters any registered allocations. More...
 
hipError_t hipDeviceGet (hipDevice_t *device, int ordinal)
 Returns a handle to a compute device. More...
 
hipError_t hipDeviceComputeCapability (int *major, int *minor, hipDevice_t device)
 Returns the compute capability of the device. More...
 
hipError_t hipDeviceGetName (char *name, int len, hipDevice_t device)
 Returns an identifer string for the device. More...
 
hipError_t hipDeviceGetPCIBusId (char *pciBusId, int len, int device)
 Returns a PCI Bus Id string for the device, overloaded to take int device ID. More...
 
hipError_t hipDeviceGetByPCIBusId (int *device, const int *pciBusId)
 Returns a handle to a compute device. More...
 
hipError_t hipDeviceTotalMem (size_t *bytes, hipDevice_t device)
 Returns the total amount of memory on the device. More...
 
hipError_t hipDriverGetVersion (int *driverVersion)
 Returns the approximate HIP driver version. More...
 
hipError_t hipRuntimeGetVersion (int *runtimeVersion)
 Returns the approximate HIP Runtime version. More...
 
hipError_t hipModuleLoad (hipModule_t *module, const char *fname)
 Loads code object from file into a hipModule_t. More...
 
hipError_t hipModuleUnload (hipModule_t module)
 Frees the module. More...
 
hipError_t hipModuleGetFunction (hipFunction_t *function, hipModule_t module, const char *kname)
 Function with kname will be extracted if present in module. More...
 
hipError_t hipModuleGetGlobal (hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name)
 returns device memory pointer and size of the kernel present in the module with symbol name More...
 
hipError_t hipModuleLoadData (hipModule_t *module, const void *image)
 builds module from code object which resides in host memory. Image is pointer to that location. More...
 
hipError_t hipModuleLaunchKernel (hipFunction_t f, unsigned int gridDimX, unsigned int gridDimY, unsigned int gridDimZ, unsigned int blockDimX, unsigned int blockDimY, unsigned int blockDimZ, unsigned int sharedMemBytes, hipStream_t stream, void **kernelParams, void **extra)
 launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra More...
 
hipError_t hipProfilerStart ()
 Start recording of profiling information When using this API, start the profiler with profiling disabled. (–startdisabled) More...
 
hipError_t hipProfilerStop ()
 Stop recording of profiling information. When using this API, start the profiler with profiling disabled. (–startdisabled) More...
 
hipError_t hipIpcGetMemHandle (hipIpcMemHandle_t *handle, void *devPtr)
 Gets an interprocess memory handle for an existing device memory allocation. More...
 
hipError_t hipIpcOpenMemHandle (void **devPtr, hipIpcMemHandle_t handle, unsigned int flags)
 Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process. More...
 
hipError_t hipIpcCloseMemHandle (void *devPtr)
 Close memory mapped with hipIpcOpenMemHandle. More...
 
+

Detailed Description

+

Contains C function APIs for HIP runtime. This file does not use any HCC builtin or special language extensions (-hc mode) ; those functions in hip_runtime.h.

+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h_source.html b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h_source.html new file mode 100644 index 0000000000..16efd64a38 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__runtime__api_8h_source.html @@ -0,0 +1,718 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_runtime_api.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_runtime_api.h
+
+
+Go to the documentation of this file.
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 //#pragma once
+
24 #ifndef HIP_RUNTIME_API_H
+
25 #define HIP_RUNTIME_API_H
+
26 
+
31 #include <stdint.h>
+
32 #include <stddef.h>
+
33 
+
34 #include <hip/hcc_detail/host_defines.h>
+
35 #include <hip/hip_runtime_api.h>
+
36 //#include "hip/hip_hcc.h"
+
37 
+
38 #if defined (__HCC__) && (__hcc_workweek__ < 16155)
+
39 #error("This version of HIP requires a newer version of HCC.");
+
40 #endif
+
41 
+
42 #define HIP_LAUNCH_PARAM_BUFFER_POINTER ((void*) 0x01)
+
43 #define HIP_LAUNCH_PARAM_BUFFER_SIZE ((void*) 0x02)
+
44 #define HIP_LAUNCH_PARAM_END ((void*) 0x03)
+
45 
+
46 // Structure definitions:
+
47 #ifdef __cplusplus
+
48 extern "C" {
+
49 #endif
+
50 
+
51 //---
+
52 //API-visible structures
+
53 typedef struct ihipCtx_t *hipCtx_t;
+
54 
+
55 // Note many APIs also use integer deviceIds as an alternative to the device pointer:
+
56 typedef struct ihipDevice_t *hipDevice_t;
+
57 
+
58 typedef struct ihipStream_t *hipStream_t;
+
59 
+
60 //TODO: IPC implementation
+
61 
+
62 #define hipIpcMemLazyEnablePeerAccess 0
+
63 
+
64 typedef struct ihipIpcMemHandle_t *hipIpcMemHandle_t;
+
65 
+
66 //TODO: IPC event handle currently unsupported
+
67 struct ihipIpcEventHandle_t;
+
68 typedef struct ihipIpcEventHandle_t *hipIpcEventHandle_t;
+
69 
+
70 
+
71 //END TODO
+
72 
+
73 typedef struct ihipModule_t *hipModule_t;
+
74 
+
75 typedef struct ihipFunction_t *hipFunction_t;
+
76 
+
77 typedef void* hipDeviceptr_t;
+
78 
+
79 typedef struct ihipEvent_t *hipEvent_t;
+
80 
+
81 enum hipLimit_t
+
82 {
+
83  hipLimitMallocHeapSize = 0x02,
+
84 };
+
85 
+
90 #define hipStreamDefault 0x00
+
92 #define hipStreamNonBlocking 0x01
+
93 
+
94 
+
96 #define hipEventDefault 0x0
+
97 #define hipEventBlockingSync 0x1
+
98 #define hipEventDisableTiming 0x2
+
99 #define hipEventInterprocess 0x4
+
100 
+
101 
+
103 #define hipHostMallocDefault 0x0
+
104 #define hipHostMallocPortable 0x1
+
105 #define hipHostMallocMapped 0x2
+
106 #define hipHostMallocWriteCombined 0x4
+
107 
+
109 #define hipHostRegisterDefault 0x0
+
110 #define hipHostRegisterPortable 0x1
+
111 #define hipHostRegisterMapped 0x2
+
112 #define hipHostRegisterIoMemory 0x4
+
113 
+
114 
+
115 #define hipDeviceScheduleAuto 0x0
+
116 #define hipDeviceScheduleSpin 0x1
+
117 #define hipDeviceScheduleYield 0x2
+
118 #define hipDeviceScheduleBlockingSync 0x4
+
119 #define hipDeviceScheduleMask 0x7
+
120 
+
121 #define hipDeviceMapHost 0x8
+
122 #define hipDeviceLmemResizeToMax 0x16
+
123 
+
124 
+
128 typedef enum hipFuncCache_t {
+
129  hipFuncCachePreferNone,
+
130  hipFuncCachePreferShared,
+
131  hipFuncCachePreferL1,
+
132  hipFuncCachePreferEqual,
+
133 } hipFuncCache_t;
+
134 
+
135 
+
139 typedef enum hipSharedMemConfig {
+
140  hipSharedMemBankSizeDefault,
+
141  hipSharedMemBankSizeFourByte,
+
142  hipSharedMemBankSizeEightByte
+
143 } hipSharedMemConfig;
+
144 
+
145 
+
146 
+
151 typedef struct dim3 {
+
152  uint32_t x;
+
153  uint32_t y;
+
154  uint32_t z;
+
155 #ifdef __cplusplus
+
156  dim3(uint32_t _x=1, uint32_t _y=1, uint32_t _z=1) : x(_x), y(_y), z(_z) {};
+
157 #endif
+
158 } dim3;
+
159 
+
160 
+
165 typedef enum hipMemcpyKind {
+
166  hipMemcpyHostToHost = 0
+
167  ,hipMemcpyHostToDevice = 1
+
168  ,hipMemcpyDeviceToHost = 2
+
169  ,hipMemcpyDeviceToDevice =3
+
170  ,hipMemcpyDefault = 4,
+
171 } hipMemcpyKind;
+
172 
+
173 
+
174 
+
175 
+
176 // Doxygen end group GlobalDefs
+
180 //-------------------------------------------------------------------------------------------------
+
181 
+
182 
+
183 // The handle allows the async commands to use the stream even if the parent hipStream_t goes out-of-scope.
+
184 //typedef class ihipStream_t * hipStream_t;
+
185 
+
186 
+
187 /*
+
188  * Opaque structure allows the true event (pointed at by the handle) to remain "live" even if the surrounding hipEvent_t goes out-of-scope.
+
189  * This is handy for cases where the hipEvent_t goes out-of-scope but the true event is being written by some async queue or device */
+
190 //typedef struct hipEvent_t {
+
191 // struct ihipEvent_t *_handle;
+
192 //} hipEvent_t;
+
193 
+
194 
+
195 
+
196 
+
197 
+
198 
+
199 
+
226 hipError_t hipDeviceSynchronize(void);
+
227 
+
228 
+
229 
+
240 hipError_t hipDeviceReset(void) ;
+
241 
+
242 
+
269 hipError_t hipSetDevice(int deviceId);
+
270 
+
271 
+
285 hipError_t hipGetDevice(int *deviceId);
+
286 
+
287 
+
299 hipError_t hipGetDeviceCount(int *count);
+
300 
+
310 hipError_t hipDeviceGetAttribute(int* pi, hipDeviceAttribute_t attr, int deviceId);
+
311 
+
325 hipError_t hipGetDeviceProperties(hipDeviceProp_t* prop, int deviceId);
+
326 
+
327 
+
337 hipError_t hipDeviceSetCacheConfig ( hipFuncCache_t cacheConfig );
+
338 
+
339 
+
349 hipError_t hipDeviceGetCacheConfig ( hipFuncCache_t *cacheConfig );
+
350 
+
361 hipError_t hipDeviceGetLimit(size_t *pValue, hipLimit_t limit);
+
362 
+
363 
+
373 hipError_t hipFuncSetCacheConfig ( hipFuncCache_t config );
+
374 
+
385 hipError_t hipDeviceGetSharedMemConfig ( hipSharedMemConfig * pConfig );
+
386 
+
387 
+
398 hipError_t hipDeviceSetSharedMemConfig ( hipSharedMemConfig config );
+
399 
+
419 hipError_t hipSetDeviceFlags ( unsigned flags);
+
420 
+
429 hipError_t hipChooseDevice(int *device, const hipDeviceProp_t* prop);
+
430 
+
431 // end doxygen Device
+
453 hipError_t hipGetLastError(void);
+
454 
+
455 
+
466 hipError_t hipPeekAtLastError(void);
+
467 
+
468 
+
469 
+
478 const char *hipGetErrorName(hipError_t hip_error);
+
479 
+
480 
+
491 const char *hipGetErrorString(hipError_t hipError);
+
492 
+
493 // end doxygen Error
+
529 hipError_t hipStreamCreate(hipStream_t *stream);
+
530 
+
531 
+
548 hipError_t hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags);
+
549 
+
550 
+
566 hipError_t hipStreamDestroy(hipStream_t stream);
+
567 
+
568 
+
581 hipError_t hipStreamQuery(hipStream_t stream);
+
582 
+
583 
+
598 hipError_t hipStreamSynchronize(hipStream_t stream);
+
599 
+
600 
+
617 hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags);
+
618 
+
619 
+
620 
+
634 hipError_t hipStreamGetFlags(hipStream_t stream, unsigned int *flags);
+
635 
+
639 typedef void(* hipStreamCallback_t)(hipStream_t stream, hipError_t status, void* userData);
+
640 
+
655 hipError_t hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags);
+
656 
+
657 
+
658 // end doxygen Stream
+
689 hipError_t hipEventCreateWithFlags(hipEvent_t* event, unsigned flags);
+
690 
+
691 
+
701 hipError_t hipEventCreate(hipEvent_t* event);
+
702 
+
703 
+
728 #ifdef __cplusplus
+
729 hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream = NULL);
+
730 #else
+
731 hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream);
+
732 #endif
+
733 
+
747 hipError_t hipEventDestroy(hipEvent_t event);
+
748 
+
749 
+
764 hipError_t hipEventSynchronize(hipEvent_t event);
+
765 
+
766 
+
793 hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop);
+
794 
+
795 
+
808 hipError_t hipEventQuery(hipEvent_t event) ;
+
809 
+
810 
+
811 // end doxygen Events
+
843 hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void* ptr);
+
844 
+
855 hipError_t hipMalloc(void** ptr, size_t size) ;
+
856 
+
867 hipError_t hipMallocHost(void** ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead"))) ;
+
868 
+
880 hipError_t hipHostMalloc(void** ptr, size_t size, unsigned int flags) ;
+
881 
+
893 hipError_t hipHostAlloc(void** ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead"))) ;
+
894 
+
906 hipError_t hipHostGetDevicePointer(void** devPtr, void* hstPtr, unsigned int flags) ;
+
907 
+
917 hipError_t hipHostGetFlags(unsigned int* flagsPtr, void* hostPtr) ;
+
918 
+
951 hipError_t hipHostRegister(void* hostPtr, size_t sizeBytes, unsigned int flags) ;
+
952 
+
961 hipError_t hipHostUnregister(void* hostPtr) ;
+
962 
+
978 hipError_t hipMallocPitch(void** ptr, size_t* pitch, size_t width, size_t height);
+
979 
+
991 hipError_t hipFree(void* ptr);
+
992 
+
1002 hipError_t hipFreeHost(void* ptr) __attribute__((deprecated("use hipHostFree instead")));
+
1003 
+
1015 hipError_t hipHostFree(void* ptr);
+
1016 
+
1038 hipError_t hipMemcpy(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind kind);
+
1039 
+
1051 hipError_t hipMemcpyHtoD(hipDeviceptr_t dst, void* src, size_t sizeBytes);
+
1052 
+
1064 hipError_t hipMemcpyDtoH(void* dst, hipDeviceptr_t src, size_t sizeBytes);
+
1065 
+
1077 hipError_t hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes);
+
1078 
+
1090 hipError_t hipMemcpyHtoDAsync(hipDeviceptr_t dst, void* src, size_t sizeBytes, hipStream_t stream);
+
1091 
+
1103 hipError_t hipMemcpyDtoHAsync(void* dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream);
+
1104 
+
1116 hipError_t hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream);
+
1117 
+
1118 
+
1135 hipError_t hipMemcpyToSymbol(const char* symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind);
+
1136 
+
1137 
+
1155 hipError_t hipMemcpyToSymbolAsync(const char* symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream);
+
1156 
+
1157 
+
1158 
+
1181 #if __cplusplus
+
1182 hipError_t hipMemcpyAsync(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream=0);
+
1183 #else
+
1184 hipError_t hipMemcpyAsync(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream);
+
1185 #endif
+
1186 
+
1199 hipError_t hipMemset(void* dst, int value, size_t sizeBytes );
+
1200 
+
1201 
+
1215 #if __cplusplus
+
1216 hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t stream = 0 );
+
1217 #else
+
1218 hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t stream);
+
1219 #endif
+
1220 
+
1229 hipError_t hipMemGetInfo (size_t * free, size_t * total) ;
+
1230 
+
1231 // doxygen end Memory
+
1264 hipError_t hipDeviceCanAccessPeer (int* canAccessPeer, int deviceId, int peerDeviceId);
+
1265 
+
1266 
+
1282 hipError_t hipDeviceEnablePeerAccess (int peerDeviceId, unsigned int flags);
+
1283 
+
1284 
+
1295 hipError_t hipDeviceDisablePeerAccess (int peerDeviceId);
+
1296 
+
1308 hipError_t hipMemGetAddressRange ( hipDeviceptr_t* pbase, size_t* psize, hipDeviceptr_t dptr );
+
1309 
+
1310 #ifndef USE_PEER_NON_UNIFIED
+
1311 #define USE_PEER_NON_UNIFIED 1
+
1312 #endif
+
1313 
+
1314 #if USE_PEER_NON_UNIFIED==1
+
1315 
+
1327 hipError_t hipMemcpyPeer (void* dst, int dstDeviceId, const void* src, int srcDeviceId, size_t sizeBytes);
+
1328 
+
1341 #if __cplusplus
+
1342 hipError_t hipMemcpyPeerAsync ( void* dst, int dstDeviceId, const void* src, int srcDevice, size_t sizeBytes, hipStream_t stream=0 );
+
1343 #else
+
1344 hipError_t hipMemcpyPeerAsync(void* dst, int dstDevice, const void* src, int srcDevice, size_t sizeBytes, hipStream_t stream);
+
1345 #endif
+
1346 #endif
+
1347 
+
1348 
+
1349 // doxygen end PeerToPeer
+
1368 // TODO-ctx - more description on error codes.
+
1369 hipError_t hipInit(unsigned int flags) ;
+
1370 
+
1371 
+
1390 hipError_t hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device);
+
1391 
+
1401 hipError_t hipCtxDestroy(hipCtx_t ctx);
+
1402 
+
1412 hipError_t hipCtxPopCurrent(hipCtx_t* ctx);
+
1413 
+
1423 hipError_t hipCtxPushCurrent(hipCtx_t ctx);
+
1424 
+
1434 hipError_t hipCtxSetCurrent(hipCtx_t ctx);
+
1435 
+
1445 hipError_t hipCtxGetCurrent(hipCtx_t* ctx);
+
1446 
+
1457 hipError_t hipCtxGetDevice(hipDevice_t *device);
+
1458 
+
1475 hipError_t hipCtxGetApiVersion (hipCtx_t ctx,int *apiVersion);
+
1476 
+
1488 hipError_t hipCtxGetCacheConfig ( hipFuncCache_t *cacheConfig );
+
1489 
+
1501 hipError_t hipCtxSetCacheConfig ( hipFuncCache_t cacheConfig );
+
1502 
+
1514 hipError_t hipCtxSetSharedMemConfig ( hipSharedMemConfig config );
+
1515 
+
1527 hipError_t hipCtxGetSharedMemConfig ( hipSharedMemConfig * pConfig );
+
1528 
+
1538 hipError_t hipCtxSynchronize ( void );
+
1539 
+
1549 hipError_t hipCtxGetFlags ( unsigned int* flags );
+
1550 
+
1567 hipError_t hipCtxEnablePeerAccess (hipCtx_t peerCtx, unsigned int flags);
+
1568 
+
1581 hipError_t hipCtxDisablePeerAccess (hipCtx_t peerCtx);
+
1582 
+
1583 // doxygen end Context Management
+
1595 hipError_t hipDeviceGet(hipDevice_t *device, int ordinal);
+
1596 
+
1605 hipError_t hipDeviceComputeCapability(int *major,int *minor,hipDevice_t device);
+
1606 
+
1615 hipError_t hipDeviceGetName(char *name,int len,hipDevice_t device);
+
1616 
+
1625 hipError_t hipDeviceGetPCIBusId (char *pciBusId,int len,int device);
+
1626 
+
1627 
+
1635 hipError_t hipDeviceGetByPCIBusId ( int* device,const int* pciBusId );
+
1636 
+
1637 
+
1645 hipError_t hipDeviceTotalMem (size_t *bytes,hipDevice_t device);
+
1646 
+
1662 hipError_t hipDriverGetVersion(int *driverVersion) ;
+
1663 
+
1676 hipError_t hipRuntimeGetVersion(int *runtimeVersion) ;
+
1677 
+
1688 hipError_t hipModuleLoad(hipModule_t *module, const char *fname);
+
1689 
+
1700 hipError_t hipModuleUnload(hipModule_t module);
+
1701 
+
1711 hipError_t hipModuleGetFunction(hipFunction_t *function, hipModule_t module, const char *kname);
+
1712 
+
1723 hipError_t hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name);
+
1724 
+
1725 
+
1734 hipError_t hipModuleLoadData(hipModule_t *module, const void *image);
+
1735 
+
1736 
+
1757 hipError_t hipModuleLaunchKernel(hipFunction_t f,
+
1758  unsigned int gridDimX,
+
1759  unsigned int gridDimY,
+
1760  unsigned int gridDimZ,
+
1761  unsigned int blockDimX,
+
1762  unsigned int blockDimY,
+
1763  unsigned int blockDimZ,
+
1764  unsigned int sharedMemBytes,
+
1765  hipStream_t stream,
+
1766  void **kernelParams,
+
1767  void **extra) ;
+
1768 
+
1769 // doxygen end Version Management
+
1787 // TODO - expand descriptions:
+
1793 hipError_t hipProfilerStart();
+
1794 
+
1795 
+
1801 hipError_t hipProfilerStop();
+
1802 
+
1803 
+
1808 //TODO: implement IPC apis
+
1809 
+
1835 hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr);
+
1836 
+
1873 hipError_t hipIpcOpenMemHandle(void **devPtr,
+
1874  hipIpcMemHandle_t handle, unsigned int flags);
+
1875 
+
1894 hipError_t hipIpcCloseMemHandle(void *devPtr);
+
1895 
+
1896 
+
1897 // hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t* handle, void* devPtr);
+
1898 // hipError_t hipIpcCloseMemHandle(void *devPtr);
+
1899 // // hipError_t hipIpcOpenEventHandle(hipEvent_t* event, hipIpcEventHandle_t handle);
+
1900 // hipError_t hipIpcOpenMemHandle(void** devPtr, hipIpcMemHandle_t handle, unsigned int flags);
+
1901 
+
1902 
+
1903 #ifdef __cplusplus
+
1904 } /* extern "c" */
+
1905 #endif
+
1906 
+
1907 #ifdef __cplusplus
+
1908 
+
1916 hipError_t hipDeviceGetPCIBusId (char *pciBusId,int len,hipDevice_t device);
+
1917 #endif
+
1918 
+
1936 // end-group HCC_Specific
+
1943 // doxygen end HIP API
+
1948 #endif
+
hipHostFree
hipError_t hipHostFree(void *ptr)
Free memory allocated by the hcc hip host memory allocation API This API performs an implicit hipDevi...
Definition: hip_memory.cpp:967
+
hipFuncCachePreferL1
prefer larger L1 cache and smaller shared memory
Definition: hip_runtime_api.h:131
+
hipModuleGetFunction
hipError_t hipModuleGetFunction(hipFunction_t *function, hipModule_t module, const char *kname)
Function with kname will be extracted if present in module.
Definition: hip_module.cpp:234
+
hipCtxDisablePeerAccess
hipError_t hipCtxDisablePeerAccess(hipCtx_t peerCtx)
Disable direct access from current context's virtual address space to memory allocations physically l...
Definition: hip_peer.cpp:222
+
hipDeviceGetByPCIBusId
hipError_t hipDeviceGetByPCIBusId(int *device, const int *pciBusId)
Returns a handle to a compute device.
Definition: hip_device.cpp:376
+
hipDeviceGetCacheConfig
hipError_t hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig)
Set Cache configuration for a specific function.
Definition: hip_device.cpp:88
+
hipMemcpyAsync
hipError_t hipMemcpyAsync(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream)
Copy data from src to dst asynchronously.
Definition: hip_memory.cpp:635
+
hipDriverGetVersion
hipError_t hipDriverGetVersion(int *driverVersion)
Returns the approximate HIP driver version.
Definition: hip_context.cpp:82
+
hipPeekAtLastError
hipError_t hipPeekAtLastError(void)
Return last error returned by any HIP runtime API call.
Definition: hip_error.cpp:42
+
ihipDevice_t
Definition: hip_hcc.h:575
+
dim3
struct dim3 dim3
+
hipDeviceComputeCapability
hipError_t hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device)
Returns the compute capability of the device.
Definition: hip_device.cpp:320
+
hipMemGetAddressRange
hipError_t hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr)
Get information on memory allocations.
Definition: hip_memory.cpp:1023
+
host_defines.h
TODO-doc.
+
hipMallocPitch
hipError_t hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height)
Definition: hip_memory.cpp:233
+
hipFuncSetCacheConfig
hipError_t hipFuncSetCacheConfig(hipFuncCache_t config)
Set Cache configuration for a specific function.
Definition: hip_device.cpp:115
+
ihipIpcMemHandle_t
Definition: hip_hcc.h:375
+
hipMemcpyToSymbol
hipError_t hipMemcpyToSymbol(const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind)
Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset...
Definition: hip_memory.cpp:432
+
hipStreamAddCallback
hipError_t hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags)
Adds a callback to be called on the host after all currently enqueued items in the stream have comple...
Definition: hip_stream.cpp:202
+
dim3::x
uint32_t x
x
Definition: hip_runtime_api.h:152
+
hipMemcpyHostToDevice
Host-to-Device Copy.
Definition: hip_runtime_api.h:167
+
hipDeviceEnablePeerAccess
hipError_t hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags)
Enable direct access from current device's virtual address space to memory allocations physically loc...
Definition: hip_peer.cpp:194
+
hipCtxPopCurrent
hipError_t hipCtxPopCurrent(hipCtx_t *ctx)
Pop the current/default context and return the popped context.
Definition: hip_context.cpp:132
+
hipGetErrorString
const char * hipGetErrorString(hipError_t hipError)
Return handy text string message to explain the error which occurred.
Definition: hip_error.cpp:57
+
hipDeviceGetSharedMemConfig
hipError_t hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig)
Returns bank width of shared memory for current device.
Definition: hip_device.cpp:133
+
hipPointerAttribute_t
Definition: hip_runtime_api.h:125
+
hipMemcpyDeviceToHost
Device-to-Host Copy.
Definition: hip_runtime_api.h:168
+
hipFuncCachePreferEqual
prefer equal size L1 cache and shared memory
Definition: hip_runtime_api.h:132
+
hipHostGetDevicePointer
hipError_t hipHostGetDevicePointer(void **devPtr, void *hstPtr, unsigned int flags)
Get Device pointer from Host Pointer allocated through hipHostMalloc.
+
hipEventSynchronize
hipError_t hipEventSynchronize(hipEvent_t event)
Wait for an event to complete.
Definition: hip_event.cpp:113
+
ihipFunction_t
Definition: hip_hcc.h:383
+
hipMemcpyHtoD
hipError_t hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes)
Copy data from Host to Device.
Definition: hip_memory.cpp:521
+
hipModuleUnload
hipError_t hipModuleUnload(hipModule_t module)
Frees the module.
Definition: hip_module.cpp:166
+
hipSetDeviceFlags
hipError_t hipSetDeviceFlags(unsigned flags)
The current device behavior is changed according the flags passed.
+
hipEventQuery
hipError_t hipEventQuery(hipEvent_t event)
Query event status.
Definition: hip_event.cpp:179
+
hipDeviceDisablePeerAccess
hipError_t hipDeviceDisablePeerAccess(int peerDeviceId)
Disable direct access from current device's virtual address space to memory allocations physically lo...
Definition: hip_peer.cpp:186
+
hipCtxGetSharedMemConfig
hipError_t hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig)
Get Shared memory bank configuration.
Definition: hip_context.cpp:247
+
hipCtxCreate
hipError_t hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device)
Create a context and set it as current/ default context.
Definition: hip_context.cpp:56
+
hipCtxSetSharedMemConfig
hipError_t hipCtxSetSharedMemConfig(hipSharedMemConfig config)
Set Shared memory bank configuration.
Definition: hip_context.cpp:238
+
hipCtxSetCurrent
hipError_t hipCtxSetCurrent(hipCtx_t ctx)
Set the passed context as current/default.
Definition: hip_context.cpp:179
+
hipMallocHost
hipError_t hipMallocHost(void **ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead")))
Allocate pinned host memory [Deprecated].
Definition: hip_memory.cpp:222
+
hipIpcGetMemHandle
hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr)
Gets an interprocess memory handle for an existing device memory allocation.
Definition: hip_memory.cpp:1042
+
hipHostMalloc
hipError_t hipHostMalloc(void **ptr, size_t size, unsigned int flags)
Allocate device accessible page locked host memory.
Definition: hip_memory.cpp:161
+
ihipCtx_t
Definition: hip_hcc.h:675
+
hipModuleLoadData
hipError_t hipModuleLoadData(hipModule_t *module, const void *image)
builds module from code object which resides in host memory. Image is pointer to that location...
Definition: hip_module.cpp:416
+
hipDeviceAttribute_t
hipDeviceAttribute_t
Definition: hip_runtime_api.h:222
+
hipEventDestroy
hipError_t hipEventDestroy(hipEvent_t event)
Destroy the specified event.
Definition: hip_event.cpp:100
+
hipStreamCreateWithFlags
hipError_t hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags)
Create an asynchronous stream.
Definition: hip_stream.cpp:63
+
hipChooseDevice
hipError_t hipChooseDevice(int *device, const hipDeviceProp_t *prop)
Device which matches hipDeviceProp_t is returned.
Definition: hip_device.cpp:395
+
hipCtxSetCacheConfig
hipError_t hipCtxSetCacheConfig(hipFuncCache_t cacheConfig)
Set L1/Shared cache partition.
Definition: hip_context.cpp:229
+
hipModuleLaunchKernel
hipError_t hipModuleLaunchKernel(hipFunction_t f, unsigned int gridDimX, unsigned int gridDimY, unsigned int gridDimZ, unsigned int blockDimX, unsigned int blockDimY, unsigned int blockDimZ, unsigned int sharedMemBytes, hipStream_t stream, void **kernelParams, void **extra)
launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelp...
+
dim3
Definition: hip_runtime_api.h:151
+
dim3::y
uint32_t y
y
Definition: hip_runtime_api.h:153
+
hipModuleLoad
hipError_t hipModuleLoad(hipModule_t *module, const char *fname)
Loads code object from file into a hipModule_t.
Definition: hip_module.cpp:106
+
hipEventCreateWithFlags
hipError_t hipEventCreateWithFlags(hipEvent_t *event, unsigned flags)
Create an event with the specified flags.
Definition: hip_event.cpp:54
+
hipEventElapsedTime
hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop)
Return the elapsed time between two events.
Definition: hip_event.cpp:137
+
hipDeviceSetCacheConfig
hipError_t hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig)
Set L1/Shared cache partition.
Definition: hip_device.cpp:79
+
hipDeviceGetLimit
hipError_t hipDeviceGetLimit(size_t *pValue, hipLimit_t limit)
Get Resource limits of current device.
Definition: hip_device.cpp:101
+
hipDeviceCanAccessPeer
hipError_t hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId)
Determine if a device can access a peer's memory.
Definition: hip_peer.cpp:179
+
hipGetDeviceCount
hipError_t hipGetDeviceCount(int *count)
Return number of compute-capable devices.
Definition: hip_device.cpp:73
+
hipMemset
hipError_t hipMemset(void *dst, int value, size_t sizeBytes)
Copy data from src to dst asynchronously.
Definition: hip_memory.cpp:854
+
hipStreamDestroy
hipError_t hipStreamDestroy(hipStream_t stream)
Destroys the specified stream.
Definition: hip_stream.cpp:157
+
hipHostGetFlags
hipError_t hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr)
Return flags associated with host pointer.
Definition: hip_memory.cpp:345
+
hipStreamSynchronize
hipError_t hipStreamSynchronize(hipStream_t stream)
Wait for all commands in stream to complete.
Definition: hip_stream.cpp:134
+
hipIpcOpenMemHandle
hipError_t hipIpcOpenMemHandle(void **devPtr, hipIpcMemHandle_t handle, unsigned int flags)
Opens an interprocess memory handle exported from another process and returns a device pointer usable...
Definition: hip_memory.cpp:1068
+
hipCtxGetCacheConfig
hipError_t hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig)
Set Cache configuration for a specific function.
Definition: hip_context.cpp:220
+
hipMemcpyDtoD
hipError_t hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes)
Copy data from Device to Device.
Definition: hip_memory.cpp:563
+
hipSharedMemBankSizeFourByte
Shared mem is banked at 4-bytes intervals and performs best when adjacent threads access data 4 bytes...
Definition: hip_runtime_api.h:141
+
hipError_t
hipError_t
Definition: hip_runtime_api.h:152
+
hipFuncCachePreferNone
no preference for shared memory or L1 (default)
Definition: hip_runtime_api.h:129
+
hipMemcpyKind
hipMemcpyKind
Definition: hip_runtime_api.h:165
+
hipCtxSynchronize
hipError_t hipCtxSynchronize(void)
Blocks until the default context has completed all preceding requested tasks.
Definition: hip_context.cpp:256
+
hipCtxGetCurrent
hipError_t hipCtxGetCurrent(hipCtx_t *ctx)
Get the handle of the current/ default context.
Definition: hip_context.cpp:166
+
hipMemcpyDtoHAsync
hipError_t hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
Copy data from Device to Host asynchronously.
Definition: hip_memory.cpp:658
+
hipDeviceSynchronize
hipError_t hipDeviceSynchronize(void)
Waits on all active streams on current device.
Definition: hip_device.cpp:153
+
hipCtxPushCurrent
hipError_t hipCtxPushCurrent(hipCtx_t ctx)
Push the context to be set as current/ default context.
Definition: hip_context.cpp:152
+
hipMemcpyDtoH
hipError_t hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes)
Copy data from Device to Host.
Definition: hip_memory.cpp:542
+
ihipModule_t
Definition: hip_hcc.h:404
+
hipDeviceGetName
hipError_t hipDeviceGetName(char *name, int len, hipDevice_t device)
Returns an identifer string for the device.
Definition: hip_device.cpp:330
+
hipGetDeviceProperties
hipError_t hipGetDeviceProperties(hipDeviceProp_t *prop, int deviceId)
Returns device properties.
Definition: hip_device.cpp:275
+
hipHostRegister
hipError_t hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags)
Register host memory so it can be accessed from the current device.
Definition: hip_memory.cpp:369
+
hipMalloc
hipError_t hipMalloc(void **ptr, size_t size)
Allocate memory on the default accelerator.
Definition: hip_memory.cpp:105
+
hipGetErrorName
const char * hipGetErrorName(hipError_t hip_error)
Return name of the specified error code in text form.
Definition: hip_error.cpp:50
+
hipFuncCache_t
hipFuncCache_t
Definition: hip_runtime_api.h:128
+
hipGetLastError
hipError_t hipGetLastError(void)
Return last error returned by any HIP runtime API call and resets the stored error code to hipSuccess...
Definition: hip_error.cpp:32
+
hipStreamWaitEvent
hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags)
Make the specified compute stream wait for an event.
Definition: hip_stream.cpp:80
+
hipStreamGetFlags
hipError_t hipStreamGetFlags(hipStream_t stream, unsigned int *flags)
Return flags associated with this stream.
Definition: hip_stream.cpp:186
+
hipHostAlloc
hipError_t hipHostAlloc(void **ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead")))
Allocate device accessible page locked host memory [Deprecated].
Definition: hip_memory.cpp:227
+
hipMemGetInfo
hipError_t hipMemGetInfo(size_t *free, size_t *total)
Query memory info. Return snapshot of free memory, and total allocatable memory on the device...
Definition: hip_memory.cpp:906
+
hipCtxGetDevice
hipError_t hipCtxGetDevice(hipDevice_t *device)
Get the handle of the device associated with current/default context.
Definition: hip_context.cpp:193
+
hipFree
hipError_t hipFree(void *ptr)
Free memory allocated by the hcc hip memory allocation API. This API performs an implicit hipDeviceSy...
Definition: hip_memory.cpp:940
+
dim3::z
uint32_t z
z
Definition: hip_runtime_api.h:154
+
hipCtxGetApiVersion
hipError_t hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion)
Returns the approximate HIP api version.
Definition: hip_context.cpp:209
+
hipDeviceReset
hipError_t hipDeviceReset(void)
The state of current device is discarded and updated to a fresh state.
Definition: hip_device.cpp:159
+
hipInit
hipError_t hipInit(unsigned int flags)
Explicitly initializes the HIP runtime.
Definition: hip_context.cpp:42
+
hipRuntimeGetVersion
hipError_t hipRuntimeGetVersion(int *runtimeVersion)
Returns the approximate HIP Runtime version.
Definition: hip_context.cpp:96
+
hipMemcpyDtoDAsync
hipError_t hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
Copy data from Device to Device asynchronously.
Definition: hip_memory.cpp:651
+
hipDeviceProp_t
Definition: hip_runtime_api.h:82
+
hipSetDevice
hipError_t hipSetDevice(int deviceId)
Set default device to be used for subsequent hip API calls from this thread.
Definition: hip_device.cpp:142
+
hipDeviceGet
hipError_t hipDeviceGet(hipDevice_t *device, int ordinal)
Returns a handle to a compute device.
Definition: hip_context.cpp:68
+
hipDeviceTotalMem
hipError_t hipDeviceTotalMem(size_t *bytes, hipDevice_t device)
Returns the total amount of memory on the device.
Definition: hip_device.cpp:368
+
hipMemsetAsync
hipError_t hipMemsetAsync(void *dst, int value, size_t sizeBytes, hipStream_t stream)
Fills the first sizeBytes bytes of the memory area pointed to by dev with the constant byte value val...
Definition: hip_memory.cpp:805
+
hipSharedMemBankSizeDefault
The compiler selects a device-specific value for the banking.
Definition: hip_runtime_api.h:140
+
hipCtxGetFlags
hipError_t hipCtxGetFlags(unsigned int *flags)
Return flags used for creating default context.
Definition: hip_context.cpp:262
+
hipMemcpyDeviceToDevice
Device-to-Device Copy.
Definition: hip_runtime_api.h:169
+
hipCtxDestroy
hipError_t hipCtxDestroy(hipCtx_t ctx)
Destroy a HIP context.
Definition: hip_context.cpp:110
+
ihipEvent_t
Definition: hip_hcc.h:558
+
hipMemcpyPeerAsync
hipError_t hipMemcpyPeerAsync(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream)
Copies memory from one device to memory on another device.
Definition: hip_peer.cpp:209
+
hipMemcpyDefault
Runtime will automatically determine copy-kind based on virtual addresses.
Definition: hip_runtime_api.h:170
+
hipSharedMemConfig
hipSharedMemConfig
Definition: hip_runtime_api.h:139
+
hipDeviceGetAttribute
hipError_t hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int deviceId)
Query for a specific device attribute.
Definition: hip_device.cpp:249
+
hipMemcpyHtoDAsync
hipError_t hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream)
Copy data from Host to Device asynchronously.
Definition: hip_memory.cpp:644
+
hipHostUnregister
hipError_t hipHostUnregister(void *hostPtr)
Un-register host pointer.
Definition: hip_memory.cpp:414
+
ihipStream_t
Definition: hip_hcc.h:463
+
hipMemcpyPeer
hipError_t hipMemcpyPeer(void *dst, int dstDeviceId, const void *src, int srcDeviceId, size_t sizeBytes)
Copies memory from one device to memory on another device.
Definition: hip_peer.cpp:202
+
hipStreamCreate
hipError_t hipStreamCreate(hipStream_t *stream)
Create an asynchronous stream.
Definition: hip_stream.cpp:72
+
hipMemcpy
hipError_t hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind)
Copy data from src to dst.
Definition: hip_memory.cpp:500
+
hipEventCreate
hipError_t hipEventCreate(hipEvent_t *event)
Definition: hip_event.cpp:61
+
hipFreeHost
hipError_t hipFreeHost(void *ptr) __attribute__((deprecated("use hipHostFree instead")))
Free memory allocated by the hcc hip host memory allocation API. [Deprecated].
Definition: hip_memory.cpp:994
+
hipStreamCallback_t
void(* hipStreamCallback_t)(hipStream_t stream, hipError_t status, void *userData)
Definition: hip_runtime_api.h:639
+
hipCtxEnablePeerAccess
hipError_t hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags)
Enables direct access to memory allocations in a peer context.
Definition: hip_peer.cpp:215
+
hipDeviceSetSharedMemConfig
hipError_t hipDeviceSetSharedMemConfig(hipSharedMemConfig config)
The bank width of shared memory on current device is set.
Definition: hip_device.cpp:124
+
hipMemcpyToSymbolAsync
hipError_t hipMemcpyToSymbolAsync(const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream)
Copies sizeBytes bytes from the memory area pointed to by src to the memory area pointed to by offset...
Definition: hip_memory.cpp:460
+
hipGetDevice
hipError_t hipGetDevice(int *deviceId)
Return the default device id for the calling host thread.
Definition: hip_device.cpp:32
+
hipEventRecord
hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream)
Record an event in the specified stream.
Definition: hip_event.cpp:68
+
hipModuleGetGlobal
hipError_t hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name)
returns device memory pointer and size of the kernel present in the module with symbol name ...
Definition: hip_module.cpp:395
+
hipMemcpyHostToHost
Host-to-Host Copy.
Definition: hip_runtime_api.h:166
+
hipDeviceGetPCIBusId
hipError_t hipDeviceGetPCIBusId(char *pciBusId, int len, int device)
Returns a PCI Bus Id string for the device, overloaded to take int device ID.
Definition: hip_device.cpp:356
+
hipFuncCachePreferShared
prefer larger shared memory and smaller L1 cache
Definition: hip_runtime_api.h:130
+
hipPointerGetAttributes
hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void *ptr)
Return attributes for the specified pointer.
Definition: hip_memory.cpp:38
+
hipStreamQuery
hipError_t hipStreamQuery(hipStream_t stream)
Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not.
Definition: hip_stream.cpp:113
+
hipSharedMemBankSizeEightByte
Shared mem is banked at 8-byte intervals and performs best when adjacent threads access data 4 bytes ...
Definition: hip_runtime_api.h:142
+
hipIpcCloseMemHandle
hipError_t hipIpcCloseMemHandle(void *devPtr)
Close memory mapped with hipIpcOpenMemHandle.
Definition: hip_memory.cpp:1087
+
hipProfilerStop
hipError_t hipProfilerStop()
Stop recording of profiling information. When using this API, start the profiler with profiling disab...
Definition: hip_hcc.cpp:1942
+
hipProfilerStart
hipError_t hipProfilerStart()
Start recording of profiling information When using this API, start the profiler with profiling disab...
Definition: hip_hcc.cpp:1931
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h.html b/docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h.html new file mode 100644 index 0000000000..5aa672dda3 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h.html @@ -0,0 +1,560 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_texture.h File Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+Classes | +Macros | +Typedefs | +Enumerations | +Functions
+
+
hip_texture.h File Reference
+
+
+ +

HIP C++ Texture API for hcc compiler. +More...

+
#include <limits.h>
+
+

Go to the source code of this file.

+ + + + + + + + +

+Classes

struct  hipChannelFormatDesc
 
struct  textureReference
 
struct  hipArray
 
+ + + + + + + +

+Macros

+#define hipTextureType1D   1
 
+#define tex1Dfetch(_tex, _addr)   (_tex._dataPtr[_addr])
 
+#define tex2D(_tex, _dx, _dy)   _tex._dataPtr[(unsigned int)_dx + (unsigned int)_dy*(_tex.width)]
 
+ + + + + + + +

+Typedefs

+typedef struct hipChannelFormatDesc hipChannelFormatDesc
 
+typedef enum hipTextureReadMode hipTextureReadMode
 
+typedef enum hipTextureFilterMode hipTextureFilterMode
 
+ + + + + + + +

+Enumerations

enum  hipChannelFormatKind { hipChannelFormatKindSigned = 0, +hipChannelFormatKindUnsigned, +hipChannelFormatKindFloat, +hipChannelFormatKindNone + }
 
enum  hipTextureReadMode { hipReadModeElementType + }
 
enum  hipTextureFilterMode { hipFilterModePoint + }
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

hipError_t hipMallocArray (hipArray **array, const hipChannelFormatDesc *desc, size_t width, size_t height=0, unsigned int flags=0)
 Allocate an array on the device. More...
 
hipError_t hipFreeArray (hipArray *array)
 Frees an array on the device. More...
 
hipError_t hipMemcpy2D (void *dst, size_t dpitch, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind)
 Copies data between host and device. More...
 
hipError_t hipMemcpy2DToArray (hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind)
 Copies data between host and device. More...
 
hipError_t hipMemcpyToArray (hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t count, hipMemcpyKind kind)
 Copies data between host and device. More...
 
hipChannelFormatDesc hipCreateChannelDesc (int x, int y, int z, int w, hipChannelFormatKind f)
 Returns a channel descriptor using the specified format. More...
 
+template<typename T >
hipChannelFormatDesc hipCreateChannelDesc ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< int > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< unsigned int > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< long > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< unsigned long > ()
 
+template<>
hipChannelFormatDesc hipCreateChannelDesc< float > ()
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipBindTexture (size_t *offset, struct texture< T, dim, readMode > &tex, const void *devPtr, const struct hipChannelFormatDesc *desc, size_t size=UINT_MAX)
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipBindTexture (size_t *offset, struct texture< T, dim, readMode > &tex, const void *devPtr, size_t size=UINT_MAX)
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipBindTextureToArray (struct texture< T, dim, readMode > &tex, hipArray *array)
 
+template<class T , int dim, enum hipTextureReadMode readMode>
hipError_t hipUnbindTexture (struct texture< T, dim, readMode > &tex)
 
+

Detailed Description

+

HIP C++ Texture API for hcc compiler.

+

Enumeration Type Documentation

+ +
+
+ + + + +
enum hipTextureFilterMode
+
+ + +
Enumerator
hipFilterModePoint  +

Point filter mode.

+
+ +
+
+ +
+
+ + + + +
enum hipTextureReadMode
+
+ + +
Enumerator
hipReadModeElementType  +

Read texture as specified element type

+
+ +
+
+

Function Documentation

+ +
+
+ + + + + + + + +
hipError_t hipFreeArray (hipArray * array)
+
+ +

Frees an array on the device.

+
Parameters
+ + +
[in]arrayPointer to array to free
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, hipErrorInitializationError
+
See Also
hipMalloc, hipMallocPitch, hipFree, hipMallocArray, hipHostMalloc, hipHostFree
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMallocArray (hipArray ** array,
const hipChannelFormatDesc * desc,
size_t width,
size_t height = 0,
unsigned int flags = 0 
)
+
+ +

Allocate an array on the device.

+
Parameters
+ + + + + + +
[out]arrayPointer to allocated array in device memory
[in]descRequested channel format
[in]widthRequested array allocation width
[in]heightRequested array allocation height
[in]flagsRequested properties of allocated array
+
+
+
Returns
hipSuccess, hipErrorMemoryAllocation
+
See Also
hipMalloc, hipMallocPitch, hipFree, hipFreeArray, hipHostMalloc, hipHostFree
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpy2D (void * dst,
size_t dpitch,
const void * src,
size_t spitch,
size_t width,
size_t height,
hipMemcpyKind kind 
)
+
+ +

Copies data between host and device.

+
Parameters
+ + + + + + + + +
[in]dstDestination memory address
[in]dpitchPitch of destination memory
[in]srcSource memory address
[in]spitchPitch of source memory
[in]widthWidth of matrix transfer (columns in bytes)
[in]heightHeight of matrix transfer (rows)
[in]kindType of transfer
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorInvalidPitchValue, hipErrorInvalidDevicePointer, hipErrorInvalidMemcpyDirection
+
See Also
hipMemcpy, hipMemcpyToArray, hipMemcpy2DToArray, hipMemcpyFromArray, hipMemcpyToSymbol, hipMemcpyAsync
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpy2DToArray (hipArray * dst,
size_t wOffset,
size_t hOffset,
const void * src,
size_t spitch,
size_t width,
size_t height,
hipMemcpyKind kind 
)
+
+ +

Copies data between host and device.

+
Parameters
+ + + + + + + + +
[in]dstDestination memory address
[in]dpitchPitch of destination memory
[in]srcSource memory address
[in]spitchPitch of source memory
[in]widthWidth of matrix transfer (columns in bytes)
[in]heightHeight of matrix transfer (rows)
[in]kindType of transfer
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorInvalidPitchValue, hipErrorInvalidDevicePointer, hipErrorInvalidMemcpyDirection
+
See Also
hipMemcpy, hipMemcpyToArray, hipMemcpy2D, hipMemcpyFromArray, hipMemcpyToSymbol, hipMemcpyAsync
+ +
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
hipError_t hipMemcpyToArray (hipArray * dst,
size_t wOffset,
size_t hOffset,
const void * src,
size_t count,
hipMemcpyKind kind 
)
+
+ +

Copies data between host and device.

+
Parameters
+ + + + + + + + +
[in]dstDestination memory address
[in]dpitchPitch of destination memory
[in]srcSource memory address
[in]spitchPitch of source memory
[in]widthWidth of matrix transfer (columns in bytes)
[in]heightHeight of matrix transfer (rows)
[in]kindType of transfer
+
+
+
Returns
hipSuccess, hipErrorInvalidValue, #hipErrorInvalidPitchValue, hipErrorInvalidDevicePointer, hipErrorInvalidMemcpyDirection
+
See Also
hipMemcpy, hipMemcpy2DToArray, hipMemcpy2D, hipMemcpyFromArray, hipMemcpyToSymbol, hipMemcpyAsync
+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h_source.html b/docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h_source.html new file mode 100644 index 0000000000..17dbbe2e7e --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__texture_8h_source.html @@ -0,0 +1,333 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_texture.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_texture.h
+
+
+Go to the documentation of this file.
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 //#pragma once
+
24 
+
25 #ifndef HIP_HCC_DETAIL_TEXTURE_H
+
26 #define HIP_HCC_DETAIL_TEXTURE_H
+
27 
+
33 #include <limits.h>
+
34 
+
35 //#include <hip/hcc_detail/hip_runtime.h>
+
36 
+
37 //----
+
38 //Texture - TODO - likely need to move this to a separate file only included with kernel compilation.
+
39 #define hipTextureType1D 1
+
40 
+
41 typedef enum {
+
42  hipChannelFormatKindSigned = 0,
+
43  hipChannelFormatKindUnsigned,
+
44  hipChannelFormatKindFloat,
+
45  hipChannelFormatKindNone
+
46 
+
47 } hipChannelFormatKind;
+
48 
+
49 typedef struct hipChannelFormatDesc {
+
50  int x;
+
51  int y;
+
52  int z;
+
53  int w;
+
54  hipChannelFormatKind f;
+
55 } hipChannelFormatDesc;
+
56 
+
57 typedef enum hipTextureReadMode
+
58 {
+
59  hipReadModeElementType,
+
60 } hipTextureReadMode;
+
62 
+
63 typedef enum hipTextureFilterMode
+
64 {
+
65  hipFilterModePoint,
+
66 } hipTextureFilterMode;
+
68 
+
69 struct textureReference {
+
70  hipTextureFilterMode filterMode;
+
71  bool normalized;
+
72  hipChannelFormatDesc channelDesc;
+
73 };
+
74 #if __cplusplus
+
75 template <class T, int texType=hipTextureType1D, enum hipTextureReadMode=hipReadModeElementType>
+
76 struct texture : public textureReference {
+
77 
+
78  const T * _dataPtr; // pointer to underlying data.
+
79 
+
80  //texture() : filterMode(hipFilterModePoint), normalized(false), _dataPtr(NULL) {};
+
81  unsigned int width;
+
82  unsigned int height;
+
83 
+
84 };
+
85 #endif
+
86 
+
87 typedef struct {
+
88  unsigned int width;
+
89  unsigned int height;
+
90  hipChannelFormatKind f;
+
91  void* data; //FIXME: generalize this
+
92 } hipArray;
+
93 
+
94 
+
95 #define tex1Dfetch(_tex, _addr) (_tex._dataPtr[_addr])
+
96 
+
97 #define tex2D(_tex, _dx, _dy) \
+
98  _tex._dataPtr[(unsigned int)_dx + (unsigned int)_dy*(_tex.width)]
+
99 
+
112 hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc,
+
113  size_t width, size_t height = 0, unsigned int flags = 0);
+
114 
+
123 hipError_t hipFreeArray(hipArray* array);
+
124 
+
139 hipError_t hipMemcpy2D(void* dst, size_t dpitch, const void* src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind);
+
140 
+
155 hipError_t hipMemcpy2DToArray(hipArray* dst, size_t wOffset, size_t hOffset, const void* src,
+
156  size_t spitch, size_t width, size_t height, hipMemcpyKind kind);
+
157 
+
172 hipError_t hipMemcpyToArray(hipArray* dst, size_t wOffset, size_t hOffset,
+
173  const void* src, size_t count, hipMemcpyKind kind);
+
174 
+
175 
+
183 // These are C++ APIs - maybe belong in separate file.
+
207 // C API:
+
208 #if 0
+
209 hipChannelFormatDesc hipBindTexture(size_t *offset, struct textureReference *tex, const void *devPtr, const struct hipChannelFormatDesc *desc, size_t size=UINT_MAX)
+
210 {
+
211  tex->_dataPtr = devPtr;
+
212 }
+
213 #endif
+
214 
+
226 hipChannelFormatDesc hipCreateChannelDesc(int x, int y, int z, int w, hipChannelFormatKind f);
+
227 
+
228 // descriptors
+
229 template <typename T> inline hipChannelFormatDesc hipCreateChannelDesc() {
+
230  return hipCreateChannelDesc(0, 0, 0, 0, hipChannelFormatKindNone);
+
231 }
+
232 template <> inline hipChannelFormatDesc hipCreateChannelDesc<int>() {
+
233  int e = (int)sizeof(int) * 8;
+
234  return hipCreateChannelDesc(e, 0, 0, 0, hipChannelFormatKindSigned);
+
235 }
+
236 template <> inline hipChannelFormatDesc hipCreateChannelDesc<unsigned int>() {
+
237  int e = (int)sizeof(unsigned int) * 8;
+
238  return hipCreateChannelDesc(e, 0, 0, 0, hipChannelFormatKindUnsigned);
+
239 }
+
240 template <> inline hipChannelFormatDesc hipCreateChannelDesc<long>() {
+
241  int e = (int)sizeof(long) * 8;
+
242  return hipCreateChannelDesc(e, 0, 0, 0, hipChannelFormatKindSigned);
+
243 }
+
244 template <> inline hipChannelFormatDesc hipCreateChannelDesc<unsigned long>() {
+
245  int e = (int)sizeof(unsigned long) * 8;
+
246  return hipCreateChannelDesc(e, 0, 0, 0, hipChannelFormatKindUnsigned);
+
247 }
+
248 template <> inline hipChannelFormatDesc hipCreateChannelDesc<float>() {
+
249  int e = (int)sizeof(float) * 8;
+
250  return hipCreateChannelDesc(e, 0, 0, 0, hipChannelFormatKindFloat);
+
251 }
+
252 
+
253 /*
+
254  * @brief hipBindTexture Binds size bytes of the memory area pointed to by @p devPtr to the texture reference tex.
+
255  *
+
256  * @p desc describes how the memory is interpreted when fetching values from the texture. The @p offset parameter is an optional byte offset as with the low-level
+
257  * hipBindTexture() function. Any memory previously bound to tex is unbound.
+
258  *
+
259  * @param[in] offset - Offset in bytes
+
260  * @param[out] tex - texture to bind
+
261  * @param[in] devPtr - Memory area on device
+
262  * @param[in] desc - Channel format
+
263  * @param[in] size - Size of the memory area pointed to by devPtr
+
264  * @return #hipSuccess, #hipErrorInvalidValue, #hipErrorMemoryFree, #hipErrorUnknown
+
265  **/
+
266 template <class T, int dim, enum hipTextureReadMode readMode>
+
267 hipError_t hipBindTexture(size_t *offset,
+
268  struct texture<T, dim, readMode> &tex,
+
269  const void *devPtr,
+
270  const struct hipChannelFormatDesc *desc,
+
271  size_t size=UINT_MAX)
+
272 {
+
273  tex._dataPtr = static_cast<const T*>(devPtr);
+
274 
+
275  return hipSuccess;
+
276 }
+
277 
+
278 /*
+
279  * @brief hipBindTexture Binds size bytes of the memory area pointed to by @p devPtr to the texture reference tex.
+
280  *
+
281  * @p desc describes how the memory is interpreted when fetching values from the texture. The @p offset parameter is an optional byte offset as with the low-level
+
282  * hipBindTexture() function. Any memory previously bound to tex is unbound.
+
283  *
+
284  * @param[in] offset - Offset in bytes
+
285  * @param[in] tex - texture to bind
+
286  * @param[in] devPtr - Memory area on device
+
287  * @param[in] size - Size of the memory area pointed to by devPtr
+
288  * @return #hipSuccess, #hipErrorInvalidValue, #hipErrorMemoryFree, #hipErrorUnknown
+
289  **/
+
290 template <class T, int dim, enum hipTextureReadMode readMode>
+
291 hipError_t hipBindTexture(size_t *offset,
+
292  struct texture<T, dim, readMode> &tex,
+
293  const void *devPtr,
+
294  size_t size=UINT_MAX)
+
295 {
+
296  return hipBindTexture(offset, tex, devPtr, &tex.channelDesc, size);
+
297 }
+
298 
+
299 template <class T, int dim, enum hipTextureReadMode readMode>
+
300 hipError_t hipBindTextureToArray(struct texture<T, dim, readMode> &tex, hipArray* array) {
+
301  tex.width = array->width;
+
302  tex.height = array->height;
+
303  tex._dataPtr = static_cast<const T*>(array->data);
+
304  return hipSuccess;
+
305 }
+
306 
+
307 /*
+
308  * @brief Unbinds the textuer bound to @p tex
+
309  *
+
310  * @param[in] tex - texture to unbind
+
311  *
+
312  * @return #hipSuccess
+
313  **/
+
314 template <class T, int dim, enum hipTextureReadMode readMode>
+
315 hipError_t hipUnbindTexture(struct texture<T, dim, readMode> &tex)
+
316 {
+
317  tex._dataPtr = NULL;
+
318 
+
319  return hipSuccess;
+
320 }
+
321 
+
322 
+
323 
+
324 // doxygen end Texture
+
330 // End doxygen API:
+
335 #endif
+
336 
+
hipFreeArray
hipError_t hipFreeArray(hipArray *array)
Frees an array on the device.
Definition: hip_memory.cpp:999
+
hipCreateChannelDesc
hipChannelFormatDesc hipCreateChannelDesc(int x, int y, int z, int w, hipChannelFormatKind f)
Returns a channel descriptor using the specified format.
Definition: hip_memory.cpp:276
+
hipSuccess
Successful completion.
Definition: hip_runtime_api.h:153
+
hipMemcpy2DToArray
hipError_t hipMemcpy2DToArray(hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind)
Copies data between host and device.
Definition: hip_memory.cpp:692
+
hipTextureFilterMode
hipTextureFilterMode
Definition: hip_texture.h:63
+
hipReadModeElementType
Definition: hip_texture.h:59
+
textureReference
Definition: hip_texture.h:69
+
hipArray
Definition: hip_texture.h:87
+
hipError_t
hipError_t
Definition: hip_runtime_api.h:152
+
hipMemcpyToArray
hipError_t hipMemcpyToArray(hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t count, hipMemcpyKind kind)
Copies data between host and device.
Definition: hip_memory.cpp:745
+
hipMemcpyKind
hipMemcpyKind
Definition: hip_runtime_api.h:165
+
hipFilterModePoint
Definition: hip_texture.h:65
+
hipChannelFormatDesc
Definition: hip_texture.h:49
+
hipTextureReadMode
hipTextureReadMode
Definition: hip_texture.h:57
+
hipMemcpy2D
hipError_t hipMemcpy2D(void *dst, size_t dpitch, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind)
Copies data between host and device.
Definition: hip_memory.cpp:666
+
hipMallocArray
hipError_t hipMallocArray(hipArray **array, const hipChannelFormatDesc *desc, size_t width, size_t height=0, unsigned int flags=0)
Allocate an array on the device.
Definition: hip_memory.cpp:284
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h.html b/docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h.html new file mode 100644 index 0000000000..80b9f0458f --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h.html @@ -0,0 +1,431 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_vector_types.h File Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+Classes | +Macros | +Functions
+
+
hip_vector_types.h File Reference
+
+
+ +

Defines the different newt vector types for HIP runtime. +More...

+
#include "hip/hcc_detail/host_defines.h"
+
+

Go to the source code of this file.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Classes

struct  char3
 
struct  uchar3
 
struct  char4
 
struct  uchar4
 
struct  short3
 
struct  ushort3
 
struct  int3
 
struct  uint3
 
struct  long3
 
struct  ulong3
 
struct  float1
 
struct  float3
 
struct  longlong3
 
struct  ulonglong3
 
struct  double1
 
struct  double3
 
+ + + + + +

+Macros

#define __hip_align(name, val, data)
 
+#define __HIP_DEVICE__   __device__ __host__
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

+struct __hip_align (char1, 1, signed char x;)
 
+struct __hip_align (uchar1, 1, unsigned char x;)
 
+struct __hip_align (char2, 2, signed char x;signed char y;)
 
+struct __hip_align (uchar2, 2, unsigned char x;unsigned char y;)
 
+struct __hip_align (short1, 2, signed short x;)
 
+struct __hip_align (ushort1, 2, unsigned short x;)
 
+struct __hip_align (short2, 4, signed short x;signed short y;)
 
+struct __hip_align (ushort2, 4, unsigned short x;unsigned short y;)
 
+struct __hip_align (short4, 8, signed short x;signed short y;signed short z;signed short w;)
 
+struct __hip_align (ushort4, 8, unsigned short x;unsigned short y;unsigned short z;unsigned short w;)
 
+struct __hip_align (int1, 4, signed int x;)
 
+struct __hip_align (uint1, 4, unsigned int x;)
 
+struct __hip_align (int2, 8, signed int x;signed int y;)
 
+struct __hip_align (uint2, 8, unsigned int x;unsigned int y;)
 
+struct __hip_align (int4, 16, signed int x;signed int y;signed int z;signed int w;)
 
+struct __hip_align (uint4, 16, unsigned int x;unsigned int y;unsigned int z;unsigned int w;)
 
+struct __hip_align (long1, 8, long int x;)
 
+struct __hip_align (ulong1, 8, unsigned long x;)
 
+struct __hip_align (long2, 16, long int x;long int y;)
 
+struct __hip_align (ulong2, 16, unsigned long x;unsigned long y;)
 
+struct __hip_align (long4, 32, long int x;long int y;long int z;long int w;)
 
+struct __hip_align (ulong4, 32, unsigned long x;unsigned long y;unsigned long z;unsigned long w;)
 
+struct __hip_align (float2, 8, float x;float y;)
 
+struct __hip_align (float4, 16, float x;float y;float z;float w;)
 
+struct __hip_align (longlong1, 16, long long int x;)
 
+struct __hip_align (ulonglong1, 16, unsigned long long int x;)
 
+struct __attribute__ ((aligned(32))) longlong2
 
+struct __attribute__ ((aligned(64))) longlong4
 
+struct __attribute__ ((aligned(16))) double2
 
+__HIP_DEVICE__ char1 make_char1 (signed char)
 
+__HIP_DEVICE__ char2 make_char2 (signed char, signed char)
 
+__HIP_DEVICE__ char3 make_char3 (signed char, signed char, signed char)
 
+__HIP_DEVICE__ char4 make_char4 (signed char, signed char, signed char, signed char)
 
+__HIP_DEVICE__ short1 make_short1 (short)
 
+__HIP_DEVICE__ short2 make_short2 (short, short)
 
+__HIP_DEVICE__ short3 make_short3 (short, short, short)
 
+__HIP_DEVICE__ short4 make_short4 (short, short, short, short)
 
+__HIP_DEVICE__ int1 make_int1 (int)
 
+__HIP_DEVICE__ int2 make_int2 (int, int)
 
+__HIP_DEVICE__ int3 make_int3 (int, int, int)
 
+__HIP_DEVICE__ int4 make_int4 (int, int, int, int)
 
+__HIP_DEVICE__ long1 make_long1 (long)
 
+__HIP_DEVICE__ long2 make_long2 (long, long)
 
+__HIP_DEVICE__ long3 make_long3 (long, long, long)
 
+__HIP_DEVICE__ long4 make_long4 (long, long, long, long)
 
+__HIP_DEVICE__ longlong1 make_longlong1 (long long)
 
+__HIP_DEVICE__ longlong2 make_longlong2 (long long, long long)
 
+__HIP_DEVICE__ longlong3 make_longlong3 (long long, long long, long long)
 
+__HIP_DEVICE__ longlong4 make_longlong4 (long long, long long, long long, long long)
 
+__HIP_DEVICE__ uchar1 make_uchar1 (unsigned char)
 
+__HIP_DEVICE__ uchar2 make_uchar2 (unsigned char, unsigned char)
 
+__HIP_DEVICE__ uchar3 make_uchar3 (unsigned char, unsigned char, unsigned char)
 
+__HIP_DEVICE__ uchar4 make_uchar4 (unsigned char, unsigned char, unsigned char, unsigned char)
 
+__HIP_DEVICE__ ushort1 make_ushort1 (unsigned short)
 
+__HIP_DEVICE__ ushort2 make_ushort2 (unsigned short, unsigned short)
 
+__HIP_DEVICE__ ushort3 make_ushort3 (unsigned short, unsigned short, unsigned short)
 
+__HIP_DEVICE__ ushort4 make_ushort4 (unsigned short, unsigned short, unsigned short, unsigned short)
 
+__HIP_DEVICE__ uint1 make_uint1 (unsigned int)
 
+__HIP_DEVICE__ uint2 make_uint2 (unsigned int, unsigned int)
 
+__HIP_DEVICE__ uint3 make_uint3 (unsigned int, unsigned int, unsigned int)
 
+__HIP_DEVICE__ uint4 make_uint4 (unsigned int, unsigned int, unsigned int, unsigned int)
 
+__HIP_DEVICE__ ulong1 make_ulong1 (unsigned long)
 
+__HIP_DEVICE__ ulong2 make_ulong2 (unsigned long, unsigned long)
 
+__HIP_DEVICE__ ulong3 make_ulong3 (unsigned long, unsigned long, unsigned long)
 
+__HIP_DEVICE__ ulong4 make_ulong4 (unsigned long, unsigned long, unsigned long, unsigned long)
 
+__HIP_DEVICE__ ulonglong1 make_ulonglong1 (unsigned long long)
 
+__HIP_DEVICE__ ulonglong2 make_ulonglong2 (unsigned long long, unsigned long long)
 
+__HIP_DEVICE__ ulonglong3 make_ulonglong3 (unsigned long long, unsigned long long, unsigned long long)
 
+__HIP_DEVICE__ ulonglong4 make_ulonglong4 (unsigned long long, unsigned long long, unsigned long long, unsigned long long)
 
+__HIP_DEVICE__ float1 make_float1 (float)
 
+__HIP_DEVICE__ float2 make_float2 (float, float)
 
+__HIP_DEVICE__ float3 make_float3 (float, float, float)
 
+__HIP_DEVICE__ float4 make_float4 (float, float, float, float)
 
+__HIP_DEVICE__ double1 make_double1 (double)
 
+__HIP_DEVICE__ double2 make_double2 (double, double)
 
+__HIP_DEVICE__ double3 make_double3 (double, double, double)
 
+__HIP_DEVICE__ double4 make_double4 (double, double, double, double)
 
+__HIP_DEVICE__ double __longlong_as_double (long long int x)
 
+__HIP_DEVICE__ long long int __double_as_longlong (double x)
 
+

Detailed Description

+

Defines the different newt vector types for HIP runtime.

+

Macro Definition Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + +
#define __hip_align( name,
 val,
 data 
)
+
+Value:
__attribute__((aligned(val))) name \
+
{ data }
+
+
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h_source.html b/docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h_source.html new file mode 100644 index 0000000000..94a8896392 --- /dev/null +++ b/docs/RuntimeAPI/html/hcc__detail_2hip__vector__types_8h_source.html @@ -0,0 +1,547 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_vector_types.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_vector_types.h
+
+
+Go to the documentation of this file.
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
28 #ifndef HIP_VECTOR_TYPES_H
+
29 #define HIP_VECTOR_TYPES_H
+
30 
+
31 #if defined (__HCC__) && (__hcc_workweek__ < 16032)
+
32 #error("This version of HIP requires a newer version of HCC.");
+
33 #endif
+
34 
+
35 #if 0
+
36 #include <hc_short_vector.hpp>
+
37 
+
38 using namespace hc::short_vector;
+
39 
+
40 
+
41 //-- Signed
+
42 // Define char vector types
+
43 typedef hc::short_vector::char1 char1;
+
44 typedef hc::short_vector::char2 char2;
+
45 typedef hc::short_vector::char3 char3;
+
46 typedef hc::short_vector::char4 char4;
+
47 
+
48 // Define short vector types
+
49 typedef hc::short_vector::short1 short1;
+
50 typedef hc::short_vector::short2 short2;
+
51 typedef hc::short_vector::short3 short3;
+
52 typedef hc::short_vector::short4 short4;
+
53 
+
54 // Define int vector types
+
55 typedef hc::short_vector::int1 int1;
+
56 typedef hc::short_vector::int2 int2;
+
57 typedef hc::short_vector::int3 int3;
+
58 typedef hc::short_vector::int4 int4;
+
59 
+
60 // Define long vector types
+
61 typedef hc::short_vector::long1 long1;
+
62 typedef hc::short_vector::long2 long2;
+
63 typedef hc::short_vector::long3 long3;
+
64 typedef hc::short_vector::long4 long4;
+
65 
+
66 // Define longlong vector types
+
67 typedef hc::short_vector::longlong1 longlong1;
+
68 typedef hc::short_vector::longlong2 longlong2;
+
69 typedef hc::short_vector::longlong3 longlong3;
+
70 typedef hc::short_vector::longlong4 longlong4;
+
71 
+
72 
+
73 //-- Unsigned
+
74 // Define uchar vector types
+
75 typedef hc::short_vector::uchar1 uchar1;
+
76 typedef hc::short_vector::uchar2 uchar2;
+
77 typedef hc::short_vector::uchar3 uchar3;
+
78 typedef hc::short_vector::uchar4 uchar4;
+
79 
+
80 // Define ushort vector types
+
81 typedef hc::short_vector::ushort1 ushort1;
+
82 typedef hc::short_vector::ushort2 ushort2;
+
83 typedef hc::short_vector::ushort3 ushort3;
+
84 typedef hc::short_vector::ushort4 ushort4;
+
85 
+
86 // Define uint vector types
+
87 typedef hc::short_vector::uint1 uint1;
+
88 typedef hc::short_vector::uint2 uint2;
+
89 typedef hc::short_vector::uint3 uint3;
+
90 typedef hc::short_vector::uint4 uint4;
+
91 
+
92 // Define ulong vector types
+
93 typedef hc::short_vector::ulong1 ulong1;
+
94 typedef hc::short_vector::ulong2 ulong2;
+
95 typedef hc::short_vector::ulong3 ulong3;
+
96 typedef hc::short_vector::ulong4 ulong4;
+
97 
+
98 // Define ulonglong vector types
+
99 typedef hc::short_vector::ulonglong1 ulonglong1;
+
100 typedef hc::short_vector::ulonglong2 ulonglong2;
+
101 typedef hc::short_vector::ulonglong3 ulonglong3;
+
102 typedef hc::short_vector::ulonglong4 ulonglong4;
+
103 
+
104 
+
105 //-- Floating point
+
106 // Define float vector types
+
107 typedef hc::short_vector::float1 float1;
+
108 typedef hc::short_vector::float2 float2;
+
109 typedef hc::short_vector::float3 float3;
+
110 typedef hc::short_vector::float4 float4;
+
111 
+
112 // Define double vector types
+
113 typedef hc::short_vector::double1 double1;
+
114 typedef hc::short_vector::double2 double2;
+
115 typedef hc::short_vector::double3 double3;
+
116 typedef hc::short_vector::double4 double4;
+
117 
+
118 #else
+
119 
+
120 #define __hip_align(name, val, data) \
+
121  __attribute__((aligned(val))) name \
+
122  { data }
+
123 
+
124 struct __hip_align(char1, 1, signed char x;);
+
125 struct __hip_align(uchar1, 1, unsigned char x;);
+
126 
+
127 struct __hip_align(char2, 2, signed char x; signed char y;);
+
128 struct __hip_align(uchar2, 2, unsigned char x; unsigned char y;);
+
129 
+
130 struct char3
+
131 {
+
132  signed char x, y, z;
+
133 };
+
134 
+
135 struct uchar3
+
136 {
+
137  unsigned char x, y, z;
+
138 };
+
139 
+
140 struct char4
+
141 {
+
142  union {
+
143  signed char x, y, z, w;
+
144  unsigned int val;
+
145  };
+
146 };
+
147 
+
148 struct uchar4
+
149 {
+
150  union {
+
151  unsigned char x, y, z, w;
+
152  unsigned int val;
+
153  };
+
154 };
+
155 
+
156 //struct __hip_align(char4, 4, signed char x; signed char y; signed char z; signed char w;);
+
157 //struct __hip_align(uchar4, 4, unsigned char x; unsigned char y; unsigned char z; unsigned char w;);
+
158 
+
159 struct __hip_align(short1, 2, signed short x;);
+
160 struct __hip_align(ushort1, 2, unsigned short x;);
+
161 
+
162 struct __hip_align(short2, 4, signed short x; signed short y;);
+
163 struct __hip_align(ushort2, 4, unsigned short x; unsigned short y;);
+
164 
+
165 struct short3
+
166 {
+
167  signed short x, y, z;
+
168 };
+
169 
+
170 struct ushort3
+
171 {
+
172  unsigned short x, y, z;
+
173 };
+
174 
+
175 struct __hip_align(short4, 8, signed short x; signed short y; signed short z; signed short w;);
+
176 struct __hip_align(ushort4, 8, unsigned short x; unsigned short y; unsigned short z; unsigned short w;);
+
177 
+
178 struct __hip_align(int1, 4, signed int x;);
+
179 struct __hip_align(uint1, 4, unsigned int x;);
+
180 
+
181 struct __hip_align(int2, 8, signed int x; signed int y;);
+
182 struct __hip_align(uint2, 8, unsigned int x; unsigned int y;);
+
183 
+
184 struct int3{
+
185  signed int x, y, z;
+
186 };
+
187 struct uint3{
+
188  unsigned int x, y, z;
+
189 };
+
190 
+
191 struct __hip_align(int4, 16, signed int x; signed int y; signed int z; signed int w;);
+
192 struct __hip_align(uint4, 16, unsigned int x; unsigned int y; unsigned int z; unsigned int w;);
+
193 
+
194 struct __hip_align(long1, 8, long int x;);
+
195 struct __hip_align(ulong1, 8, unsigned long x;);
+
196 
+
197 struct __hip_align(long2, 16, long int x; long int y;);
+
198 struct __hip_align(ulong2, 16, unsigned long x; unsigned long y;);
+
199 
+
200 struct long3{
+
201  long int x, y, z;
+
202 };
+
203 struct ulong3{
+
204  unsigned long x, y, z;
+
205 };
+
206 
+
207 struct __hip_align(long4, 32, long int x; long int y; long int z; long int w;);
+
208 struct __hip_align(ulong4, 32, unsigned long x; unsigned long y; unsigned long z; unsigned long w;);
+
209 
+
210 struct float1
+
211 {
+
212  float x;
+
213 };
+
214 
+
215 struct __hip_align(float2, 8, float x; float y;);
+
216 
+
217 struct float3
+
218 {
+
219  float x, y, z;
+
220 };
+
221 
+
222 struct __hip_align(float4, 16, float x; float y; float z; float w;);
+
223 
+
224 struct __hip_align(longlong1, 16, long long int x;);
+
225 struct __hip_align(ulonglong1, 16, unsigned long long int x;);
+
226 
+
227 struct __attribute__((aligned(32))) longlong2
+
228 {
+
229  long long int x, y;
+
230 };
+
231 
+
232 struct __attribute__((aligned(32))) ulonglong2
+
233 {
+
234  unsigned long long int x, y;
+
235 };
+
236 
+
237 struct longlong3
+
238 {
+
239  long long int x, y, z;
+
240 };
+
241 
+
242 struct ulonglong3
+
243 {
+
244  unsigned long long int x, y, z;
+
245 };
+
246 
+
247 struct __attribute__((aligned(64))) longlong4
+
248 {
+
249  long long int x, y, z, w;
+
250 };
+
251 
+
252 struct __attribute__((aligned(64))) ulonglong4
+
253 {
+
254  unsigned long long int x, y, z, w;
+
255 };
+
256 
+
257 struct double1
+
258 {
+
259  double x;
+
260 };
+
261 
+
262 struct __attribute__((aligned(16))) double2
+
263 {
+
264  double x, y;
+
265 };
+
266 
+
267 struct double3
+
268 {
+
269  double x, y, z;
+
270 };
+
271 
+
272 struct __attribute__((aligned(32))) double4
+
273 {
+
274  double x, y, z, w;
+
275 };
+
276 
+
277 #endif
+
278 
+
279 #if __HCC__
+
280 #include"hip/hcc_detail/host_defines.h"
+
281 #define __HIP_DEVICE__ __device__ __host__
+
282 #else
+
283 #define __HIP_DEVICE__
+
284 #endif
+
285 
+
286 __HIP_DEVICE__ char1 make_char1(signed char );
+
287 __HIP_DEVICE__ char2 make_char2(signed char, signed char );
+
288 __HIP_DEVICE__ char3 make_char3(signed char, signed char, signed char );
+
289 __HIP_DEVICE__ char4 make_char4(signed char, signed char, signed char, signed char );
+
290 
+
291 __HIP_DEVICE__ short1 make_short1(short );
+
292 __HIP_DEVICE__ short2 make_short2(short, short );
+
293 __HIP_DEVICE__ short3 make_short3(short, short, short );
+
294 __HIP_DEVICE__ short4 make_short4(short, short, short, short );
+
295 
+
296 __HIP_DEVICE__ int1 make_int1(int );
+
297 __HIP_DEVICE__ int2 make_int2(int, int );
+
298 __HIP_DEVICE__ int3 make_int3(int, int, int );
+
299 __HIP_DEVICE__ int4 make_int4(int, int, int, int );
+
300 
+
301 __HIP_DEVICE__ long1 make_long1(long );
+
302 __HIP_DEVICE__ long2 make_long2(long, long );
+
303 __HIP_DEVICE__ long3 make_long3(long, long, long );
+
304 __HIP_DEVICE__ long4 make_long4(long, long, long, long );
+
305 
+
306 __HIP_DEVICE__ longlong1 make_longlong1(long long );
+
307 __HIP_DEVICE__ longlong2 make_longlong2(long long, long long );
+
308 __HIP_DEVICE__ longlong3 make_longlong3(long long, long long, long long );
+
309 __HIP_DEVICE__ longlong4 make_longlong4(long long, long long, long long, long long );
+
310 
+
311 __HIP_DEVICE__ uchar1 make_uchar1(unsigned char );
+
312 __HIP_DEVICE__ uchar2 make_uchar2(unsigned char, unsigned char );
+
313 __HIP_DEVICE__ uchar3 make_uchar3(unsigned char, unsigned char, unsigned char );
+
314 __HIP_DEVICE__ uchar4 make_uchar4(unsigned char, unsigned char, unsigned char, unsigned char );
+
315 
+
316 __HIP_DEVICE__ ushort1 make_ushort1(unsigned short );
+
317 __HIP_DEVICE__ ushort2 make_ushort2(unsigned short, unsigned short );
+
318 __HIP_DEVICE__ ushort3 make_ushort3(unsigned short, unsigned short, unsigned short );
+
319 __HIP_DEVICE__ ushort4 make_ushort4(unsigned short, unsigned short, unsigned short, unsigned short );
+
320 
+
321 __HIP_DEVICE__ uint1 make_uint1(unsigned int );
+
322 __HIP_DEVICE__ uint2 make_uint2(unsigned int, unsigned int );
+
323 __HIP_DEVICE__ uint3 make_uint3(unsigned int, unsigned int, unsigned int );
+
324 __HIP_DEVICE__ uint4 make_uint4(unsigned int, unsigned int, unsigned int, unsigned int );
+
325 
+
326 __HIP_DEVICE__ ulong1 make_ulong1(unsigned long );
+
327 __HIP_DEVICE__ ulong2 make_ulong2(unsigned long, unsigned long );
+
328 __HIP_DEVICE__ ulong3 make_ulong3(unsigned long, unsigned long, unsigned long );
+
329 __HIP_DEVICE__ ulong4 make_ulong4(unsigned long, unsigned long, unsigned long, unsigned long );
+
330 
+
331 __HIP_DEVICE__ ulonglong1 make_ulonglong1(unsigned long long );
+
332 __HIP_DEVICE__ ulonglong2 make_ulonglong2(unsigned long long, unsigned long long);
+
333 __HIP_DEVICE__ ulonglong3 make_ulonglong3(unsigned long long, unsigned long long, unsigned long long);
+
334 __HIP_DEVICE__ ulonglong4 make_ulonglong4(unsigned long long, unsigned long long, unsigned long long, unsigned long long );
+
335 
+
336 __HIP_DEVICE__ float1 make_float1(float );
+
337 __HIP_DEVICE__ float2 make_float2(float, float );
+
338 __HIP_DEVICE__ float3 make_float3(float, float, float );
+
339 __HIP_DEVICE__ float4 make_float4(float, float, float, float );
+
340 
+
341 __HIP_DEVICE__ double1 make_double1(double );
+
342 __HIP_DEVICE__ double2 make_double2(double, double );
+
343 __HIP_DEVICE__ double3 make_double3(double, double, double );
+
344 __HIP_DEVICE__ double4 make_double4(double, double, double, double );
+
345 
+
346 extern __HIP_DEVICE__ double __longlong_as_double(long long int x);
+
347 extern __HIP_DEVICE__ long long int __double_as_longlong(double x);
+
348 
+
349 
+
350 /*
+
352 // Inline functions for creating vector types from basic types
+
353 #define ONE_COMPONENT_ACCESS(T, VT) inline VT make_ ##VT [[hc]] [[cpu]] (T x) { VT t; t.x = x; return t; };
+
354 #define TWO_COMPONENT_ACCESS(T, VT) inline VT make_ ##VT [[hc]] [[cpu]] (T x, T y) { VT t; t.x=x; t.y=y; return t; };
+
355 #define THREE_COMPONENT_ACCESS(T, VT) inline VT make_ ##VT [[hc]] [[cpu]] (T x, T y, T z) { VT t; t.x=x; t.y=y; t.z=z; return t; };
+
356 #define FOUR_COMPONENT_ACCESS(T, VT) inline VT make_ ##VT [[hc]] [[cpu]] (T x, T y, T z, T w) { VT t; t.x=x; t.y=y; t.z=z; t.w=w; return t; };
+
357 
+
358 
+
359 //signed:
+
360 ONE_COMPONENT_ACCESS (signed char, char1);
+
361 TWO_COMPONENT_ACCESS (signed char, char2);
+
362 THREE_COMPONENT_ACCESS(signed char, char3);
+
363 FOUR_COMPONENT_ACCESS (signed char, char4);
+
364 
+
365 ONE_COMPONENT_ACCESS (short, short1);
+
366 TWO_COMPONENT_ACCESS (short, short2);
+
367 THREE_COMPONENT_ACCESS(short, short3);
+
368 FOUR_COMPONENT_ACCESS (short, short4);
+
369 
+
370 ONE_COMPONENT_ACCESS (int, int1);
+
371 TWO_COMPONENT_ACCESS (int, int2);
+
372 THREE_COMPONENT_ACCESS(int, int3);
+
373 FOUR_COMPONENT_ACCESS (int, int4);
+
374 
+
375 ONE_COMPONENT_ACCESS (long int, long1);
+
376 TWO_COMPONENT_ACCESS (long int, long2);
+
377 THREE_COMPONENT_ACCESS(long int, long3);
+
378 FOUR_COMPONENT_ACCESS (long int, long4);
+
379 
+
380 ONE_COMPONENT_ACCESS (long long int, ulong1);
+
381 TWO_COMPONENT_ACCESS (long long int, ulong2);
+
382 THREE_COMPONENT_ACCESS(long long int, ulong3);
+
383 FOUR_COMPONENT_ACCESS (long long int, ulong4);
+
384 
+
385 ONE_COMPONENT_ACCESS (long long int, longlong1);
+
386 TWO_COMPONENT_ACCESS (long long int, longlong2);
+
387 THREE_COMPONENT_ACCESS(long long int, longlong3);
+
388 FOUR_COMPONENT_ACCESS (long long int, longlong4);
+
389 
+
390 
+
391 // unsigned:
+
392 ONE_COMPONENT_ACCESS (unsigned char, uchar1);
+
393 TWO_COMPONENT_ACCESS (unsigned char, uchar2);
+
394 THREE_COMPONENT_ACCESS(unsigned char, uchar3);
+
395 FOUR_COMPONENT_ACCESS (unsigned char, uchar4);
+
396 
+
397 ONE_COMPONENT_ACCESS (unsigned short, ushort1);
+
398 TWO_COMPONENT_ACCESS (unsigned short, ushort2);
+
399 THREE_COMPONENT_ACCESS(unsigned short, ushort3);
+
400 FOUR_COMPONENT_ACCESS (unsigned short, ushort4);
+
401 
+
402 ONE_COMPONENT_ACCESS (unsigned int, uint1);
+
403 TWO_COMPONENT_ACCESS (unsigned int, uint2);
+
404 THREE_COMPONENT_ACCESS(unsigned int, uint3);
+
405 FOUR_COMPONENT_ACCESS (unsigned int, uint4);
+
406 
+
407 ONE_COMPONENT_ACCESS (unsigned long int, ulong1);
+
408 TWO_COMPONENT_ACCESS (unsigned long int, ulong2);
+
409 THREE_COMPONENT_ACCESS(unsigned long int, ulong3);
+
410 FOUR_COMPONENT_ACCESS (unsigned long int, ulong4);
+
411 
+
412 ONE_COMPONENT_ACCESS (unsigned long long int, ulong1);
+
413 TWO_COMPONENT_ACCESS (unsigned long long int, ulong2);
+
414 THREE_COMPONENT_ACCESS(unsigned long long int, ulong3);
+
415 FOUR_COMPONENT_ACCESS (unsigned long long int, ulong4);
+
416 
+
417 ONE_COMPONENT_ACCESS (unsigned long long int, ulonglong1);
+
418 TWO_COMPONENT_ACCESS (unsigned long long int, ulonglong2);
+
419 THREE_COMPONENT_ACCESS(unsigned long long int, ulonglong3);
+
420 FOUR_COMPONENT_ACCESS (unsigned long long int, ulonglong4);
+
421 
+
422 
+
423 //Floating point
+
424 ONE_COMPONENT_ACCESS (float, float1);
+
425 TWO_COMPONENT_ACCESS (float, float2);
+
426 THREE_COMPONENT_ACCESS(float, float3);
+
427 FOUR_COMPONENT_ACCESS (float, float4);
+
428 
+
429 ONE_COMPONENT_ACCESS (double, double1);
+
430 TWO_COMPONENT_ACCESS (double, double2);
+
431 THREE_COMPONENT_ACCESS(double, double3);
+
432 FOUR_COMPONENT_ACCESS (double, double4);
+
433 */
+
434 
+
435 #endif
+
436 
+
uchar3
Definition: hip_vector_types.h:135
+
uint3
Definition: hip_vector_types.h:187
+
ulong3
Definition: hip_vector_types.h:203
+
double1
Definition: hip_vector_types.h:257
+
host_defines.h
TODO-doc.
+
float3
Definition: hip_vector_types.h:217
+
short3
Definition: hip_vector_types.h:165
+
longlong3
Definition: hip_vector_types.h:237
+
int3
Definition: hip_vector_types.h:184
+
ulonglong3
Definition: hip_vector_types.h:242
+
char4
Definition: hip_vector_types.h:140
+
long3
Definition: hip_vector_types.h:200
+
double3
Definition: hip_vector_types.h:267
+
ushort3
Definition: hip_vector_types.h:170
+
char3
Definition: hip_vector_types.h:130
+
uchar4
Definition: hip_vector_types.h:148
+
float1
Definition: hip_vector_types.h:210
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hierarchy.html b/docs/RuntimeAPI/html/hierarchy.html new file mode 100644 index 0000000000..7a51e86a9f --- /dev/null +++ b/docs/RuntimeAPI/html/hierarchy.html @@ -0,0 +1,150 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Class Hierarchy + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + +
+ + + + +
+ +
+ +
+
+
Class Hierarchy
+
+
+
This inheritance list is sorted roughly, but not completely, alphabetically:
+
[detail level 12]
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
oC__half
oCchar3
oCchar4
oCDbName
oCdim3
oCdouble1
oCdouble3
oCexception
|\CihipException
oCFakeMutex
oCfloat1
oCfloat3
oChipArray
oChipChannelFormatDesc
oChipDeviceArch_t
oChipDeviceProp_t
oChipDoubleComplex
oChipFloatComplex
oChipPointerAttribute_t
oCihipCtx_t
oCihipDevice_t
oCihipEvent_t
oCihipFunction_t
oCihipIpcMemHandle_t
oCihipModule_t
oCihipStream_t
oCint3
oCLockedAccessor< T >
oCLockedBase< MUTEX_TYPE >
|oCihipCtxCriticalBase_t< MUTEX_TYPE >
|\CihipStreamCriticalBase_t< MUTEX_TYPE >
oCLockedBase< CtxMutex >
|\CihipCtxCriticalBase_t< CtxMutex >
oCLockedBase< StreamMutex >
|\CihipStreamCriticalBase_t< StreamMutex >
oClong3
oClonglong3
oCProfTrigger
oCshort3
oCShortTid
oCstruct_float
oCtextureReference
oCuchar3
oCuchar4
oCuint3
oCulong3
oCulonglong3
\Cushort3
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__common_8h_source.html b/docs/RuntimeAPI/html/hip__common_8h_source.html new file mode 100644 index 0000000000..fbe04a41b0 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__common_8h_source.html @@ -0,0 +1,186 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_common.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_common.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #pragma once
+
24 
+
25 // Common code included at start of every hip file.
+
26 // Auto enable __HIP_PLATFORM_HCC__ if compiling with HCC
+
27 // Other compiler (GCC,ICC,etc) need to set one of these macros explicitly
+
28 #if defined(__HCC__)
+
29 #define __HIP_PLATFORM_HCC__
+
30 #define __HIPCC__
+
31 
+
32 #if defined(__HCC_ACCELERATOR__) && (__HCC_ACCELERATOR__ != 0)
+
33 #define __HIP_DEVICE_COMPILE__ 1
+
34 #else
+
35 #define __HIP_DEVICE_COMPILE__ 0
+
36 #endif
+
37 #endif
+
38 
+
39 // Auto enable __HIP_PLATFORM_NVCC__ if compiling with NVCC
+
40 #if defined(__NVCC__)
+
41 #define __HIP_PLATFORM_NVCC__
+
42 # ifdef __CUDACC__
+
43 # define __HIPCC__
+
44 # endif
+
45 
+
46 #if defined(__CUDA_ARCH__) && (__CUDA_ARCH__ != 0)
+
47 #define __HIP_DEVICE_COMPILE__ 1
+
48 #else
+
49 #define __HIP_DEVICE_COMPILE__ 0
+
50 #endif
+
51 
+
52 #endif
+
53 
+
54 
+
55 
+
56 
+
57 #if __HIP_DEVICE_COMPILE__ == 0
+
58 // 32-bit Atomics
+
59 #define __HIP_ARCH_HAS_GLOBAL_INT32_ATOMICS__ (0)
+
60 #define __HIP_ARCH_HAS_GLOBAL_FLOAT_ATOMIC_EXCH__ (0)
+
61 #define __HIP_ARCH_HAS_SHARED_INT32_ATOMICS__ (0)
+
62 #define __HIP_ARCH_HAS_SHARED_FLOAT_ATOMIC_EXCH__ (0)
+
63 #define __HIP_ARCH_HAS_FLOAT_ATOMIC_ADD__ (0)
+
64 
+
65 // 64-bit Atomics
+
66 #define __HIP_ARCH_HAS_GLOBAL_INT64_ATOMICS__ (0)
+
67 #define __HIP_ARCH_HAS_SHARED_INT64_ATOMICS__ (0)
+
68 
+
69 // Doubles
+
70 #define __HIP_ARCH_HAS_DOUBLES__ (0)
+
71 
+
72 // Warp cross-lane operations
+
73 #define __HIP_ARCH_HAS_WARP_VOTE__ (0)
+
74 #define __HIP_ARCH_HAS_WARP_BALLOT__ (0)
+
75 #define __HIP_ARCH_HAS_WARP_SHUFFLE__ (0)
+
76 #define __HIP_ARCH_HAS_WARP_FUNNEL_SHIFT__ (0)
+
77 
+
78 // Sync
+
79 #define __HIP_ARCH_HAS_THREAD_FENCE_SYSTEM__ (0)
+
80 #define __HIP_ARCH_HAS_SYNC_THREAD_EXT__ (0)
+
81 
+
82 // Misc
+
83 #define __HIP_ARCH_HAS_SURFACE_FUNCS__ (0)
+
84 #define __HIP_ARCH_HAS_3DGRID__ (0)
+
85 #define __HIP_ARCH_HAS_DYNAMIC_PARALLEL__ (0)
+
86 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__complex_8h_source.html b/docs/RuntimeAPI/html/hip__complex_8h_source.html new file mode 100644 index 0000000000..94525b02a3 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__complex_8h_source.html @@ -0,0 +1,134 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_complex.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_complex.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #pragma once
+
24 
+
25 #include <hip/hip_common.h>
+
26 
+
27 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
28 #include <hip/hcc_detail/hip_complex.h>
+
29 #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__)
+
30 #include <hip/nvcc_detail/hip_complex.h>
+
31 #else
+
32 #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__");
+
33 #endif
+
34 
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__fp16_8h_source.html b/docs/RuntimeAPI/html/hip__fp16_8h_source.html new file mode 100644 index 0000000000..219cf2d03c --- /dev/null +++ b/docs/RuntimeAPI/html/hip__fp16_8h_source.html @@ -0,0 +1,133 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_fp16.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_fp16.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #pragma once
+
24 
+
25 #include <hip/hip_common.h>
+
26 
+
27 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
28 #include <hip/hcc_detail/hip_fp16.h>
+
29 #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__)
+
30 #include "cuda_fp16.h"
+
31 #else
+
32 #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__");
+
33 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__hcc_8cpp.html b/docs/RuntimeAPI/html/hip__hcc_8cpp.html new file mode 100644 index 0000000000..7a1c376dc0 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__hcc_8cpp.html @@ -0,0 +1,486 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/src/hip_hcc.cpp File Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+Macros | +Functions | +Variables
+
+
hip_hcc.cpp File Reference
+
+
+
#include <assert.h>
+#include <stdint.h>
+#include <iostream>
+#include <sstream>
+#include <list>
+#include <sys/types.h>
+#include <unistd.h>
+#include <deque>
+#include <vector>
+#include <algorithm>
+#include <atomic>
+#include <hc.hpp>
+#include <hc_am.hpp>
+#include "hsa/hsa_ext_amd.h"
+#include "libhsakmt/hsakmt.h"
+#include "hip/hip_runtime.h"
+#include "hip_hcc.h"
+#include "trace_helper.h"
+
+ + + + + + + + + + + + + + + + + +

+Macros

+#define USE_COPY_EXT_V2   1
 
+#define HIP_USE_PRODUCT_NAME   1
 
+#define ErrorCheck(x)   error_check(x, __LINE__, __FILE__)
 
+#define DeviceErrorCheck(x)   if (x != HSA_STATUS_SUCCESS) { return hipErrorInvalidDevice; }
 
#define READ_ENV_I(_build, _ENV_VAR, _ENV_VAR2, _description)
 
#define READ_ENV_S(_build, _ENV_VAR, _ENV_VAR2, _description)
 
#define READ_ENV_C(_build, _ENV_VAR, _ENV_VAR2, _description, _callback)
 
+#define CASE_STRING(X)   case X: return #X ;break;
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Functions

+std::string HIP_TRACE_API_COLOR ("green")
 
+std::atomic< int > g_lastShortTid (1)
 
+void recordApiTrace (std::string *fullStr, const std::string &apiStr)
 
+ihipDevice_t * ihipGetDevice (int deviceIndex)
 
+ihipCtx_t * ihipGetPrimaryCtx (unsigned deviceIndex)
 
+void ihipSetTlsDefaultCtx (ihipCtx_t *ctx)
 
+ihipCtx_t * ihipGetTlsDefaultCtx ()
 
+hipError_t ihipSynchronize (void)
 
+void error_check (hsa_status_t hsa_error_code, int line_num, std::string str)
 
+hsa_status_t FindGpuDevice (hsa_agent_t agent, void *data)
 
+hsa_status_t GetDevicePool (hsa_amd_memory_pool_t pool, void *data)
 
+int checkAccess (hsa_agent_t agent, hsa_amd_memory_pool_t pool)
 
+hsa_status_t get_region_info (hsa_region_t region, void *data)
 
+void ihipReadEnv_I (int *var_ptr, const char *var_name1, const char *var_name2, const char *description)
 
+void ihipReadEnv_S (std::string *var_ptr, const char *var_name1, const char *var_name2, const char *description)
 
+void ihipReadEnv_Callback (void *var_ptr, const char *var_name1, const char *var_name2, const char *description, std::string(*setterCallback)(void *var_ptr, const char *env))
 
+void parseTrigger (std::string triggerString, std::vector< ProfTrigger > &profTriggers)
 
+std::string HIP_DB_string (unsigned db)
 
+std::string HIP_DB_callback (void *var_ptr, const char *envVarString)
 
+std::string HIP_VISIBLE_DEVICES_callback (void *var_ptr, const char *envVarString)
 
+void ihipInit ()
 
+hipStream_t ihipSyncAndResolveStream (hipStream_t stream)
 
+void ihipPrintKernelLaunch (const char *kernelName, const grid_launch_parm *lp, const hipStream_t stream)
 
+hipStream_t ihipPreLaunchKernel (hipStream_t stream, dim3 grid, dim3 block, grid_launch_parm *lp, const char *kernelNameStr)
 
+hipStream_t ihipPreLaunchKernel (hipStream_t stream, size_t grid, dim3 block, grid_launch_parm *lp, const char *kernelNameStr)
 
+hipStream_t ihipPreLaunchKernel (hipStream_t stream, dim3 grid, size_t block, grid_launch_parm *lp, const char *kernelNameStr)
 
+hipStream_t ihipPreLaunchKernel (hipStream_t stream, size_t grid, size_t block, grid_launch_parm *lp, const char *kernelNameStr)
 
+void ihipPostLaunchKernel (const char *kernelName, hipStream_t stream, grid_launch_parm &lp)
 
+const char * ihipErrorString (hipError_t hip_error)
 
+void ihipSetTs (hipEvent_t e)
 
+const char * hipMemcpyStr (unsigned memKind)
 
+const char * hcMemcpyStr (hc::hcCommandKind memKind)
 
hipError_t hipProfilerStart ()
 Start recording of profiling information When using this API, start the profiler with profiling disabled. (–startdisabled) More...
 
hipError_t hipProfilerStop ()
 Stop recording of profiling information. When using this API, start the profiler with profiling disabled. (–startdisabled) More...
 
+hipError_t hipHccGetAccelerator (int deviceId, hc::accelerator *acc)
 
+hipError_t hipHccGetAcceleratorView (hipStream_t stream, hc::accelerator_view **av)
 
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Variables

+const int release = 1
 
+const char * API_COLOR = KGRN
 
+const char * API_COLOR_END = KNRM
 
+int HIP_LAUNCH_BLOCKING = 0
 
+std::string HIP_LAUNCH_BLOCKING_KERNELS
 
+std::vector< std::string > g_hipLaunchBlockingKernels
 
+int HIP_API_BLOCKING = 0
 
+int HIP_PRINT_ENV = 0
 
+int HIP_TRACE_API = 0
 
+int HIP_PROFILE_API = 0
 
+std::string HIP_DB_START_API
 
+std::string HIP_DB_STOP_API
 
+int HIP_DB = 0
 
+int HIP_VISIBLE_DEVICES = 0
 
+int HIP_NUM_KERNELS_INFLIGHT = 128
 
+int HIP_WAIT_MODE = 0
 
+int HIP_FORCE_P2P_HOST = 0
 
+int HIP_DENY_PEER_ACCESS = 0
 
+int HIP_FORCE_SYNC_COPY = 0
 
+int HIP_COHERENT_HOST_ALLOC = 0
 
+std::once_flag hip_initialized
 
+ihipDevice_t ** g_deviceArray
 
+bool g_visible_device = false
 
+unsigned g_deviceCnt
 
+std::vector< int > g_hip_visible_devices
 
+hsa_agent_t g_cpu_agent
 
+unsigned g_numLogicalThreads
 
+std::vector< ProfTrigger > g_dbStartTriggers
 
+std::vector< ProfTrigger > g_dbStopTriggers
 
+thread_local hipError_t tls_lastHipError = hipSuccess
 
+thread_local ShortTid tls_shortTid
 
+

Detailed Description

+

Contains definitions for functions that are large enough that we don't want to inline them everywhere. This file is compiled and linked into apps running HIP / HCC path.

+

Macro Definition Documentation

+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
#define READ_ENV_C( _build,
 _ENV_VAR,
 _ENV_VAR2,
 _description,
 _callback 
)
+
+Value:
if (_build == release) {\
+
ihipReadEnv_Callback(&_ENV_VAR, #_ENV_VAR, #_ENV_VAR2, _description, _callback);\
+
};
+
+
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
#define READ_ENV_I( _build,
 _ENV_VAR,
 _ENV_VAR2,
 _description 
)
+
+Value:
if (_build == release) {\
+
ihipReadEnv_I(&_ENV_VAR, #_ENV_VAR, #_ENV_VAR2, _description);\
+
};
+
+
+
+ +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
#define READ_ENV_S( _build,
 _ENV_VAR,
 _ENV_VAR2,
 _description 
)
+
+Value:
if (_build == release) {\
+
ihipReadEnv_S(&_ENV_VAR, #_ENV_VAR, #_ENV_VAR2, _description);\
+
};
+
+
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__hcc_8h_source.html b/docs/RuntimeAPI/html/hip__hcc_8h_source.html new file mode 100644 index 0000000000..47908b7e34 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__hcc_8h_source.html @@ -0,0 +1,922 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/src/hip_hcc.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_hcc.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HIP_HCC_H
+
24 #define HIP_HCC_H
+
25 
+
26 #include <hc.hpp>
+
27 #include <hsa/hsa.h>
+
28 #include "hsa/hsa_ext_amd.h"
+
29 #include "hip_util.h"
+
30 
+
31 
+
32 #if defined(__HCC__) && (__hcc_workweek__ < 16354)
+
33 #error("This version of HIP requires a newer version of HCC.");
+
34 #endif
+
35 
+
36 #define USE_DISPATCH_HSA_KERNEL 1
+
37 //
+
38 
+
39 
+
40 //---
+
41 // Environment variables:
+
42 
+
43 // Intended to distinguish whether an environment variable should be visible only in debug mode, or in debug+release.
+
44 //static const int debug = 0;
+
45 extern const int release;
+
46 
+
47 // TODO - this blocks both kernels and memory ops. Perhaps should have separate env var for kernels?
+
48 extern int HIP_LAUNCH_BLOCKING;
+
49 extern int HIP_API_BLOCKING;
+
50 
+
51 extern int HIP_PRINT_ENV;
+
52 extern int HIP_PROFILE_API;
+
53 //extern int HIP_TRACE_API;
+
54 extern int HIP_ATP;
+
55 extern int HIP_DB;
+
56 extern int HIP_STAGING_SIZE; /* size of staging buffers, in KB */
+
57 extern int HIP_STREAM_SIGNALS; /* number of signals to allocate at stream creation */
+
58 extern int HIP_VISIBLE_DEVICES; /* Contains a comma-separated sequence of GPU identifiers */
+
59 extern int HIP_FORCE_P2P_HOST;
+
60 
+
61 extern int HIP_COHERENT_HOST_ALLOC;
+
62 
+
63 
+
64 //---
+
65 // Chicken bits for disabling functionality to work around potential issues:
+
66 extern int HIP_DISABLE_HW_KERNEL_DEP;
+
67 
+
68 
+
69 // Class to assign a short TID to each new thread, for HIP debugging purposes.
+
70 class ShortTid {
+
71 public:
+
72 
+
73  ShortTid() ;
+
74 
+
75  int tid() const { return _shortTid; };
+
76  uint64_t incApiSeqNum() { return ++_apiSeqNum; };
+
77  uint64_t apiSeqNum() const { return _apiSeqNum; };
+
78 
+
79 private:
+
80  int _shortTid;
+
81 
+
82  // monotonically increasing API sequence number for this threa.
+
83  uint64_t _apiSeqNum;
+
84 };
+
85 
+
86 struct ProfTrigger {
+
87 
+
88  static const uint64_t MAX_TRIGGER = std::numeric_limits<uint64_t>::max();
+
89 
+
90  void print (int tid) {
+
91  std::cout << "Enabling tracing for ";
+
92  for (auto iter=_profTrigger.begin(); iter != _profTrigger.end(); iter++) {
+
93  std::cout << "tid:" << tid << "." << *iter << ",";
+
94  }
+
95  std::cout << "\n";
+
96  };
+
97 
+
98  uint64_t nextTrigger() { return _profTrigger.empty() ? MAX_TRIGGER : _profTrigger.back(); };
+
99  void add(uint64_t trigger) { _profTrigger.push_back(trigger); };
+
100  void sort() { std::sort (_profTrigger.begin(), _profTrigger.end(), std::greater<int>()); };
+
101 private:
+
102  std::vector<uint64_t> _profTrigger;
+
103 };
+
104 
+
105 
+
106 
+
107 //---
+
108 //Extern tls
+
109 extern thread_local hipError_t tls_lastHipError;
+
110 extern thread_local ShortTid tls_shortTid;
+
111 
+
112 extern std::vector<ProfTrigger> g_dbStartTriggers;
+
113 extern std::vector<ProfTrigger> g_dbStopTriggers;
+
114 
+
115 //---
+
116 //Forward defs:
+
117 class ihipStream_t;
+
118 class ihipDevice_t;
+
119 class ihipCtx_t;
+
120 
+
121 // Color defs for debug messages:
+
122 #define KNRM "\x1B[0m"
+
123 #define KRED "\x1B[31m"
+
124 #define KGRN "\x1B[32m"
+
125 #define KYEL "\x1B[33m"
+
126 #define KBLU "\x1B[34m"
+
127 #define KMAG "\x1B[35m"
+
128 #define KCYN "\x1B[36m"
+
129 #define KWHT "\x1B[37m"
+
130 
+
131 extern const char *API_COLOR;
+
132 extern const char *API_COLOR_END;
+
133 
+
134 
+
135 // If set, thread-safety is enforced on all stream functions.
+
136 // Stream functions will acquire a mutex before entering critical sections.
+
137 #define STREAM_THREAD_SAFE 1
+
138 
+
139 
+
140 #define CTX_THREAD_SAFE 1
+
141 
+
142 
+
143 // Compile debug trace mode - this prints debug messages to stderr when env var HIP_DB is set.
+
144 // May be set to 0 to remove debug if checks - possible code size and performance difference?
+
145 #define COMPILE_HIP_DB 1
+
146 
+
147 
+
148 // Compile HIP tracing capability.
+
149 // 0x1 = print a string at function entry with arguments.
+
150 // 0x2 = prints a simple message with function name + return code when function exits.
+
151 // 0x3 = print both.
+
152 // Must be enabled at runtime with HIP_TRACE_API
+
153 #define COMPILE_HIP_TRACE_API 0x3
+
154 
+
155 
+
156 // Compile code that generates trace markers for CodeXL ATP at HIP function begin/end.
+
157 // ATP is standard CodeXL format that includes timestamps for kernels, HSA RT APIs, and HIP APIs.
+
158 #ifndef COMPILE_HIP_ATP_MARKER
+
159 #define COMPILE_HIP_ATP_MARKER 0
+
160 #endif
+
161 
+
162 
+
163 
+
164 
+
165 // Compile support for trace markers that are displayed on CodeXL GUI at start/stop of each function boundary.
+
166 // TODO - currently we print the trace message at the beginning. if we waited, we could also include return codes, and any values returned
+
167 // through ptr-to-args (ie the pointers allocated by hipMalloc).
+
168 #if COMPILE_HIP_ATP_MARKER
+
169 #include "CXLActivityLogger.h"
+
170 #define MARKER_BEGIN(markerName,group) amdtBeginMarker(markerName, group, nullptr);
+
171 #define MARKER_END() amdtEndMarker();
+
172 #define RESUME_PROFILING amdtResumeProfiling(AMDT_ALL_PROFILING);
+
173 #define STOP_PROFILING amdtStopProfiling(AMDT_ALL_PROFILING);
+
174 #else
+
175 // Swallow scoped markers:
+
176 #define MARKER_BEGIN(markerName,group)
+
177 #define MARKER_END()
+
178 #define RESUME_PROFILING
+
179 #define STOP_PROFILING
+
180 #endif
+
181 
+
182 
+
183 extern void recordApiTrace(std::string *fullStr, const std::string &apiStr);
+
184 
+
185 #if COMPILE_HIP_ATP_MARKER || (COMPILE_HIP_TRACE_API & 0x1)
+
186 #define API_TRACE(...)\
+
187 {\
+
188  if (HIP_PROFILE_API || (COMPILE_HIP_DB && HIP_TRACE_API)) {\
+
189  std::string apiStr = std::string(__func__) + " (" + ToString(__VA_ARGS__) + ')';\
+
190  std::string fullStr;\
+
191  recordApiTrace(&fullStr, apiStr);\
+
192  if (HIP_PROFILE_API == 0x1) {MARKER_BEGIN(__func__, "HIP") }\
+
193  else if (HIP_PROFILE_API == 0x2) {MARKER_BEGIN(fullStr.c_str(), "HIP"); }\
+
194  }\
+
195 }
+
196 #else
+
197 // Swallow API_TRACE
+
198 #define API_TRACE(...)
+
199 #endif
+
200 
+
201 
+
202 // Just initialize the HIP runtime, but don't log any trace information.
+
203 #define HIP_INIT()\
+
204  std::call_once(hip_initialized, ihipInit);\
+
205  ihipCtxStackUpdate();
+
206 
+
207 
+
208 // This macro should be called at the beginning of every HIP API.
+
209 // It initialies the hip runtime (exactly once), and
+
210 // generate trace string that can be output to stderr or to ATP file.
+
211 #define HIP_INIT_API(...) \
+
212  HIP_INIT()\
+
213  API_TRACE(__VA_ARGS__);
+
214 
+
215 #define ihipLogStatus(hipStatus) \
+
216  ({\
+
217  hipError_t localHipStatus = hipStatus; /*local copy so hipStatus only evaluated once*/ \
+
218  tls_lastHipError = localHipStatus;\
+
219  \
+
220  if ((COMPILE_HIP_TRACE_API & 0x2) && HIP_TRACE_API) {\
+
221  fprintf(stderr, " %ship-api tid:%d.%lu %-30s ret=%2d (%s)>>%s\n", (localHipStatus == 0) ? API_COLOR:KRED, tls_shortTid.tid(),tls_shortTid.apiSeqNum(), __func__, localHipStatus, ihipErrorString(localHipStatus), API_COLOR_END);\
+
222  }\
+
223  if (HIP_PROFILE_API) { MARKER_END(); }\
+
224  localHipStatus;\
+
225  })
+
226 
+
227 
+
228 
+
229 
+
230 //---
+
231 //HIP_DB Debug flags:
+
232 #define DB_API 0 /* 0x01 - shortcut to enable HIP_TRACE_API on single switch */
+
233 #define DB_SYNC 1 /* 0x02 - trace synchronization pieces */
+
234 #define DB_MEM 2 /* 0x04 - trace memory allocation / deallocation */
+
235 #define DB_COPY 3 /* 0x08 - trace memory copy and peer commands. . */
+
236 #define DB_SIGNAL 4 /* 0x10 - trace signal pool commands */
+
237 #define DB_MAX_FLAG 5
+
238 // When adding a new debug flag, also add to the char name table below.
+
239 //
+
240 
+
241 struct DbName {
+
242  const char *_color;
+
243  const char *_shortName;
+
244 };
+
245 
+
246 static const DbName dbName [] =
+
247 {
+
248  {KGRN, "api"}, // not used,
+
249  {KYEL, "sync"},
+
250  {KCYN, "mem"},
+
251  {KMAG, "copy"},
+
252  {KRED, "signal"},
+
253 };
+
254 
+
255 
+
256 
+
257 #if COMPILE_HIP_DB
+
258 #define tprintf(trace_level, ...) {\
+
259  if (HIP_DB & (1<<(trace_level))) {\
+
260  char msgStr[1000];\
+
261  snprintf(msgStr, 2000, __VA_ARGS__);\
+
262  fprintf (stderr, " %ship-%s tid:%d:%s%s", dbName[trace_level]._color, dbName[trace_level]._shortName, tls_shortTid.tid(), msgStr, KNRM); \
+
263  }\
+
264 }
+
265 #else
+
266 /* Compile to empty code */
+
267 #define tprintf(trace_level, ...)
+
268 #endif
+
269 
+
270 
+
271 
+
272 
+
273 
+
274 class ihipException : public std::exception
+
275 {
+
276 public:
+
277  ihipException(hipError_t e) : _code(e) {};
+
278 
+
279  hipError_t _code;
+
280 };
+
281 
+
282 
+
283 #ifdef __cplusplus
+
284 extern "C" {
+
285 #endif
+
286 
+
287 
+
288 #ifdef __cplusplus
+
289 }
+
290 #endif
+
291 
+
292 const hipStream_t hipStreamNull = 0x0;
+
293 
+
294 
+
295 // Used to remove lock, for performance or stimulating bugs.
+
296 class FakeMutex
+
297 {
+
298  public:
+
299  void lock() { }
+
300  bool try_lock() {return true; }
+
301  void unlock() { }
+
302 };
+
303 
+
304 
+
305 #if STREAM_THREAD_SAFE
+
306 typedef std::mutex StreamMutex;
+
307 #else
+
308 #warning "Stream thread-safe disabled"
+
309 typedef FakeMutex StreamMutex;
+
310 #endif
+
311 
+
312 // Pair Device and Ctx together, these could also be toggled separately if desired.
+
313 #if CTX_THREAD_SAFE
+
314 typedef std::mutex CtxMutex;
+
315 #else
+
316 typedef FakeMutex CtxMutex;
+
317 #warning "Device thread-safe disabled"
+
318 #endif
+
319 
+
320 //
+
321 //---
+
322 // Protects access to the member _data with a lock acquired on contruction/destruction.
+
323 // T must contain a _mutex field which meets the BasicLockable requirements (lock/unlock)
+
324 template<typename T>
+
325 class LockedAccessor
+
326 {
+
327 public:
+
328  LockedAccessor(T &criticalData, bool autoUnlock=true) :
+
329  _criticalData(&criticalData),
+
330  _autoUnlock(autoUnlock)
+
331 
+
332  {
+
333  tprintf(DB_SYNC, "lock critical data %s.%p\n", typeid(T).name(), _criticalData);
+
334  _criticalData->_mutex.lock();
+
335  };
+
336 
+
337  ~LockedAccessor()
+
338  {
+
339  if (_autoUnlock) {
+
340  tprintf(DB_SYNC, "auto-unlock critical data %s.%p\n",typeid(T).name(), _criticalData);
+
341  _criticalData->_mutex.unlock();
+
342  }
+
343  }
+
344 
+
345  void unlock()
+
346  {
+
347  tprintf(DB_SYNC, "unlock critical data %s.%p\n", typeid(T).name(), _criticalData);
+
348  _criticalData->_mutex.unlock();
+
349  }
+
350 
+
351  // Syntactic sugar so -> can be used to get the underlying type.
+
352  T *operator->() { return _criticalData; };
+
353 
+
354 private:
+
355  T *_criticalData;
+
356  bool _autoUnlock;
+
357 };
+
358 
+
359 
+
360 template <typename MUTEX_TYPE>
+
361 struct LockedBase {
+
362 
+
363  // Experts-only interface for explicit locking.
+
364  // Most uses should use the lock-accessor.
+
365  void lock() { _mutex.lock(); }
+
366  void unlock() { _mutex.unlock(); }
+
367 
+
368  MUTEX_TYPE _mutex;
+
369 };
+
370 
+
374 #define HIP_IPC_HANDLE_SIZE 64
+
375 class ihipIpcMemHandle_t
+
376 {
+
377 public:
+
378  hsa_amd_ipc_memory_t ipc_handle;
+
379  char reserved[HIP_IPC_HANDLE_SIZE];
+
380  size_t psize;
+
381 };
+
382 
+
383 class ihipFunction_t{
+
384 public:
+
385  ihipFunction_t(const char *name) {
+
386  size_t nameSz = strlen(name);
+
387  char *kernelName = (char*)malloc(nameSz);
+
388  strncpy(kernelName, name, nameSz);
+
389  _kernelName = kernelName;
+
390  };
+
391 
+
392  ~ihipFunction_t() {
+
393  if (_kernelName) {
+
394  free((void*)_kernelName);
+
395  _kernelName = NULL;
+
396  };
+
397  };
+
398 public:
+
399  const char *_kernelName;
+
400  hsa_executable_symbol_t _kernelSymbol;
+
401  uint64_t _kernel;
+
402 };
+
403 
+
404 class ihipModule_t {
+
405 public:
+
406  hsa_executable_t executable;
+
407  hsa_code_object_t object;
+
408  std::string fileName;
+
409  void *ptr;
+
410  size_t size;
+
411 
+
412  ihipModule_t() : executable(), object(), fileName(), ptr(nullptr), size(0), hipFunctionTable() {}
+
413  ~ihipModule_t() {
+
414  for (int i = 0; i < hipFunctionTable.size(); ++i) {
+
415  ihipFunction_t *func = hipFunctionTable[i];
+
416  delete func;
+
417  }
+
418  hipFunctionTable.clear();
+
419  }
+
420 
+
421  void registerFunction(ihipFunction_t* func) {
+
422  hipFunctionTable.push_back(func);
+
423  }
+
424 private:
+
425  std::vector<ihipFunction_t*> hipFunctionTable;
+
426 };
+
427 
+
428 template <typename MUTEX_TYPE>
+
429 class ihipStreamCriticalBase_t : public LockedBase<MUTEX_TYPE>
+
430 {
+
431 public:
+
432  ihipStreamCriticalBase_t(hc::accelerator_view av) :
+
433  _kernelCnt(0),
+
434  _av(av)
+
435  {
+
436  };
+
437 
+
438  ~ihipStreamCriticalBase_t() {
+
439  }
+
440 
+
441  ihipStreamCriticalBase_t<StreamMutex> * mlock() { LockedBase<MUTEX_TYPE>::lock(); return this;};
+
442 
+
443 public:
+
444  // TODO - remove _kernelCnt mechanism:
+
445  uint32_t _kernelCnt; // Count of inflight kernels in this stream. Reset at ::wait().
+
446  hc::accelerator_view _av;
+
447 };
+
448 
+
449 
+
450 // if HIP code needs to acquire locks for both ihipCtx_t and ihipStream_t, it should first acquire the lock
+
451 // for the ihipCtx_t and then for the individual streams. The locks should not be acquired in reverse order
+
452 // or deadlock may occur. In some cases, it may be possible to reduce the range where the locks must be held.
+
453 // HIP routines should avoid acquiring and releasing the same lock during the execution of a single HIP API.
+
454 
+
455 
+
456 typedef ihipStreamCriticalBase_t<StreamMutex> ihipStreamCritical_t;
+
457 typedef LockedAccessor<ihipStreamCritical_t> LockedAccessor_StreamCrit_t;
+
458 
+
459 
+
460 
+
461 //---
+
462 // Internal stream structure.
+
463 class ihipStream_t {
+
464 public:
+
465  enum ScheduleMode {Auto, Spin, Yield};
+
466  typedef uint64_t SeqNum_t ;
+
467 
+
468  ihipStream_t(ihipCtx_t *ctx, hc::accelerator_view av, unsigned int flags);
+
469  ~ihipStream_t();
+
470 
+
471  // kind is hipMemcpyKind
+
472  void locked_copySync (void* dst, const void* src, size_t sizeBytes, unsigned kind, bool resolveOn = true);
+
473 
+
474 
+
475  void locked_copyAsync(void* dst, const void* src, size_t sizeBytes, unsigned kind);
+
476 
+
477 
+
478  //---
+
479  // Member functions that begin with locked_ are thread-safe accessors - these acquire / release the critical mutex.
+
480  LockedAccessor_StreamCrit_t lockopen_preKernelCommand();
+
481  void lockclose_postKernelCommand(const char *kernelName, hc::accelerator_view *av);
+
482 
+
483 
+
484  void locked_wait(bool assertQueueEmpty=false);
+
485 
+
486  hc::accelerator_view* locked_getAv() { LockedAccessor_StreamCrit_t crit(_criticalData); return &(crit->_av); };
+
487 
+
488  void locked_waitEvent(hipEvent_t event);
+
489  void locked_recordEvent(hipEvent_t event);
+
490 
+
491 
+
492  //---
+
493 
+
494  // Use this if we already have the stream critical data mutex:
+
495  void wait(LockedAccessor_StreamCrit_t &crit, bool assertQueueEmpty=false);
+
496 
+
497  void launchModuleKernel(hc::accelerator_view av, hsa_signal_t signal,
+
498  uint32_t blockDimX, uint32_t blockDimY, uint32_t blockDimZ,
+
499  uint32_t gridDimX, uint32_t gridDimY, uint32_t gridDimZ,
+
500  uint32_t groupSegmentSize, uint32_t sharedMemBytes,
+
501  void *kernarg, size_t kernSize, uint64_t kernel);
+
502 
+
503 
+
504 
+
505  //-- Non-racy accessors:
+
506  // These functions access fields set at initialization time and are non-racy (so do not acquire mutex)
+
507  const ihipDevice_t * getDevice() const;
+
508  ihipCtx_t * getCtx() const;
+
509 
+
510 
+
511 public:
+
512  //---
+
513  //Public member vars - these are set at initialization and never change:
+
514  SeqNum_t _id; // monotonic sequence ID
+
515  unsigned _flags;
+
516 
+
517 
+
518 private:
+
519 
+
520 
+
521  // The unsigned return is hipMemcpyKind
+
522  unsigned resolveMemcpyDirection(bool srcInDeviceMem, bool dstInDeviceMem);
+
523  void resolveHcMemcpyDirection(unsigned hipMemKind,
+
524  const hc::AmPointerInfo *dstPtrInfo, const hc::AmPointerInfo *srcPtrInfo,
+
525  hc::hcCommandKind *hcCopyDir,
+
526  ihipCtx_t **copyDevice,
+
527  bool *forceUnpinnedCopy);
+
528 
+
529  bool canSeeMemory(const ihipCtx_t *thisCtx, const hc::AmPointerInfo *dstInfo, const hc::AmPointerInfo *srcInfo);
+
530 
+
531 
+
532 private: // Data
+
533  // Critical Data - MUST be accessed through LockedAccessor_StreamCrit_t
+
534  ihipStreamCritical_t _criticalData;
+
535 
+
536  ihipCtx_t *_ctx; // parent context that owns this stream.
+
537 
+
538  // Friends:
+
539  friend std::ostream& operator<<(std::ostream& os, const ihipStream_t& s);
+
540  friend hipError_t hipStreamQuery(hipStream_t);
+
541 
+
542  ScheduleMode _scheduleMode;
+
543 };
+
544 
+
545 
+
546 
+
547 //----
+
548 // Internal event structure:
+
549 enum hipEventStatus_t {
+
550  hipEventStatusUnitialized = 0, // event is unutilized, must be "Created" before use.
+
551  hipEventStatusCreated = 1,
+
552  hipEventStatusRecording = 2, // event has been enqueued to record something.
+
553  hipEventStatusRecorded = 3, // event has been recorded - timestamps are valid.
+
554 } ;
+
555 
+
556 
+
557 // internal hip event structure.
+
558 struct ihipEvent_t {
+
559  hipEventStatus_t _state;
+
560 
+
561  hipStream_t _stream; // Stream where the event is recorded, or NULL if all streams.
+
562  unsigned _flags;
+
563 
+
564  hc::completion_future _marker;
+
565  uint64_t _timestamp; // store timestamp, may be set on host or by marker.
+
566 } ;
+
567 
+
568 
+
569 
+
570 
+
571 
+
572 //----
+
573 // Properties of the HIP device.
+
574 // Multiple contexts can point to same device.
+
575 class ihipDevice_t
+
576 {
+
577 public:
+
578  ihipDevice_t(unsigned deviceId, unsigned deviceCnt, hc::accelerator &acc);
+
579  ~ihipDevice_t();
+
580 
+
581  // Accessors:
+
582  ihipCtx_t *getPrimaryCtx() const { return _primaryCtx; };
+
583 
+
584 public:
+
585  unsigned _deviceId; // device ID
+
586 
+
587  hc::accelerator _acc;
+
588  hsa_agent_t _hsaAgent; // hsa agent handle
+
589 
+
591  unsigned _computeUnits;
+
592  hipDeviceProp_t _props; // saved device properties.
+
593 
+
594  // TODO - report this through device properties, base on HCC API call.
+
595  int _isLargeBar;
+
596 
+
597  ihipCtx_t *_primaryCtx;
+
598 
+
599 private:
+
600  hipError_t initProperties(hipDeviceProp_t* prop);
+
601 };
+
602 //=============================================================================
+
603 
+
604 
+
605 
+
606 //=============================================================================
+
607 //class ihipCtxCriticalBase_t
+
608 template <typename MUTEX_TYPE>
+
609 class ihipCtxCriticalBase_t : LockedBase<MUTEX_TYPE>
+
610 {
+
611 public:
+
612  ihipCtxCriticalBase_t(unsigned deviceCnt) :
+
613  _peerCnt(0)
+
614  {
+
615  _peerAgents = new hsa_agent_t[deviceCnt];
+
616  };
+
617 
+
618  ~ihipCtxCriticalBase_t() {
+
619  if (_peerAgents != nullptr) {
+
620  delete _peerAgents;
+
621  _peerAgents = nullptr;
+
622  }
+
623  _peerCnt = 0;
+
624  }
+
625 
+
626  // Streams:
+
627  void addStream(ihipStream_t *stream);
+
628  std::list<ihipStream_t*> &streams() { return _streams; };
+
629  const std::list<ihipStream_t*> &const_streams() const { return _streams; };
+
630 
+
631 
+
632  // Peer Accessor classes:
+
633  bool isPeerWatcher(const ihipCtx_t *peer); // returns True if peer has access to memory physically located on this device.
+
634  bool addPeerWatcher(const ihipCtx_t *thisCtx, ihipCtx_t *peer);
+
635  bool removePeerWatcher(const ihipCtx_t *thisCtx, ihipCtx_t *peer);
+
636  void resetPeerWatchers(ihipCtx_t *thisDevice);
+
637  void printPeerWatchers(FILE *f) const;
+
638 
+
639  uint32_t peerCnt() const { return _peerCnt; };
+
640  hsa_agent_t *peerAgents() const { return _peerAgents; };
+
641 
+
642 
+
643  // TODO - move private
+
644  std::list<ihipCtx_t*> _peers; // list of enabled peer devices.
+
645 
+
646  friend class LockedAccessor<ihipCtxCriticalBase_t>;
+
647 private:
+
648  //--- Stream Tracker:
+
649  std::list< ihipStream_t* > _streams; // streams associated with this device.
+
650 
+
651 
+
652  //--- Peer Tracker:
+
653  // These reflect the currently Enabled set of peers for this GPU:
+
654  // Enabled peers have permissions to access the memory physically allocated on this device.
+
655  // Note the peers always contain the self agent for easy interfacing with HSA APIs.
+
656  uint32_t _peerCnt; // number of enabled peers
+
657  hsa_agent_t *_peerAgents; // efficient packed array of enabled agents (to use for allocations.)
+
658 private:
+
659  void recomputePeerAgents();
+
660 };
+
661 // Note Mutex type Real/Fake selected based on CtxMutex
+
662 typedef ihipCtxCriticalBase_t<CtxMutex> ihipCtxCritical_t;
+
663 
+
664 // This type is used by functions that need access to the critical device structures.
+
665 typedef LockedAccessor<ihipCtxCritical_t> LockedAccessor_CtxCrit_t;
+
666 //=============================================================================
+
667 
+
668 
+
669 //=============================================================================
+
670 //class ihipCtx_t:
+
671 // A HIP CTX (context) points at one of the existing devices and contains the streams,
+
672 // peer-to-peer mappings, creation flags. Multiple contexts can point to the same
+
673 // device.
+
674 //
+
675 class ihipCtx_t
+
676 {
+
677 public: // Functions:
+
678  ihipCtx_t(ihipDevice_t *device, unsigned deviceCnt, unsigned flags); // note: calls constructor for _criticalData
+
679  ~ihipCtx_t();
+
680 
+
681  // Functions which read or write the critical data are named locked_.
+
682  // ihipCtx_t does not use recursive locks so the ihip implementation must avoid calling a locked_ function from within a locked_ function.
+
683  // External functions which call several locked_ functions will acquire and release the lock for each function. if this occurs in
+
684  // performance-sensitive code we may want to refactor by adding non-locked functions and creating a new locked_ member function to call them all.
+
685  void locked_addStream(ihipStream_t *s);
+
686  void locked_removeStream(ihipStream_t *s);
+
687  void locked_reset();
+
688  void locked_waitAllStreams();
+
689  void locked_syncDefaultStream(bool waitOnSelf);
+
690 
+
691  ihipCtxCritical_t &criticalData() { return _criticalData; }; // TODO, move private. Fix P2P.
+
692 
+
693  const ihipDevice_t *getDevice() const { return _device; };
+
694  int getDeviceNum() const { return _device->_deviceId; };
+
695 
+
696  // TODO - review uses of getWriteableDevice(), can these be converted to getDevice()
+
697  ihipDevice_t *getWriteableDevice() const { return _device; };
+
698 
+
699  std::string toString() const;
+
700 
+
701 public: // Data
+
702  // The NULL stream is used if no other stream is specified.
+
703  // Default stream has special synchronization properties with other streams.
+
704  ihipStream_t *_defaultStream;
+
705 
+
706  // Flags specified when the context is created:
+
707  unsigned _ctxFlags;
+
708 
+
709 private:
+
710  ihipDevice_t *_device;
+
711 
+
712 
+
713 private: // Critical data, protected with locked access:
+
714  // Members of _protected data MUST be accessed through the LockedAccessor.
+
715  // Search for LockedAccessor<ihipCtxCritical_t> for examples; do not access _criticalData directly.
+
716  ihipCtxCritical_t _criticalData;
+
717 
+
718 };
+
719 
+
720 
+
721 
+
722 //=================================================================================================
+
723 // Global variable definition:
+
724 extern std::once_flag hip_initialized;
+
725 extern unsigned g_deviceCnt;
+
726 extern hsa_agent_t g_cpu_agent ; // the CPU agent.
+
727 
+
728 //=================================================================================================
+
729 // Extern functions:
+
730 extern void ihipInit();
+
731 extern const char *ihipErrorString(hipError_t);
+
732 extern ihipCtx_t *ihipGetTlsDefaultCtx();
+
733 extern void ihipSetTlsDefaultCtx(ihipCtx_t *ctx);
+
734 extern hipError_t ihipSynchronize(void);
+
735 extern void ihipCtxStackUpdate();
+
736 
+
737 extern ihipDevice_t *ihipGetDevice(int);
+
738 ihipCtx_t * ihipGetPrimaryCtx(unsigned deviceIndex);
+
739 
+
740 extern void ihipSetTs(hipEvent_t e);
+
741 
+
742 
+
743 hipStream_t ihipSyncAndResolveStream(hipStream_t);
+
744 
+
745 // Stream printf functions:
+
746 inline std::ostream& operator<<(std::ostream& os, const ihipStream_t& s)
+
747 {
+
748  os << "stream#";
+
749  os << s.getDevice()->_deviceId;;
+
750  os << '.';
+
751  os << s._id;
+
752  return os;
+
753 }
+
754 
+
755 inline std::ostream & operator<<(std::ostream& os, const dim3& s)
+
756 {
+
757  os << '{';
+
758  os << s.x;
+
759  os << ',';
+
760  os << s.y;
+
761  os << ',';
+
762  os << s.z;
+
763  os << '}';
+
764  return os;
+
765 }
+
766 
+
767 inline std::ostream & operator<<(std::ostream& os, const gl_dim3& s)
+
768 {
+
769  os << '{';
+
770  os << s.x;
+
771  os << ',';
+
772  os << s.y;
+
773  os << ',';
+
774  os << s.z;
+
775  os << '}';
+
776  return os;
+
777 }
+
778 
+
779 // Stream printf functions:
+
780 inline std::ostream& operator<<(std::ostream& os, const hipEvent_t& e)
+
781 {
+
782  os << "event:" << std::hex << static_cast<void*> (e);
+
783  return os;
+
784 }
+
785 
+
786 inline std::ostream& operator<<(std::ostream& os, const ihipCtx_t* c)
+
787 {
+
788  os << "ctx:" << static_cast<const void*> (c)
+
789  << ".dev:" << c->getDevice()->_deviceId;
+
790  return os;
+
791 }
+
792 
+
793 
+
794 // Helper functions that are used across src files:
+
795 namespace hip_internal {
+
796  hipError_t memcpyAsync (void* dst, const void* src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream);
+
797 };
+
798 
+
799 
+
800 #endif
+
DbName
Definition: hip_hcc.h:241
+
ihipDevice_t
Definition: hip_hcc.h:575
+
LockedBase
Definition: hip_hcc.h:361
+
ProfTrigger
Definition: hip_hcc.h:86
+
ihipStream_t::hipStreamQuery
friend hipError_t hipStreamQuery(hipStream_t)
Return hipSuccess if all of the operations in the specified stream have completed, or hipErrorNotReady if not.
Definition: hip_stream.cpp:113
+
FakeMutex
Definition: hip_hcc.h:296
+
ihipIpcMemHandle_t
Definition: hip_hcc.h:375
+
dim3::x
uint32_t x
x
Definition: hip_runtime_api.h:152
+
ihipFunction_t
Definition: hip_hcc.h:383
+
ihipCtx_t
Definition: hip_hcc.h:675
+
dim3
Definition: hip_runtime_api.h:151
+
dim3::y
uint32_t y
y
Definition: hip_runtime_api.h:153
+
hipError_t
hipError_t
Definition: hip_runtime_api.h:152
+
hipMemcpyKind
hipMemcpyKind
Definition: hip_runtime_api.h:165
+
ihipModule_t
Definition: hip_hcc.h:404
+
ihipDevice_t::_computeUnits
unsigned _computeUnits
Number of compute units supported by the device:
Definition: hip_hcc.h:591
+
dim3::z
uint32_t z
z
Definition: hip_runtime_api.h:154
+
ShortTid
Definition: hip_hcc.h:70
+
hipDeviceProp_t
Definition: hip_runtime_api.h:82
+
ihipEvent_t
Definition: hip_hcc.h:558
+
ihipException
Definition: hip_hcc.h:274
+
ihipCtxCriticalBase_t
Definition: hip_hcc.h:609
+
ihipStream_t
Definition: hip_hcc.h:463
+
ihipStreamCriticalBase_t
Definition: hip_hcc.h:429
+
ihipIpcMemHandle_t::ipc_handle
hsa_amd_ipc_memory_t ipc_handle
ipc memory handle on ROCr
Definition: hip_hcc.h:378
+
LockedAccessor
Definition: hip_hcc.h:325
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__ldg_8h_source.html b/docs/RuntimeAPI/html/hip__ldg_8h_source.html new file mode 100644 index 0000000000..3fe22d965f --- /dev/null +++ b/docs/RuntimeAPI/html/hip__ldg_8h_source.html @@ -0,0 +1,181 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/hip_ldg.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_ldg.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HIP_LDG_H
+
24 #define HIP_LDG_H
+
25 
+
26 #if __HCC__
+
27 #if __hcc_workweek__ >= 16164
+
28 #include "hip/hip_vector_types.h"
+
29 #include "hip/hcc_detail/host_defines.h"
+
30 
+
31 
+
32 __device__ char __ldg(const char* );
+
33 __device__ char2 __ldg(const char2* );
+
34 __device__ char4 __ldg(const char4* );
+
35 __device__ signed char __ldg(const signed char* );
+
36 __device__ unsigned char __ldg(const unsigned char* );
+
37 
+
38 __device__ short __ldg(const short* );
+
39 __device__ short2 __ldg(const short2* );
+
40 __device__ short4 __ldg(const short4* );
+
41 __device__ unsigned short __ldg(const unsigned short* );
+
42 
+
43 __device__ int __ldg(const int* );
+
44 __device__ int2 __ldg(const int2* );
+
45 __device__ int4 __ldg(const int4* );
+
46 __device__ unsigned int __ldg(const unsigned int* );
+
47 
+
48 
+
49 __device__ long __ldg(const long* );
+
50 __device__ unsigned long __ldg(const unsigned long* );
+
51 
+
52 __device__ long long __ldg(const long long* );
+
53 __device__ longlong2 __ldg(const longlong2* );
+
54 __device__ unsigned long long __ldg(const unsigned long long* );
+
55 
+
56 __device__ uchar2 __ldg(const uchar2* );
+
57 __device__ uchar4 __ldg(const uchar4* );
+
58 
+
59 __device__ ushort2 __ldg(const ushort2* );
+
60 
+
61 __device__ uint2 __ldg(const uint2* );
+
62 __device__ uint4 __ldg(const uint4* );
+
63 
+
64 __device__ ulonglong2 __ldg(const ulonglong2* );
+
65 
+
66 __device__ float __ldg(const float* );
+
67 __device__ float2 __ldg(const float2* );
+
68 __device__ float4 __ldg(const float4* );
+
69 
+
70 __device__ double __ldg(const double* );
+
71 __device__ double2 __ldg(const double2* );
+
72 
+
73 #endif // __hcc_workweek__
+
74 
+
75 #endif // __HCC__
+
76 
+
77 #endif // HIP_LDG_H
+
78 
+
host_defines.h
TODO-doc.
+
char4
Definition: hip_vector_types.h:140
+
uchar4
Definition: hip_vector_types.h:148
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__profile_8h_source.html b/docs/RuntimeAPI/html/hip__profile_8h_source.html new file mode 100644 index 0000000000..e11e0a46f3 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__profile_8h_source.html @@ -0,0 +1,138 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_profile.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_profile.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #pragma once
+
24 
+
25 #if not defined (ENABLE_HIP_PROFILE)
+
26 #define ENABLE_HIP_PROFILE 1
+
27 #endif
+
28 
+
29 #if defined(__HIP_PLATFORM_HCC__) and (ENABLE_HIP_PROFILE==1)
+
30 #include <CXLActivityLogger.h>
+
31 #define HIP_SCOPED_MARKER(markerName, group) amdtScopedMarker __scopedMarker(markerName, group, nullptr);
+
32 #define HIP_BEGIN_MARKER(markerName, group) amdtBeginMarker(markerName, group, nullptr);
+
33 #define HIP_END_MARKER() amdtEndMarker();
+
34 #else
+
35 #define HIP_SCOPED_MARKER(markerName, group)
+
36 #define HIP_BEGIN_MARKER(markerName, group)
+
37 #define HIP_END_MARKER()
+
38 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__runtime_8h_source.html b/docs/RuntimeAPI/html/hip__runtime_8h_source.html new file mode 100644 index 0000000000..6989a57f94 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__runtime_8h_source.html @@ -0,0 +1,155 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_runtime.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_runtime.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
28 //
+
32 //
+
35 
+
36 #pragma once
+
37 
+
38 // Some standard header files, these are included by hc.hpp and so want to make them avail on both
+
39 // paths to provide a consistent include env and avoid "missing symbol" errors that only appears
+
40 // on NVCC path:
+
41 #include <stdint.h>
+
42 #include <stdio.h>
+
43 #include <stdlib.h>
+
44 #include <assert.h>
+
45 
+
46 #if __cplusplus > 199711L
+
47 #include <thread>
+
48 #endif
+
49 
+
50 
+
51 #include <hip/hip_common.h>
+
52 
+
53 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
54 #include <hip/hcc_detail/hip_runtime.h>
+
55 #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__)
+
56 #include <hip/nvcc_detail/hip_runtime.h>
+
57 #else
+
58 #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__");
+
59 #endif
+
60 
+
61 
+
62 #include <hip/hip_runtime_api.h>
+
63 #include <hip/hip_vector_types.h>
+
64 
+
hip_runtime.h
Contains definitions of APIs for HIP runtime.
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__runtime__api_8h_source.html b/docs/RuntimeAPI/html/hip__runtime__api_8h_source.html new file mode 100644 index 0000000000..61781f127d --- /dev/null +++ b/docs/RuntimeAPI/html/hip__runtime__api_8h_source.html @@ -0,0 +1,431 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_runtime_api.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_runtime_api.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
30 #pragma once
+
31 
+
32 
+
33 #include <string.h> // for getDeviceProp
+
34 #include <hip/hip_common.h>
+
35 
+
36 enum {
+
37 HIP_SUCCESS = 0,
+
38 HIP_ERROR_INVALID_VALUE,
+
39 HIP_ERROR_NOT_INITIALIZED,
+
40 HIP_ERROR_LAUNCH_OUT_OF_RESOURCES
+
41 };
+
42 
+
43 typedef struct {
+
44  // 32-bit Atomics
+
45  unsigned hasGlobalInt32Atomics : 1;
+
46  unsigned hasGlobalFloatAtomicExch : 1;
+
47  unsigned hasSharedInt32Atomics : 1;
+
48  unsigned hasSharedFloatAtomicExch : 1;
+
49  unsigned hasFloatAtomicAdd : 1;
+
50 
+
51  // 64-bit Atomics
+
52  unsigned hasGlobalInt64Atomics : 1;
+
53  unsigned hasSharedInt64Atomics : 1;
+
54 
+
55  // Doubles
+
56  unsigned hasDoubles : 1;
+
57 
+
58  // Warp cross-lane operations
+
59  unsigned hasWarpVote : 1;
+
60  unsigned hasWarpBallot : 1;
+
61  unsigned hasWarpShuffle : 1;
+
62  unsigned hasFunnelShift : 1;
+
63 
+
64  // Sync
+
65  unsigned hasThreadFenceSystem : 1;
+
66  unsigned hasSyncThreadsExt : 1;
+
67 
+
68  // Misc
+
69  unsigned hasSurfaceFuncs : 1;
+
70  unsigned has3dGrid : 1;
+
71  unsigned hasDynamicParallelism : 1;
+
72 } hipDeviceArch_t;
+
73 
+
74 
+
75 //---
+
76 // Common headers for both NVCC and HCC paths:
+
77 
+
82 typedef struct hipDeviceProp_t {
+
83  char name[256];
+
84  size_t totalGlobalMem;
+
85  size_t sharedMemPerBlock;
+
86  int regsPerBlock;
+
87  int warpSize;
+
88  int maxThreadsPerBlock;
+
89  int maxThreadsDim[3];
+
90  int maxGridSize[3];
+
91  int clockRate;
+
92  int memoryClockRate;
+
93  int memoryBusWidth;
+
94  size_t totalConstMem;
+
95  int major;
+
96  int minor;
+
97  int multiProcessorCount;
+
98  int l2CacheSize;
+
99  int maxThreadsPerMultiProcessor;
+
100  int computeMode;
+
101  int clockInstructionRate;
+
102  hipDeviceArch_t arch;
+
103  int concurrentKernels;
+
104  int pciBusID;
+
105  int pciDeviceID;
+
106  size_t maxSharedMemoryPerMultiProcessor;
+
107  int isMultiGpuBoard;
+
108  int canMapHostMemory;
+
109  } hipDeviceProp_t;
+
110 
+
111 
+
115 enum hipMemoryType {
+
116  hipMemoryTypeHost,
+
117  hipMemoryTypeDevice
+
118 };
+
119 
+
120 
+
121 
+
125 typedef struct hipPointerAttribute_t {
+
126  enum hipMemoryType memoryType;
+
127  int device;
+
128  void *devicePointer;
+
129  void *hostPointer;
+
130  int isManaged;
+
131  unsigned allocationFlags; /* flags specified when memory was allocated*/
+
132  /* peers? */
+
133 } hipPointerAttribute_t;
+
134 
+
135 
+
136 // hack to get these to show up in Doxygen:
+
144 /*
+
145  * @brief hipError_t
+
146  * @enum
+
147  * @ingroup Enumerations
+
148  */
+
149 // Developer note - when updating these, update the hipErrorName and hipErrorString functions in NVCC and HCC paths
+
150 // Also update the hipCUDAErrorTohipError function in NVCC path.
+
151 
+
152 typedef enum hipError_t {
+
153  hipSuccess = 0,
+
154  hipErrorOutOfMemory = 2,
+
155  hipErrorNotInitialized = 3,
+
156  hipErrorDeinitialized = 4,
+
157  hipErrorProfilerDisabled = 5,
+
158  hipErrorProfilerNotInitialized = 6,
+
159  hipErrorProfilerAlreadyStarted = 7,
+
160  hipErrorProfilerAlreadyStopped = 8,
+
161  hipErrorInvalidImage = 200,
+
162  hipErrorInvalidContext = 201,
+
163  hipErrorContextAlreadyCurrent = 202,
+
164  hipErrorMapFailed = 205,
+
165  hipErrorUnmapFailed = 206,
+
166  hipErrorArrayIsMapped = 207,
+
167  hipErrorAlreadyMapped = 208,
+
168  hipErrorNoBinaryForGpu = 209,
+
169  hipErrorAlreadyAcquired = 210,
+
170  hipErrorNotMapped = 211,
+
171  hipErrorNotMappedAsArray = 212,
+
172  hipErrorNotMappedAsPointer = 213,
+
173  hipErrorECCNotCorrectable = 214,
+
174  hipErrorUnsupportedLimit = 215,
+
175  hipErrorContextAlreadyInUse = 216,
+
176  hipErrorPeerAccessUnsupported = 217,
+
177  hipErrorInvalidKernelFile = 218,
+
178  hipErrorInvalidGraphicsContext = 219,
+
179  hipErrorInvalidSource = 300,
+
180  hipErrorFileNotFound = 301,
+
181  hipErrorSharedObjectSymbolNotFound = 302,
+
182  hipErrorSharedObjectInitFailed = 303,
+
183  hipErrorOperatingSystem = 304,
+
184  hipErrorInvalidHandle = 400,
+
185  hipErrorNotFound = 500,
+
186  hipErrorIllegalAddress = 700,
+
187  hipErrorInvalidSymbol = 701,
+
188 // Runtime Error Codes start here.
+
189  hipErrorMissingConfiguration = 1001,
+
190  hipErrorMemoryAllocation = 1002,
+
191  hipErrorInitializationError = 1003,
+
192  hipErrorLaunchFailure = 1004,
+
193  hipErrorPriorLaunchFailure = 1005,
+
194  hipErrorLaunchTimeOut = 1006,
+
195  hipErrorLaunchOutOfResources = 1007,
+
196  hipErrorInvalidDeviceFunction = 1008,
+
197  hipErrorInvalidConfiguration = 1009,
+
198  hipErrorInvalidDevice = 1010,
+
199  hipErrorInvalidValue = 1011,
+
200  hipErrorInvalidDevicePointer = 1017,
+
201  hipErrorInvalidMemcpyDirection = 1021,
+
202  hipErrorUnknown = 1030,
+
203  hipErrorInvalidResourceHandle = 1033,
+
204  hipErrorNotReady = 1034,
+
205  hipErrorNoDevice = 1038,
+
206  hipErrorPeerAccessAlreadyEnabled = 1050,
+
207 
+
208  hipErrorPeerAccessNotEnabled = 1051,
+
209  hipErrorRuntimeMemory = 1052,
+
210  hipErrorRuntimeOther = 1053,
+
211  hipErrorHostMemoryAlreadyRegistered = 1061,
+
212  hipErrorHostMemoryNotRegistered = 1062,
+
213  hipErrorMapBufferObjectFailed = 1071,
+
214  hipErrorTbd
+
215 } hipError_t;
+
216 
+
217 /*
+
218  * @brief hipDeviceAttribute_t
+
219  * @enum
+
220  * @ingroup Enumerations
+
221  */
+
222 typedef enum hipDeviceAttribute_t {
+
223  hipDeviceAttributeMaxThreadsPerBlock,
+
224  hipDeviceAttributeMaxBlockDimX,
+
225  hipDeviceAttributeMaxBlockDimY,
+
226  hipDeviceAttributeMaxBlockDimZ,
+
227  hipDeviceAttributeMaxGridDimX,
+
228  hipDeviceAttributeMaxGridDimY,
+
229  hipDeviceAttributeMaxGridDimZ,
+
230  hipDeviceAttributeMaxSharedMemoryPerBlock,
+
231  hipDeviceAttributeTotalConstantMemory,
+
232  hipDeviceAttributeWarpSize,
+
233  hipDeviceAttributeMaxRegistersPerBlock,
+
234  hipDeviceAttributeClockRate,
+
235  hipDeviceAttributeMemoryClockRate,
+
236  hipDeviceAttributeMemoryBusWidth,
+
237  hipDeviceAttributeMultiprocessorCount,
+
238  hipDeviceAttributeComputeMode,
+
239  hipDeviceAttributeL2CacheSize,
+
240  hipDeviceAttributeMaxThreadsPerMultiProcessor,
+
241  hipDeviceAttributeComputeCapabilityMajor,
+
242  hipDeviceAttributeComputeCapabilityMinor,
+
243  hipDeviceAttributeConcurrentKernels,
+
244  hipDeviceAttributePciBusId,
+
245  hipDeviceAttributePciDeviceId,
+
246  hipDeviceAttributeMaxSharedMemoryPerMultiprocessor,
+
247  hipDeviceAttributeIsMultiGpuBoard,
+
248 } hipDeviceAttribute_t;
+
249 
+
254 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
255 #include "hip/hcc_detail/hip_runtime_api.h"
+
256 #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__)
+
257 #include "hip/nvcc_detail/hip_runtime_api.h"
+
258 #else
+
259 #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__");
+
260 #endif
+
261 
+
262 
+
270 #ifdef __cplusplus
+
271 template<class T>
+
272 static inline hipError_t hipMalloc ( T** devPtr, size_t size)
+
273 {
+
274  return hipMalloc((void**)devPtr, size);
+
275 }
+
276 
+
277 // Provide an override to automatically typecast the pointer type from void**, and also provide a default for the flags.
+
278 template<class T>
+
279 static inline hipError_t hipHostMalloc( T** ptr, size_t size, unsigned int flags = hipHostMallocDefault)
+
280 {
+
281  return hipHostMalloc((void**)ptr, size, flags);
+
282 }
+
283 #endif
+
hipErrorNoDevice
Call to hipGetDeviceCount returned 0 devices.
Definition: hip_runtime_api.h:205
+
hipDeviceProp_t::totalConstMem
size_t totalConstMem
Size of shared memory region (in bytes).
Definition: hip_runtime_api.h:94
+
hipDeviceAttributeMaxSharedMemoryPerMultiprocessor
Maximum Shared Memory Per Multiprocessor.
Definition: hip_runtime_api.h:246
+
hipDeviceAttributeMaxBlockDimX
Maximum x-dimension of a block.
Definition: hip_runtime_api.h:224
+
hipDeviceAttributeMaxGridDimX
Maximum x-dimension of a grid.
Definition: hip_runtime_api.h:227
+
hipErrorPeerAccessAlreadyEnabled
Peer access was already enabled from the current device.
Definition: hip_runtime_api.h:206
+
hipErrorRuntimeMemory
HSA runtime memory call returned error. Typically not seen in production systems. ...
Definition: hip_runtime_api.h:209
+
hipDeviceAttributeMemoryBusWidth
Global memory bus width in bits.
Definition: hip_runtime_api.h:236
+
hipErrorMapBufferObjectFailed
Produced when the IPC memory attach failed from ROCr.
Definition: hip_runtime_api.h:213
+
hipSuccess
Successful completion.
Definition: hip_runtime_api.h:153
+
hipDeviceProp_t::minor
int minor
Minor compute capability. On HCC, this is an approximation and features may differ from CUDA CC...
Definition: hip_runtime_api.h:96
+
hipErrorInvalidKernelFile
In CUDA DRV, it is CUDA_ERROR_INVALID_PTX.
Definition: hip_runtime_api.h:177
+
hipDeviceProp_t::canMapHostMemory
int canMapHostMemory
Check whether HIP can map host memory.
Definition: hip_runtime_api.h:108
+
hipDeviceAttributeMaxRegistersPerBlock
Maximum number of 32-bit registers available to a thread block. This number is shared by all thread b...
Definition: hip_runtime_api.h:233
+
hipDeviceProp_t::regsPerBlock
int regsPerBlock
Registers per block.
Definition: hip_runtime_api.h:86
+
hipDeviceAttributeL2CacheSize
Size of L2 cache in bytes. 0 if the device doesn't have L2 cache.
Definition: hip_runtime_api.h:239
+
hipHostMallocDefault
#define hipHostMallocDefault
Flags that can be used with hipHostMalloc.
Definition: hip_runtime_api.h:103
+
hipErrorRuntimeOther
HSA runtime call other than memory returned error. Typically not seen in production systems...
Definition: hip_runtime_api.h:210
+
hipDeviceProp_t::isMultiGpuBoard
int isMultiGpuBoard
1 if device is on a multi-GPU board, 0 if not.
Definition: hip_runtime_api.h:107
+
hipErrorInvalidDevice
DeviceID must be in range 0...#compute-devices.
Definition: hip_runtime_api.h:198
+
hipDeviceAttributeClockRate
Peak clock frequency in kilohertz.
Definition: hip_runtime_api.h:234
+
hipPointerAttribute_t
Definition: hip_runtime_api.h:125
+
hipDeviceProp_t::clockRate
int clockRate
Max clock frequency of the multiProcessors in khz.
Definition: hip_runtime_api.h:91
+
hipDeviceAttributeMaxGridDimZ
Maximum z-dimension of a grid.
Definition: hip_runtime_api.h:229
+
hipDeviceAttributeComputeCapabilityMinor
Minor compute capability version number.
Definition: hip_runtime_api.h:242
+
hipDeviceAttributeMaxSharedMemoryPerBlock
Maximum shared memory available per block in bytes.
Definition: hip_runtime_api.h:230
+
hipDeviceProp_t::pciBusID
int pciBusID
PCI Bus ID.
Definition: hip_runtime_api.h:104
+
hipDeviceAttributeMaxGridDimY
Maximum y-dimension of a grid.
Definition: hip_runtime_api.h:228
+
hipDeviceAttributeIsMultiGpuBoard
Multiple GPU devices.
Definition: hip_runtime_api.h:247
+
hipErrorUnknown
Unknown error.
Definition: hip_runtime_api.h:202
+
hipDeviceProp_t::maxThreadsPerBlock
int maxThreadsPerBlock
Max work items per work group or workgroup max size.
Definition: hip_runtime_api.h:88
+
hipDeviceAttributeMaxBlockDimY
Maximum y-dimension of a block.
Definition: hip_runtime_api.h:225
+
hipHostMalloc
hipError_t hipHostMalloc(void **ptr, size_t size, unsigned int flags)
Allocate device accessible page locked host memory.
Definition: hip_memory.cpp:161
+
hipDeviceProp_t::sharedMemPerBlock
size_t sharedMemPerBlock
Size of shared memory region (in bytes).
Definition: hip_runtime_api.h:85
+
hipDeviceProp_t::maxThreadsPerMultiProcessor
int maxThreadsPerMultiProcessor
Maximum resident threads per multi-processor.
Definition: hip_runtime_api.h:99
+
hipErrorHostMemoryAlreadyRegistered
Produced when trying to lock a page-locked memory.
Definition: hip_runtime_api.h:211
+
hipDeviceProp_t::l2CacheSize
int l2CacheSize
L2 cache size.
Definition: hip_runtime_api.h:98
+
hipDeviceAttribute_t
hipDeviceAttribute_t
Definition: hip_runtime_api.h:222
+
hipDeviceAttributeComputeCapabilityMajor
Major compute capability version number.
Definition: hip_runtime_api.h:241
+
hipErrorPeerAccessNotEnabled
Peer access was never enabled from the current device.
Definition: hip_runtime_api.h:208
+
hipDeviceAttributeMaxThreadsPerBlock
Maximum number of threads per block.
Definition: hip_runtime_api.h:223
+
hipErrorInvalidResourceHandle
Resource handle (hipEvent_t or hipStream_t) invalid.
Definition: hip_runtime_api.h:203
+
hipErrorMemoryAllocation
Memory allocation error.
Definition: hip_runtime_api.h:190
+
hipDeviceProp_t::arch
hipDeviceArch_t arch
Architectural feature flags. New for HIP.
Definition: hip_runtime_api.h:102
+
hipDeviceProp_t::maxGridSize
int maxGridSize[3]
Max grid dimensions (XYZ).
Definition: hip_runtime_api.h:90
+
hipDeviceProp_t::computeMode
int computeMode
Compute mode.
Definition: hip_runtime_api.h:100
+
hipDeviceAttributeMaxBlockDimZ
Maximum z-dimension of a block.
Definition: hip_runtime_api.h:226
+
hipDeviceAttributePciBusId
PCI Bus ID.
Definition: hip_runtime_api.h:244
+
hipErrorInvalidMemcpyDirection
Invalid memory copy direction.
Definition: hip_runtime_api.h:201
+
hipErrorTbd
Marker that more error codes are needed.
Definition: hip_runtime_api.h:214
+
hipDeviceAttributeWarpSize
Warp size in threads.
Definition: hip_runtime_api.h:232
+
hipDeviceProp_t::major
int major
Major compute capability. On HCC, this is an approximation and features may differ from CUDA CC...
Definition: hip_runtime_api.h:95
+
hipDeviceAttributeMemoryClockRate
Peak memory clock frequency in kilohertz.
Definition: hip_runtime_api.h:235
+
hipDeviceAttributeMaxThreadsPerMultiProcessor
Maximum resident threads per multiprocessor.
Definition: hip_runtime_api.h:240
+
hipError_t
hipError_t
Definition: hip_runtime_api.h:152
+
hipDeviceProp_t::clockInstructionRate
int clockInstructionRate
Frequency in khz of the timer used by the device-side "clock*" instructions. New for HIP...
Definition: hip_runtime_api.h:101
+
hipDeviceAttributeTotalConstantMemory
Constant memory size in bytes.
Definition: hip_runtime_api.h:231
+
hipDeviceProp_t::warpSize
int warpSize
Warp size.
Definition: hip_runtime_api.h:87
+
hipDeviceProp_t::concurrentKernels
int concurrentKernels
Device can possibly execute multiple kernels concurrently.
Definition: hip_runtime_api.h:103
+
hipDeviceProp_t::totalGlobalMem
size_t totalGlobalMem
Size of global memory region (in bytes).
Definition: hip_runtime_api.h:84
+
hipErrorInvalidDevicePointer
Invalid Device Pointer.
Definition: hip_runtime_api.h:200
+
hipErrorLaunchFailure
An exception occurred on the device while executing a kernel.
Definition: hip_runtime_api.h:192
+
hipMalloc
hipError_t hipMalloc(void **ptr, size_t size)
Allocate memory on the default accelerator.
Definition: hip_memory.cpp:105
+
hipDeviceAttributeComputeMode
Compute mode that device is currently in.
Definition: hip_runtime_api.h:238
+
hipDeviceAttributePciDeviceId
PCI Device ID.
Definition: hip_runtime_api.h:245
+
hipDeviceProp_t::maxThreadsDim
int maxThreadsDim[3]
Max number of threads in each dimension (XYZ) of a block.
Definition: hip_runtime_api.h:89
+
hipDeviceAttributeMultiprocessorCount
Number of multiprocessors on the device.
Definition: hip_runtime_api.h:237
+
hipDeviceProp_t::memoryBusWidth
int memoryBusWidth
Global memory bus width in bits.
Definition: hip_runtime_api.h:93
+
hipErrorInvalidValue
One or more of the parameters passed to the API call is NULL or not in an acceptable range...
Definition: hip_runtime_api.h:199
+
hipDeviceProp_t
Definition: hip_runtime_api.h:82
+
hipErrorNotReady
Indicates that asynchronous operations enqueued earlier are not ready. This is not actually an error...
Definition: hip_runtime_api.h:204
+
hipDeviceProp_t::maxSharedMemoryPerMultiProcessor
size_t maxSharedMemoryPerMultiProcessor
Maximum Shared Memory Per Multiprocessor.
Definition: hip_runtime_api.h:106
+
hipDeviceProp_t::pciDeviceID
int pciDeviceID
PCI Device ID.
Definition: hip_runtime_api.h:105
+
hipDeviceProp_t::name
char name[256]
Device name.
Definition: hip_runtime_api.h:83
+
hipErrorHostMemoryNotRegistered
Produced when trying to unlock a non-page-locked memory.
Definition: hip_runtime_api.h:212
+
hipDeviceArch_t
Definition: hip_runtime_api.h:43
+
hip_runtime_api.h
Contains C function APIs for HIP runtime. This file does not use any HCC builtin or special language ...
+
hipDeviceProp_t::memoryClockRate
int memoryClockRate
Max global memory clock frequency in khz.
Definition: hip_runtime_api.h:92
+
hipErrorInitializationError
TODO comment from hipErrorInitializationError.
Definition: hip_runtime_api.h:191
+
hipErrorInvalidContext
Produced when input context is invalid.
Definition: hip_runtime_api.h:162
+
hipDeviceAttributeConcurrentKernels
Device can possibly execute multiple kernels concurrently.
Definition: hip_runtime_api.h:243
+
hipDeviceProp_t::multiProcessorCount
int multiProcessorCount
Number of multi-processors (compute units).
Definition: hip_runtime_api.h:97
+
hipErrorLaunchOutOfResources
Out of resources error.
Definition: hip_runtime_api.h:195
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__texture_8h_source.html b/docs/RuntimeAPI/html/hip__texture_8h_source.html new file mode 100644 index 0000000000..6fdfd354c0 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__texture_8h_source.html @@ -0,0 +1,138 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_texture.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_texture.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 
+
24 
+
25 #ifndef HIP_HIP_TEXTURE_H
+
26 #define HIP_HIP_TEXTURE_H
+
27 
+
28 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
29 #include <hip/hcc_detail/hip_texture.h>
+
30 #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__)
+
31 #include <hip/nvcc_detail/hip_texture.h>
+
32 #else
+
33 #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__");
+
34 #endif
+
35 
+
36 
+
37 #endif
+
hip_texture.h
HIP C++ Texture API for hcc compiler.
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__util_8h_source.html b/docs/RuntimeAPI/html/hip__util_8h_source.html new file mode 100644 index 0000000000..b3ecdd89b1 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__util_8h_source.html @@ -0,0 +1,138 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/src/hip_util.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_util.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 #ifndef HIP_UTIL_H
+
24 #define HIP_UTIL_H
+
25 
+
26 #include <assert.h>
+
27 #include <stdint.h>
+
28 #include <iostream>
+
29 #include <sstream>
+
30 #include <list>
+
31 #include <sys/types.h>
+
32 #include <unistd.h>
+
33 #include <deque>
+
34 #include <vector>
+
35 #include <algorithm>
+
36 
+
37 
+
38 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/hip__vector__types_8h_source.html b/docs/RuntimeAPI/html/hip__vector__types_8h_source.html new file mode 100644 index 0000000000..dcc0703001 --- /dev/null +++ b/docs/RuntimeAPI/html/hip__vector__types_8h_source.html @@ -0,0 +1,138 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hip_vector_types.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
hip_vector_types.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
24 
+
25 #pragma once
+
26 
+
27 #include <hip/hip_common.h>
+
28 
+
29 
+
30 #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__)
+
31 #if __cplusplus
+
32 #include <hip/hcc_detail/hip_vector_types.h>
+
33 #endif
+
34 #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__)
+
35 #include <vector_types.h>
+
36 #else
+
37 #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__");
+
38 #endif
+
hip_vector_types.h
Defines the different newt vector types for HIP runtime.
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/host__defines_8h.html b/docs/RuntimeAPI/html/host__defines_8h.html new file mode 100644 index 0000000000..9b8eb086e1 --- /dev/null +++ b/docs/RuntimeAPI/html/host__defines_8h.html @@ -0,0 +1,147 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/host_defines.h File Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+Macros
+
+
host_defines.h File Reference
+
+
+ +

TODO-doc. +More...

+ +

Go to the source code of this file.

+ + + + + + + + + + + + + + + + +

+Macros

#define __host__   __attribute__((cpu))
 
+#define __device__   __attribute__((hc))
 
+#define __global__   __attribute__((hc_grid_launch))
 
+#define __noinline__   __attribute__((noinline))
 
+#define __forceinline__   __attribute__((always_inline))
 
+#define __shared__   tile_static
 
+#define __constant__   __attribute__((address_space(1)))
 
+

Detailed Description

+

TODO-doc.

+

Macro Definition Documentation

+ +
+
+ + + + +
#define __host__   __attribute__((cpu))
+
+

Function and kernel markers

+ +
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/host__defines_8h_source.html b/docs/RuntimeAPI/html/host__defines_8h_source.html new file mode 100644 index 0000000000..f8eac1b2a7 --- /dev/null +++ b/docs/RuntimeAPI/html/host__defines_8h_source.html @@ -0,0 +1,160 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/include/hip/hcc_detail/host_defines.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
host_defines.h
+
+
+Go to the documentation of this file.
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
28 #ifndef HOST_DEFINES_H
+
29 #define HOST_DEFINES_H
+
30 
+
31 #ifdef __HCC__
+
32 
+
35 #define __host__ __attribute__((cpu))
+
36 #define __device__ __attribute__((hc))
+
37 
+
38 #define __global__ __attribute__((hc_grid_launch))
+
39 
+
40 #define __noinline__ __attribute__((noinline))
+
41 #define __forceinline__ __attribute__((always_inline))
+
42 
+
43 
+
44 
+
45 /*
+
46  * Variable Type Qualifiers:
+
47  */
+
48 // _restrict is supported by the compiler
+
49 #define __shared__ tile_static
+
50 #define __constant__ __attribute__((address_space(1)))
+
51 
+
52 #else
+
53 // Non-HCC compiler
+
57 #define __host__
+
58 #define __device__
+
59 
+
60 #define __global__
+
61 
+
62 #define __noinline__
+
63 #define __forceinline__
+
64 
+
65 #define __shared__
+
66 #define __constant__
+
67 
+
68 #endif
+
69 
+
70 #endif
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/index.html b/docs/RuntimeAPI/html/index.html new file mode 100644 index 0000000000..a5823c259c --- /dev/null +++ b/docs/RuntimeAPI/html/index.html @@ -0,0 +1,99 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Heterogeneous-computing Interface for Portability (HIP) + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+
Heterogeneous-computing Interface for Portability (HIP)
+
+
+

The HIP interface makes it very easy to port existing CUDA apps to run on AMD GPUs, or to develop new apps that can run on either CUDA or AMD GPUs from a common source base.

+
    +
  • HIP is very thin and has little or no performance impact over coding directly in CUDA NVCC or HCC HC mode.
    • +
  • HIP allows developers to use the "best" development environment and tools on each target platform.
    • +
  • HIP allows coding in a single-source C++ programming language including features such as templates, C++11 lambdas,and more.
    • +
  • "hipify" tool automatically converts source from CUDA to HIP.
    • +
  • Developers can specialize for CUDA or HCC to tune for performance or handle tricky cases with #ifdef.
    • +
  • See the HIP API.
    • +
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/jquery.js b/docs/RuntimeAPI/html/jquery.js new file mode 100644 index 0000000000..c197801c5d --- /dev/null +++ b/docs/RuntimeAPI/html/jquery.js @@ -0,0 +1,31 @@ +/*! + * jQuery JavaScript Library v1.7.1 + * http://jquery.com/ + * + * Copyright 2011, John Resig + * Dual licensed under the MIT or GPL Version 2 licenses. + * http://jquery.org/license + * + * Includes Sizzle.js + * http://sizzlejs.com/ + * Copyright 2011, The Dojo Foundation + * Released under the MIT, BSD, and GPL Licenses. + * + * Date: Mon Nov 21 21:11:03 2011 -0500 + */ +(function(bb,L){var av=bb.document,bu=bb.navigator,bl=bb.location;var b=(function(){var bF=function(b0,b1){return new bF.fn.init(b0,b1,bD)},bU=bb.jQuery,bH=bb.$,bD,bY=/^(?:[^#<]*(<[\w\W]+>)[^>]*$|#([\w\-]*)$)/,bM=/\S/,bI=/^\s+/,bE=/\s+$/,bA=/^<(\w+)\s*\/?>(?:<\/\1>)?$/,bN=/^[\],:{}\s]*$/,bW=/\\(?:["\\\/bfnrt]|u[0-9a-fA-F]{4})/g,bP=/"[^"\\\n\r]*"|true|false|null|-?\d+(?:\.\d*)?(?:[eE][+\-]?\d+)?/g,bJ=/(?:^|:|,)(?:\s*\[)+/g,by=/(webkit)[ \/]([\w.]+)/,bR=/(opera)(?:.*version)?[ \/]([\w.]+)/,bQ=/(msie) ([\w.]+)/,bS=/(mozilla)(?:.*? rv:([\w.]+))?/,bB=/-([a-z]|[0-9])/ig,bZ=/^-ms-/,bT=function(b0,b1){return(b1+"").toUpperCase()},bX=bu.userAgent,bV,bC,e,bL=Object.prototype.toString,bG=Object.prototype.hasOwnProperty,bz=Array.prototype.push,bK=Array.prototype.slice,bO=String.prototype.trim,bv=Array.prototype.indexOf,bx={};bF.fn=bF.prototype={constructor:bF,init:function(b0,b4,b3){var b2,b5,b1,b6;if(!b0){return this}if(b0.nodeType){this.context=this[0]=b0;this.length=1;return this}if(b0==="body"&&!b4&&av.body){this.context=av;this[0]=av.body;this.selector=b0;this.length=1;return this}if(typeof b0==="string"){if(b0.charAt(0)==="<"&&b0.charAt(b0.length-1)===">"&&b0.length>=3){b2=[null,b0,null]}else{b2=bY.exec(b0)}if(b2&&(b2[1]||!b4)){if(b2[1]){b4=b4 instanceof bF?b4[0]:b4;b6=(b4?b4.ownerDocument||b4:av);b1=bA.exec(b0);if(b1){if(bF.isPlainObject(b4)){b0=[av.createElement(b1[1])];bF.fn.attr.call(b0,b4,true)}else{b0=[b6.createElement(b1[1])]}}else{b1=bF.buildFragment([b2[1]],[b6]);b0=(b1.cacheable?bF.clone(b1.fragment):b1.fragment).childNodes}return bF.merge(this,b0)}else{b5=av.getElementById(b2[2]);if(b5&&b5.parentNode){if(b5.id!==b2[2]){return b3.find(b0)}this.length=1;this[0]=b5}this.context=av;this.selector=b0;return this}}else{if(!b4||b4.jquery){return(b4||b3).find(b0)}else{return this.constructor(b4).find(b0)}}}else{if(bF.isFunction(b0)){return b3.ready(b0)}}if(b0.selector!==L){this.selector=b0.selector;this.context=b0.context}return bF.makeArray(b0,this)},selector:"",jquery:"1.7.1",length:0,size:function(){return this.length},toArray:function(){return bK.call(this,0)},get:function(b0){return b0==null?this.toArray():(b0<0?this[this.length+b0]:this[b0])},pushStack:function(b1,b3,b0){var b2=this.constructor();if(bF.isArray(b1)){bz.apply(b2,b1)}else{bF.merge(b2,b1)}b2.prevObject=this;b2.context=this.context;if(b3==="find"){b2.selector=this.selector+(this.selector?" ":"")+b0}else{if(b3){b2.selector=this.selector+"."+b3+"("+b0+")"}}return b2},each:function(b1,b0){return bF.each(this,b1,b0)},ready:function(b0){bF.bindReady();bC.add(b0);return this},eq:function(b0){b0=+b0;return b0===-1?this.slice(b0):this.slice(b0,b0+1)},first:function(){return this.eq(0)},last:function(){return this.eq(-1)},slice:function(){return this.pushStack(bK.apply(this,arguments),"slice",bK.call(arguments).join(","))},map:function(b0){return this.pushStack(bF.map(this,function(b2,b1){return b0.call(b2,b1,b2)}))},end:function(){return this.prevObject||this.constructor(null)},push:bz,sort:[].sort,splice:[].splice};bF.fn.init.prototype=bF.fn;bF.extend=bF.fn.extend=function(){var b9,b2,b0,b1,b6,b7,b5=arguments[0]||{},b4=1,b3=arguments.length,b8=false;if(typeof b5==="boolean"){b8=b5;b5=arguments[1]||{};b4=2}if(typeof b5!=="object"&&!bF.isFunction(b5)){b5={}}if(b3===b4){b5=this;--b4}for(;b40){return}bC.fireWith(av,[bF]);if(bF.fn.trigger){bF(av).trigger("ready").off("ready")}}},bindReady:function(){if(bC){return}bC=bF.Callbacks("once memory");if(av.readyState==="complete"){return setTimeout(bF.ready,1)}if(av.addEventListener){av.addEventListener("DOMContentLoaded",e,false);bb.addEventListener("load",bF.ready,false)}else{if(av.attachEvent){av.attachEvent("onreadystatechange",e);bb.attachEvent("onload",bF.ready);var b0=false;try{b0=bb.frameElement==null}catch(b1){}if(av.documentElement.doScroll&&b0){bw()}}}},isFunction:function(b0){return bF.type(b0)==="function"},isArray:Array.isArray||function(b0){return bF.type(b0)==="array"},isWindow:function(b0){return b0&&typeof b0==="object"&&"setInterval" in b0},isNumeric:function(b0){return !isNaN(parseFloat(b0))&&isFinite(b0)},type:function(b0){return b0==null?String(b0):bx[bL.call(b0)]||"object"},isPlainObject:function(b2){if(!b2||bF.type(b2)!=="object"||b2.nodeType||bF.isWindow(b2)){return false}try{if(b2.constructor&&!bG.call(b2,"constructor")&&!bG.call(b2.constructor.prototype,"isPrototypeOf")){return false}}catch(b1){return false}var b0;for(b0 in b2){}return b0===L||bG.call(b2,b0)},isEmptyObject:function(b1){for(var b0 in b1){return false}return true},error:function(b0){throw new Error(b0)},parseJSON:function(b0){if(typeof b0!=="string"||!b0){return null}b0=bF.trim(b0);if(bb.JSON&&bb.JSON.parse){return bb.JSON.parse(b0)}if(bN.test(b0.replace(bW,"@").replace(bP,"]").replace(bJ,""))){return(new Function("return "+b0))()}bF.error("Invalid JSON: "+b0)},parseXML:function(b2){var b0,b1;try{if(bb.DOMParser){b1=new DOMParser();b0=b1.parseFromString(b2,"text/xml")}else{b0=new ActiveXObject("Microsoft.XMLDOM");b0.async="false";b0.loadXML(b2)}}catch(b3){b0=L}if(!b0||!b0.documentElement||b0.getElementsByTagName("parsererror").length){bF.error("Invalid XML: "+b2)}return b0},noop:function(){},globalEval:function(b0){if(b0&&bM.test(b0)){(bb.execScript||function(b1){bb["eval"].call(bb,b1)})(b0)}},camelCase:function(b0){return b0.replace(bZ,"ms-").replace(bB,bT)},nodeName:function(b1,b0){return b1.nodeName&&b1.nodeName.toUpperCase()===b0.toUpperCase()},each:function(b3,b6,b2){var b1,b4=0,b5=b3.length,b0=b5===L||bF.isFunction(b3);if(b2){if(b0){for(b1 in b3){if(b6.apply(b3[b1],b2)===false){break}}}else{for(;b40&&b0[0]&&b0[b1-1])||b1===0||bF.isArray(b0));if(b3){for(;b21?aJ.call(arguments,0):bG;if(!(--bw)){bC.resolveWith(bC,bx)}}}function bz(bF){return function(bG){bB[bF]=arguments.length>1?aJ.call(arguments,0):bG;bC.notifyWith(bE,bB)}}if(e>1){for(;bv
a";bI=bv.getElementsByTagName("*");bF=bv.getElementsByTagName("a")[0];if(!bI||!bI.length||!bF){return{}}bG=av.createElement("select");bx=bG.appendChild(av.createElement("option"));bE=bv.getElementsByTagName("input")[0];bJ={leadingWhitespace:(bv.firstChild.nodeType===3),tbody:!bv.getElementsByTagName("tbody").length,htmlSerialize:!!bv.getElementsByTagName("link").length,style:/top/.test(bF.getAttribute("style")),hrefNormalized:(bF.getAttribute("href")==="/a"),opacity:/^0.55/.test(bF.style.opacity),cssFloat:!!bF.style.cssFloat,checkOn:(bE.value==="on"),optSelected:bx.selected,getSetAttribute:bv.className!=="t",enctype:!!av.createElement("form").enctype,html5Clone:av.createElement("nav").cloneNode(true).outerHTML!=="<:nav>",submitBubbles:true,changeBubbles:true,focusinBubbles:false,deleteExpando:true,noCloneEvent:true,inlineBlockNeedsLayout:false,shrinkWrapBlocks:false,reliableMarginRight:true};bE.checked=true;bJ.noCloneChecked=bE.cloneNode(true).checked;bG.disabled=true;bJ.optDisabled=!bx.disabled;try{delete bv.test}catch(bC){bJ.deleteExpando=false}if(!bv.addEventListener&&bv.attachEvent&&bv.fireEvent){bv.attachEvent("onclick",function(){bJ.noCloneEvent=false});bv.cloneNode(true).fireEvent("onclick")}bE=av.createElement("input");bE.value="t";bE.setAttribute("type","radio");bJ.radioValue=bE.value==="t";bE.setAttribute("checked","checked");bv.appendChild(bE);bD=av.createDocumentFragment();bD.appendChild(bv.lastChild);bJ.checkClone=bD.cloneNode(true).cloneNode(true).lastChild.checked;bJ.appendChecked=bE.checked;bD.removeChild(bE);bD.appendChild(bv);bv.innerHTML="";if(bb.getComputedStyle){bA=av.createElement("div");bA.style.width="0";bA.style.marginRight="0";bv.style.width="2px";bv.appendChild(bA);bJ.reliableMarginRight=(parseInt((bb.getComputedStyle(bA,null)||{marginRight:0}).marginRight,10)||0)===0}if(bv.attachEvent){for(by in {submit:1,change:1,focusin:1}){bB="on"+by;bw=(bB in bv);if(!bw){bv.setAttribute(bB,"return;");bw=(typeof bv[bB]==="function")}bJ[by+"Bubbles"]=bw}}bD.removeChild(bv);bD=bG=bx=bA=bv=bE=null;b(function(){var bM,bU,bV,bT,bN,bO,bL,bS,bR,e,bP,bQ=av.getElementsByTagName("body")[0];if(!bQ){return}bL=1;bS="position:absolute;top:0;left:0;width:1px;height:1px;margin:0;";bR="visibility:hidden;border:0;";e="style='"+bS+"border:5px solid #000;padding:0;'";bP="
";bM=av.createElement("div");bM.style.cssText=bR+"width:0;height:0;position:static;top:0;margin-top:"+bL+"px";bQ.insertBefore(bM,bQ.firstChild);bv=av.createElement("div");bM.appendChild(bv);bv.innerHTML="
t
";bz=bv.getElementsByTagName("td");bw=(bz[0].offsetHeight===0);bz[0].style.display="";bz[1].style.display="none";bJ.reliableHiddenOffsets=bw&&(bz[0].offsetHeight===0);bv.innerHTML="";bv.style.width=bv.style.paddingLeft="1px";b.boxModel=bJ.boxModel=bv.offsetWidth===2;if(typeof bv.style.zoom!=="undefined"){bv.style.display="inline";bv.style.zoom=1;bJ.inlineBlockNeedsLayout=(bv.offsetWidth===2);bv.style.display="";bv.innerHTML="
";bJ.shrinkWrapBlocks=(bv.offsetWidth!==2)}bv.style.cssText=bS+bR;bv.innerHTML=bP;bU=bv.firstChild;bV=bU.firstChild;bN=bU.nextSibling.firstChild.firstChild;bO={doesNotAddBorder:(bV.offsetTop!==5),doesAddBorderForTableAndCells:(bN.offsetTop===5)};bV.style.position="fixed";bV.style.top="20px";bO.fixedPosition=(bV.offsetTop===20||bV.offsetTop===15);bV.style.position=bV.style.top="";bU.style.overflow="hidden";bU.style.position="relative";bO.subtractsBorderForOverflowNotVisible=(bV.offsetTop===-5);bO.doesNotIncludeMarginInBodyOffset=(bQ.offsetTop!==bL);bQ.removeChild(bM);bv=bM=null;b.extend(bJ,bO)});return bJ})();var aS=/^(?:\{.*\}|\[.*\])$/,aA=/([A-Z])/g;b.extend({cache:{},uuid:0,expando:"jQuery"+(b.fn.jquery+Math.random()).replace(/\D/g,""),noData:{embed:true,object:"clsid:D27CDB6E-AE6D-11cf-96B8-444553540000",applet:true},hasData:function(e){e=e.nodeType?b.cache[e[b.expando]]:e[b.expando];return !!e&&!S(e)},data:function(bx,bv,bz,by){if(!b.acceptData(bx)){return}var bG,bA,bD,bE=b.expando,bC=typeof bv==="string",bF=bx.nodeType,e=bF?b.cache:bx,bw=bF?bx[bE]:bx[bE]&&bE,bB=bv==="events";if((!bw||!e[bw]||(!bB&&!by&&!e[bw].data))&&bC&&bz===L){return}if(!bw){if(bF){bx[bE]=bw=++b.uuid}else{bw=bE}}if(!e[bw]){e[bw]={};if(!bF){e[bw].toJSON=b.noop}}if(typeof bv==="object"||typeof bv==="function"){if(by){e[bw]=b.extend(e[bw],bv)}else{e[bw].data=b.extend(e[bw].data,bv)}}bG=bA=e[bw];if(!by){if(!bA.data){bA.data={}}bA=bA.data}if(bz!==L){bA[b.camelCase(bv)]=bz}if(bB&&!bA[bv]){return bG.events}if(bC){bD=bA[bv];if(bD==null){bD=bA[b.camelCase(bv)]}}else{bD=bA}return bD},removeData:function(bx,bv,by){if(!b.acceptData(bx)){return}var bB,bA,bz,bC=b.expando,bD=bx.nodeType,e=bD?b.cache:bx,bw=bD?bx[bC]:bC;if(!e[bw]){return}if(bv){bB=by?e[bw]:e[bw].data;if(bB){if(!b.isArray(bv)){if(bv in bB){bv=[bv]}else{bv=b.camelCase(bv);if(bv in bB){bv=[bv]}else{bv=bv.split(" ")}}}for(bA=0,bz=bv.length;bA-1){return true}}return false},val:function(bx){var e,bv,by,bw=this[0];if(!arguments.length){if(bw){e=b.valHooks[bw.nodeName.toLowerCase()]||b.valHooks[bw.type];if(e&&"get" in e&&(bv=e.get(bw,"value"))!==L){return bv}bv=bw.value;return typeof bv==="string"?bv.replace(aU,""):bv==null?"":bv}return}by=b.isFunction(bx);return this.each(function(bA){var bz=b(this),bB;if(this.nodeType!==1){return}if(by){bB=bx.call(this,bA,bz.val())}else{bB=bx}if(bB==null){bB=""}else{if(typeof bB==="number"){bB+=""}else{if(b.isArray(bB)){bB=b.map(bB,function(bC){return bC==null?"":bC+""})}}}e=b.valHooks[this.nodeName.toLowerCase()]||b.valHooks[this.type];if(!e||!("set" in e)||e.set(this,bB,"value")===L){this.value=bB}})}});b.extend({valHooks:{option:{get:function(e){var bv=e.attributes.value;return !bv||bv.specified?e.value:e.text}},select:{get:function(e){var bA,bv,bz,bx,by=e.selectedIndex,bB=[],bC=e.options,bw=e.type==="select-one";if(by<0){return null}bv=bw?by:0;bz=bw?by+1:bC.length;for(;bv=0});if(!e.length){bv.selectedIndex=-1}return e}}},attrFn:{val:true,css:true,html:true,text:true,data:true,width:true,height:true,offset:true},attr:function(bA,bx,bB,bz){var bw,e,by,bv=bA.nodeType; +if(!bA||bv===3||bv===8||bv===2){return}if(bz&&bx in b.attrFn){return b(bA)[bx](bB)}if(typeof bA.getAttribute==="undefined"){return b.prop(bA,bx,bB)}by=bv!==1||!b.isXMLDoc(bA);if(by){bx=bx.toLowerCase();e=b.attrHooks[bx]||(ao.test(bx)?aY:be)}if(bB!==L){if(bB===null){b.removeAttr(bA,bx);return}else{if(e&&"set" in e&&by&&(bw=e.set(bA,bB,bx))!==L){return bw}else{bA.setAttribute(bx,""+bB);return bB}}}else{if(e&&"get" in e&&by&&(bw=e.get(bA,bx))!==null){return bw}else{bw=bA.getAttribute(bx);return bw===null?L:bw}}},removeAttr:function(bx,bz){var by,bA,bv,e,bw=0;if(bz&&bx.nodeType===1){bA=bz.toLowerCase().split(af);e=bA.length;for(;bw=0)}}})});var bd=/^(?:textarea|input|select)$/i,n=/^([^\.]*)?(?:\.(.+))?$/,J=/\bhover(\.\S+)?\b/,aO=/^key/,bf=/^(?:mouse|contextmenu)|click/,T=/^(?:focusinfocus|focusoutblur)$/,U=/^(\w*)(?:#([\w\-]+))?(?:\.([\w\-]+))?$/,Y=function(e){var bv=U.exec(e);if(bv){bv[1]=(bv[1]||"").toLowerCase();bv[3]=bv[3]&&new RegExp("(?:^|\\s)"+bv[3]+"(?:\\s|$)")}return bv},j=function(bw,e){var bv=bw.attributes||{};return((!e[1]||bw.nodeName.toLowerCase()===e[1])&&(!e[2]||(bv.id||{}).value===e[2])&&(!e[3]||e[3].test((bv["class"]||{}).value)))},bt=function(e){return b.event.special.hover?e:e.replace(J,"mouseenter$1 mouseleave$1")};b.event={add:function(bx,bC,bJ,bA,by){var bD,bB,bK,bI,bH,bF,e,bG,bv,bz,bw,bE;if(bx.nodeType===3||bx.nodeType===8||!bC||!bJ||!(bD=b._data(bx))){return}if(bJ.handler){bv=bJ;bJ=bv.handler}if(!bJ.guid){bJ.guid=b.guid++}bK=bD.events;if(!bK){bD.events=bK={}}bB=bD.handle;if(!bB){bD.handle=bB=function(bL){return typeof b!=="undefined"&&(!bL||b.event.triggered!==bL.type)?b.event.dispatch.apply(bB.elem,arguments):L};bB.elem=bx}bC=b.trim(bt(bC)).split(" ");for(bI=0;bI=0){bG=bG.slice(0,-1);bw=true}if(bG.indexOf(".")>=0){bx=bG.split(".");bG=bx.shift();bx.sort()}if((!bA||b.event.customEvent[bG])&&!b.event.global[bG]){return}bv=typeof bv==="object"?bv[b.expando]?bv:new b.Event(bG,bv):new b.Event(bG);bv.type=bG;bv.isTrigger=true;bv.exclusive=bw;bv.namespace=bx.join(".");bv.namespace_re=bv.namespace?new RegExp("(^|\\.)"+bx.join("\\.(?:.*\\.)?")+"(\\.|$)"):null;by=bG.indexOf(":")<0?"on"+bG:"";if(!bA){e=b.cache;for(bC in e){if(e[bC].events&&e[bC].events[bG]){b.event.trigger(bv,bD,e[bC].handle.elem,true)}}return}bv.result=L;if(!bv.target){bv.target=bA}bD=bD!=null?b.makeArray(bD):[];bD.unshift(bv);bF=b.event.special[bG]||{};if(bF.trigger&&bF.trigger.apply(bA,bD)===false){return}bB=[[bA,bF.bindType||bG]];if(!bJ&&!bF.noBubble&&!b.isWindow(bA)){bI=bF.delegateType||bG;bH=T.test(bI+bG)?bA:bA.parentNode;bz=null;for(;bH;bH=bH.parentNode){bB.push([bH,bI]);bz=bH}if(bz&&bz===bA.ownerDocument){bB.push([bz.defaultView||bz.parentWindow||bb,bI])}}for(bC=0;bCbA){bH.push({elem:this,matches:bz.slice(bA)})}for(bC=0;bC0?this.on(e,null,bx,bw):this.trigger(e)};if(b.attrFn){b.attrFn[e]=true}if(aO.test(e)){b.event.fixHooks[e]=b.event.keyHooks}if(bf.test(e)){b.event.fixHooks[e]=b.event.mouseHooks}}); +/*! + * Sizzle CSS Selector Engine + * Copyright 2011, The Dojo Foundation + * Released under the MIT, BSD, and GPL Licenses. + * More information: http://sizzlejs.com/ + */ +(function(){var bH=/((?:\((?:\([^()]+\)|[^()]+)+\)|\[(?:\[[^\[\]]*\]|['"][^'"]*['"]|[^\[\]'"]+)+\]|\\.|[^ >+~,(\[\\]+)+|[>+~])(\s*,\s*)?((?:.|\r|\n)*)/g,bC="sizcache"+(Math.random()+"").replace(".",""),bI=0,bL=Object.prototype.toString,bB=false,bA=true,bK=/\\/g,bO=/\r\n/g,bQ=/\W/;[0,0].sort(function(){bA=false;return 0});var by=function(bV,e,bY,bZ){bY=bY||[];e=e||av;var b1=e;if(e.nodeType!==1&&e.nodeType!==9){return[]}if(!bV||typeof bV!=="string"){return bY}var bS,b3,b6,bR,b2,b5,b4,bX,bU=true,bT=by.isXML(e),bW=[],b0=bV;do{bH.exec("");bS=bH.exec(b0);if(bS){b0=bS[3];bW.push(bS[1]);if(bS[2]){bR=bS[3];break}}}while(bS);if(bW.length>1&&bD.exec(bV)){if(bW.length===2&&bE.relative[bW[0]]){b3=bM(bW[0]+bW[1],e,bZ)}else{b3=bE.relative[bW[0]]?[e]:by(bW.shift(),e);while(bW.length){bV=bW.shift();if(bE.relative[bV]){bV+=bW.shift()}b3=bM(bV,b3,bZ)}}}else{if(!bZ&&bW.length>1&&e.nodeType===9&&!bT&&bE.match.ID.test(bW[0])&&!bE.match.ID.test(bW[bW.length-1])){b2=by.find(bW.shift(),e,bT);e=b2.expr?by.filter(b2.expr,b2.set)[0]:b2.set[0]}if(e){b2=bZ?{expr:bW.pop(),set:bF(bZ)}:by.find(bW.pop(),bW.length===1&&(bW[0]==="~"||bW[0]==="+")&&e.parentNode?e.parentNode:e,bT);b3=b2.expr?by.filter(b2.expr,b2.set):b2.set;if(bW.length>0){b6=bF(b3)}else{bU=false}while(bW.length){b5=bW.pop();b4=b5;if(!bE.relative[b5]){b5=""}else{b4=bW.pop()}if(b4==null){b4=e}bE.relative[b5](b6,b4,bT)}}else{b6=bW=[]}}if(!b6){b6=b3}if(!b6){by.error(b5||bV)}if(bL.call(b6)==="[object Array]"){if(!bU){bY.push.apply(bY,b6)}else{if(e&&e.nodeType===1){for(bX=0;b6[bX]!=null;bX++){if(b6[bX]&&(b6[bX]===true||b6[bX].nodeType===1&&by.contains(e,b6[bX]))){bY.push(b3[bX])}}}else{for(bX=0;b6[bX]!=null;bX++){if(b6[bX]&&b6[bX].nodeType===1){bY.push(b3[bX])}}}}}else{bF(b6,bY)}if(bR){by(bR,b1,bY,bZ);by.uniqueSort(bY)}return bY};by.uniqueSort=function(bR){if(bJ){bB=bA;bR.sort(bJ);if(bB){for(var e=1;e0};by.find=function(bX,e,bY){var bW,bS,bU,bT,bV,bR;if(!bX){return[]}for(bS=0,bU=bE.order.length;bS":function(bW,bR){var bV,bU=typeof bR==="string",bS=0,e=bW.length;if(bU&&!bQ.test(bR)){bR=bR.toLowerCase();for(;bS=0)){if(!bS){e.push(bV)}}else{if(bS){bR[bU]=false}}}}return false},ID:function(e){return e[1].replace(bK,"")},TAG:function(bR,e){return bR[1].replace(bK,"").toLowerCase()},CHILD:function(e){if(e[1]==="nth"){if(!e[2]){by.error(e[0])}e[2]=e[2].replace(/^\+|\s*/g,"");var bR=/(-?)(\d*)(?:n([+\-]?\d*))?/.exec(e[2]==="even"&&"2n"||e[2]==="odd"&&"2n+1"||!/\D/.test(e[2])&&"0n+"+e[2]||e[2]);e[2]=(bR[1]+(bR[2]||1))-0;e[3]=bR[3]-0}else{if(e[2]){by.error(e[0])}}e[0]=bI++;return e},ATTR:function(bU,bR,bS,e,bV,bW){var bT=bU[1]=bU[1].replace(bK,"");if(!bW&&bE.attrMap[bT]){bU[1]=bE.attrMap[bT]}bU[4]=(bU[4]||bU[5]||"").replace(bK,"");if(bU[2]==="~="){bU[4]=" "+bU[4]+" "}return bU},PSEUDO:function(bU,bR,bS,e,bV){if(bU[1]==="not"){if((bH.exec(bU[3])||"").length>1||/^\w/.test(bU[3])){bU[3]=by(bU[3],null,null,bR)}else{var bT=by.filter(bU[3],bR,bS,true^bV);if(!bS){e.push.apply(e,bT)}return false}}else{if(bE.match.POS.test(bU[0])||bE.match.CHILD.test(bU[0])){return true}}return bU},POS:function(e){e.unshift(true);return e}},filters:{enabled:function(e){return e.disabled===false&&e.type!=="hidden"},disabled:function(e){return e.disabled===true},checked:function(e){return e.checked===true},selected:function(e){if(e.parentNode){e.parentNode.selectedIndex}return e.selected===true},parent:function(e){return !!e.firstChild},empty:function(e){return !e.firstChild},has:function(bS,bR,e){return !!by(e[3],bS).length},header:function(e){return(/h\d/i).test(e.nodeName)},text:function(bS){var e=bS.getAttribute("type"),bR=bS.type;return bS.nodeName.toLowerCase()==="input"&&"text"===bR&&(e===bR||e===null)},radio:function(e){return e.nodeName.toLowerCase()==="input"&&"radio"===e.type},checkbox:function(e){return e.nodeName.toLowerCase()==="input"&&"checkbox"===e.type},file:function(e){return e.nodeName.toLowerCase()==="input"&&"file"===e.type},password:function(e){return e.nodeName.toLowerCase()==="input"&&"password"===e.type},submit:function(bR){var e=bR.nodeName.toLowerCase();return(e==="input"||e==="button")&&"submit"===bR.type},image:function(e){return e.nodeName.toLowerCase()==="input"&&"image"===e.type},reset:function(bR){var e=bR.nodeName.toLowerCase();return(e==="input"||e==="button")&&"reset"===bR.type},button:function(bR){var e=bR.nodeName.toLowerCase();return e==="input"&&"button"===bR.type||e==="button"},input:function(e){return(/input|select|textarea|button/i).test(e.nodeName)},focus:function(e){return e===e.ownerDocument.activeElement}},setFilters:{first:function(bR,e){return e===0},last:function(bS,bR,e,bT){return bR===bT.length-1},even:function(bR,e){return e%2===0},odd:function(bR,e){return e%2===1 +},lt:function(bS,bR,e){return bRe[3]-0},nth:function(bS,bR,e){return e[3]-0===bR},eq:function(bS,bR,e){return e[3]-0===bR}},filter:{PSEUDO:function(bS,bX,bW,bY){var e=bX[1],bR=bE.filters[e];if(bR){return bR(bS,bW,bX,bY)}else{if(e==="contains"){return(bS.textContent||bS.innerText||bw([bS])||"").indexOf(bX[3])>=0}else{if(e==="not"){var bT=bX[3];for(var bV=0,bU=bT.length;bV=0)}}},ID:function(bR,e){return bR.nodeType===1&&bR.getAttribute("id")===e},TAG:function(bR,e){return(e==="*"&&bR.nodeType===1)||!!bR.nodeName&&bR.nodeName.toLowerCase()===e},CLASS:function(bR,e){return(" "+(bR.className||bR.getAttribute("class"))+" ").indexOf(e)>-1},ATTR:function(bV,bT){var bS=bT[1],e=by.attr?by.attr(bV,bS):bE.attrHandle[bS]?bE.attrHandle[bS](bV):bV[bS]!=null?bV[bS]:bV.getAttribute(bS),bW=e+"",bU=bT[2],bR=bT[4];return e==null?bU==="!=":!bU&&by.attr?e!=null:bU==="="?bW===bR:bU==="*="?bW.indexOf(bR)>=0:bU==="~="?(" "+bW+" ").indexOf(bR)>=0:!bR?bW&&e!==false:bU==="!="?bW!==bR:bU==="^="?bW.indexOf(bR)===0:bU==="$="?bW.substr(bW.length-bR.length)===bR:bU==="|="?bW===bR||bW.substr(0,bR.length+1)===bR+"-":false},POS:function(bU,bR,bS,bV){var e=bR[2],bT=bE.setFilters[e];if(bT){return bT(bU,bS,bR,bV)}}}};var bD=bE.match.POS,bx=function(bR,e){return"\\"+(e-0+1)};for(var bz in bE.match){bE.match[bz]=new RegExp(bE.match[bz].source+(/(?![^\[]*\])(?![^\(]*\))/.source));bE.leftMatch[bz]=new RegExp(/(^(?:.|\r|\n)*?)/.source+bE.match[bz].source.replace(/\\(\d+)/g,bx))}var bF=function(bR,e){bR=Array.prototype.slice.call(bR,0);if(e){e.push.apply(e,bR);return e}return bR};try{Array.prototype.slice.call(av.documentElement.childNodes,0)[0].nodeType}catch(bP){bF=function(bU,bT){var bS=0,bR=bT||[];if(bL.call(bU)==="[object Array]"){Array.prototype.push.apply(bR,bU)}else{if(typeof bU.length==="number"){for(var e=bU.length;bS";e.insertBefore(bR,e.firstChild);if(av.getElementById(bS)){bE.find.ID=function(bU,bV,bW){if(typeof bV.getElementById!=="undefined"&&!bW){var bT=bV.getElementById(bU[1]);return bT?bT.id===bU[1]||typeof bT.getAttributeNode!=="undefined"&&bT.getAttributeNode("id").nodeValue===bU[1]?[bT]:L:[]}};bE.filter.ID=function(bV,bT){var bU=typeof bV.getAttributeNode!=="undefined"&&bV.getAttributeNode("id");return bV.nodeType===1&&bU&&bU.nodeValue===bT}}e.removeChild(bR);e=bR=null})();(function(){var e=av.createElement("div");e.appendChild(av.createComment(""));if(e.getElementsByTagName("*").length>0){bE.find.TAG=function(bR,bV){var bU=bV.getElementsByTagName(bR[1]);if(bR[1]==="*"){var bT=[];for(var bS=0;bU[bS];bS++){if(bU[bS].nodeType===1){bT.push(bU[bS])}}bU=bT}return bU}}e.innerHTML="";if(e.firstChild&&typeof e.firstChild.getAttribute!=="undefined"&&e.firstChild.getAttribute("href")!=="#"){bE.attrHandle.href=function(bR){return bR.getAttribute("href",2)}}e=null})();if(av.querySelectorAll){(function(){var e=by,bT=av.createElement("div"),bS="__sizzle__";bT.innerHTML="

";if(bT.querySelectorAll&&bT.querySelectorAll(".TEST").length===0){return}by=function(b4,bV,bZ,b3){bV=bV||av;if(!b3&&!by.isXML(bV)){var b2=/^(\w+$)|^\.([\w\-]+$)|^#([\w\-]+$)/.exec(b4);if(b2&&(bV.nodeType===1||bV.nodeType===9)){if(b2[1]){return bF(bV.getElementsByTagName(b4),bZ)}else{if(b2[2]&&bE.find.CLASS&&bV.getElementsByClassName){return bF(bV.getElementsByClassName(b2[2]),bZ)}}}if(bV.nodeType===9){if(b4==="body"&&bV.body){return bF([bV.body],bZ)}else{if(b2&&b2[3]){var bY=bV.getElementById(b2[3]);if(bY&&bY.parentNode){if(bY.id===b2[3]){return bF([bY],bZ)}}else{return bF([],bZ)}}}try{return bF(bV.querySelectorAll(b4),bZ)}catch(b0){}}else{if(bV.nodeType===1&&bV.nodeName.toLowerCase()!=="object"){var bW=bV,bX=bV.getAttribute("id"),bU=bX||bS,b6=bV.parentNode,b5=/^\s*[+~]/.test(b4);if(!bX){bV.setAttribute("id",bU)}else{bU=bU.replace(/'/g,"\\$&")}if(b5&&b6){bV=bV.parentNode}try{if(!b5||b6){return bF(bV.querySelectorAll("[id='"+bU+"'] "+b4),bZ)}}catch(b1){}finally{if(!bX){bW.removeAttribute("id")}}}}}return e(b4,bV,bZ,b3)};for(var bR in e){by[bR]=e[bR]}bT=null})()}(function(){var e=av.documentElement,bS=e.matchesSelector||e.mozMatchesSelector||e.webkitMatchesSelector||e.msMatchesSelector;if(bS){var bU=!bS.call(av.createElement("div"),"div"),bR=false;try{bS.call(av.documentElement,"[test!='']:sizzle")}catch(bT){bR=true}by.matchesSelector=function(bW,bY){bY=bY.replace(/\=\s*([^'"\]]*)\s*\]/g,"='$1']");if(!by.isXML(bW)){try{if(bR||!bE.match.PSEUDO.test(bY)&&!/!=/.test(bY)){var bV=bS.call(bW,bY);if(bV||!bU||bW.document&&bW.document.nodeType!==11){return bV}}}catch(bX){}}return by(bY,null,null,[bW]).length>0}}})();(function(){var e=av.createElement("div");e.innerHTML="
";if(!e.getElementsByClassName||e.getElementsByClassName("e").length===0){return}e.lastChild.className="e";if(e.getElementsByClassName("e").length===1){return}bE.order.splice(1,0,"CLASS");bE.find.CLASS=function(bR,bS,bT){if(typeof bS.getElementsByClassName!=="undefined"&&!bT){return bS.getElementsByClassName(bR[1])}};e=null})();function bv(bR,bW,bV,bZ,bX,bY){for(var bT=0,bS=bZ.length;bT0){bU=e;break}}}e=e[bR]}bZ[bT]=bU}}}if(av.documentElement.contains){by.contains=function(bR,e){return bR!==e&&(bR.contains?bR.contains(e):true)}}else{if(av.documentElement.compareDocumentPosition){by.contains=function(bR,e){return !!(bR.compareDocumentPosition(e)&16)}}else{by.contains=function(){return false}}}by.isXML=function(e){var bR=(e?e.ownerDocument||e:0).documentElement;return bR?bR.nodeName!=="HTML":false};var bM=function(bS,e,bW){var bV,bX=[],bU="",bY=e.nodeType?[e]:e;while((bV=bE.match.PSEUDO.exec(bS))){bU+=bV[0];bS=bS.replace(bE.match.PSEUDO,"")}bS=bE.relative[bS]?bS+"*":bS;for(var bT=0,bR=bY.length;bT0){for(bB=bA;bB=0:b.filter(e,this).length>0:this.filter(e).length>0)},closest:function(by,bx){var bv=[],bw,e,bz=this[0];if(b.isArray(by)){var bB=1;while(bz&&bz.ownerDocument&&bz!==bx){for(bw=0;bw-1:b.find.matchesSelector(bz,by)){bv.push(bz);break}else{bz=bz.parentNode;if(!bz||!bz.ownerDocument||bz===bx||bz.nodeType===11){break}}}}bv=bv.length>1?b.unique(bv):bv;return this.pushStack(bv,"closest",by)},index:function(e){if(!e){return(this[0]&&this[0].parentNode)?this.prevAll().length:-1}if(typeof e==="string"){return b.inArray(this[0],b(e))}return b.inArray(e.jquery?e[0]:e,this)},add:function(e,bv){var bx=typeof e==="string"?b(e,bv):b.makeArray(e&&e.nodeType?[e]:e),bw=b.merge(this.get(),bx);return this.pushStack(C(bx[0])||C(bw[0])?bw:b.unique(bw))},andSelf:function(){return this.add(this.prevObject)}});function C(e){return !e||!e.parentNode||e.parentNode.nodeType===11}b.each({parent:function(bv){var e=bv.parentNode;return e&&e.nodeType!==11?e:null},parents:function(e){return b.dir(e,"parentNode")},parentsUntil:function(bv,e,bw){return b.dir(bv,"parentNode",bw)},next:function(e){return b.nth(e,2,"nextSibling")},prev:function(e){return b.nth(e,2,"previousSibling")},nextAll:function(e){return b.dir(e,"nextSibling")},prevAll:function(e){return b.dir(e,"previousSibling")},nextUntil:function(bv,e,bw){return b.dir(bv,"nextSibling",bw)},prevUntil:function(bv,e,bw){return b.dir(bv,"previousSibling",bw)},siblings:function(e){return b.sibling(e.parentNode.firstChild,e)},children:function(e){return b.sibling(e.firstChild)},contents:function(e){return b.nodeName(e,"iframe")?e.contentDocument||e.contentWindow.document:b.makeArray(e.childNodes)}},function(e,bv){b.fn[e]=function(by,bw){var bx=b.map(this,bv,by);if(!ab.test(e)){bw=by}if(bw&&typeof bw==="string"){bx=b.filter(bw,bx)}bx=this.length>1&&!ay[e]?b.unique(bx):bx;if((this.length>1||a9.test(bw))&&aq.test(e)){bx=bx.reverse()}return this.pushStack(bx,e,P.call(arguments).join(","))}});b.extend({filter:function(bw,e,bv){if(bv){bw=":not("+bw+")"}return e.length===1?b.find.matchesSelector(e[0],bw)?[e[0]]:[]:b.find.matches(bw,e)},dir:function(bw,bv,by){var e=[],bx=bw[bv];while(bx&&bx.nodeType!==9&&(by===L||bx.nodeType!==1||!b(bx).is(by))){if(bx.nodeType===1){e.push(bx)}bx=bx[bv]}return e},nth:function(by,e,bw,bx){e=e||1;var bv=0;for(;by;by=by[bw]){if(by.nodeType===1&&++bv===e){break}}return by},sibling:function(bw,bv){var e=[];for(;bw;bw=bw.nextSibling){if(bw.nodeType===1&&bw!==bv){e.push(bw)}}return e}});function aG(bx,bw,e){bw=bw||0;if(b.isFunction(bw)){return b.grep(bx,function(bz,by){var bA=!!bw.call(bz,by,bz);return bA===e})}else{if(bw.nodeType){return b.grep(bx,function(bz,by){return(bz===bw)===e})}else{if(typeof bw==="string"){var bv=b.grep(bx,function(by){return by.nodeType===1});if(bp.test(bw)){return b.filter(bw,bv,!e)}else{bw=b.filter(bw,bv)}}}}return b.grep(bx,function(bz,by){return(b.inArray(bz,bw)>=0)===e})}function a(e){var bw=aR.split("|"),bv=e.createDocumentFragment();if(bv.createElement){while(bw.length){bv.createElement(bw.pop())}}return bv}var aR="abbr|article|aside|audio|canvas|datalist|details|figcaption|figure|footer|header|hgroup|mark|meter|nav|output|progress|section|summary|time|video",ag=/ jQuery\d+="(?:\d+|null)"/g,ar=/^\s+/,R=/<(?!area|br|col|embed|hr|img|input|link|meta|param)(([\w:]+)[^>]*)\/>/ig,d=/<([\w:]+)/,w=/",""],legend:[1,"
","
"],thead:[1,"","
"],tr:[2,"","
"],td:[3,"","
"],col:[2,"","
"],area:[1,"",""],_default:[0,"",""]},ac=a(av); +ax.optgroup=ax.option;ax.tbody=ax.tfoot=ax.colgroup=ax.caption=ax.thead;ax.th=ax.td;if(!b.support.htmlSerialize){ax._default=[1,"div
","
"]}b.fn.extend({text:function(e){if(b.isFunction(e)){return this.each(function(bw){var bv=b(this);bv.text(e.call(this,bw,bv.text()))})}if(typeof e!=="object"&&e!==L){return this.empty().append((this[0]&&this[0].ownerDocument||av).createTextNode(e))}return b.text(this)},wrapAll:function(e){if(b.isFunction(e)){return this.each(function(bw){b(this).wrapAll(e.call(this,bw))})}if(this[0]){var bv=b(e,this[0].ownerDocument).eq(0).clone(true);if(this[0].parentNode){bv.insertBefore(this[0])}bv.map(function(){var bw=this;while(bw.firstChild&&bw.firstChild.nodeType===1){bw=bw.firstChild}return bw}).append(this)}return this},wrapInner:function(e){if(b.isFunction(e)){return this.each(function(bv){b(this).wrapInner(e.call(this,bv))})}return this.each(function(){var bv=b(this),bw=bv.contents();if(bw.length){bw.wrapAll(e)}else{bv.append(e)}})},wrap:function(e){var bv=b.isFunction(e);return this.each(function(bw){b(this).wrapAll(bv?e.call(this,bw):e)})},unwrap:function(){return this.parent().each(function(){if(!b.nodeName(this,"body")){b(this).replaceWith(this.childNodes)}}).end()},append:function(){return this.domManip(arguments,true,function(e){if(this.nodeType===1){this.appendChild(e)}})},prepend:function(){return this.domManip(arguments,true,function(e){if(this.nodeType===1){this.insertBefore(e,this.firstChild)}})},before:function(){if(this[0]&&this[0].parentNode){return this.domManip(arguments,false,function(bv){this.parentNode.insertBefore(bv,this)})}else{if(arguments.length){var e=b.clean(arguments);e.push.apply(e,this.toArray());return this.pushStack(e,"before",arguments)}}},after:function(){if(this[0]&&this[0].parentNode){return this.domManip(arguments,false,function(bv){this.parentNode.insertBefore(bv,this.nextSibling)})}else{if(arguments.length){var e=this.pushStack(this,"after",arguments);e.push.apply(e,b.clean(arguments));return e}}},remove:function(e,bx){for(var bv=0,bw;(bw=this[bv])!=null;bv++){if(!e||b.filter(e,[bw]).length){if(!bx&&bw.nodeType===1){b.cleanData(bw.getElementsByTagName("*"));b.cleanData([bw])}if(bw.parentNode){bw.parentNode.removeChild(bw)}}}return this},empty:function(){for(var e=0,bv;(bv=this[e])!=null;e++){if(bv.nodeType===1){b.cleanData(bv.getElementsByTagName("*"))}while(bv.firstChild){bv.removeChild(bv.firstChild)}}return this},clone:function(bv,e){bv=bv==null?false:bv;e=e==null?bv:e;return this.map(function(){return b.clone(this,bv,e)})},html:function(bx){if(bx===L){return this[0]&&this[0].nodeType===1?this[0].innerHTML.replace(ag,""):null}else{if(typeof bx==="string"&&!ae.test(bx)&&(b.support.leadingWhitespace||!ar.test(bx))&&!ax[(d.exec(bx)||["",""])[1].toLowerCase()]){bx=bx.replace(R,"<$1>");try{for(var bw=0,bv=this.length;bw1&&bw0?this.clone(true):this).get();b(bC[bA])[bv](by);bz=bz.concat(by)}return this.pushStack(bz,e,bC.selector)}}});function bg(e){if(typeof e.getElementsByTagName!=="undefined"){return e.getElementsByTagName("*")}else{if(typeof e.querySelectorAll!=="undefined"){return e.querySelectorAll("*")}else{return[]}}}function az(e){if(e.type==="checkbox"||e.type==="radio"){e.defaultChecked=e.checked}}function E(e){var bv=(e.nodeName||"").toLowerCase();if(bv==="input"){az(e)}else{if(bv!=="script"&&typeof e.getElementsByTagName!=="undefined"){b.grep(e.getElementsByTagName("input"),az)}}}function al(e){var bv=av.createElement("div");ac.appendChild(bv);bv.innerHTML=e.outerHTML;return bv.firstChild}b.extend({clone:function(by,bA,bw){var e,bv,bx,bz=b.support.html5Clone||!ah.test("<"+by.nodeName)?by.cloneNode(true):al(by);if((!b.support.noCloneEvent||!b.support.noCloneChecked)&&(by.nodeType===1||by.nodeType===11)&&!b.isXMLDoc(by)){ai(by,bz);e=bg(by);bv=bg(bz);for(bx=0;e[bx];++bx){if(bv[bx]){ai(e[bx],bv[bx])}}}if(bA){t(by,bz);if(bw){e=bg(by);bv=bg(bz);for(bx=0;e[bx];++bx){t(e[bx],bv[bx])}}}e=bv=null;return bz},clean:function(bw,by,bH,bA){var bF;by=by||av;if(typeof by.createElement==="undefined"){by=by.ownerDocument||by[0]&&by[0].ownerDocument||av}var bI=[],bB;for(var bE=0,bz;(bz=bw[bE])!=null;bE++){if(typeof bz==="number"){bz+=""}if(!bz){continue}if(typeof bz==="string"){if(!W.test(bz)){bz=by.createTextNode(bz)}else{bz=bz.replace(R,"<$1>");var bK=(d.exec(bz)||["",""])[1].toLowerCase(),bx=ax[bK]||ax._default,bD=bx[0],bv=by.createElement("div");if(by===av){ac.appendChild(bv)}else{a(by).appendChild(bv)}bv.innerHTML=bx[1]+bz+bx[2];while(bD--){bv=bv.lastChild}if(!b.support.tbody){var e=w.test(bz),bC=bK==="table"&&!e?bv.firstChild&&bv.firstChild.childNodes:bx[1]===""&&!e?bv.childNodes:[];for(bB=bC.length-1;bB>=0;--bB){if(b.nodeName(bC[bB],"tbody")&&!bC[bB].childNodes.length){bC[bB].parentNode.removeChild(bC[bB])}}}if(!b.support.leadingWhitespace&&ar.test(bz)){bv.insertBefore(by.createTextNode(ar.exec(bz)[0]),bv.firstChild)}bz=bv.childNodes}}var bG;if(!b.support.appendChecked){if(bz[0]&&typeof(bG=bz.length)==="number"){for(bB=0;bB=0){return bx+"px"}}else{return bx}}}});if(!b.support.opacity){b.cssHooks.opacity={get:function(bv,e){return au.test((e&&bv.currentStyle?bv.currentStyle.filter:bv.style.filter)||"")?(parseFloat(RegExp.$1)/100)+"":e?"1":""},set:function(by,bz){var bx=by.style,bv=by.currentStyle,e=b.isNumeric(bz)?"alpha(opacity="+bz*100+")":"",bw=bv&&bv.filter||bx.filter||"";bx.zoom=1;if(bz>=1&&b.trim(bw.replace(ak,""))===""){bx.removeAttribute("filter");if(bv&&!bv.filter){return}}bx.filter=ak.test(bw)?bw.replace(ak,e):bw+" "+e}}}b(function(){if(!b.support.reliableMarginRight){b.cssHooks.marginRight={get:function(bw,bv){var e;b.swap(bw,{display:"inline-block"},function(){if(bv){e=Z(bw,"margin-right","marginRight")}else{e=bw.style.marginRight}});return e}}}});if(av.defaultView&&av.defaultView.getComputedStyle){aI=function(by,bw){var bv,bx,e;bw=bw.replace(z,"-$1").toLowerCase();if((bx=by.ownerDocument.defaultView)&&(e=bx.getComputedStyle(by,null))){bv=e.getPropertyValue(bw);if(bv===""&&!b.contains(by.ownerDocument.documentElement,by)){bv=b.style(by,bw)}}return bv}}if(av.documentElement.currentStyle){aX=function(bz,bw){var bA,e,by,bv=bz.currentStyle&&bz.currentStyle[bw],bx=bz.style;if(bv===null&&bx&&(by=bx[bw])){bv=by}if(!bc.test(bv)&&bn.test(bv)){bA=bx.left;e=bz.runtimeStyle&&bz.runtimeStyle.left;if(e){bz.runtimeStyle.left=bz.currentStyle.left}bx.left=bw==="fontSize"?"1em":(bv||0);bv=bx.pixelLeft+"px";bx.left=bA;if(e){bz.runtimeStyle.left=e}}return bv===""?"auto":bv}}Z=aI||aX;function p(by,bw,bv){var bA=bw==="width"?by.offsetWidth:by.offsetHeight,bz=bw==="width"?an:a1,bx=0,e=bz.length; +if(bA>0){if(bv!=="border"){for(;bx)<[^<]*)*<\/script>/gi,q=/^(?:select|textarea)/i,h=/\s+/,br=/([?&])_=[^&]*/,K=/^([\w\+\.\-]+:)(?:\/\/([^\/?#:]*)(?::(\d+))?)?/,A=b.fn.load,aa={},r={},aE,s,aV=["*/"]+["*"];try{aE=bl.href}catch(aw){aE=av.createElement("a");aE.href="";aE=aE.href}s=K.exec(aE.toLowerCase())||[];function f(e){return function(by,bA){if(typeof by!=="string"){bA=by;by="*"}if(b.isFunction(bA)){var bx=by.toLowerCase().split(h),bw=0,bz=bx.length,bv,bB,bC;for(;bw=0){var e=bw.slice(by,bw.length);bw=bw.slice(0,by)}var bx="GET";if(bz){if(b.isFunction(bz)){bA=bz;bz=L}else{if(typeof bz==="object"){bz=b.param(bz,b.ajaxSettings.traditional);bx="POST"}}}var bv=this;b.ajax({url:bw,type:bx,dataType:"html",data:bz,complete:function(bC,bB,bD){bD=bC.responseText;if(bC.isResolved()){bC.done(function(bE){bD=bE});bv.html(e?b("
").append(bD.replace(a6,"")).find(e):bD)}if(bA){bv.each(bA,[bD,bB,bC])}}});return this},serialize:function(){return b.param(this.serializeArray())},serializeArray:function(){return this.map(function(){return this.elements?b.makeArray(this.elements):this}).filter(function(){return this.name&&!this.disabled&&(this.checked||q.test(this.nodeName)||aZ.test(this.type))}).map(function(e,bv){var bw=b(this).val();return bw==null?null:b.isArray(bw)?b.map(bw,function(by,bx){return{name:bv.name,value:by.replace(bs,"\r\n")}}):{name:bv.name,value:bw.replace(bs,"\r\n")}}).get()}});b.each("ajaxStart ajaxStop ajaxComplete ajaxError ajaxSuccess ajaxSend".split(" "),function(e,bv){b.fn[bv]=function(bw){return this.on(bv,bw)}});b.each(["get","post"],function(e,bv){b[bv]=function(bw,by,bz,bx){if(b.isFunction(by)){bx=bx||bz;bz=by;by=L}return b.ajax({type:bv,url:bw,data:by,success:bz,dataType:bx})}});b.extend({getScript:function(e,bv){return b.get(e,L,bv,"script")},getJSON:function(e,bv,bw){return b.get(e,bv,bw,"json")},ajaxSetup:function(bv,e){if(e){am(bv,b.ajaxSettings)}else{e=bv;bv=b.ajaxSettings}am(bv,e);return bv},ajaxSettings:{url:aE,isLocal:aM.test(s[1]),global:true,type:"GET",contentType:"application/x-www-form-urlencoded",processData:true,async:true,accepts:{xml:"application/xml, text/xml",html:"text/html",text:"text/plain",json:"application/json, text/javascript","*":aV},contents:{xml:/xml/,html:/html/,json:/json/},responseFields:{xml:"responseXML",text:"responseText"},converters:{"* text":bb.String,"text html":true,"text json":b.parseJSON,"text xml":b.parseXML},flatOptions:{context:true,url:true}},ajaxPrefilter:f(aa),ajaxTransport:f(r),ajax:function(bz,bx){if(typeof bz==="object"){bx=bz;bz=L}bx=bx||{};var bD=b.ajaxSetup({},bx),bS=bD.context||bD,bG=bS!==bD&&(bS.nodeType||bS instanceof b)?b(bS):b.event,bR=b.Deferred(),bN=b.Callbacks("once memory"),bB=bD.statusCode||{},bC,bH={},bO={},bQ,by,bL,bE,bI,bA=0,bw,bK,bJ={readyState:0,setRequestHeader:function(bT,bU){if(!bA){var e=bT.toLowerCase();bT=bO[e]=bO[e]||bT;bH[bT]=bU}return this},getAllResponseHeaders:function(){return bA===2?bQ:null},getResponseHeader:function(bT){var e;if(bA===2){if(!by){by={};while((e=aD.exec(bQ))){by[e[1].toLowerCase()]=e[2]}}e=by[bT.toLowerCase()]}return e===L?null:e},overrideMimeType:function(e){if(!bA){bD.mimeType=e}return this},abort:function(e){e=e||"abort";if(bL){bL.abort(e)}bF(0,e);return this}};function bF(bZ,bU,b0,bW){if(bA===2){return}bA=2;if(bE){clearTimeout(bE)}bL=L;bQ=bW||"";bJ.readyState=bZ>0?4:0;var bT,b4,b3,bX=bU,bY=b0?bj(bD,bJ,b0):L,bV,b2;if(bZ>=200&&bZ<300||bZ===304){if(bD.ifModified){if((bV=bJ.getResponseHeader("Last-Modified"))){b.lastModified[bC]=bV}if((b2=bJ.getResponseHeader("Etag"))){b.etag[bC]=b2}}if(bZ===304){bX="notmodified";bT=true}else{try{b4=G(bD,bY);bX="success";bT=true}catch(b1){bX="parsererror";b3=b1}}}else{b3=bX;if(!bX||bZ){bX="error";if(bZ<0){bZ=0}}}bJ.status=bZ;bJ.statusText=""+(bU||bX);if(bT){bR.resolveWith(bS,[b4,bX,bJ])}else{bR.rejectWith(bS,[bJ,bX,b3])}bJ.statusCode(bB);bB=L;if(bw){bG.trigger("ajax"+(bT?"Success":"Error"),[bJ,bD,bT?b4:b3])}bN.fireWith(bS,[bJ,bX]);if(bw){bG.trigger("ajaxComplete",[bJ,bD]);if(!(--b.active)){b.event.trigger("ajaxStop")}}}bR.promise(bJ);bJ.success=bJ.done;bJ.error=bJ.fail;bJ.complete=bN.add;bJ.statusCode=function(bT){if(bT){var e;if(bA<2){for(e in bT){bB[e]=[bB[e],bT[e]]}}else{e=bT[bJ.status];bJ.then(e,e)}}return this};bD.url=((bz||bD.url)+"").replace(bq,"").replace(c,s[1]+"//");bD.dataTypes=b.trim(bD.dataType||"*").toLowerCase().split(h);if(bD.crossDomain==null){bI=K.exec(bD.url.toLowerCase());bD.crossDomain=!!(bI&&(bI[1]!=s[1]||bI[2]!=s[2]||(bI[3]||(bI[1]==="http:"?80:443))!=(s[3]||(s[1]==="http:"?80:443))))}if(bD.data&&bD.processData&&typeof bD.data!=="string"){bD.data=b.param(bD.data,bD.traditional)}aW(aa,bD,bx,bJ);if(bA===2){return false}bw=bD.global;bD.type=bD.type.toUpperCase();bD.hasContent=!aQ.test(bD.type);if(bw&&b.active++===0){b.event.trigger("ajaxStart")}if(!bD.hasContent){if(bD.data){bD.url+=(M.test(bD.url)?"&":"?")+bD.data;delete bD.data}bC=bD.url;if(bD.cache===false){var bv=b.now(),bP=bD.url.replace(br,"$1_="+bv);bD.url=bP+((bP===bD.url)?(M.test(bD.url)?"&":"?")+"_="+bv:"")}}if(bD.data&&bD.hasContent&&bD.contentType!==false||bx.contentType){bJ.setRequestHeader("Content-Type",bD.contentType)}if(bD.ifModified){bC=bC||bD.url;if(b.lastModified[bC]){bJ.setRequestHeader("If-Modified-Since",b.lastModified[bC])}if(b.etag[bC]){bJ.setRequestHeader("If-None-Match",b.etag[bC])}}bJ.setRequestHeader("Accept",bD.dataTypes[0]&&bD.accepts[bD.dataTypes[0]]?bD.accepts[bD.dataTypes[0]]+(bD.dataTypes[0]!=="*"?", "+aV+"; q=0.01":""):bD.accepts["*"]);for(bK in bD.headers){bJ.setRequestHeader(bK,bD.headers[bK])}if(bD.beforeSend&&(bD.beforeSend.call(bS,bJ,bD)===false||bA===2)){bJ.abort();return false}for(bK in {success:1,error:1,complete:1}){bJ[bK](bD[bK])}bL=aW(r,bD,bx,bJ);if(!bL){bF(-1,"No Transport")}else{bJ.readyState=1;if(bw){bG.trigger("ajaxSend",[bJ,bD])}if(bD.async&&bD.timeout>0){bE=setTimeout(function(){bJ.abort("timeout")},bD.timeout)}try{bA=1;bL.send(bH,bF)}catch(bM){if(bA<2){bF(-1,bM)}else{throw bM}}}return bJ},param:function(e,bw){var bv=[],by=function(bz,bA){bA=b.isFunction(bA)?bA():bA;bv[bv.length]=encodeURIComponent(bz)+"="+encodeURIComponent(bA)};if(bw===L){bw=b.ajaxSettings.traditional}if(b.isArray(e)||(e.jquery&&!b.isPlainObject(e))){b.each(e,function(){by(this.name,this.value)})}else{for(var bx in e){v(bx,e[bx],bw,by)}}return bv.join("&").replace(k,"+")}});function v(bw,by,bv,bx){if(b.isArray(by)){b.each(by,function(bA,bz){if(bv||ap.test(bw)){bx(bw,bz)}else{v(bw+"["+(typeof bz==="object"||b.isArray(bz)?bA:"")+"]",bz,bv,bx)}})}else{if(!bv&&by!=null&&typeof by==="object"){for(var e in by){v(bw+"["+e+"]",by[e],bv,bx)}}else{bx(bw,by)}}}b.extend({active:0,lastModified:{},etag:{}});function bj(bD,bC,bz){var bv=bD.contents,bB=bD.dataTypes,bw=bD.responseFields,by,bA,bx,e;for(bA in bw){if(bA in bz){bC[bw[bA]]=bz[bA]}}while(bB[0]==="*"){bB.shift();if(by===L){by=bD.mimeType||bC.getResponseHeader("content-type")}}if(by){for(bA in bv){if(bv[bA]&&bv[bA].test(by)){bB.unshift(bA);break}}}if(bB[0] in bz){bx=bB[0]}else{for(bA in bz){if(!bB[0]||bD.converters[bA+" "+bB[0]]){bx=bA;break}if(!e){e=bA}}bx=bx||e}if(bx){if(bx!==bB[0]){bB.unshift(bx)}return bz[bx]}}function G(bH,bz){if(bH.dataFilter){bz=bH.dataFilter(bz,bH.dataType)}var bD=bH.dataTypes,bG={},bA,bE,bw=bD.length,bB,bC=bD[0],bx,by,bF,bv,e;for(bA=1;bA=bw.duration+this.startTime){this.now=this.end;this.pos=this.state=1;this.update();bw.animatedProperties[this.prop]=true;for(bA in bw.animatedProperties){if(bw.animatedProperties[bA]!==true){e=false}}if(e){if(bw.overflow!=null&&!b.support.shrinkWrapBlocks){b.each(["","X","Y"],function(bC,bD){bz.style["overflow"+bD]=bw.overflow[bC]})}if(bw.hide){b(bz).hide()}if(bw.hide||bw.show){for(bA in bw.animatedProperties){b.style(bz,bA,bw.orig[bA]);b.removeData(bz,"fxshow"+bA,true);b.removeData(bz,"toggle"+bA,true)}}bv=bw.complete;if(bv){bw.complete=false;bv.call(bz)}}return false}else{if(bw.duration==Infinity){this.now=bx}else{bB=bx-this.startTime;this.state=bB/bw.duration;this.pos=b.easing[bw.animatedProperties[this.prop]](this.state,bB,0,1,bw.duration);this.now=this.start+((this.end-this.start)*this.pos)}this.update()}return true}};b.extend(b.fx,{tick:function(){var bw,bv=b.timers,e=0;for(;e").appendTo(e),bw=bv.css("display");bv.remove();if(bw==="none"||bw===""){if(!a8){a8=av.createElement("iframe");a8.frameBorder=a8.width=a8.height=0}e.appendChild(a8);if(!m||!a8.createElement){m=(a8.contentWindow||a8.contentDocument).document;m.write((av.compatMode==="CSS1Compat"?"":"")+"");m.close()}bv=m.createElement(bx);m.body.appendChild(bv);bw=b.css(bv,"display");e.removeChild(a8)}Q[bx]=bw}return Q[bx]}var V=/^t(?:able|d|h)$/i,ad=/^(?:body|html)$/i;if("getBoundingClientRect" in av.documentElement){b.fn.offset=function(bI){var by=this[0],bB;if(bI){return this.each(function(e){b.offset.setOffset(this,bI,e)})}if(!by||!by.ownerDocument){return null}if(by===by.ownerDocument.body){return b.offset.bodyOffset(by)}try{bB=by.getBoundingClientRect()}catch(bF){}var bH=by.ownerDocument,bw=bH.documentElement;if(!bB||!b.contains(bw,by)){return bB?{top:bB.top,left:bB.left}:{top:0,left:0}}var bC=bH.body,bD=aK(bH),bA=bw.clientTop||bC.clientTop||0,bE=bw.clientLeft||bC.clientLeft||0,bv=bD.pageYOffset||b.support.boxModel&&bw.scrollTop||bC.scrollTop,bz=bD.pageXOffset||b.support.boxModel&&bw.scrollLeft||bC.scrollLeft,bG=bB.top+bv-bA,bx=bB.left+bz-bE;return{top:bG,left:bx}}}else{b.fn.offset=function(bF){var bz=this[0];if(bF){return this.each(function(bG){b.offset.setOffset(this,bF,bG)})}if(!bz||!bz.ownerDocument){return null}if(bz===bz.ownerDocument.body){return b.offset.bodyOffset(bz)}var bC,bw=bz.offsetParent,bv=bz,bE=bz.ownerDocument,bx=bE.documentElement,bA=bE.body,bB=bE.defaultView,e=bB?bB.getComputedStyle(bz,null):bz.currentStyle,bD=bz.offsetTop,by=bz.offsetLeft;while((bz=bz.parentNode)&&bz!==bA&&bz!==bx){if(b.support.fixedPosition&&e.position==="fixed"){break}bC=bB?bB.getComputedStyle(bz,null):bz.currentStyle;bD-=bz.scrollTop;by-=bz.scrollLeft;if(bz===bw){bD+=bz.offsetTop;by+=bz.offsetLeft;if(b.support.doesNotAddBorder&&!(b.support.doesAddBorderForTableAndCells&&V.test(bz.nodeName))){bD+=parseFloat(bC.borderTopWidth)||0;by+=parseFloat(bC.borderLeftWidth)||0}bv=bw;bw=bz.offsetParent}if(b.support.subtractsBorderForOverflowNotVisible&&bC.overflow!=="visible"){bD+=parseFloat(bC.borderTopWidth)||0;by+=parseFloat(bC.borderLeftWidth)||0}e=bC}if(e.position==="relative"||e.position==="static"){bD+=bA.offsetTop;by+=bA.offsetLeft}if(b.support.fixedPosition&&e.position==="fixed"){bD+=Math.max(bx.scrollTop,bA.scrollTop);by+=Math.max(bx.scrollLeft,bA.scrollLeft)}return{top:bD,left:by}}}b.offset={bodyOffset:function(e){var bw=e.offsetTop,bv=e.offsetLeft;if(b.support.doesNotIncludeMarginInBodyOffset){bw+=parseFloat(b.css(e,"marginTop"))||0;bv+=parseFloat(b.css(e,"marginLeft"))||0}return{top:bw,left:bv}},setOffset:function(bx,bG,bA){var bB=b.css(bx,"position");if(bB==="static"){bx.style.position="relative"}var bz=b(bx),bv=bz.offset(),e=b.css(bx,"top"),bE=b.css(bx,"left"),bF=(bB==="absolute"||bB==="fixed")&&b.inArray("auto",[e,bE])>-1,bD={},bC={},bw,by;if(bF){bC=bz.position();bw=bC.top;by=bC.left}else{bw=parseFloat(e)||0;by=parseFloat(bE)||0}if(b.isFunction(bG)){bG=bG.call(bx,bA,bv)}if(bG.top!=null){bD.top=(bG.top-bv.top)+bw}if(bG.left!=null){bD.left=(bG.left-bv.left)+by}if("using" in bG){bG.using.call(bx,bD)}else{bz.css(bD)}}};b.fn.extend({position:function(){if(!this[0]){return null}var bw=this[0],bv=this.offsetParent(),bx=this.offset(),e=ad.test(bv[0].nodeName)?{top:0,left:0}:bv.offset();bx.top-=parseFloat(b.css(bw,"marginTop"))||0;bx.left-=parseFloat(b.css(bw,"marginLeft"))||0;e.top+=parseFloat(b.css(bv[0],"borderTopWidth"))||0;e.left+=parseFloat(b.css(bv[0],"borderLeftWidth"))||0;return{top:bx.top-e.top,left:bx.left-e.left}},offsetParent:function(){return this.map(function(){var e=this.offsetParent||av.body;while(e&&(!ad.test(e.nodeName)&&b.css(e,"position")==="static")){e=e.offsetParent}return e})}});b.each(["Left","Top"],function(bv,e){var bw="scroll"+e;b.fn[bw]=function(bz){var bx,by;if(bz===L){bx=this[0];if(!bx){return null}by=aK(bx);return by?("pageXOffset" in by)?by[bv?"pageYOffset":"pageXOffset"]:b.support.boxModel&&by.document.documentElement[bw]||by.document.body[bw]:bx[bw]}return this.each(function(){by=aK(this);if(by){by.scrollTo(!bv?bz:b(by).scrollLeft(),bv?bz:b(by).scrollTop())}else{this[bw]=bz}})}});function aK(e){return b.isWindow(e)?e:e.nodeType===9?e.defaultView||e.parentWindow:false}b.each(["Height","Width"],function(bv,e){var bw=e.toLowerCase();b.fn["inner"+e]=function(){var bx=this[0];return bx?bx.style?parseFloat(b.css(bx,bw,"padding")):this[bw]():null};b.fn["outer"+e]=function(by){var bx=this[0];return bx?bx.style?parseFloat(b.css(bx,bw,by?"margin":"border")):this[bw]():null};b.fn[bw]=function(bz){var bA=this[0];if(!bA){return bz==null?null:this}if(b.isFunction(bz)){return this.each(function(bE){var bD=b(this);bD[bw](bz.call(this,bE,bD[bw]()))})}if(b.isWindow(bA)){var bB=bA.document.documentElement["client"+e],bx=bA.document.body;return bA.document.compatMode==="CSS1Compat"&&bB||bx&&bx["client"+e]||bB}else{if(bA.nodeType===9){return Math.max(bA.documentElement["client"+e],bA.body["scroll"+e],bA.documentElement["scroll"+e],bA.body["offset"+e],bA.documentElement["offset"+e])}else{if(bz===L){var bC=b.css(bA,bw),by=parseFloat(bC);return b.isNumeric(by)?by:bC}else{return this.css(bw,typeof bz==="string"?bz:bz+"px")}}}}});bb.jQuery=bb.$=b;if(typeof define==="function"&&define.amd&&define.amd.jQuery){define("jquery",[],function(){return b +})}})(window); diff --git a/docs/RuntimeAPI/html/modules.html b/docs/RuntimeAPI/html/modules.html new file mode 100644 index 0000000000..ba3bae37e0 --- /dev/null +++ b/docs/RuntimeAPI/html/modules.html @@ -0,0 +1,109 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Modules + + + + + + + + + +
+
+
+ + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+ + + + + + + + + + +
+ +
+ +
+
+
Modules
+
+
+
Here is a list of all modules:
+
[detail level 123]
+ + + + + + + + + + + + + + + +
oGlobal enum and defines
oFence Functions
oHIP Environment Variables
\HIP API
 oDevice Management
 oError Handling
 oStream Management
 oEvent Management
 oMemory Management
 oDevice Memory Access
 oInitialization and Version
 |\Management
 oControl
 oHCC-Specific Accessors
 \Texture Reference Management
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/nav_f.png b/docs/RuntimeAPI/html/nav_f.png new file mode 100644 index 0000000000000000000000000000000000000000..72a58a529ed3a9ed6aa0c51a79cf207e026deee2 GIT binary patch literal 153 zcmeAS@N?(olHy`uVBq!ia0vp^j6iI`!2~2XGqLUlQVE_ejv*C{Z|{2ZH7M}7UYxc) zn!W8uqtnIQ>_z8U literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/nav_g.png b/docs/RuntimeAPI/html/nav_g.png new file mode 100644 index 0000000000000000000000000000000000000000..2093a237a94f6c83e19ec6e5fd42f7ddabdafa81 GIT binary patch literal 95 zcmeAS@N?(olHy`uVBq!ia0vp^j6lrB!3HFm1ilyoDK$?Q$B+ufw|5PB85lU25BhtE tr?otc=hd~V+ws&_A@j8Fiv!KF$B+ufw|5=67#uj90@pIL wZ=Q8~_Ju`#59=RjDrmm`tMD@M=!-l18IR?&vFVdQ&MBb@0HFXL1|%O$WD@{VPM$7~Ar*{o?;hlAFyLXmaDC0y znK1_#cQqJWPES%4Uujug^TE?jMft$}Eq^WaR~)%f)vSNs&gek&x%A9X9sM + + + + + +HIP: Heterogenous-computing Interface for Portability: Related Pages + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + +
+ + + + +
+ +
+ +
+
+
Related Pages
+
+
+
Here is a list of all related documentation pages:
+ + + + +
oSynchonization
oDeprecated List
\Bug List
+
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/search/all_0.html b/docs/RuntimeAPI/html/search/all_0.html new file mode 100644 index 0000000000..17b6da85df --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_0.js b/docs/RuntimeAPI/html/search/all_0.js new file mode 100644 index 0000000000..2901eb97b7 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_0.js @@ -0,0 +1,9 @@ +var searchData= +[ + ['_5f_5fhalf',['__half',['../struct____half.html',1,'']]], + ['_5f_5fhost_5f_5f',['__host__',['../host__defines_8h.html#a803050db3c78e0db3ea59a0c35499622',1,'host_defines.h']]], + ['_5f_5fthreadfence',['__threadfence',['../group__Fence.html#ga8ed093b29d5c05d46a1240e7a35a6d13',1,'hip_runtime.h']]], + ['_5f_5fthreadfence_5fblock',['__threadfence_block',['../group__Fence.html#ga675c8823d497b2fd91abc42085e5ae25',1,'hip_runtime.h']]], + ['_5f_5fthreadfence_5fsystem',['__threadfence_system',['../group__Fence.html#ga21e7328a2daaf9e376d10b5c6b3de0ae',1,'__threadfence_system(void): device_util.cpp'],['../group__Fence.html#ga21e7328a2daaf9e376d10b5c6b3de0ae',1,'__threadfence_system(void): device_util.cpp']]], + ['_5fcomputeunits',['_computeUnits',['../classihipDevice__t.html#a655e03136394df32571a52707aa371c5',1,'ihipDevice_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_1.html b/docs/RuntimeAPI/html/search/all_1.html new file mode 100644 index 0000000000..e2906449f2 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_1.js b/docs/RuntimeAPI/html/search/all_1.js new file mode 100644 index 0000000000..2fba32867b --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_1.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['arch',['arch',['../structhipDeviceProp__t.html#afc58158e44bef6ad26f2be401434b049',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_10.html b/docs/RuntimeAPI/html/search/all_10.html new file mode 100644 index 0000000000..c55c8367e4 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_10.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_10.js b/docs/RuntimeAPI/html/search/all_10.js new file mode 100644 index 0000000000..bb8f6295d1 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_10.js @@ -0,0 +1,7 @@ +var searchData= +[ + ['texture_20reference_20management',['Texture Reference Management',['../group__Texture.html',1,'']]], + ['texturereference',['textureReference',['../structtextureReference.html',1,'']]], + ['totalconstmem',['totalConstMem',['../structhipDeviceProp__t.html#a29880232c56120be3455ce00d5379665',1,'hipDeviceProp_t']]], + ['totalglobalmem',['totalGlobalMem',['../structhipDeviceProp__t.html#acedd6a2d23423441e4bf51c4a1b719f9',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_11.html b/docs/RuntimeAPI/html/search/all_11.html new file mode 100644 index 0000000000..6f3943a941 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_11.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_11.js b/docs/RuntimeAPI/html/search/all_11.js new file mode 100644 index 0000000000..e1eb527d76 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_11.js @@ -0,0 +1,9 @@ +var searchData= +[ + ['uchar3',['uchar3',['../structuchar3.html',1,'']]], + ['uchar4',['uchar4',['../structuchar4.html',1,'']]], + ['uint3',['uint3',['../structuint3.html',1,'']]], + ['ulong3',['ulong3',['../structulong3.html',1,'']]], + ['ulonglong3',['ulonglong3',['../structulonglong3.html',1,'']]], + ['ushort3',['ushort3',['../structushort3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_12.html b/docs/RuntimeAPI/html/search/all_12.html new file mode 100644 index 0000000000..3c7c89ef60 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_12.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_12.js b/docs/RuntimeAPI/html/search/all_12.js new file mode 100644 index 0000000000..46a1400a7b --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_12.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['warpsize',['warpSize',['../structhipDeviceProp__t.html#af3357d33c004608bf05bc21a352be81b',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_13.html b/docs/RuntimeAPI/html/search/all_13.html new file mode 100644 index 0000000000..0bd629b86b --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_13.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_13.js b/docs/RuntimeAPI/html/search/all_13.js new file mode 100644 index 0000000000..250c203caf --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_13.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['x',['x',['../structdim3.html#ac866c05f83a28dac20a153fc65b3b16c',1,'dim3']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_14.html b/docs/RuntimeAPI/html/search/all_14.html new file mode 100644 index 0000000000..2ad638b277 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_14.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_14.js b/docs/RuntimeAPI/html/search/all_14.js new file mode 100644 index 0000000000..133dd9dc6e --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_14.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['y',['y',['../structdim3.html#a83e60e072f7e8bdfde6ac05053cbb370',1,'dim3']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_15.html b/docs/RuntimeAPI/html/search/all_15.html new file mode 100644 index 0000000000..d3b5274ba7 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_15.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_15.js b/docs/RuntimeAPI/html/search/all_15.js new file mode 100644 index 0000000000..e8bf38b99c --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_15.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['z',['z',['../structdim3.html#a866e38993ecc4e76fd47311236c16b04',1,'dim3']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_2.html b/docs/RuntimeAPI/html/search/all_2.html new file mode 100644 index 0000000000..95ded1222d --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_2.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_2.js b/docs/RuntimeAPI/html/search/all_2.js new file mode 100644 index 0000000000..8d42650d49 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_2.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['bug_20list',['Bug List',['../bug.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_3.html b/docs/RuntimeAPI/html/search/all_3.html new file mode 100644 index 0000000000..4d312d03dd --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_3.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_3.js b/docs/RuntimeAPI/html/search/all_3.js new file mode 100644 index 0000000000..eed6c54048 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_3.js @@ -0,0 +1,11 @@ +var searchData= +[ + ['canmaphostmemory',['canMapHostMemory',['../structhipDeviceProp__t.html#ac2143f5448607d1a02a9e8783fcf06a1',1,'hipDeviceProp_t']]], + ['char3',['char3',['../structchar3.html',1,'']]], + ['char4',['char4',['../structchar4.html',1,'']]], + ['clockinstructionrate',['clockInstructionRate',['../structhipDeviceProp__t.html#a6fbf3b08a1a08ae700f1a06265f6666b',1,'hipDeviceProp_t']]], + ['clockrate',['clockRate',['../structhipDeviceProp__t.html#a1dd15bee43692b8649dfbdc1adbaaf96',1,'hipDeviceProp_t']]], + ['computemode',['computeMode',['../structhipDeviceProp__t.html#ae7d9216f8583a703359d0b9373823f5d',1,'hipDeviceProp_t']]], + ['concurrentkernels',['concurrentKernels',['../structhipDeviceProp__t.html#ad8461a28caf9c38c58cf358583b5bee3',1,'hipDeviceProp_t']]], + ['control',['Control',['../group__Profiler.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_4.html b/docs/RuntimeAPI/html/search/all_4.html new file mode 100644 index 0000000000..d72a9104cd --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_4.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_4.js b/docs/RuntimeAPI/html/search/all_4.js new file mode 100644 index 0000000000..07d1b6e566 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_4.js @@ -0,0 +1,10 @@ +var searchData= +[ + ['dbname',['DbName',['../structDbName.html',1,'']]], + ['deprecated_20list',['Deprecated List',['../deprecated.html',1,'']]], + ['device_20management',['Device Management',['../group__Device.html',1,'']]], + ['dim3',['dim3',['../structdim3.html',1,'dim3'],['../group__GlobalDefs.html#gacb37281795c3567d0b10a61c056d512b',1,'dim3(): hip_runtime_api.h']]], + ['double1',['double1',['../structdouble1.html',1,'']]], + ['double3',['double3',['../structdouble3.html',1,'']]], + ['device_20memory_20access',['Device Memory Access',['../group__PeerToPeer.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_5.html b/docs/RuntimeAPI/html/search/all_5.html new file mode 100644 index 0000000000..99ef726704 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_5.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_5.js b/docs/RuntimeAPI/html/search/all_5.js new file mode 100644 index 0000000000..c9fd17e83e --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_5.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['error_20handling',['Error Handling',['../group__Error.html',1,'']]], + ['event_20management',['Event Management',['../group__Event.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_6.html b/docs/RuntimeAPI/html/search/all_6.html new file mode 100644 index 0000000000..6133ab3af1 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_6.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_6.js b/docs/RuntimeAPI/html/search/all_6.js new file mode 100644 index 0000000000..feba6529f6 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_6.js @@ -0,0 +1,7 @@ +var searchData= +[ + ['fakemutex',['FakeMutex',['../classFakeMutex.html',1,'']]], + ['fence_20functions',['Fence Functions',['../group__Fence.html',1,'']]], + ['float1',['float1',['../structfloat1.html',1,'']]], + ['float3',['float3',['../structfloat3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_7.html b/docs/RuntimeAPI/html/search/all_7.html new file mode 100644 index 0000000000..574812594f --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_7.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_7.js b/docs/RuntimeAPI/html/search/all_7.js new file mode 100644 index 0000000000..529e9394c3 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_7.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['global_20enum_20and_20defines',['Global enum and defines',['../group__GlobalDefs.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_8.html b/docs/RuntimeAPI/html/search/all_8.html new file mode 100644 index 0000000000..0179bdd463 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_8.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_8.js b/docs/RuntimeAPI/html/search/all_8.js new file mode 100644 index 0000000000..1e1570aaa0 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_8.js @@ -0,0 +1,230 @@ +var searchData= +[ + ['hip_20api',['HIP API',['../group__API.html',1,'']]], + ['has3dgrid',['has3dGrid',['../structhipDeviceArch__t.html#aa5e22d295cce0d9a34ee6e7e7e378c26',1,'hipDeviceArch_t']]], + ['hasdoubles',['hasDoubles',['../structhipDeviceArch__t.html#a2d624e3d85e615b71d1182f8912893b4',1,'hipDeviceArch_t']]], + ['hasdynamicparallelism',['hasDynamicParallelism',['../structhipDeviceArch__t.html#a9114bfc718bf0648b54ff9a319a36b35',1,'hipDeviceArch_t']]], + ['hasfloatatomicadd',['hasFloatAtomicAdd',['../structhipDeviceArch__t.html#a2321d6ef74aac91c044f3289d25b2d41',1,'hipDeviceArch_t']]], + ['hasfunnelshift',['hasFunnelShift',['../structhipDeviceArch__t.html#aaec4f2d983d4602858fae8e9ddeee3ff',1,'hipDeviceArch_t']]], + ['hasglobalfloatatomicexch',['hasGlobalFloatAtomicExch',['../structhipDeviceArch__t.html#a8f213ae9a4729dff1c636ac5de0e2fa2',1,'hipDeviceArch_t']]], + ['hasglobalint32atomics',['hasGlobalInt32Atomics',['../structhipDeviceArch__t.html#a8d00c3ab98869b602c714fe7abe68e93',1,'hipDeviceArch_t']]], + ['hasglobalint64atomics',['hasGlobalInt64Atomics',['../structhipDeviceArch__t.html#ad5aa54dbab22dbcd8cf98f57a96c6636',1,'hipDeviceArch_t']]], + ['hassharedfloatatomicexch',['hasSharedFloatAtomicExch',['../structhipDeviceArch__t.html#aff005558b4edabd27b27f286ac5b2f2b',1,'hipDeviceArch_t']]], + ['hassharedint32atomics',['hasSharedInt32Atomics',['../structhipDeviceArch__t.html#a1596330b1cb9cc73f142aee11b2ab853',1,'hipDeviceArch_t']]], + ['hassharedint64atomics',['hasSharedInt64Atomics',['../structhipDeviceArch__t.html#a00c2b930fcdcad9ea7b54b449db13966',1,'hipDeviceArch_t']]], + ['hassurfacefuncs',['hasSurfaceFuncs',['../structhipDeviceArch__t.html#a9eb2462148686d4c048b69b6e09f835e',1,'hipDeviceArch_t']]], + ['hassyncthreadsext',['hasSyncThreadsExt',['../structhipDeviceArch__t.html#ade6a3b21ad5f344dcd92c52102c274ba',1,'hipDeviceArch_t']]], + ['hasthreadfencesystem',['hasThreadFenceSystem',['../structhipDeviceArch__t.html#ac2818e3b91cba8beb36741e9867bb887',1,'hipDeviceArch_t']]], + ['haswarpballot',['hasWarpBallot',['../structhipDeviceArch__t.html#af1e934a8a5106995bcc256287585564c',1,'hipDeviceArch_t']]], + ['haswarpshuffle',['hasWarpShuffle',['../structhipDeviceArch__t.html#a3d922e8fc97ca1e8ecc39600b138fa2d',1,'hipDeviceArch_t']]], + ['haswarpvote',['hasWarpVote',['../structhipDeviceArch__t.html#a35bde017352eca1d4e0eceb3bf79f274',1,'hipDeviceArch_t']]], + ['hcc_2dspecific_20accessors',['HCC-Specific Accessors',['../group__HCC__Specific.html',1,'']]], + ['hip_20environment_20variables',['HIP Environment Variables',['../group__HIP-ENV.html',1,'']]], + ['hip_5fdynamic_5fshared',['HIP_DYNAMIC_SHARED',['../hcc__detail_2hip__runtime_8h.html#a1e7ec14fb6b74c03b6ec804d6ef7e0ea',1,'hip_runtime.h']]], + ['hip_5fhcc_2ecpp',['hip_hcc.cpp',['../hip__hcc_8cpp.html',1,'']]], + ['hip_5fruntime_2eh',['hip_runtime.h',['../hcc__detail_2hip__runtime_8h.html',1,'']]], + ['hip_5fruntime_5fapi_2eh',['hip_runtime_api.h',['../hcc__detail_2hip__runtime__api_8h.html',1,'']]], + ['hip_5ftexture_2eh',['hip_texture.h',['../hcc__detail_2hip__texture_8h.html',1,'']]], + ['hip_5fvector_5ftypes_2eh',['hip_vector_types.h',['../hcc__detail_2hip__vector__types_8h.html',1,'']]], + ['hiparray',['hipArray',['../structhipArray.html',1,'']]], + ['hipchannelformatdesc',['hipChannelFormatDesc',['../structhipChannelFormatDesc.html',1,'']]], + ['hipchoosedevice',['hipChooseDevice',['../group__Device.html#gaf1e365e1d17cf40644d1470de4817c8e',1,'hipChooseDevice(int *device, const hipDeviceProp_t *prop): hip_device.cpp'],['../group__Device.html#gaf1e365e1d17cf40644d1470de4817c8e',1,'hipChooseDevice(int *device, const hipDeviceProp_t *prop): hip_device.cpp']]], + ['hipcreatechanneldesc',['hipCreateChannelDesc',['../group__Texture.html#gac9b0ae733fa849842342fab6d42c7b6f',1,'hipCreateChannelDesc(int x, int y, int z, int w, hipChannelFormatKind f): hip_memory.cpp'],['../group__Texture.html#gac9b0ae733fa849842342fab6d42c7b6f',1,'hipCreateChannelDesc(int x, int y, int z, int w, hipChannelFormatKind f): hip_memory.cpp']]], + ['hipctxcreate',['hipCtxCreate',['../group__Context.html#gab6dbcff5c5b1249a5ac5cf39ae9d08bc',1,'hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device): hip_context.cpp'],['../group__Context.html#gab6dbcff5c5b1249a5ac5cf39ae9d08bc',1,'hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device): hip_context.cpp']]], + ['hipctxdestroy',['hipCtxDestroy',['../group__Context.html#ga9a65fe43238ef303a6d97826c05fd14e',1,'hipCtxDestroy(hipCtx_t ctx): hip_context.cpp'],['../group__Context.html#ga9a65fe43238ef303a6d97826c05fd14e',1,'hipCtxDestroy(hipCtx_t ctx): hip_context.cpp']]], + ['hipctxdisablepeeraccess',['hipCtxDisablePeerAccess',['../group__Context.html#gaf48e4e9c3b6bbad5deaeff10d2e28b31',1,'hipCtxDisablePeerAccess(hipCtx_t peerCtx): hip_peer.cpp'],['../group__Context.html#gaf48e4e9c3b6bbad5deaeff10d2e28b31',1,'hipCtxDisablePeerAccess(hipCtx_t peerCtx): hip_peer.cpp']]], + ['hipctxenablepeeraccess',['hipCtxEnablePeerAccess',['../group__Context.html#ga834dfd99d72082fe8770142fa30b30e2',1,'hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags): hip_peer.cpp'],['../group__Context.html#ga834dfd99d72082fe8770142fa30b30e2',1,'hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags): hip_peer.cpp']]], + ['hipctxgetapiversion',['hipCtxGetApiVersion',['../group__Context.html#gabe5941e6798c71a8351b7d51311e7e8f',1,'hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion): hip_context.cpp'],['../group__Context.html#gabe5941e6798c71a8351b7d51311e7e8f',1,'hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion): hip_context.cpp']]], + ['hipctxgetcacheconfig',['hipCtxGetCacheConfig',['../group__Context.html#gab10373068faafd3042c9003e2e6d905a',1,'hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig): hip_context.cpp'],['../group__Context.html#gab10373068faafd3042c9003e2e6d905a',1,'hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig): hip_context.cpp']]], + ['hipctxgetcurrent',['hipCtxGetCurrent',['../group__Context.html#ga741786101d348fdbfa1f64546860357a',1,'hipCtxGetCurrent(hipCtx_t *ctx): hip_context.cpp'],['../group__Context.html#ga741786101d348fdbfa1f64546860357a',1,'hipCtxGetCurrent(hipCtx_t *ctx): hip_context.cpp']]], + ['hipctxgetdevice',['hipCtxGetDevice',['../group__Context.html#ga8aa32cf64272da929f23ecbafefefcee',1,'hipCtxGetDevice(hipDevice_t *device): hip_context.cpp'],['../group__Context.html#ga8aa32cf64272da929f23ecbafefefcee',1,'hipCtxGetDevice(hipDevice_t *device): hip_context.cpp']]], + ['hipctxgetflags',['hipCtxGetFlags',['../group__Context.html#ga12a7a993e10f775fbf61a0b14288ed1b',1,'hipCtxGetFlags(unsigned int *flags): hip_context.cpp'],['../group__Context.html#ga12a7a993e10f775fbf61a0b14288ed1b',1,'hipCtxGetFlags(unsigned int *flags): hip_context.cpp']]], + ['hipctxgetsharedmemconfig',['hipCtxGetSharedMemConfig',['../group__Context.html#ga3c78b22dd03435a7ca88621a45409565',1,'hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_context.cpp'],['../group__Context.html#ga3c78b22dd03435a7ca88621a45409565',1,'hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_context.cpp']]], + ['hipctxpopcurrent',['hipCtxPopCurrent',['../group__Context.html#ga2cc4ea5a6b78d9d7990a88a7863467d4',1,'hipCtxPopCurrent(hipCtx_t *ctx): hip_context.cpp'],['../group__Context.html#ga2cc4ea5a6b78d9d7990a88a7863467d4',1,'hipCtxPopCurrent(hipCtx_t *ctx): hip_context.cpp']]], + ['hipctxpushcurrent',['hipCtxPushCurrent',['../group__Context.html#ga9c9d92f29d68cacdea4c062c97e50a8a',1,'hipCtxPushCurrent(hipCtx_t ctx): hip_context.cpp'],['../group__Context.html#ga9c9d92f29d68cacdea4c062c97e50a8a',1,'hipCtxPushCurrent(hipCtx_t ctx): hip_context.cpp']]], + ['hipctxsetcacheconfig',['hipCtxSetCacheConfig',['../group__Context.html#ga7eacc656f1d1b6f5a683bea31be67a2f',1,'hipCtxSetCacheConfig(hipFuncCache_t cacheConfig): hip_context.cpp'],['../group__Context.html#ga7eacc656f1d1b6f5a683bea31be67a2f',1,'hipCtxSetCacheConfig(hipFuncCache_t cacheConfig): hip_context.cpp']]], + ['hipctxsetcurrent',['hipCtxSetCurrent',['../group__Context.html#ga834a192f70c2bfc0269c309436776feb',1,'hipCtxSetCurrent(hipCtx_t ctx): hip_context.cpp'],['../group__Context.html#ga834a192f70c2bfc0269c309436776feb',1,'hipCtxSetCurrent(hipCtx_t ctx): hip_context.cpp']]], + ['hipctxsetsharedmemconfig',['hipCtxSetSharedMemConfig',['../group__Context.html#gad5c1d4ced27f584a74ed550dd002fa5a',1,'hipCtxSetSharedMemConfig(hipSharedMemConfig config): hip_context.cpp'],['../group__Context.html#gad5c1d4ced27f584a74ed550dd002fa5a',1,'hipCtxSetSharedMemConfig(hipSharedMemConfig config): hip_context.cpp']]], + ['hipctxsynchronize',['hipCtxSynchronize',['../group__Context.html#gad45cd968e8e3dcfd24ef050cab2f41c8',1,'hipCtxSynchronize(void): hip_context.cpp'],['../group__Context.html#gad45cd968e8e3dcfd24ef050cab2f41c8',1,'hipCtxSynchronize(void): hip_context.cpp']]], + ['hipdevicearch_5ft',['hipDeviceArch_t',['../structhipDeviceArch__t.html',1,'']]], + ['hipdeviceattribute_5ft',['hipDeviceAttribute_t',['../group__GlobalDefs.html#gacc0acd7b9bda126c6bb3dfd6e2796d7c',1,'hip_runtime_api.h']]], + ['hipdeviceattributeclockrate',['hipDeviceAttributeClockRate',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca2300e077e020e7967592065561373b00',1,'hip_runtime_api.h']]], + ['hipdeviceattributecomputecapabilitymajor',['hipDeviceAttributeComputeCapabilityMajor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca2735739cf977b7d303266f6781131e8d',1,'hip_runtime_api.h']]], + ['hipdeviceattributecomputecapabilityminor',['hipDeviceAttributeComputeCapabilityMinor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca38edc4fcae456e47160d349da3249b85',1,'hip_runtime_api.h']]], + ['hipdeviceattributecomputemode',['hipDeviceAttributeComputeMode',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca4d0369a6ef7bd7890fdcabc16ed3385d',1,'hip_runtime_api.h']]], + ['hipdeviceattributeconcurrentkernels',['hipDeviceAttributeConcurrentKernels',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cad9f45254d0d048677f560032532d5504',1,'hip_runtime_api.h']]], + ['hipdeviceattributeismultigpuboard',['hipDeviceAttributeIsMultiGpuBoard',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca6129311200a17dcc5fa8d2256874ae3d',1,'hip_runtime_api.h']]], + ['hipdeviceattributel2cachesize',['hipDeviceAttributeL2CacheSize',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca582ae5a26a7148504878890028e4b64c',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxblockdimx',['hipDeviceAttributeMaxBlockDimX',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cac1e4ac589db0d8adbbc241e3d0fcd594',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxblockdimy',['hipDeviceAttributeMaxBlockDimY',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca187dbffe12db09a56c0f75c340d879c9',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxblockdimz',['hipDeviceAttributeMaxBlockDimZ',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7caf811f51e03d1ffb025d80ac1da088675',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxgriddimx',['hipDeviceAttributeMaxGridDimX',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca03db8df0e7a9fbdaae683d97e8ac9c87',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxgriddimy',['hipDeviceAttributeMaxGridDimY',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca5b5cc49972679c5ccf62b79425ee99df',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxgriddimz',['hipDeviceAttributeMaxGridDimZ',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca6c206ac083999caf4640e5d91dae24f7',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxregistersperblock',['hipDeviceAttributeMaxRegistersPerBlock',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca82289b170192b6ea742be0efc6f95107',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxsharedmemoryperblock',['hipDeviceAttributeMaxSharedMemoryPerBlock',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca7bca3aa18b26d40eba043ae93e15c7e5',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxsharedmemorypermultiprocessor',['hipDeviceAttributeMaxSharedMemoryPerMultiprocessor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cad3e7f3d01533b32e12211172fcf410ba',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxthreadsperblock',['hipDeviceAttributeMaxThreadsPerBlock',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca8327aa23782d9c994bdef33a6d62e02e',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxthreadspermultiprocessor',['hipDeviceAttributeMaxThreadsPerMultiProcessor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7caddc08922b491eb1f6a583833cbf4e2f0',1,'hip_runtime_api.h']]], + ['hipdeviceattributememorybuswidth',['hipDeviceAttributeMemoryBusWidth',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca848c1396fab6f20463c6aefb828b0870',1,'hip_runtime_api.h']]], + ['hipdeviceattributememoryclockrate',['hipDeviceAttributeMemoryClockRate',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca6b68deafd65f036b30dc8051573eb000',1,'hip_runtime_api.h']]], + ['hipdeviceattributemultiprocessorcount',['hipDeviceAttributeMultiprocessorCount',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca5c1519870733ccf0b83f722678240e5f',1,'hip_runtime_api.h']]], + ['hipdeviceattributepcibusid',['hipDeviceAttributePciBusId',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca572b29c44f1322aa7657fdd784832f88',1,'hip_runtime_api.h']]], + ['hipdeviceattributepcideviceid',['hipDeviceAttributePciDeviceId',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca955d90286e87be9e3528f0b817ab32ff',1,'hip_runtime_api.h']]], + ['hipdeviceattributetotalconstantmemory',['hipDeviceAttributeTotalConstantMemory',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cac6089ac3a0f9c77cc382fb0eaa73ae9c',1,'hip_runtime_api.h']]], + ['hipdeviceattributewarpsize',['hipDeviceAttributeWarpSize',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7caffd94133e823247a6f1215343232f6ec',1,'hip_runtime_api.h']]], + ['hipdevicecanaccesspeer',['hipDeviceCanAccessPeer',['../group__PeerToPeer.html#ga0a1c9ccd775758d9d7d5b5a1f525b719',1,'hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId): hip_peer.cpp'],['../group__PeerToPeer.html#ga0a1c9ccd775758d9d7d5b5a1f525b719',1,'hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId): hip_peer.cpp']]], + ['hipdevicecomputecapability',['hipDeviceComputeCapability',['../group__Driver.html#ga0a1cf94d2b571ca3279577d5af0d1672',1,'hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device): hip_device.cpp'],['../group__Driver.html#ga0a1cf94d2b571ca3279577d5af0d1672',1,'hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device): hip_device.cpp']]], + ['hipdevicedisablepeeraccess',['hipDeviceDisablePeerAccess',['../group__PeerToPeer.html#ga85030c72824fb60aaddc7374ab60481b',1,'hipDeviceDisablePeerAccess(int peerDeviceId): hip_peer.cpp'],['../group__PeerToPeer.html#ga85030c72824fb60aaddc7374ab60481b',1,'hipDeviceDisablePeerAccess(int peerDeviceId): hip_peer.cpp']]], + ['hipdeviceenablepeeraccess',['hipDeviceEnablePeerAccess',['../group__PeerToPeer.html#ga0caca59034134d7a7bb893cc1caa653e',1,'hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags): hip_peer.cpp'],['../group__PeerToPeer.html#ga0caca59034134d7a7bb893cc1caa653e',1,'hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags): hip_peer.cpp']]], + ['hipdeviceget',['hipDeviceGet',['../group__Driver.html#gadf6c74aaf7c22fea80e9ac3400d43704',1,'hipDeviceGet(hipDevice_t *device, int ordinal): hip_context.cpp'],['../group__Driver.html#gadf6c74aaf7c22fea80e9ac3400d43704',1,'hipDeviceGet(hipDevice_t *device, int deviceId): hip_context.cpp']]], + ['hipdevicegetattribute',['hipDeviceGetAttribute',['../group__Device.html#ga7080a145a4239a7276e0dc22062026c1',1,'hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int deviceId): hip_device.cpp'],['../group__Device.html#ga7080a145a4239a7276e0dc22062026c1',1,'hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int device): hip_device.cpp']]], + ['hipdevicegetbypcibusid',['hipDeviceGetByPCIBusId',['../group__Driver.html#gafe08d5632ebd24edc26a19da380b2120',1,'hipDeviceGetByPCIBusId(int *device, const int *pciBusId): hip_device.cpp'],['../group__Driver.html#gafe08d5632ebd24edc26a19da380b2120',1,'hipDeviceGetByPCIBusId(int *device, const int *pciBusId): hip_device.cpp']]], + ['hipdevicegetcacheconfig',['hipDeviceGetCacheConfig',['../group__Device.html#ga37057f9830ad6fab7ce5f05f6d3c89ab',1,'hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig): hip_device.cpp'],['../group__Device.html#ga37057f9830ad6fab7ce5f05f6d3c89ab',1,'hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig): hip_device.cpp']]], + ['hipdevicegetlimit',['hipDeviceGetLimit',['../group__Device.html#gaa72baeb520f1d52fe17817e852937218',1,'hipDeviceGetLimit(size_t *pValue, hipLimit_t limit): hip_device.cpp'],['../group__Device.html#gaa72baeb520f1d52fe17817e852937218',1,'hipDeviceGetLimit(size_t *pValue, hipLimit_t limit): hip_device.cpp']]], + ['hipdevicegetname',['hipDeviceGetName',['../group__Driver.html#ga24df15e180a7b2b351cd362e5b7d2dac',1,'hipDeviceGetName(char *name, int len, hipDevice_t device): hip_device.cpp'],['../group__Driver.html#ga24df15e180a7b2b351cd362e5b7d2dac',1,'hipDeviceGetName(char *name, int len, hipDevice_t device): hip_device.cpp']]], + ['hipdevicegetpcibusid',['hipDeviceGetPCIBusId',['../group__Driver.html#gaaa961c8ec8047903617f4245fa50256a',1,'hipDeviceGetPCIBusId(char *pciBusId, int len, int device): hip_device.cpp'],['../group__Driver.html#gaaa961c8ec8047903617f4245fa50256a',1,'hipDeviceGetPCIBusId(char *pciBusId, int len, int device): hip_device.cpp']]], + ['hipdevicegetsharedmemconfig',['hipDeviceGetSharedMemConfig',['../group__Device.html#ga1bb08f774a34a468d969a8a04791c9bb',1,'hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_device.cpp'],['../group__Device.html#ga1bb08f774a34a468d969a8a04791c9bb',1,'hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_device.cpp']]], + ['hipdeviceprop_5ft',['hipDeviceProp_t',['../structhipDeviceProp__t.html',1,'']]], + ['hipdevicereset',['hipDeviceReset',['../group__Device.html#ga8d57161ae56a8edc46eeda447417bf6c',1,'hipDeviceReset(void): hip_device.cpp'],['../group__Device.html#ga8d57161ae56a8edc46eeda447417bf6c',1,'hipDeviceReset(void): hip_device.cpp']]], + ['hipdevicescheduleauto',['hipDeviceScheduleAuto',['../group__GlobalDefs.html#ga9032d35eb7383948828ee48e1e19f5fd',1,'hip_runtime_api.h']]], + ['hipdeviceschedulespin',['hipDeviceScheduleSpin',['../group__GlobalDefs.html#ga713d962bafb7758dc1ff0675e4239453',1,'hip_runtime_api.h']]], + ['hipdevicescheduleyield',['hipDeviceScheduleYield',['../group__GlobalDefs.html#gaaf1e2706430c06601aa12a8af2a0ba5a',1,'hip_runtime_api.h']]], + ['hipdevicesetcacheconfig',['hipDeviceSetCacheConfig',['../group__Device.html#gaada3d30a46ae06f68cf1574f496b86ee',1,'hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig): hip_device.cpp'],['../group__Device.html#gaada3d30a46ae06f68cf1574f496b86ee',1,'hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig): hip_device.cpp']]], + ['hipdevicesetsharedmemconfig',['hipDeviceSetSharedMemConfig',['../group__Device.html#ga9b1f279084e76691cedfbfadf9c717ee',1,'hipDeviceSetSharedMemConfig(hipSharedMemConfig config): hip_device.cpp'],['../group__Device.html#ga9b1f279084e76691cedfbfadf9c717ee',1,'hipDeviceSetSharedMemConfig(hipSharedMemConfig config): hip_device.cpp']]], + ['hipdevicesynchronize',['hipDeviceSynchronize',['../group__Device.html#gaefdc2847fb1d6c3fb1354e827a191ebd',1,'hipDeviceSynchronize(void): hip_device.cpp'],['../group__Device.html#gaefdc2847fb1d6c3fb1354e827a191ebd',1,'hipDeviceSynchronize(void): hip_device.cpp']]], + ['hipdevicetotalmem',['hipDeviceTotalMem',['../group__Driver.html#ga8991e535d0ef1ead0524e73364623041',1,'hipDeviceTotalMem(size_t *bytes, hipDevice_t device): hip_device.cpp'],['../group__Driver.html#ga8991e535d0ef1ead0524e73364623041',1,'hipDeviceTotalMem(size_t *bytes, hipDevice_t device): hip_device.cpp']]], + ['hipdoublecomplex',['hipDoubleComplex',['../structhipDoubleComplex.html',1,'']]], + ['hipdrivergetversion',['hipDriverGetVersion',['../group__Driver.html#gaf6c342f52d2a29a0aca5cdd89b4dd47c',1,'hipDriverGetVersion(int *driverVersion): hip_context.cpp'],['../group__Driver.html#gaf6c342f52d2a29a0aca5cdd89b4dd47c',1,'hipDriverGetVersion(int *driverVersion): hip_context.cpp']]], + ['hiperror_5ft',['hipError_t',['../group__GlobalDefs.html#gadf5010f6e140a53ecbdf949e73e87594',1,'hip_runtime_api.h']]], + ['hiperrorhostmemoryalreadyregistered',['hipErrorHostMemoryAlreadyRegistered',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a9d7173cea72aace620a83d502569de1b',1,'hip_runtime_api.h']]], + ['hiperrorhostmemorynotregistered',['hipErrorHostMemoryNotRegistered',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a6901476ca88eed786fb8be003d9661d9',1,'hip_runtime_api.h']]], + ['hiperrorinitializationerror',['hipErrorInitializationError',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a7e935ae88ee1f9ff3920156ac6864520',1,'hip_runtime_api.h']]], + ['hiperrorinvalidcontext',['hipErrorInvalidContext',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ad9d895f7f4dbf63a4586b2aab593c9fe',1,'hip_runtime_api.h']]], + ['hiperrorinvaliddevice',['hipErrorInvalidDevice',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a07ab9b704ea693c1781a52741c60cd0d',1,'hip_runtime_api.h']]], + ['hiperrorinvaliddevicepointer',['hipErrorInvalidDevicePointer',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a37a93fcd2b0aed9bf52b82fa26031e6f',1,'hip_runtime_api.h']]], + ['hiperrorinvalidkernelfile',['hipErrorInvalidKernelFile',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ae508537d5e9a293bb1e63802d4a61380',1,'hip_runtime_api.h']]], + ['hiperrorinvalidmemcpydirection',['hipErrorInvalidMemcpyDirection',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ae35ad9fb66b7c33f21a021efe89975b0',1,'hip_runtime_api.h']]], + ['hiperrorinvalidresourcehandle',['hipErrorInvalidResourceHandle',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a88e525a7c8f35552dfada58e9f2f6d3a',1,'hip_runtime_api.h']]], + ['hiperrorinvalidvalue',['hipErrorInvalidValue',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a1e8215fe1108a508bad3944bce7b4d83',1,'hip_runtime_api.h']]], + ['hiperrorlaunchfailure',['hipErrorLaunchFailure',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a3df72146e762c5a901d3af06027a1de9',1,'hip_runtime_api.h']]], + ['hiperrorlaunchoutofresources',['hipErrorLaunchOutOfResources',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a3ea57f211997a84fa1c582414bcabc47',1,'hip_runtime_api.h']]], + ['hiperrormapbufferobjectfailed',['hipErrorMapBufferObjectFailed',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ab4de2174991cc16a3e1986f5da25262b',1,'hip_runtime_api.h']]], + ['hiperrormemoryallocation',['hipErrorMemoryAllocation',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a8293288a10109874749afe2562db09f2',1,'hip_runtime_api.h']]], + ['hiperrornodevice',['hipErrorNoDevice',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ad4406972c318df36d231310a15131c24',1,'hip_runtime_api.h']]], + ['hiperrornotready',['hipErrorNotReady',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594aa9638063c8746a9d1fda2b2069a0a9f1',1,'hip_runtime_api.h']]], + ['hiperrorpeeraccessalreadyenabled',['hipErrorPeerAccessAlreadyEnabled',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a5399c146f91246f8b77abcd4ea30e7ac',1,'hip_runtime_api.h']]], + ['hiperrorpeeraccessnotenabled',['hipErrorPeerAccessNotEnabled',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a2ee0bf2e64840f253e4a1b12bbaf2d09',1,'hip_runtime_api.h']]], + ['hiperrorruntimememory',['hipErrorRuntimeMemory',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a1159eb9a6be66bea740a8bfb61378723',1,'hip_runtime_api.h']]], + ['hiperrorruntimeother',['hipErrorRuntimeOther',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a44f847c5914be2177feac107dcb096d1',1,'hip_runtime_api.h']]], + ['hiperrortbd',['hipErrorTbd',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ab556409e11ddb0c4cf77a2f4fc91ea9e',1,'hip_runtime_api.h']]], + ['hiperrorunknown',['hipErrorUnknown',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594aa74e64c5b2f5fb0d6a92681f5b234073',1,'hip_runtime_api.h']]], + ['hipeventblockingsync',['hipEventBlockingSync',['../group__GlobalDefs.html#gafa1c076a5b991763a98695063f1ea11d',1,'hip_runtime_api.h']]], + ['hipeventcreate',['hipEventCreate',['../group__Event.html#ga5df2309c9f29ca4c8e669db658d411b4',1,'hipEventCreate(hipEvent_t *event): hip_event.cpp'],['../group__Event.html#ga5df2309c9f29ca4c8e669db658d411b4',1,'hipEventCreate(hipEvent_t *event): hip_event.cpp']]], + ['hipeventcreatewithflags',['hipEventCreateWithFlags',['../group__Event.html#gae86a5acb1b22b61bc9ecb9c28fc71b75',1,'hipEventCreateWithFlags(hipEvent_t *event, unsigned flags): hip_event.cpp'],['../group__Event.html#gae86a5acb1b22b61bc9ecb9c28fc71b75',1,'hipEventCreateWithFlags(hipEvent_t *event, unsigned flags): hip_event.cpp']]], + ['hipeventdefault',['hipEventDefault',['../group__GlobalDefs.html#ga122a5853359eba97cf047ddd153740f0',1,'hip_runtime_api.h']]], + ['hipeventdestroy',['hipEventDestroy',['../group__Event.html#ga83260357dce0c39e8c6a3c74ec97484c',1,'hipEventDestroy(hipEvent_t event): hip_event.cpp'],['../group__Event.html#ga83260357dce0c39e8c6a3c74ec97484c',1,'hipEventDestroy(hipEvent_t event): hip_event.cpp']]], + ['hipeventdisabletiming',['hipEventDisableTiming',['../group__GlobalDefs.html#ga3c0f44a85e36a4c67671da6bcdad0351',1,'hip_runtime_api.h']]], + ['hipeventelapsedtime',['hipEventElapsedTime',['../group__Event.html#gad4128b815cb475c8e13c7e66ff6250b7',1,'hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop): hip_event.cpp'],['../group__Event.html#gad4128b815cb475c8e13c7e66ff6250b7',1,'hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop): hip_event.cpp']]], + ['hipeventinterprocess',['hipEventInterprocess',['../group__GlobalDefs.html#ga0f01d74059baa704e42aeff8222166bb',1,'hip_runtime_api.h']]], + ['hipeventquery',['hipEventQuery',['../group__Event.html#ga5d12d7b798b5ceb5932d1ac21f5ac776',1,'hipEventQuery(hipEvent_t event): hip_event.cpp'],['../group__Event.html#ga5d12d7b798b5ceb5932d1ac21f5ac776',1,'hipEventQuery(hipEvent_t event): hip_event.cpp']]], + ['hipeventrecord',['hipEventRecord',['../group__Event.html#ga553b6f7a8e7b7dd9536d8a64c24d7e29',1,'hipEventRecord(hipEvent_t event, hipStream_t stream): hip_event.cpp'],['../group__Event.html#ga553b6f7a8e7b7dd9536d8a64c24d7e29',1,'hipEventRecord(hipEvent_t event, hipStream_t stream): hip_event.cpp']]], + ['hipeventsynchronize',['hipEventSynchronize',['../group__Event.html#ga1f72d98ba5d6f7dc3da54e0c41fe38b1',1,'hipEventSynchronize(hipEvent_t event): hip_event.cpp'],['../group__Event.html#ga1f72d98ba5d6f7dc3da54e0c41fe38b1',1,'hipEventSynchronize(hipEvent_t event): hip_event.cpp']]], + ['hipfiltermodepoint',['hipFilterModePoint',['../hcc__detail_2hip__texture_8h.html#aa2f0b6002b81d0a43a808cb880bb21e6a56ede038ab7c805ec4b5b61d2b678dfc',1,'hip_texture.h']]], + ['hipfloatcomplex',['hipFloatComplex',['../structhipFloatComplex.html',1,'']]], + ['hipfree',['hipFree',['../group__Memory.html#ga740d08da65cae1441ba32f8fedb863d1',1,'hipFree(void *ptr): hip_memory.cpp'],['../group__Memory.html#ga740d08da65cae1441ba32f8fedb863d1',1,'hipFree(void *ptr): hip_memory.cpp']]], + ['hipfreearray',['hipFreeArray',['../hcc__detail_2hip__texture_8h.html#ae29935073517245c69bc64790dd2747a',1,'hip_memory.cpp']]], + ['hipfreehost',['hipFreeHost',['../group__Memory.html#gad2164cc3d49da53052f4b83b789e90c9',1,'hipFreeHost(void *ptr) __attribute__((deprecated("use hipHostFree instead"))): hip_memory.cpp'],['../group__Memory.html#gad2164cc3d49da53052f4b83b789e90c9',1,'hipFreeHost(void *ptr): hip_memory.cpp']]], + ['hipfunccache_5ft',['hipFuncCache_t',['../group__GlobalDefs.html#ga86e5c5692645963a9a673e1aa88ba6ca',1,'hipFuncCache_t(): hip_runtime_api.h'],['../group__GlobalDefs.html#ga7effbca2af70714feaa3330bf1a77a72',1,'hipFuncCache_t(): hip_runtime_api.h']]], + ['hipfunccachepreferequal',['hipFuncCachePreferEqual',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa0ddab0e840107634a152033103be44d7',1,'hip_runtime_api.h']]], + ['hipfunccachepreferl1',['hipFuncCachePreferL1',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa636a3c140db6b9d4a8bf7d5a61c398c5',1,'hip_runtime_api.h']]], + ['hipfunccacheprefernone',['hipFuncCachePreferNone',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa0813fbaa008ce1231ff9fed3911eb3af',1,'hip_runtime_api.h']]], + ['hipfunccacheprefershared',['hipFuncCachePreferShared',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa9b34337dfbadba25ed2aa270bbcabc43',1,'hip_runtime_api.h']]], + ['hipfuncsetcacheconfig',['hipFuncSetCacheConfig',['../group__Device.html#ga82f0637f09eb1082f6fd025f76c2c6f5',1,'hipFuncSetCacheConfig(hipFuncCache_t config): hip_device.cpp'],['../group__Device.html#ga82f0637f09eb1082f6fd025f76c2c6f5',1,'hipFuncSetCacheConfig(hipFuncCache_t cacheConfig): hip_device.cpp']]], + ['hipgetdevice',['hipGetDevice',['../group__Device.html#ga7e0e2e8c5f78e3c7449764657c254e0a',1,'hipGetDevice(int *deviceId): hip_device.cpp'],['../group__Device.html#ga7e0e2e8c5f78e3c7449764657c254e0a',1,'hipGetDevice(int *deviceId): hip_device.cpp']]], + ['hipgetdevicecount',['hipGetDeviceCount',['../group__Device.html#ga8555d5c76d88c50ddbf54ae70b568394',1,'hipGetDeviceCount(int *count): hip_device.cpp'],['../group__Device.html#ga8555d5c76d88c50ddbf54ae70b568394',1,'hipGetDeviceCount(int *count): hip_device.cpp']]], + ['hipgetdeviceproperties',['hipGetDeviceProperties',['../group__Device.html#ga32208513b7cd491f0cb5fc884053f790',1,'hipGetDeviceProperties(hipDeviceProp_t *prop, int deviceId): hip_device.cpp'],['../group__Device.html#ga32208513b7cd491f0cb5fc884053f790',1,'hipGetDeviceProperties(hipDeviceProp_t *props, int device): hip_device.cpp']]], + ['hipgeterrorname',['hipGetErrorName',['../group__Error.html#ga88c474d77635523dbf6ca67be7b56999',1,'hipGetErrorName(hipError_t hip_error): hip_error.cpp'],['../group__Error.html#ga88c474d77635523dbf6ca67be7b56999',1,'hipGetErrorName(hipError_t hip_error): hip_error.cpp']]], + ['hipgeterrorstring',['hipGetErrorString',['../group__Error.html#ga1421a2ee561f30d5b1dbed3ecb28b264',1,'hipGetErrorString(hipError_t hipError): hip_error.cpp'],['../group__Error.html#ga1421a2ee561f30d5b1dbed3ecb28b264',1,'hipGetErrorString(hipError_t hip_error): hip_error.cpp']]], + ['hipgetlasterror',['hipGetLastError',['../group__Error.html#ga533daeb9114d7fc2db8d867adf9e419b',1,'hipGetLastError(void): hip_error.cpp'],['../group__Error.html#ga533daeb9114d7fc2db8d867adf9e419b',1,'hipGetLastError(): hip_error.cpp']]], + ['hiphostalloc',['hipHostAlloc',['../group__Memory.html#gac5ae7f11cefc7085f39ee7868e71724a',1,'hipHostAlloc(void **ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead"))): hip_memory.cpp'],['../group__Memory.html#gac5ae7f11cefc7085f39ee7868e71724a',1,'hipHostAlloc(void **ptr, size_t sizeBytes, unsigned int flags): hip_memory.cpp']]], + ['hiphostfree',['hipHostFree',['../group__Memory.html#ga2e543f58ee4544e317cd695d6d82e0a3',1,'hipHostFree(void *ptr): hip_memory.cpp'],['../group__Memory.html#ga2e543f58ee4544e317cd695d6d82e0a3',1,'hipHostFree(void *ptr): hip_memory.cpp']]], + ['hiphostgetdevicepointer',['hipHostGetDevicePointer',['../group__Memory.html#ga8fa7a0478020b835a24785cd6bb89725',1,'hip_runtime_api.h']]], + ['hiphostgetflags',['hipHostGetFlags',['../group__Memory.html#ga4d26915873b3e3534ceb4dc310f8709a',1,'hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr): hip_memory.cpp'],['../group__Memory.html#ga4d26915873b3e3534ceb4dc310f8709a',1,'hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr): hip_memory.cpp']]], + ['hiphostmalloc',['hipHostMalloc',['../group__Memory.html#gaad40bc7d97ccc799403ef5a9a8c246e1',1,'hipHostMalloc(void **ptr, size_t size, unsigned int flags): hip_memory.cpp'],['../group__Memory.html#gaad40bc7d97ccc799403ef5a9a8c246e1',1,'hipHostMalloc(void **ptr, size_t sizeBytes, unsigned int flags): hip_memory.cpp']]], + ['hiphostmallocdefault',['hipHostMallocDefault',['../group__GlobalDefs.html#gad594ec51cb5b5e946c1e354bf80bddc7',1,'hip_runtime_api.h']]], + ['hiphostregister',['hipHostRegister',['../group__Memory.html#gab8258f051e1a1f7385f794a15300e674',1,'hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags): hip_memory.cpp'],['../group__Memory.html#gab8258f051e1a1f7385f794a15300e674',1,'hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags): hip_memory.cpp']]], + ['hiphostregisterdefault',['hipHostRegisterDefault',['../group__GlobalDefs.html#gac7c100d241ff84ad10109bb00b7b25dc',1,'hip_runtime_api.h']]], + ['hiphostregisteriomemory',['hipHostRegisterIoMemory',['../group__GlobalDefs.html#gaefa79f1b4481d6a1d1091c14b24f33d0',1,'hip_runtime_api.h']]], + ['hiphostregistermapped',['hipHostRegisterMapped',['../group__GlobalDefs.html#gacfa4edcfcb39fc61bff6bdecb14d7618',1,'hip_runtime_api.h']]], + ['hiphostregisterportable',['hipHostRegisterPortable',['../group__GlobalDefs.html#ga2db444f2315d412d3c7ba80ec6049583',1,'hip_runtime_api.h']]], + ['hiphostunregister',['hipHostUnregister',['../group__Memory.html#ga4c9e1810b9f5858d36c4d28c91c86924',1,'hipHostUnregister(void *hostPtr): hip_memory.cpp'],['../group__Memory.html#ga4c9e1810b9f5858d36c4d28c91c86924',1,'hipHostUnregister(void *hostPtr): hip_memory.cpp']]], + ['hipinit',['hipInit',['../group__Driver.html#ga01baa652dda5815c594d047060496caa',1,'hipInit(unsigned int flags): hip_context.cpp'],['../group__Driver.html#ga01baa652dda5815c594d047060496caa',1,'hipInit(unsigned int flags): hip_context.cpp']]], + ['hipipcclosememhandle',['hipIpcCloseMemHandle',['../group__API.html#gac2db0688a6a471e17ca631977e199da7',1,'hipIpcCloseMemHandle(void *devPtr): hip_memory.cpp'],['../group__API.html#gac2db0688a6a471e17ca631977e199da7',1,'hipIpcCloseMemHandle(void *devPtr): hip_memory.cpp']]], + ['hipipcgetmemhandle',['hipIpcGetMemHandle',['../group__API.html#gafd8c80f7e3b6426a630fff768409be70',1,'hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr): hip_memory.cpp'],['../group__API.html#gafd8c80f7e3b6426a630fff768409be70',1,'hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr): hip_memory.cpp']]], + ['hipipcopenmemhandle',['hipIpcOpenMemHandle',['../group__API.html#ga2ada334c986e10805d58167e260cb0df',1,'hipIpcOpenMemHandle(void **devPtr, hipIpcMemHandle_t handle, unsigned int flags): hip_memory.cpp'],['../group__API.html#ga2ada334c986e10805d58167e260cb0df',1,'hipIpcOpenMemHandle(void **devPtr, hipIpcMemHandle_t handle, unsigned int flags): hip_memory.cpp']]], + ['hipmalloc',['hipMalloc',['../group__Memory.html#ga4c6fcfe80010069d2792780d00dcead2',1,'hipMalloc(void **ptr, size_t size): hip_memory.cpp'],['../group__Memory.html#ga4c6fcfe80010069d2792780d00dcead2',1,'hipMalloc(void **ptr, size_t sizeBytes): hip_memory.cpp']]], + ['hipmallocarray',['hipMallocArray',['../hcc__detail_2hip__texture_8h.html#aea2f071ae2e4a0847f6aa42d1ef56129',1,'hip_memory.cpp']]], + ['hipmallochost',['hipMallocHost',['../group__Memory.html#gad3d3cdf82eb0058fc9eac1f939cd9d30',1,'hipMallocHost(void **ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead"))): hip_memory.cpp'],['../group__Memory.html#gad3d3cdf82eb0058fc9eac1f939cd9d30',1,'hipMallocHost(void **ptr, size_t sizeBytes): hip_memory.cpp']]], + ['hipmallocpitch',['hipMallocPitch',['../group__Memory.html#ga805c7320498926e444616fe090c727ee',1,'hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height): hip_memory.cpp'],['../group__Memory.html#ga805c7320498926e444616fe090c727ee',1,'hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height): hip_memory.cpp']]], + ['hipmemcpy',['hipMemcpy',['../group__Memory.html#gac1a055d288302edd641c6d7416858e1e',1,'hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind): hip_memory.cpp'],['../group__Memory.html#gac1a055d288302edd641c6d7416858e1e',1,'hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind): hip_memory.cpp']]], + ['hipmemcpy2d',['hipMemcpy2D',['../hcc__detail_2hip__texture_8h.html#a8af4597ff0cd17247d8a857c4d8bfa8a',1,'hip_memory.cpp']]], + ['hipmemcpy2dtoarray',['hipMemcpy2DToArray',['../hcc__detail_2hip__texture_8h.html#ab08b0bd6cdbb57e79df40ef1cd0374e0',1,'hip_memory.cpp']]], + ['hipmemcpyasync',['hipMemcpyAsync',['../group__Memory.html#gad55fa9f5980b711bc93c52820149ba18',1,'hipMemcpyAsync(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gad55fa9f5980b711bc93c52820149ba18',1,'hipMemcpyAsync(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpydefault',['hipMemcpyDefault',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18a4e37107e416f79a2edf2b6534163c823',1,'hip_runtime_api.h']]], + ['hipmemcpydevicetodevice',['hipMemcpyDeviceToDevice',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18abd05a09d3105e0ce25b34dd91cf83f88',1,'hip_runtime_api.h']]], + ['hipmemcpydevicetohost',['hipMemcpyDeviceToHost',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18aba2505e9ce1e5382f17730bc670917d1',1,'hip_runtime_api.h']]], + ['hipmemcpydtod',['hipMemcpyDtoD',['../group__Memory.html#ga814f245b8918f173c1f2f8c1480f7f93',1,'hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#ga814f245b8918f173c1f2f8c1480f7f93',1,'hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemcpydtodasync',['hipMemcpyDtoDAsync',['../group__Memory.html#gad3bba6016cba62f0e933cdfb4c312d27',1,'hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gad3bba6016cba62f0e933cdfb4c312d27',1,'hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpydtoh',['hipMemcpyDtoH',['../group__Memory.html#gae61f4e35ff1b9643c6328bc45d091c3f',1,'hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#gae61f4e35ff1b9643c6328bc45d091c3f',1,'hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemcpydtohasync',['hipMemcpyDtoHAsync',['../group__Memory.html#gad69da1994a646b843fb1fa465dbeb623',1,'hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gad69da1994a646b843fb1fa465dbeb623',1,'hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpyhosttodevice',['hipMemcpyHostToDevice',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18aff32175ecb0c7113200286eff8211008',1,'hip_runtime_api.h']]], + ['hipmemcpyhosttohost',['hipMemcpyHostToHost',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18a9d66b705aa85a9c83f0f533cef70d0af',1,'hip_runtime_api.h']]], + ['hipmemcpyhtod',['hipMemcpyHtoD',['../group__Memory.html#gaa415ecdd2535e1eb618ed8a31a6f03af',1,'hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#gaa415ecdd2535e1eb618ed8a31a6f03af',1,'hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemcpyhtodasync',['hipMemcpyHtoDAsync',['../group__Memory.html#gae3e76acb34bed5e3b19b4c1822975bcd',1,'hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gae3e76acb34bed5e3b19b4c1822975bcd',1,'hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpykind',['hipMemcpyKind',['../group__GlobalDefs.html#ga232e222db36b1fc672ba98054d036a18',1,'hipMemcpyKind(): hip_runtime_api.h'],['../group__GlobalDefs.html#ga0c04e67413ce030817361f02673e5c85',1,'hipMemcpyKind(): hip_runtime_api.h']]], + ['hipmemcpypeer',['hipMemcpyPeer',['../group__PeerToPeer.html#ga5512f45e25c08052667c8ffe7162333b',1,'hipMemcpyPeer(void *dst, int dstDeviceId, const void *src, int srcDeviceId, size_t sizeBytes): hip_peer.cpp'],['../group__PeerToPeer.html#ga5512f45e25c08052667c8ffe7162333b',1,'hipMemcpyPeer(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes): hip_peer.cpp']]], + ['hipmemcpypeerasync',['hipMemcpyPeerAsync',['../group__PeerToPeer.html#ga216f951370c931d22e80c089ab724ed9',1,'hipMemcpyPeerAsync(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream): hip_peer.cpp'],['../group__PeerToPeer.html#ga216f951370c931d22e80c089ab724ed9',1,'hipMemcpyPeerAsync(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream): hip_peer.cpp']]], + ['hipmemcpytoarray',['hipMemcpyToArray',['../hcc__detail_2hip__texture_8h.html#a2e631f0b6757b47b74cdaa2911784be9',1,'hip_memory.cpp']]], + ['hipmemcpytosymbol',['hipMemcpyToSymbol',['../group__Memory.html#ga131ac5c1ba04e186112491cb9bf964bc',1,'hipMemcpyToSymbol(const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind): hip_memory.cpp'],['../group__Memory.html#ga131ac5c1ba04e186112491cb9bf964bc',1,'hipMemcpyToSymbol(const char *symbolName, const void *src, size_t count, size_t offset, hipMemcpyKind kind): hip_memory.cpp']]], + ['hipmemcpytosymbolasync',['hipMemcpyToSymbolAsync',['../group__Memory.html#ga7dd66b9e8e0e7159177b42a250fe23c4',1,'hipMemcpyToSymbolAsync(const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#ga7dd66b9e8e0e7159177b42a250fe23c4',1,'hipMemcpyToSymbolAsync(const char *symbolName, const void *src, size_t count, size_t offset, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp']]], + ['hipmemgetaddressrange',['hipMemGetAddressRange',['../group__PeerToPeer.html#gac7d9132f6e3d102e9b512020e5654f38',1,'hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr): hip_memory.cpp'],['../group__PeerToPeer.html#gac7d9132f6e3d102e9b512020e5654f38',1,'hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr): hip_memory.cpp']]], + ['hipmemgetinfo',['hipMemGetInfo',['../group__Memory.html#ga311c3e246a21590de14478b8bd063be2',1,'hipMemGetInfo(size_t *free, size_t *total): hip_memory.cpp'],['../group__Memory.html#ga311c3e246a21590de14478b8bd063be2',1,'hipMemGetInfo(size_t *free, size_t *total): hip_memory.cpp']]], + ['hipmemset',['hipMemset',['../group__Memory.html#gac7441e74affcce4b8b69dba996c5ebc4',1,'hipMemset(void *dst, int value, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#gac7441e74affcce4b8b69dba996c5ebc4',1,'hipMemset(void *dst, int value, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemsetasync',['hipMemsetAsync',['../group__Memory.html#gae7d90e14c387e49f10db597f12915c54',1,'hipMemsetAsync(void *dst, int value, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gae7d90e14c387e49f10db597f12915c54',1,'hipMemsetAsync(void *dst, int value, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmodulegetfunction',['hipModuleGetFunction',['../group__Driver.html#ga9648b457bb837838cb936b417b56a65d',1,'hipModuleGetFunction(hipFunction_t *function, hipModule_t module, const char *kname): hip_module.cpp'],['../group__Driver.html#ga9648b457bb837838cb936b417b56a65d',1,'hipModuleGetFunction(hipFunction_t *hfunc, hipModule_t hmod, const char *name): hip_module.cpp']]], + ['hipmodulegetglobal',['hipModuleGetGlobal',['../group__Driver.html#ga3e425a680285f495e776f096e9632c89',1,'hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name): hip_module.cpp'],['../group__Driver.html#ga3e425a680285f495e776f096e9632c89',1,'hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name): hip_module.cpp']]], + ['hipmodulelaunchkernel',['hipModuleLaunchKernel',['../group__Driver.html#ga2e4de5937aa8171e9eda16c881ed0674',1,'hip_runtime_api.h']]], + ['hipmoduleload',['hipModuleLoad',['../group__Driver.html#ga31d806d976e91d36bd990ae3004d8760',1,'hipModuleLoad(hipModule_t *module, const char *fname): hip_module.cpp'],['../group__Driver.html#ga31d806d976e91d36bd990ae3004d8760',1,'hipModuleLoad(hipModule_t *module, const char *fname): hip_module.cpp']]], + ['hipmoduleloaddata',['hipModuleLoadData',['../group__Driver.html#gaabdbd73e952a741e861d01109c4790f3',1,'hipModuleLoadData(hipModule_t *module, const void *image): hip_module.cpp'],['../group__Driver.html#gaabdbd73e952a741e861d01109c4790f3',1,'hipModuleLoadData(hipModule_t *module, const void *image): hip_module.cpp']]], + ['hipmoduleunload',['hipModuleUnload',['../group__Driver.html#gae58e345f55bb3ec13dca80d2df88e0ed',1,'hipModuleUnload(hipModule_t module): hip_module.cpp'],['../group__Driver.html#gae58e345f55bb3ec13dca80d2df88e0ed',1,'hipModuleUnload(hipModule_t hmod): hip_module.cpp']]], + ['hippeekatlasterror',['hipPeekAtLastError',['../group__Error.html#ga1dd660bc739f7e13edd34615660f0148',1,'hipPeekAtLastError(void): hip_error.cpp'],['../group__Error.html#ga1dd660bc739f7e13edd34615660f0148',1,'hipPeekAtLastError(): hip_error.cpp']]], + ['hippointerattribute_5ft',['hipPointerAttribute_t',['../structhipPointerAttribute__t.html',1,'']]], + ['hippointergetattributes',['hipPointerGetAttributes',['../group__Memory.html#ga3d68ba64959615d4ab84f10caa12433b',1,'hipPointerGetAttributes(hipPointerAttribute_t *attributes, void *ptr): hip_memory.cpp'],['../group__Memory.html#ga3d68ba64959615d4ab84f10caa12433b',1,'hipPointerGetAttributes(hipPointerAttribute_t *attributes, void *ptr): hip_memory.cpp']]], + ['hipprofilerstart',['hipProfilerStart',['../group__Profiler.html#ga40aa20d731f9c8f0586127d589759e1d',1,'hipProfilerStart(): hip_hcc.cpp'],['../group__Profiler.html#ga40aa20d731f9c8f0586127d589759e1d',1,'hipProfilerStart(): hip_hcc.cpp']]], + ['hipprofilerstop',['hipProfilerStop',['../group__Profiler.html#ga777994cae0f7bdef24131fc1f3ed626e',1,'hipProfilerStop(): hip_hcc.cpp'],['../group__Profiler.html#ga777994cae0f7bdef24131fc1f3ed626e',1,'hipProfilerStop(): hip_hcc.cpp']]], + ['hipreadmodeelementtype',['hipReadModeElementType',['../hcc__detail_2hip__texture_8h.html#a442e950774f7306dc33692e358c92c94a829645801202174d052d667ffa4e1b8d',1,'hip_texture.h']]], + ['hipruntimegetversion',['hipRuntimeGetVersion',['../group__Driver.html#gae8b7ba34d2e11e334650aa51a4dd87ee',1,'hipRuntimeGetVersion(int *runtimeVersion): hip_context.cpp'],['../group__Driver.html#gae8b7ba34d2e11e334650aa51a4dd87ee',1,'hipRuntimeGetVersion(int *runtimeVersion): hip_context.cpp']]], + ['hipsetdevice',['hipSetDevice',['../group__Device.html#ga43c1e7f15925eeb762195ccb5e063eae',1,'hipSetDevice(int deviceId): hip_device.cpp'],['../group__Device.html#ga43c1e7f15925eeb762195ccb5e063eae',1,'hipSetDevice(int deviceId): hip_device.cpp']]], + ['hipsetdeviceflags',['hipSetDeviceFlags',['../group__Device.html#ga6e54db382768827e84725632018307aa',1,'hip_runtime_api.h']]], + ['hipsharedmembanksizedefault',['hipSharedMemBankSizeDefault',['../group__GlobalDefs.html#gga2e17b71d94ac350f2ccd914fd49d104eaf5b325c9b7bde878913f768eaba5014d',1,'hip_runtime_api.h']]], + ['hipsharedmembanksizeeightbyte',['hipSharedMemBankSizeEightByte',['../group__GlobalDefs.html#gga2e17b71d94ac350f2ccd914fd49d104ea64518b4f5a25f536c883330167e79258',1,'hip_runtime_api.h']]], + ['hipsharedmembanksizefourbyte',['hipSharedMemBankSizeFourByte',['../group__GlobalDefs.html#gga2e17b71d94ac350f2ccd914fd49d104ea0a95a6e0c33106c42d66ab9476ff954a',1,'hip_runtime_api.h']]], + ['hipsharedmemconfig',['hipSharedMemConfig',['../group__GlobalDefs.html#ga2e17b71d94ac350f2ccd914fd49d104e',1,'hipSharedMemConfig(): hip_runtime_api.h'],['../group__GlobalDefs.html#ga6b1ca424fa26a5fb718937d662eaee7f',1,'hipSharedMemConfig(): hip_runtime_api.h']]], + ['hipstreamaddcallback',['hipStreamAddCallback',['../group__Stream.html#ga3e098cd7478828b2104abb41a7bb00d3',1,'hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags): hip_stream.cpp'],['../group__Stream.html#ga3e098cd7478828b2104abb41a7bb00d3',1,'hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags): hip_stream.cpp']]], + ['hipstreamcallback_5ft',['hipStreamCallback_t',['../group__Stream.html#gadd389da2e3926a8ff17658d029015dcd',1,'hip_runtime_api.h']]], + ['hipstreamcreate',['hipStreamCreate',['../group__Stream.html#gaff5b62d6e9502d80879f7176f4d03102',1,'hipStreamCreate(hipStream_t *stream): hip_stream.cpp'],['../group__Stream.html#gaff5b62d6e9502d80879f7176f4d03102',1,'hipStreamCreate(hipStream_t *stream): hip_stream.cpp']]], + ['hipstreamcreatewithflags',['hipStreamCreateWithFlags',['../group__Stream.html#gaf2382e3cc6632332a8983a0f58e43494',1,'hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags): hip_stream.cpp'],['../group__Stream.html#gaf2382e3cc6632332a8983a0f58e43494',1,'hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags): hip_stream.cpp']]], + ['hipstreamdefault',['hipStreamDefault',['../group__GlobalDefs.html#ga6df5f70eb976836ab3598cacf0ffcdf9',1,'hip_runtime_api.h']]], + ['hipstreamdestroy',['hipStreamDestroy',['../group__Stream.html#ga3076a3499ed2c7821311006100bb95ec',1,'hipStreamDestroy(hipStream_t stream): hip_stream.cpp'],['../group__Stream.html#ga3076a3499ed2c7821311006100bb95ec',1,'hipStreamDestroy(hipStream_t stream): hip_stream.cpp']]], + ['hipstreamgetflags',['hipStreamGetFlags',['../group__Stream.html#ga3249555a26439591b8873f70b39bb116',1,'hipStreamGetFlags(hipStream_t stream, unsigned int *flags): hip_stream.cpp'],['../group__Stream.html#ga3249555a26439591b8873f70b39bb116',1,'hipStreamGetFlags(hipStream_t stream, unsigned int *flags): hip_stream.cpp']]], + ['hipstreamnonblocking',['hipStreamNonBlocking',['../group__GlobalDefs.html#gaaba9ae995d9b43b7d1ee70c6fa12c57d',1,'hip_runtime_api.h']]], + ['hipstreamquery',['hipStreamQuery',['../classihipStream__t.html#a46bba2e3bfbb0915a08fb608aed31858',1,'ihipStream_t::hipStreamQuery()'],['../group__Stream.html#ga925b39ff78d3b5fd458bd9e2cade9f4e',1,'hipStreamQuery(hipStream_t stream): hip_stream.cpp'],['../group__Stream.html#ga925b39ff78d3b5fd458bd9e2cade9f4e',1,'hipStreamQuery(hipStream_t stream): hip_stream.cpp']]], + ['hipstreamsynchronize',['hipStreamSynchronize',['../group__Stream.html#gabbfb9f573a6ebe8c478605ecb5504a74',1,'hipStreamSynchronize(hipStream_t stream): hip_stream.cpp'],['../group__Stream.html#gabbfb9f573a6ebe8c478605ecb5504a74',1,'hipStreamSynchronize(hipStream_t stream): hip_stream.cpp']]], + ['hipstreamwaitevent',['hipStreamWaitEvent',['../group__Stream.html#gacdd84c8f8ef1539c96c57c1d5bcae633',1,'hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags): hip_stream.cpp'],['../group__Stream.html#gacdd84c8f8ef1539c96c57c1d5bcae633',1,'hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags): hip_stream.cpp']]], + ['hipsuccess',['hipSuccess',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594aadfbdb847b149723c684ebd764556063',1,'hip_runtime_api.h']]], + ['hiptexturefiltermode',['hipTextureFilterMode',['../hcc__detail_2hip__texture_8h.html#aa2f0b6002b81d0a43a808cb880bb21e6',1,'hip_texture.h']]], + ['hiptexturereadmode',['hipTextureReadMode',['../hcc__detail_2hip__texture_8h.html#a442e950774f7306dc33692e358c92c94',1,'hip_texture.h']]], + ['host_5fdefines_2eh',['host_defines.h',['../host__defines_8h.html',1,'']]], + ['heterogeneous_2dcomputing_20interface_20for_20portability_20_28hip_29',['Heterogeneous-computing Interface for Portability (HIP)',['../index.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_9.html b/docs/RuntimeAPI/html/search/all_9.html new file mode 100644 index 0000000000..cd46d44056 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_9.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_9.js b/docs/RuntimeAPI/html/search/all_9.js new file mode 100644 index 0000000000..33821caa67 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_9.js @@ -0,0 +1,19 @@ +var searchData= +[ + ['initialization_20and_20version',['Initialization and Version',['../group__Driver.html',1,'']]], + ['ihipctx_5ft',['ihipCtx_t',['../classihipCtx__t.html',1,'']]], + ['ihipctxcriticalbase_5ft',['ihipCtxCriticalBase_t',['../classihipCtxCriticalBase__t.html',1,'']]], + ['ihipctxcriticalbase_5ft_3c_20ctxmutex_20_3e',['ihipCtxCriticalBase_t< CtxMutex >',['../classihipCtxCriticalBase__t.html',1,'']]], + ['ihipdevice_5ft',['ihipDevice_t',['../classihipDevice__t.html',1,'']]], + ['ihipevent_5ft',['ihipEvent_t',['../structihipEvent__t.html',1,'']]], + ['ihipexception',['ihipException',['../classihipException.html',1,'']]], + ['ihipfunction_5ft',['ihipFunction_t',['../classihipFunction__t.html',1,'']]], + ['ihipipcmemhandle_5ft',['ihipIpcMemHandle_t',['../classihipIpcMemHandle__t.html',1,'']]], + ['ihipmodule_5ft',['ihipModule_t',['../classihipModule__t.html',1,'']]], + ['ihipstream_5ft',['ihipStream_t',['../classihipStream__t.html',1,'']]], + ['ihipstreamcriticalbase_5ft',['ihipStreamCriticalBase_t',['../classihipStreamCriticalBase__t.html',1,'']]], + ['ihipstreamcriticalbase_5ft_3c_20streammutex_20_3e',['ihipStreamCriticalBase_t< StreamMutex >',['../classihipStreamCriticalBase__t.html',1,'']]], + ['int3',['int3',['../structint3.html',1,'']]], + ['ipc_5fhandle',['ipc_handle',['../classihipIpcMemHandle__t.html#af2142ab7d9f820acbad7638428509d42',1,'ihipIpcMemHandle_t']]], + ['ismultigpuboard',['isMultiGpuBoard',['../structhipDeviceProp__t.html#a9bb19b2b0cdee8977ed63964532d639d',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_a.html b/docs/RuntimeAPI/html/search/all_a.html new file mode 100644 index 0000000000..eab655301b --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_a.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_a.js b/docs/RuntimeAPI/html/search/all_a.js new file mode 100644 index 0000000000..8be519d6d6 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_a.js @@ -0,0 +1,10 @@ +var searchData= +[ + ['l2cachesize',['l2CacheSize',['../structhipDeviceProp__t.html#a24404decccc16833973c803ced6f3a51',1,'hipDeviceProp_t']]], + ['lockedaccessor',['LockedAccessor',['../classLockedAccessor.html',1,'']]], + ['lockedbase',['LockedBase',['../structLockedBase.html',1,'']]], + ['lockedbase_3c_20ctxmutex_20_3e',['LockedBase< CtxMutex >',['../structLockedBase.html',1,'']]], + ['lockedbase_3c_20streammutex_20_3e',['LockedBase< StreamMutex >',['../structLockedBase.html',1,'']]], + ['long3',['long3',['../structlong3.html',1,'']]], + ['longlong3',['longlong3',['../structlonglong3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_b.html b/docs/RuntimeAPI/html/search/all_b.html new file mode 100644 index 0000000000..a2c161e0bd --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_b.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_b.js b/docs/RuntimeAPI/html/search/all_b.js new file mode 100644 index 0000000000..462c78d2d5 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_b.js @@ -0,0 +1,15 @@ +var searchData= +[ + ['management',['Management',['../group__Context.html',1,'']]], + ['major',['major',['../structhipDeviceProp__t.html#aec9e4173c2e34cc232300c415dbd5e4f',1,'hipDeviceProp_t']]], + ['maxgridsize',['maxGridSize',['../structhipDeviceProp__t.html#ae529c23929f592120081fed31d877a55',1,'hipDeviceProp_t']]], + ['maxsharedmemorypermultiprocessor',['maxSharedMemoryPerMultiProcessor',['../structhipDeviceProp__t.html#aa1a32a7f387f6da845db7b228711fce8',1,'hipDeviceProp_t']]], + ['maxthreadsdim',['maxThreadsDim',['../structhipDeviceProp__t.html#a8ebba6fc12f80c9a9cf9b9193f0da465',1,'hipDeviceProp_t']]], + ['maxthreadsperblock',['maxThreadsPerBlock',['../structhipDeviceProp__t.html#af971cf1ca3ec1f68ad09036c0cc672e0',1,'hipDeviceProp_t']]], + ['maxthreadspermultiprocessor',['maxThreadsPerMultiProcessor',['../structhipDeviceProp__t.html#a23a39f4fd795addb3b125e9c3f6295ea',1,'hipDeviceProp_t']]], + ['memory_20management',['Memory Management',['../group__Memory.html',1,'']]], + ['memorybuswidth',['memoryBusWidth',['../structhipDeviceProp__t.html#a1d710ce8bd34daf66f892a66a2273132',1,'hipDeviceProp_t']]], + ['memoryclockrate',['memoryClockRate',['../structhipDeviceProp__t.html#a6db0ab8e7e8cc13c84d7bb7f70226d5e',1,'hipDeviceProp_t']]], + ['minor',['minor',['../structhipDeviceProp__t.html#abb51208e2509a7a1d107f0da69108938',1,'hipDeviceProp_t']]], + ['multiprocessorcount',['multiProcessorCount',['../structhipDeviceProp__t.html#add8d9d2ad52aece9fd1dbe25c18d9d57',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_c.html b/docs/RuntimeAPI/html/search/all_c.html new file mode 100644 index 0000000000..bdd3ee2c50 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_c.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_c.js b/docs/RuntimeAPI/html/search/all_c.js new file mode 100644 index 0000000000..124bf0ddb8 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_c.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['name',['name',['../structhipDeviceProp__t.html#a5b44bf8fa46faefcde989942b1d11a5e',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_d.html b/docs/RuntimeAPI/html/search/all_d.html new file mode 100644 index 0000000000..d51093367e --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_d.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_d.js b/docs/RuntimeAPI/html/search/all_d.js new file mode 100644 index 0000000000..0584a45193 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_d.js @@ -0,0 +1,6 @@ +var searchData= +[ + ['pcibusid',['pciBusID',['../structhipDeviceProp__t.html#a1350f64d49b717ed3a06458f7549ccb0',1,'hipDeviceProp_t']]], + ['pcideviceid',['pciDeviceID',['../structhipDeviceProp__t.html#ae6aa845dc2d540f85098ea30be35f4eb',1,'hipDeviceProp_t']]], + ['proftrigger',['ProfTrigger',['../structProfTrigger.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_e.html b/docs/RuntimeAPI/html/search/all_e.html new file mode 100644 index 0000000000..3cda0172ed --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_e.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_e.js b/docs/RuntimeAPI/html/search/all_e.js new file mode 100644 index 0000000000..44ba50e0b7 --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_e.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['regsperblock',['regsPerBlock',['../structhipDeviceProp__t.html#a73c1c21648a901799ff6bef83c11135b',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/all_f.html b/docs/RuntimeAPI/html/search/all_f.html new file mode 100644 index 0000000000..7419b0297e --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_f.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/all_f.js b/docs/RuntimeAPI/html/search/all_f.js new file mode 100644 index 0000000000..0656e6726f --- /dev/null +++ b/docs/RuntimeAPI/html/search/all_f.js @@ -0,0 +1,8 @@ +var searchData= +[ + ['sharedmemperblock',['sharedMemPerBlock',['../structhipDeviceProp__t.html#a3b9138678a0795c2677eddcfb1c67156',1,'hipDeviceProp_t']]], + ['short3',['short3',['../structshort3.html',1,'']]], + ['shorttid',['ShortTid',['../classShortTid.html',1,'']]], + ['stream_20management',['Stream Management',['../group__Stream.html',1,'']]], + ['struct_5ffloat',['struct_float',['../structstruct__float.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_0.html b/docs/RuntimeAPI/html/search/classes_0.html new file mode 100644 index 0000000000..fabdc03679 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_0.js b/docs/RuntimeAPI/html/search/classes_0.js new file mode 100644 index 0000000000..f1a5d5e882 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['_5f_5fhalf',['__half',['../struct____half.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_1.html b/docs/RuntimeAPI/html/search/classes_1.html new file mode 100644 index 0000000000..800b6ae1bf --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_1.js b/docs/RuntimeAPI/html/search/classes_1.js new file mode 100644 index 0000000000..31c2da2c30 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_1.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['char3',['char3',['../structchar3.html',1,'']]], + ['char4',['char4',['../structchar4.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_2.html b/docs/RuntimeAPI/html/search/classes_2.html new file mode 100644 index 0000000000..f65d263c0c --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_2.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_2.js b/docs/RuntimeAPI/html/search/classes_2.js new file mode 100644 index 0000000000..f0c767feb0 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_2.js @@ -0,0 +1,7 @@ +var searchData= +[ + ['dbname',['DbName',['../structDbName.html',1,'']]], + ['dim3',['dim3',['../structdim3.html',1,'']]], + ['double1',['double1',['../structdouble1.html',1,'']]], + ['double3',['double3',['../structdouble3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_3.html b/docs/RuntimeAPI/html/search/classes_3.html new file mode 100644 index 0000000000..927e837f11 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_3.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_3.js b/docs/RuntimeAPI/html/search/classes_3.js new file mode 100644 index 0000000000..5149e86e09 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_3.js @@ -0,0 +1,6 @@ +var searchData= +[ + ['fakemutex',['FakeMutex',['../classFakeMutex.html',1,'']]], + ['float1',['float1',['../structfloat1.html',1,'']]], + ['float3',['float3',['../structfloat3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_4.html b/docs/RuntimeAPI/html/search/classes_4.html new file mode 100644 index 0000000000..a447290ed6 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_4.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_4.js b/docs/RuntimeAPI/html/search/classes_4.js new file mode 100644 index 0000000000..059435714b --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_4.js @@ -0,0 +1,10 @@ +var searchData= +[ + ['hiparray',['hipArray',['../structhipArray.html',1,'']]], + ['hipchannelformatdesc',['hipChannelFormatDesc',['../structhipChannelFormatDesc.html',1,'']]], + ['hipdevicearch_5ft',['hipDeviceArch_t',['../structhipDeviceArch__t.html',1,'']]], + ['hipdeviceprop_5ft',['hipDeviceProp_t',['../structhipDeviceProp__t.html',1,'']]], + ['hipdoublecomplex',['hipDoubleComplex',['../structhipDoubleComplex.html',1,'']]], + ['hipfloatcomplex',['hipFloatComplex',['../structhipFloatComplex.html',1,'']]], + ['hippointerattribute_5ft',['hipPointerAttribute_t',['../structhipPointerAttribute__t.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_5.html b/docs/RuntimeAPI/html/search/classes_5.html new file mode 100644 index 0000000000..63484c451e --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_5.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_5.js b/docs/RuntimeAPI/html/search/classes_5.js new file mode 100644 index 0000000000..5a5c313721 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_5.js @@ -0,0 +1,16 @@ +var searchData= +[ + ['ihipctx_5ft',['ihipCtx_t',['../classihipCtx__t.html',1,'']]], + ['ihipctxcriticalbase_5ft',['ihipCtxCriticalBase_t',['../classihipCtxCriticalBase__t.html',1,'']]], + ['ihipctxcriticalbase_5ft_3c_20ctxmutex_20_3e',['ihipCtxCriticalBase_t< CtxMutex >',['../classihipCtxCriticalBase__t.html',1,'']]], + ['ihipdevice_5ft',['ihipDevice_t',['../classihipDevice__t.html',1,'']]], + ['ihipevent_5ft',['ihipEvent_t',['../structihipEvent__t.html',1,'']]], + ['ihipexception',['ihipException',['../classihipException.html',1,'']]], + ['ihipfunction_5ft',['ihipFunction_t',['../classihipFunction__t.html',1,'']]], + ['ihipipcmemhandle_5ft',['ihipIpcMemHandle_t',['../classihipIpcMemHandle__t.html',1,'']]], + ['ihipmodule_5ft',['ihipModule_t',['../classihipModule__t.html',1,'']]], + ['ihipstream_5ft',['ihipStream_t',['../classihipStream__t.html',1,'']]], + ['ihipstreamcriticalbase_5ft',['ihipStreamCriticalBase_t',['../classihipStreamCriticalBase__t.html',1,'']]], + ['ihipstreamcriticalbase_5ft_3c_20streammutex_20_3e',['ihipStreamCriticalBase_t< StreamMutex >',['../classihipStreamCriticalBase__t.html',1,'']]], + ['int3',['int3',['../structint3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_6.html b/docs/RuntimeAPI/html/search/classes_6.html new file mode 100644 index 0000000000..f3d70354c6 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_6.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_6.js b/docs/RuntimeAPI/html/search/classes_6.js new file mode 100644 index 0000000000..ddc2606e32 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_6.js @@ -0,0 +1,9 @@ +var searchData= +[ + ['lockedaccessor',['LockedAccessor',['../classLockedAccessor.html',1,'']]], + ['lockedbase',['LockedBase',['../structLockedBase.html',1,'']]], + ['lockedbase_3c_20ctxmutex_20_3e',['LockedBase< CtxMutex >',['../structLockedBase.html',1,'']]], + ['lockedbase_3c_20streammutex_20_3e',['LockedBase< StreamMutex >',['../structLockedBase.html',1,'']]], + ['long3',['long3',['../structlong3.html',1,'']]], + ['longlong3',['longlong3',['../structlonglong3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_7.html b/docs/RuntimeAPI/html/search/classes_7.html new file mode 100644 index 0000000000..9e5f5c9861 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_7.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_7.js b/docs/RuntimeAPI/html/search/classes_7.js new file mode 100644 index 0000000000..750e29e1e6 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_7.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['proftrigger',['ProfTrigger',['../structProfTrigger.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_8.html b/docs/RuntimeAPI/html/search/classes_8.html new file mode 100644 index 0000000000..82c35b32ef --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_8.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_8.js b/docs/RuntimeAPI/html/search/classes_8.js new file mode 100644 index 0000000000..8440e320e4 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_8.js @@ -0,0 +1,6 @@ +var searchData= +[ + ['short3',['short3',['../structshort3.html',1,'']]], + ['shorttid',['ShortTid',['../classShortTid.html',1,'']]], + ['struct_5ffloat',['struct_float',['../structstruct__float.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_9.html b/docs/RuntimeAPI/html/search/classes_9.html new file mode 100644 index 0000000000..4e83ac829b --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_9.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_9.js b/docs/RuntimeAPI/html/search/classes_9.js new file mode 100644 index 0000000000..8ced671759 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_9.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['texturereference',['textureReference',['../structtextureReference.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/classes_a.html b/docs/RuntimeAPI/html/search/classes_a.html new file mode 100644 index 0000000000..616feb69a2 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_a.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/classes_a.js b/docs/RuntimeAPI/html/search/classes_a.js new file mode 100644 index 0000000000..e1eb527d76 --- /dev/null +++ b/docs/RuntimeAPI/html/search/classes_a.js @@ -0,0 +1,9 @@ +var searchData= +[ + ['uchar3',['uchar3',['../structuchar3.html',1,'']]], + ['uchar4',['uchar4',['../structuchar4.html',1,'']]], + ['uint3',['uint3',['../structuint3.html',1,'']]], + ['ulong3',['ulong3',['../structulong3.html',1,'']]], + ['ulonglong3',['ulonglong3',['../structulonglong3.html',1,'']]], + ['ushort3',['ushort3',['../structushort3.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/close.png b/docs/RuntimeAPI/html/search/close.png new file mode 100644 index 0000000000000000000000000000000000000000..9342d3dfeea7b7c4ee610987e717804b5a42ceb9 GIT binary patch literal 273 zcmV+s0q*{ZP)4(RlMby96)VwnbG{ zbe&}^BDn7x>$<{ck4zAK-=nT;=hHG)kmplIF${xqm8db3oX6wT3bvp`TE@m0cg;b) zBuSL}5?N7O(iZLdAlz@)b)Rd~DnSsSX&P5qC`XwuFwcAYLC+d2>+1(8on;wpt8QIC X2MT$R4iQDd00000NkvXXu0mjfia~GN literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/search/defines_0.html b/docs/RuntimeAPI/html/search/defines_0.html new file mode 100644 index 0000000000..c2f0b46cfb --- /dev/null +++ b/docs/RuntimeAPI/html/search/defines_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/defines_0.js b/docs/RuntimeAPI/html/search/defines_0.js new file mode 100644 index 0000000000..00fe08ccff --- /dev/null +++ b/docs/RuntimeAPI/html/search/defines_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['_5f_5fhost_5f_5f',['__host__',['../host__defines_8h.html#a803050db3c78e0db3ea59a0c35499622',1,'host_defines.h']]] +]; diff --git a/docs/RuntimeAPI/html/search/defines_1.html b/docs/RuntimeAPI/html/search/defines_1.html new file mode 100644 index 0000000000..9e1d9fa116 --- /dev/null +++ b/docs/RuntimeAPI/html/search/defines_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/defines_1.js b/docs/RuntimeAPI/html/search/defines_1.js new file mode 100644 index 0000000000..b74e8321a7 --- /dev/null +++ b/docs/RuntimeAPI/html/search/defines_1.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['hip_5fdynamic_5fshared',['HIP_DYNAMIC_SHARED',['../hcc__detail_2hip__runtime_8h.html#a1e7ec14fb6b74c03b6ec804d6ef7e0ea',1,'hip_runtime.h']]] +]; diff --git a/docs/RuntimeAPI/html/search/enums_0.html b/docs/RuntimeAPI/html/search/enums_0.html new file mode 100644 index 0000000000..b4cbe1e3ca --- /dev/null +++ b/docs/RuntimeAPI/html/search/enums_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/enums_0.js b/docs/RuntimeAPI/html/search/enums_0.js new file mode 100644 index 0000000000..bce6de432f --- /dev/null +++ b/docs/RuntimeAPI/html/search/enums_0.js @@ -0,0 +1,10 @@ +var searchData= +[ + ['hipdeviceattribute_5ft',['hipDeviceAttribute_t',['../group__GlobalDefs.html#gacc0acd7b9bda126c6bb3dfd6e2796d7c',1,'hip_runtime_api.h']]], + ['hiperror_5ft',['hipError_t',['../group__GlobalDefs.html#gadf5010f6e140a53ecbdf949e73e87594',1,'hip_runtime_api.h']]], + ['hipfunccache_5ft',['hipFuncCache_t',['../group__GlobalDefs.html#ga86e5c5692645963a9a673e1aa88ba6ca',1,'hip_runtime_api.h']]], + ['hipmemcpykind',['hipMemcpyKind',['../group__GlobalDefs.html#ga232e222db36b1fc672ba98054d036a18',1,'hip_runtime_api.h']]], + ['hipsharedmemconfig',['hipSharedMemConfig',['../group__GlobalDefs.html#ga2e17b71d94ac350f2ccd914fd49d104e',1,'hip_runtime_api.h']]], + ['hiptexturefiltermode',['hipTextureFilterMode',['../hcc__detail_2hip__texture_8h.html#aa2f0b6002b81d0a43a808cb880bb21e6',1,'hip_texture.h']]], + ['hiptexturereadmode',['hipTextureReadMode',['../hcc__detail_2hip__texture_8h.html#a442e950774f7306dc33692e358c92c94',1,'hip_texture.h']]] +]; diff --git a/docs/RuntimeAPI/html/search/enumvalues_0.html b/docs/RuntimeAPI/html/search/enumvalues_0.html new file mode 100644 index 0000000000..3e00fcf61a --- /dev/null +++ b/docs/RuntimeAPI/html/search/enumvalues_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/enumvalues_0.js b/docs/RuntimeAPI/html/search/enumvalues_0.js new file mode 100644 index 0000000000..55b989c53e --- /dev/null +++ b/docs/RuntimeAPI/html/search/enumvalues_0.js @@ -0,0 +1,65 @@ +var searchData= +[ + ['hipdeviceattributeclockrate',['hipDeviceAttributeClockRate',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca2300e077e020e7967592065561373b00',1,'hip_runtime_api.h']]], + ['hipdeviceattributecomputecapabilitymajor',['hipDeviceAttributeComputeCapabilityMajor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca2735739cf977b7d303266f6781131e8d',1,'hip_runtime_api.h']]], + ['hipdeviceattributecomputecapabilityminor',['hipDeviceAttributeComputeCapabilityMinor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca38edc4fcae456e47160d349da3249b85',1,'hip_runtime_api.h']]], + ['hipdeviceattributecomputemode',['hipDeviceAttributeComputeMode',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca4d0369a6ef7bd7890fdcabc16ed3385d',1,'hip_runtime_api.h']]], + ['hipdeviceattributeconcurrentkernels',['hipDeviceAttributeConcurrentKernels',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cad9f45254d0d048677f560032532d5504',1,'hip_runtime_api.h']]], + ['hipdeviceattributeismultigpuboard',['hipDeviceAttributeIsMultiGpuBoard',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca6129311200a17dcc5fa8d2256874ae3d',1,'hip_runtime_api.h']]], + ['hipdeviceattributel2cachesize',['hipDeviceAttributeL2CacheSize',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca582ae5a26a7148504878890028e4b64c',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxblockdimx',['hipDeviceAttributeMaxBlockDimX',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cac1e4ac589db0d8adbbc241e3d0fcd594',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxblockdimy',['hipDeviceAttributeMaxBlockDimY',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca187dbffe12db09a56c0f75c340d879c9',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxblockdimz',['hipDeviceAttributeMaxBlockDimZ',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7caf811f51e03d1ffb025d80ac1da088675',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxgriddimx',['hipDeviceAttributeMaxGridDimX',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca03db8df0e7a9fbdaae683d97e8ac9c87',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxgriddimy',['hipDeviceAttributeMaxGridDimY',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca5b5cc49972679c5ccf62b79425ee99df',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxgriddimz',['hipDeviceAttributeMaxGridDimZ',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca6c206ac083999caf4640e5d91dae24f7',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxregistersperblock',['hipDeviceAttributeMaxRegistersPerBlock',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca82289b170192b6ea742be0efc6f95107',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxsharedmemoryperblock',['hipDeviceAttributeMaxSharedMemoryPerBlock',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca7bca3aa18b26d40eba043ae93e15c7e5',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxsharedmemorypermultiprocessor',['hipDeviceAttributeMaxSharedMemoryPerMultiprocessor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cad3e7f3d01533b32e12211172fcf410ba',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxthreadsperblock',['hipDeviceAttributeMaxThreadsPerBlock',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca8327aa23782d9c994bdef33a6d62e02e',1,'hip_runtime_api.h']]], + ['hipdeviceattributemaxthreadspermultiprocessor',['hipDeviceAttributeMaxThreadsPerMultiProcessor',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7caddc08922b491eb1f6a583833cbf4e2f0',1,'hip_runtime_api.h']]], + ['hipdeviceattributememorybuswidth',['hipDeviceAttributeMemoryBusWidth',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca848c1396fab6f20463c6aefb828b0870',1,'hip_runtime_api.h']]], + ['hipdeviceattributememoryclockrate',['hipDeviceAttributeMemoryClockRate',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca6b68deafd65f036b30dc8051573eb000',1,'hip_runtime_api.h']]], + ['hipdeviceattributemultiprocessorcount',['hipDeviceAttributeMultiprocessorCount',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca5c1519870733ccf0b83f722678240e5f',1,'hip_runtime_api.h']]], + ['hipdeviceattributepcibusid',['hipDeviceAttributePciBusId',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca572b29c44f1322aa7657fdd784832f88',1,'hip_runtime_api.h']]], + ['hipdeviceattributepcideviceid',['hipDeviceAttributePciDeviceId',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7ca955d90286e87be9e3528f0b817ab32ff',1,'hip_runtime_api.h']]], + ['hipdeviceattributetotalconstantmemory',['hipDeviceAttributeTotalConstantMemory',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7cac6089ac3a0f9c77cc382fb0eaa73ae9c',1,'hip_runtime_api.h']]], + ['hipdeviceattributewarpsize',['hipDeviceAttributeWarpSize',['../group__GlobalDefs.html#ggacc0acd7b9bda126c6bb3dfd6e2796d7caffd94133e823247a6f1215343232f6ec',1,'hip_runtime_api.h']]], + ['hiperrorhostmemoryalreadyregistered',['hipErrorHostMemoryAlreadyRegistered',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a9d7173cea72aace620a83d502569de1b',1,'hip_runtime_api.h']]], + ['hiperrorhostmemorynotregistered',['hipErrorHostMemoryNotRegistered',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a6901476ca88eed786fb8be003d9661d9',1,'hip_runtime_api.h']]], + ['hiperrorinitializationerror',['hipErrorInitializationError',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a7e935ae88ee1f9ff3920156ac6864520',1,'hip_runtime_api.h']]], + ['hiperrorinvalidcontext',['hipErrorInvalidContext',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ad9d895f7f4dbf63a4586b2aab593c9fe',1,'hip_runtime_api.h']]], + ['hiperrorinvaliddevice',['hipErrorInvalidDevice',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a07ab9b704ea693c1781a52741c60cd0d',1,'hip_runtime_api.h']]], + ['hiperrorinvaliddevicepointer',['hipErrorInvalidDevicePointer',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a37a93fcd2b0aed9bf52b82fa26031e6f',1,'hip_runtime_api.h']]], + ['hiperrorinvalidkernelfile',['hipErrorInvalidKernelFile',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ae508537d5e9a293bb1e63802d4a61380',1,'hip_runtime_api.h']]], + ['hiperrorinvalidmemcpydirection',['hipErrorInvalidMemcpyDirection',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ae35ad9fb66b7c33f21a021efe89975b0',1,'hip_runtime_api.h']]], + ['hiperrorinvalidresourcehandle',['hipErrorInvalidResourceHandle',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a88e525a7c8f35552dfada58e9f2f6d3a',1,'hip_runtime_api.h']]], + ['hiperrorinvalidvalue',['hipErrorInvalidValue',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a1e8215fe1108a508bad3944bce7b4d83',1,'hip_runtime_api.h']]], + ['hiperrorlaunchfailure',['hipErrorLaunchFailure',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a3df72146e762c5a901d3af06027a1de9',1,'hip_runtime_api.h']]], + ['hiperrorlaunchoutofresources',['hipErrorLaunchOutOfResources',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a3ea57f211997a84fa1c582414bcabc47',1,'hip_runtime_api.h']]], + ['hiperrormapbufferobjectfailed',['hipErrorMapBufferObjectFailed',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ab4de2174991cc16a3e1986f5da25262b',1,'hip_runtime_api.h']]], + ['hiperrormemoryallocation',['hipErrorMemoryAllocation',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a8293288a10109874749afe2562db09f2',1,'hip_runtime_api.h']]], + ['hiperrornodevice',['hipErrorNoDevice',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ad4406972c318df36d231310a15131c24',1,'hip_runtime_api.h']]], + ['hiperrornotready',['hipErrorNotReady',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594aa9638063c8746a9d1fda2b2069a0a9f1',1,'hip_runtime_api.h']]], + ['hiperrorpeeraccessalreadyenabled',['hipErrorPeerAccessAlreadyEnabled',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a5399c146f91246f8b77abcd4ea30e7ac',1,'hip_runtime_api.h']]], + ['hiperrorpeeraccessnotenabled',['hipErrorPeerAccessNotEnabled',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a2ee0bf2e64840f253e4a1b12bbaf2d09',1,'hip_runtime_api.h']]], + ['hiperrorruntimememory',['hipErrorRuntimeMemory',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a1159eb9a6be66bea740a8bfb61378723',1,'hip_runtime_api.h']]], + ['hiperrorruntimeother',['hipErrorRuntimeOther',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594a44f847c5914be2177feac107dcb096d1',1,'hip_runtime_api.h']]], + ['hiperrortbd',['hipErrorTbd',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594ab556409e11ddb0c4cf77a2f4fc91ea9e',1,'hip_runtime_api.h']]], + ['hiperrorunknown',['hipErrorUnknown',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594aa74e64c5b2f5fb0d6a92681f5b234073',1,'hip_runtime_api.h']]], + ['hipfiltermodepoint',['hipFilterModePoint',['../hcc__detail_2hip__texture_8h.html#aa2f0b6002b81d0a43a808cb880bb21e6a56ede038ab7c805ec4b5b61d2b678dfc',1,'hip_texture.h']]], + ['hipfunccachepreferequal',['hipFuncCachePreferEqual',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa0ddab0e840107634a152033103be44d7',1,'hip_runtime_api.h']]], + ['hipfunccachepreferl1',['hipFuncCachePreferL1',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa636a3c140db6b9d4a8bf7d5a61c398c5',1,'hip_runtime_api.h']]], + ['hipfunccacheprefernone',['hipFuncCachePreferNone',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa0813fbaa008ce1231ff9fed3911eb3af',1,'hip_runtime_api.h']]], + ['hipfunccacheprefershared',['hipFuncCachePreferShared',['../group__GlobalDefs.html#gga86e5c5692645963a9a673e1aa88ba6caa9b34337dfbadba25ed2aa270bbcabc43',1,'hip_runtime_api.h']]], + ['hipmemcpydefault',['hipMemcpyDefault',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18a4e37107e416f79a2edf2b6534163c823',1,'hip_runtime_api.h']]], + ['hipmemcpydevicetodevice',['hipMemcpyDeviceToDevice',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18abd05a09d3105e0ce25b34dd91cf83f88',1,'hip_runtime_api.h']]], + ['hipmemcpydevicetohost',['hipMemcpyDeviceToHost',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18aba2505e9ce1e5382f17730bc670917d1',1,'hip_runtime_api.h']]], + ['hipmemcpyhosttodevice',['hipMemcpyHostToDevice',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18aff32175ecb0c7113200286eff8211008',1,'hip_runtime_api.h']]], + ['hipmemcpyhosttohost',['hipMemcpyHostToHost',['../group__GlobalDefs.html#gga232e222db36b1fc672ba98054d036a18a9d66b705aa85a9c83f0f533cef70d0af',1,'hip_runtime_api.h']]], + ['hipreadmodeelementtype',['hipReadModeElementType',['../hcc__detail_2hip__texture_8h.html#a442e950774f7306dc33692e358c92c94a829645801202174d052d667ffa4e1b8d',1,'hip_texture.h']]], + ['hipsharedmembanksizedefault',['hipSharedMemBankSizeDefault',['../group__GlobalDefs.html#gga2e17b71d94ac350f2ccd914fd49d104eaf5b325c9b7bde878913f768eaba5014d',1,'hip_runtime_api.h']]], + ['hipsharedmembanksizeeightbyte',['hipSharedMemBankSizeEightByte',['../group__GlobalDefs.html#gga2e17b71d94ac350f2ccd914fd49d104ea64518b4f5a25f536c883330167e79258',1,'hip_runtime_api.h']]], + ['hipsharedmembanksizefourbyte',['hipSharedMemBankSizeFourByte',['../group__GlobalDefs.html#gga2e17b71d94ac350f2ccd914fd49d104ea0a95a6e0c33106c42d66ab9476ff954a',1,'hip_runtime_api.h']]], + ['hipsuccess',['hipSuccess',['../group__GlobalDefs.html#ggadf5010f6e140a53ecbdf949e73e87594aadfbdb847b149723c684ebd764556063',1,'hip_runtime_api.h']]] +]; diff --git a/docs/RuntimeAPI/html/search/files_0.html b/docs/RuntimeAPI/html/search/files_0.html new file mode 100644 index 0000000000..c7aa36c96f --- /dev/null +++ b/docs/RuntimeAPI/html/search/files_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/files_0.js b/docs/RuntimeAPI/html/search/files_0.js new file mode 100644 index 0000000000..31d22e4c27 --- /dev/null +++ b/docs/RuntimeAPI/html/search/files_0.js @@ -0,0 +1,9 @@ +var searchData= +[ + ['hip_5fhcc_2ecpp',['hip_hcc.cpp',['../hip__hcc_8cpp.html',1,'']]], + ['hip_5fruntime_2eh',['hip_runtime.h',['../hcc__detail_2hip__runtime_8h.html',1,'']]], + ['hip_5fruntime_5fapi_2eh',['hip_runtime_api.h',['../hcc__detail_2hip__runtime__api_8h.html',1,'']]], + ['hip_5ftexture_2eh',['hip_texture.h',['../hcc__detail_2hip__texture_8h.html',1,'']]], + ['hip_5fvector_5ftypes_2eh',['hip_vector_types.h',['../hcc__detail_2hip__vector__types_8h.html',1,'']]], + ['host_5fdefines_2eh',['host_defines.h',['../host__defines_8h.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/functions_0.html b/docs/RuntimeAPI/html/search/functions_0.html new file mode 100644 index 0000000000..16a5a527a8 --- /dev/null +++ b/docs/RuntimeAPI/html/search/functions_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/functions_0.js b/docs/RuntimeAPI/html/search/functions_0.js new file mode 100644 index 0000000000..81e400d4c8 --- /dev/null +++ b/docs/RuntimeAPI/html/search/functions_0.js @@ -0,0 +1,6 @@ +var searchData= +[ + ['_5f_5fthreadfence',['__threadfence',['../group__Fence.html#ga8ed093b29d5c05d46a1240e7a35a6d13',1,'hip_runtime.h']]], + ['_5f_5fthreadfence_5fblock',['__threadfence_block',['../group__Fence.html#ga675c8823d497b2fd91abc42085e5ae25',1,'hip_runtime.h']]], + ['_5f_5fthreadfence_5fsystem',['__threadfence_system',['../group__Fence.html#ga21e7328a2daaf9e376d10b5c6b3de0ae',1,'__threadfence_system(void): device_util.cpp'],['../group__Fence.html#ga21e7328a2daaf9e376d10b5c6b3de0ae',1,'__threadfence_system(void): device_util.cpp']]] +]; diff --git a/docs/RuntimeAPI/html/search/functions_1.html b/docs/RuntimeAPI/html/search/functions_1.html new file mode 100644 index 0000000000..3b4eacfebf --- /dev/null +++ b/docs/RuntimeAPI/html/search/functions_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/functions_1.js b/docs/RuntimeAPI/html/search/functions_1.js new file mode 100644 index 0000000000..d70b5aa8c0 --- /dev/null +++ b/docs/RuntimeAPI/html/search/functions_1.js @@ -0,0 +1,111 @@ +var searchData= +[ + ['hipchoosedevice',['hipChooseDevice',['../group__Device.html#gaf1e365e1d17cf40644d1470de4817c8e',1,'hipChooseDevice(int *device, const hipDeviceProp_t *prop): hip_device.cpp'],['../group__Device.html#gaf1e365e1d17cf40644d1470de4817c8e',1,'hipChooseDevice(int *device, const hipDeviceProp_t *prop): hip_device.cpp']]], + ['hipcreatechanneldesc',['hipCreateChannelDesc',['../group__Texture.html#gac9b0ae733fa849842342fab6d42c7b6f',1,'hipCreateChannelDesc(int x, int y, int z, int w, hipChannelFormatKind f): hip_memory.cpp'],['../group__Texture.html#gac9b0ae733fa849842342fab6d42c7b6f',1,'hipCreateChannelDesc(int x, int y, int z, int w, hipChannelFormatKind f): hip_memory.cpp']]], + ['hipctxcreate',['hipCtxCreate',['../group__Context.html#gab6dbcff5c5b1249a5ac5cf39ae9d08bc',1,'hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device): hip_context.cpp'],['../group__Context.html#gab6dbcff5c5b1249a5ac5cf39ae9d08bc',1,'hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device): hip_context.cpp']]], + ['hipctxdestroy',['hipCtxDestroy',['../group__Context.html#ga9a65fe43238ef303a6d97826c05fd14e',1,'hipCtxDestroy(hipCtx_t ctx): hip_context.cpp'],['../group__Context.html#ga9a65fe43238ef303a6d97826c05fd14e',1,'hipCtxDestroy(hipCtx_t ctx): hip_context.cpp']]], + ['hipctxdisablepeeraccess',['hipCtxDisablePeerAccess',['../group__Context.html#gaf48e4e9c3b6bbad5deaeff10d2e28b31',1,'hipCtxDisablePeerAccess(hipCtx_t peerCtx): hip_peer.cpp'],['../group__Context.html#gaf48e4e9c3b6bbad5deaeff10d2e28b31',1,'hipCtxDisablePeerAccess(hipCtx_t peerCtx): hip_peer.cpp']]], + ['hipctxenablepeeraccess',['hipCtxEnablePeerAccess',['../group__Context.html#ga834dfd99d72082fe8770142fa30b30e2',1,'hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags): hip_peer.cpp'],['../group__Context.html#ga834dfd99d72082fe8770142fa30b30e2',1,'hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags): hip_peer.cpp']]], + ['hipctxgetapiversion',['hipCtxGetApiVersion',['../group__Context.html#gabe5941e6798c71a8351b7d51311e7e8f',1,'hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion): hip_context.cpp'],['../group__Context.html#gabe5941e6798c71a8351b7d51311e7e8f',1,'hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion): hip_context.cpp']]], + ['hipctxgetcacheconfig',['hipCtxGetCacheConfig',['../group__Context.html#gab10373068faafd3042c9003e2e6d905a',1,'hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig): hip_context.cpp'],['../group__Context.html#gab10373068faafd3042c9003e2e6d905a',1,'hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig): hip_context.cpp']]], + ['hipctxgetcurrent',['hipCtxGetCurrent',['../group__Context.html#ga741786101d348fdbfa1f64546860357a',1,'hipCtxGetCurrent(hipCtx_t *ctx): hip_context.cpp'],['../group__Context.html#ga741786101d348fdbfa1f64546860357a',1,'hipCtxGetCurrent(hipCtx_t *ctx): hip_context.cpp']]], + ['hipctxgetdevice',['hipCtxGetDevice',['../group__Context.html#ga8aa32cf64272da929f23ecbafefefcee',1,'hipCtxGetDevice(hipDevice_t *device): hip_context.cpp'],['../group__Context.html#ga8aa32cf64272da929f23ecbafefefcee',1,'hipCtxGetDevice(hipDevice_t *device): hip_context.cpp']]], + ['hipctxgetflags',['hipCtxGetFlags',['../group__Context.html#ga12a7a993e10f775fbf61a0b14288ed1b',1,'hipCtxGetFlags(unsigned int *flags): hip_context.cpp'],['../group__Context.html#ga12a7a993e10f775fbf61a0b14288ed1b',1,'hipCtxGetFlags(unsigned int *flags): hip_context.cpp']]], + ['hipctxgetsharedmemconfig',['hipCtxGetSharedMemConfig',['../group__Context.html#ga3c78b22dd03435a7ca88621a45409565',1,'hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_context.cpp'],['../group__Context.html#ga3c78b22dd03435a7ca88621a45409565',1,'hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_context.cpp']]], + ['hipctxpopcurrent',['hipCtxPopCurrent',['../group__Context.html#ga2cc4ea5a6b78d9d7990a88a7863467d4',1,'hipCtxPopCurrent(hipCtx_t *ctx): hip_context.cpp'],['../group__Context.html#ga2cc4ea5a6b78d9d7990a88a7863467d4',1,'hipCtxPopCurrent(hipCtx_t *ctx): hip_context.cpp']]], + ['hipctxpushcurrent',['hipCtxPushCurrent',['../group__Context.html#ga9c9d92f29d68cacdea4c062c97e50a8a',1,'hipCtxPushCurrent(hipCtx_t ctx): hip_context.cpp'],['../group__Context.html#ga9c9d92f29d68cacdea4c062c97e50a8a',1,'hipCtxPushCurrent(hipCtx_t ctx): hip_context.cpp']]], + ['hipctxsetcacheconfig',['hipCtxSetCacheConfig',['../group__Context.html#ga7eacc656f1d1b6f5a683bea31be67a2f',1,'hipCtxSetCacheConfig(hipFuncCache_t cacheConfig): hip_context.cpp'],['../group__Context.html#ga7eacc656f1d1b6f5a683bea31be67a2f',1,'hipCtxSetCacheConfig(hipFuncCache_t cacheConfig): hip_context.cpp']]], + ['hipctxsetcurrent',['hipCtxSetCurrent',['../group__Context.html#ga834a192f70c2bfc0269c309436776feb',1,'hipCtxSetCurrent(hipCtx_t ctx): hip_context.cpp'],['../group__Context.html#ga834a192f70c2bfc0269c309436776feb',1,'hipCtxSetCurrent(hipCtx_t ctx): hip_context.cpp']]], + ['hipctxsetsharedmemconfig',['hipCtxSetSharedMemConfig',['../group__Context.html#gad5c1d4ced27f584a74ed550dd002fa5a',1,'hipCtxSetSharedMemConfig(hipSharedMemConfig config): hip_context.cpp'],['../group__Context.html#gad5c1d4ced27f584a74ed550dd002fa5a',1,'hipCtxSetSharedMemConfig(hipSharedMemConfig config): hip_context.cpp']]], + ['hipctxsynchronize',['hipCtxSynchronize',['../group__Context.html#gad45cd968e8e3dcfd24ef050cab2f41c8',1,'hipCtxSynchronize(void): hip_context.cpp'],['../group__Context.html#gad45cd968e8e3dcfd24ef050cab2f41c8',1,'hipCtxSynchronize(void): hip_context.cpp']]], + ['hipdevicecanaccesspeer',['hipDeviceCanAccessPeer',['../group__PeerToPeer.html#ga0a1c9ccd775758d9d7d5b5a1f525b719',1,'hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId): hip_peer.cpp'],['../group__PeerToPeer.html#ga0a1c9ccd775758d9d7d5b5a1f525b719',1,'hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId): hip_peer.cpp']]], + ['hipdevicecomputecapability',['hipDeviceComputeCapability',['../group__Driver.html#ga0a1cf94d2b571ca3279577d5af0d1672',1,'hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device): hip_device.cpp'],['../group__Driver.html#ga0a1cf94d2b571ca3279577d5af0d1672',1,'hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device): hip_device.cpp']]], + ['hipdevicedisablepeeraccess',['hipDeviceDisablePeerAccess',['../group__PeerToPeer.html#ga85030c72824fb60aaddc7374ab60481b',1,'hipDeviceDisablePeerAccess(int peerDeviceId): hip_peer.cpp'],['../group__PeerToPeer.html#ga85030c72824fb60aaddc7374ab60481b',1,'hipDeviceDisablePeerAccess(int peerDeviceId): hip_peer.cpp']]], + ['hipdeviceenablepeeraccess',['hipDeviceEnablePeerAccess',['../group__PeerToPeer.html#ga0caca59034134d7a7bb893cc1caa653e',1,'hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags): hip_peer.cpp'],['../group__PeerToPeer.html#ga0caca59034134d7a7bb893cc1caa653e',1,'hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags): hip_peer.cpp']]], + ['hipdeviceget',['hipDeviceGet',['../group__Driver.html#gadf6c74aaf7c22fea80e9ac3400d43704',1,'hipDeviceGet(hipDevice_t *device, int ordinal): hip_context.cpp'],['../group__Driver.html#gadf6c74aaf7c22fea80e9ac3400d43704',1,'hipDeviceGet(hipDevice_t *device, int deviceId): hip_context.cpp']]], + ['hipdevicegetattribute',['hipDeviceGetAttribute',['../group__Device.html#ga7080a145a4239a7276e0dc22062026c1',1,'hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int deviceId): hip_device.cpp'],['../group__Device.html#ga7080a145a4239a7276e0dc22062026c1',1,'hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int device): hip_device.cpp']]], + ['hipdevicegetbypcibusid',['hipDeviceGetByPCIBusId',['../group__Driver.html#gafe08d5632ebd24edc26a19da380b2120',1,'hipDeviceGetByPCIBusId(int *device, const int *pciBusId): hip_device.cpp'],['../group__Driver.html#gafe08d5632ebd24edc26a19da380b2120',1,'hipDeviceGetByPCIBusId(int *device, const int *pciBusId): hip_device.cpp']]], + ['hipdevicegetcacheconfig',['hipDeviceGetCacheConfig',['../group__Device.html#ga37057f9830ad6fab7ce5f05f6d3c89ab',1,'hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig): hip_device.cpp'],['../group__Device.html#ga37057f9830ad6fab7ce5f05f6d3c89ab',1,'hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig): hip_device.cpp']]], + ['hipdevicegetlimit',['hipDeviceGetLimit',['../group__Device.html#gaa72baeb520f1d52fe17817e852937218',1,'hipDeviceGetLimit(size_t *pValue, hipLimit_t limit): hip_device.cpp'],['../group__Device.html#gaa72baeb520f1d52fe17817e852937218',1,'hipDeviceGetLimit(size_t *pValue, hipLimit_t limit): hip_device.cpp']]], + ['hipdevicegetname',['hipDeviceGetName',['../group__Driver.html#ga24df15e180a7b2b351cd362e5b7d2dac',1,'hipDeviceGetName(char *name, int len, hipDevice_t device): hip_device.cpp'],['../group__Driver.html#ga24df15e180a7b2b351cd362e5b7d2dac',1,'hipDeviceGetName(char *name, int len, hipDevice_t device): hip_device.cpp']]], + ['hipdevicegetpcibusid',['hipDeviceGetPCIBusId',['../group__Driver.html#gaaa961c8ec8047903617f4245fa50256a',1,'hipDeviceGetPCIBusId(char *pciBusId, int len, int device): hip_device.cpp'],['../group__Driver.html#gaaa961c8ec8047903617f4245fa50256a',1,'hipDeviceGetPCIBusId(char *pciBusId, int len, int device): hip_device.cpp']]], + ['hipdevicegetsharedmemconfig',['hipDeviceGetSharedMemConfig',['../group__Device.html#ga1bb08f774a34a468d969a8a04791c9bb',1,'hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_device.cpp'],['../group__Device.html#ga1bb08f774a34a468d969a8a04791c9bb',1,'hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig): hip_device.cpp']]], + ['hipdevicereset',['hipDeviceReset',['../group__Device.html#ga8d57161ae56a8edc46eeda447417bf6c',1,'hipDeviceReset(void): hip_device.cpp'],['../group__Device.html#ga8d57161ae56a8edc46eeda447417bf6c',1,'hipDeviceReset(void): hip_device.cpp']]], + ['hipdevicesetcacheconfig',['hipDeviceSetCacheConfig',['../group__Device.html#gaada3d30a46ae06f68cf1574f496b86ee',1,'hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig): hip_device.cpp'],['../group__Device.html#gaada3d30a46ae06f68cf1574f496b86ee',1,'hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig): hip_device.cpp']]], + ['hipdevicesetsharedmemconfig',['hipDeviceSetSharedMemConfig',['../group__Device.html#ga9b1f279084e76691cedfbfadf9c717ee',1,'hipDeviceSetSharedMemConfig(hipSharedMemConfig config): hip_device.cpp'],['../group__Device.html#ga9b1f279084e76691cedfbfadf9c717ee',1,'hipDeviceSetSharedMemConfig(hipSharedMemConfig config): hip_device.cpp']]], + ['hipdevicesynchronize',['hipDeviceSynchronize',['../group__Device.html#gaefdc2847fb1d6c3fb1354e827a191ebd',1,'hipDeviceSynchronize(void): hip_device.cpp'],['../group__Device.html#gaefdc2847fb1d6c3fb1354e827a191ebd',1,'hipDeviceSynchronize(void): hip_device.cpp']]], + ['hipdevicetotalmem',['hipDeviceTotalMem',['../group__Driver.html#ga8991e535d0ef1ead0524e73364623041',1,'hipDeviceTotalMem(size_t *bytes, hipDevice_t device): hip_device.cpp'],['../group__Driver.html#ga8991e535d0ef1ead0524e73364623041',1,'hipDeviceTotalMem(size_t *bytes, hipDevice_t device): hip_device.cpp']]], + ['hipdrivergetversion',['hipDriverGetVersion',['../group__Driver.html#gaf6c342f52d2a29a0aca5cdd89b4dd47c',1,'hipDriverGetVersion(int *driverVersion): hip_context.cpp'],['../group__Driver.html#gaf6c342f52d2a29a0aca5cdd89b4dd47c',1,'hipDriverGetVersion(int *driverVersion): hip_context.cpp']]], + ['hipeventcreate',['hipEventCreate',['../group__Event.html#ga5df2309c9f29ca4c8e669db658d411b4',1,'hipEventCreate(hipEvent_t *event): hip_event.cpp'],['../group__Event.html#ga5df2309c9f29ca4c8e669db658d411b4',1,'hipEventCreate(hipEvent_t *event): hip_event.cpp']]], + ['hipeventcreatewithflags',['hipEventCreateWithFlags',['../group__Event.html#gae86a5acb1b22b61bc9ecb9c28fc71b75',1,'hipEventCreateWithFlags(hipEvent_t *event, unsigned flags): hip_event.cpp'],['../group__Event.html#gae86a5acb1b22b61bc9ecb9c28fc71b75',1,'hipEventCreateWithFlags(hipEvent_t *event, unsigned flags): hip_event.cpp']]], + ['hipeventdestroy',['hipEventDestroy',['../group__Event.html#ga83260357dce0c39e8c6a3c74ec97484c',1,'hipEventDestroy(hipEvent_t event): hip_event.cpp'],['../group__Event.html#ga83260357dce0c39e8c6a3c74ec97484c',1,'hipEventDestroy(hipEvent_t event): hip_event.cpp']]], + ['hipeventelapsedtime',['hipEventElapsedTime',['../group__Event.html#gad4128b815cb475c8e13c7e66ff6250b7',1,'hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop): hip_event.cpp'],['../group__Event.html#gad4128b815cb475c8e13c7e66ff6250b7',1,'hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop): hip_event.cpp']]], + ['hipeventquery',['hipEventQuery',['../group__Event.html#ga5d12d7b798b5ceb5932d1ac21f5ac776',1,'hipEventQuery(hipEvent_t event): hip_event.cpp'],['../group__Event.html#ga5d12d7b798b5ceb5932d1ac21f5ac776',1,'hipEventQuery(hipEvent_t event): hip_event.cpp']]], + ['hipeventrecord',['hipEventRecord',['../group__Event.html#ga553b6f7a8e7b7dd9536d8a64c24d7e29',1,'hipEventRecord(hipEvent_t event, hipStream_t stream): hip_event.cpp'],['../group__Event.html#ga553b6f7a8e7b7dd9536d8a64c24d7e29',1,'hipEventRecord(hipEvent_t event, hipStream_t stream): hip_event.cpp']]], + ['hipeventsynchronize',['hipEventSynchronize',['../group__Event.html#ga1f72d98ba5d6f7dc3da54e0c41fe38b1',1,'hipEventSynchronize(hipEvent_t event): hip_event.cpp'],['../group__Event.html#ga1f72d98ba5d6f7dc3da54e0c41fe38b1',1,'hipEventSynchronize(hipEvent_t event): hip_event.cpp']]], + ['hipfree',['hipFree',['../group__Memory.html#ga740d08da65cae1441ba32f8fedb863d1',1,'hipFree(void *ptr): hip_memory.cpp'],['../group__Memory.html#ga740d08da65cae1441ba32f8fedb863d1',1,'hipFree(void *ptr): hip_memory.cpp']]], + ['hipfreearray',['hipFreeArray',['../hcc__detail_2hip__texture_8h.html#ae29935073517245c69bc64790dd2747a',1,'hip_memory.cpp']]], + ['hipfreehost',['hipFreeHost',['../group__Memory.html#gad2164cc3d49da53052f4b83b789e90c9',1,'hipFreeHost(void *ptr) __attribute__((deprecated("use hipHostFree instead"))): hip_memory.cpp'],['../group__Memory.html#gad2164cc3d49da53052f4b83b789e90c9',1,'hipFreeHost(void *ptr): hip_memory.cpp']]], + ['hipfuncsetcacheconfig',['hipFuncSetCacheConfig',['../group__Device.html#ga82f0637f09eb1082f6fd025f76c2c6f5',1,'hipFuncSetCacheConfig(hipFuncCache_t config): hip_device.cpp'],['../group__Device.html#ga82f0637f09eb1082f6fd025f76c2c6f5',1,'hipFuncSetCacheConfig(hipFuncCache_t cacheConfig): hip_device.cpp']]], + ['hipgetdevice',['hipGetDevice',['../group__Device.html#ga7e0e2e8c5f78e3c7449764657c254e0a',1,'hipGetDevice(int *deviceId): hip_device.cpp'],['../group__Device.html#ga7e0e2e8c5f78e3c7449764657c254e0a',1,'hipGetDevice(int *deviceId): hip_device.cpp']]], + ['hipgetdevicecount',['hipGetDeviceCount',['../group__Device.html#ga8555d5c76d88c50ddbf54ae70b568394',1,'hipGetDeviceCount(int *count): hip_device.cpp'],['../group__Device.html#ga8555d5c76d88c50ddbf54ae70b568394',1,'hipGetDeviceCount(int *count): hip_device.cpp']]], + ['hipgetdeviceproperties',['hipGetDeviceProperties',['../group__Device.html#ga32208513b7cd491f0cb5fc884053f790',1,'hipGetDeviceProperties(hipDeviceProp_t *prop, int deviceId): hip_device.cpp'],['../group__Device.html#ga32208513b7cd491f0cb5fc884053f790',1,'hipGetDeviceProperties(hipDeviceProp_t *props, int device): hip_device.cpp']]], + ['hipgeterrorname',['hipGetErrorName',['../group__Error.html#ga88c474d77635523dbf6ca67be7b56999',1,'hipGetErrorName(hipError_t hip_error): hip_error.cpp'],['../group__Error.html#ga88c474d77635523dbf6ca67be7b56999',1,'hipGetErrorName(hipError_t hip_error): hip_error.cpp']]], + ['hipgeterrorstring',['hipGetErrorString',['../group__Error.html#ga1421a2ee561f30d5b1dbed3ecb28b264',1,'hipGetErrorString(hipError_t hipError): hip_error.cpp'],['../group__Error.html#ga1421a2ee561f30d5b1dbed3ecb28b264',1,'hipGetErrorString(hipError_t hip_error): hip_error.cpp']]], + ['hipgetlasterror',['hipGetLastError',['../group__Error.html#ga533daeb9114d7fc2db8d867adf9e419b',1,'hipGetLastError(void): hip_error.cpp'],['../group__Error.html#ga533daeb9114d7fc2db8d867adf9e419b',1,'hipGetLastError(): hip_error.cpp']]], + ['hiphostalloc',['hipHostAlloc',['../group__Memory.html#gac5ae7f11cefc7085f39ee7868e71724a',1,'hipHostAlloc(void **ptr, size_t size, unsigned int flags) __attribute__((deprecated("use hipHostMalloc instead"))): hip_memory.cpp'],['../group__Memory.html#gac5ae7f11cefc7085f39ee7868e71724a',1,'hipHostAlloc(void **ptr, size_t sizeBytes, unsigned int flags): hip_memory.cpp']]], + ['hiphostfree',['hipHostFree',['../group__Memory.html#ga2e543f58ee4544e317cd695d6d82e0a3',1,'hipHostFree(void *ptr): hip_memory.cpp'],['../group__Memory.html#ga2e543f58ee4544e317cd695d6d82e0a3',1,'hipHostFree(void *ptr): hip_memory.cpp']]], + ['hiphostgetdevicepointer',['hipHostGetDevicePointer',['../group__Memory.html#ga8fa7a0478020b835a24785cd6bb89725',1,'hip_runtime_api.h']]], + ['hiphostgetflags',['hipHostGetFlags',['../group__Memory.html#ga4d26915873b3e3534ceb4dc310f8709a',1,'hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr): hip_memory.cpp'],['../group__Memory.html#ga4d26915873b3e3534ceb4dc310f8709a',1,'hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr): hip_memory.cpp']]], + ['hiphostmalloc',['hipHostMalloc',['../group__Memory.html#gaad40bc7d97ccc799403ef5a9a8c246e1',1,'hipHostMalloc(void **ptr, size_t size, unsigned int flags): hip_memory.cpp'],['../group__Memory.html#gaad40bc7d97ccc799403ef5a9a8c246e1',1,'hipHostMalloc(void **ptr, size_t sizeBytes, unsigned int flags): hip_memory.cpp']]], + ['hiphostregister',['hipHostRegister',['../group__Memory.html#gab8258f051e1a1f7385f794a15300e674',1,'hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags): hip_memory.cpp'],['../group__Memory.html#gab8258f051e1a1f7385f794a15300e674',1,'hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags): hip_memory.cpp']]], + ['hiphostunregister',['hipHostUnregister',['../group__Memory.html#ga4c9e1810b9f5858d36c4d28c91c86924',1,'hipHostUnregister(void *hostPtr): hip_memory.cpp'],['../group__Memory.html#ga4c9e1810b9f5858d36c4d28c91c86924',1,'hipHostUnregister(void *hostPtr): hip_memory.cpp']]], + ['hipinit',['hipInit',['../group__Driver.html#ga01baa652dda5815c594d047060496caa',1,'hipInit(unsigned int flags): hip_context.cpp'],['../group__Driver.html#ga01baa652dda5815c594d047060496caa',1,'hipInit(unsigned int flags): hip_context.cpp']]], + ['hipipcclosememhandle',['hipIpcCloseMemHandle',['../group__API.html#gac2db0688a6a471e17ca631977e199da7',1,'hipIpcCloseMemHandle(void *devPtr): hip_memory.cpp'],['../group__API.html#gac2db0688a6a471e17ca631977e199da7',1,'hipIpcCloseMemHandle(void *devPtr): hip_memory.cpp']]], + ['hipipcgetmemhandle',['hipIpcGetMemHandle',['../group__API.html#gafd8c80f7e3b6426a630fff768409be70',1,'hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr): hip_memory.cpp'],['../group__API.html#gafd8c80f7e3b6426a630fff768409be70',1,'hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr): hip_memory.cpp']]], + ['hipipcopenmemhandle',['hipIpcOpenMemHandle',['../group__API.html#ga2ada334c986e10805d58167e260cb0df',1,'hipIpcOpenMemHandle(void **devPtr, hipIpcMemHandle_t handle, unsigned int flags): hip_memory.cpp'],['../group__API.html#ga2ada334c986e10805d58167e260cb0df',1,'hipIpcOpenMemHandle(void **devPtr, hipIpcMemHandle_t handle, unsigned int flags): hip_memory.cpp']]], + ['hipmalloc',['hipMalloc',['../group__Memory.html#ga4c6fcfe80010069d2792780d00dcead2',1,'hipMalloc(void **ptr, size_t size): hip_memory.cpp'],['../group__Memory.html#ga4c6fcfe80010069d2792780d00dcead2',1,'hipMalloc(void **ptr, size_t sizeBytes): hip_memory.cpp']]], + ['hipmallocarray',['hipMallocArray',['../hcc__detail_2hip__texture_8h.html#aea2f071ae2e4a0847f6aa42d1ef56129',1,'hip_memory.cpp']]], + ['hipmallochost',['hipMallocHost',['../group__Memory.html#gad3d3cdf82eb0058fc9eac1f939cd9d30',1,'hipMallocHost(void **ptr, size_t size) __attribute__((deprecated("use hipHostMalloc instead"))): hip_memory.cpp'],['../group__Memory.html#gad3d3cdf82eb0058fc9eac1f939cd9d30',1,'hipMallocHost(void **ptr, size_t sizeBytes): hip_memory.cpp']]], + ['hipmallocpitch',['hipMallocPitch',['../group__Memory.html#ga805c7320498926e444616fe090c727ee',1,'hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height): hip_memory.cpp'],['../group__Memory.html#ga805c7320498926e444616fe090c727ee',1,'hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height): hip_memory.cpp']]], + ['hipmemcpy',['hipMemcpy',['../group__Memory.html#gac1a055d288302edd641c6d7416858e1e',1,'hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind): hip_memory.cpp'],['../group__Memory.html#gac1a055d288302edd641c6d7416858e1e',1,'hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind): hip_memory.cpp']]], + ['hipmemcpy2d',['hipMemcpy2D',['../hcc__detail_2hip__texture_8h.html#a8af4597ff0cd17247d8a857c4d8bfa8a',1,'hip_memory.cpp']]], + ['hipmemcpy2dtoarray',['hipMemcpy2DToArray',['../hcc__detail_2hip__texture_8h.html#ab08b0bd6cdbb57e79df40ef1cd0374e0',1,'hip_memory.cpp']]], + ['hipmemcpyasync',['hipMemcpyAsync',['../group__Memory.html#gad55fa9f5980b711bc93c52820149ba18',1,'hipMemcpyAsync(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gad55fa9f5980b711bc93c52820149ba18',1,'hipMemcpyAsync(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpydtod',['hipMemcpyDtoD',['../group__Memory.html#ga814f245b8918f173c1f2f8c1480f7f93',1,'hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#ga814f245b8918f173c1f2f8c1480f7f93',1,'hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemcpydtodasync',['hipMemcpyDtoDAsync',['../group__Memory.html#gad3bba6016cba62f0e933cdfb4c312d27',1,'hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gad3bba6016cba62f0e933cdfb4c312d27',1,'hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpydtoh',['hipMemcpyDtoH',['../group__Memory.html#gae61f4e35ff1b9643c6328bc45d091c3f',1,'hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#gae61f4e35ff1b9643c6328bc45d091c3f',1,'hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemcpydtohasync',['hipMemcpyDtoHAsync',['../group__Memory.html#gad69da1994a646b843fb1fa465dbeb623',1,'hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gad69da1994a646b843fb1fa465dbeb623',1,'hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpyhtod',['hipMemcpyHtoD',['../group__Memory.html#gaa415ecdd2535e1eb618ed8a31a6f03af',1,'hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#gaa415ecdd2535e1eb618ed8a31a6f03af',1,'hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemcpyhtodasync',['hipMemcpyHtoDAsync',['../group__Memory.html#gae3e76acb34bed5e3b19b4c1822975bcd',1,'hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gae3e76acb34bed5e3b19b4c1822975bcd',1,'hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmemcpypeer',['hipMemcpyPeer',['../group__PeerToPeer.html#ga5512f45e25c08052667c8ffe7162333b',1,'hipMemcpyPeer(void *dst, int dstDeviceId, const void *src, int srcDeviceId, size_t sizeBytes): hip_peer.cpp'],['../group__PeerToPeer.html#ga5512f45e25c08052667c8ffe7162333b',1,'hipMemcpyPeer(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes): hip_peer.cpp']]], + ['hipmemcpypeerasync',['hipMemcpyPeerAsync',['../group__PeerToPeer.html#ga216f951370c931d22e80c089ab724ed9',1,'hipMemcpyPeerAsync(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream): hip_peer.cpp'],['../group__PeerToPeer.html#ga216f951370c931d22e80c089ab724ed9',1,'hipMemcpyPeerAsync(void *dst, int dstDevice, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream): hip_peer.cpp']]], + ['hipmemcpytoarray',['hipMemcpyToArray',['../hcc__detail_2hip__texture_8h.html#a2e631f0b6757b47b74cdaa2911784be9',1,'hip_memory.cpp']]], + ['hipmemcpytosymbol',['hipMemcpyToSymbol',['../group__Memory.html#ga131ac5c1ba04e186112491cb9bf964bc',1,'hipMemcpyToSymbol(const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind): hip_memory.cpp'],['../group__Memory.html#ga131ac5c1ba04e186112491cb9bf964bc',1,'hipMemcpyToSymbol(const char *symbolName, const void *src, size_t count, size_t offset, hipMemcpyKind kind): hip_memory.cpp']]], + ['hipmemcpytosymbolasync',['hipMemcpyToSymbolAsync',['../group__Memory.html#ga7dd66b9e8e0e7159177b42a250fe23c4',1,'hipMemcpyToSymbolAsync(const char *symbolName, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#ga7dd66b9e8e0e7159177b42a250fe23c4',1,'hipMemcpyToSymbolAsync(const char *symbolName, const void *src, size_t count, size_t offset, hipMemcpyKind kind, hipStream_t stream): hip_memory.cpp']]], + ['hipmemgetaddressrange',['hipMemGetAddressRange',['../group__PeerToPeer.html#gac7d9132f6e3d102e9b512020e5654f38',1,'hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr): hip_memory.cpp'],['../group__PeerToPeer.html#gac7d9132f6e3d102e9b512020e5654f38',1,'hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr): hip_memory.cpp']]], + ['hipmemgetinfo',['hipMemGetInfo',['../group__Memory.html#ga311c3e246a21590de14478b8bd063be2',1,'hipMemGetInfo(size_t *free, size_t *total): hip_memory.cpp'],['../group__Memory.html#ga311c3e246a21590de14478b8bd063be2',1,'hipMemGetInfo(size_t *free, size_t *total): hip_memory.cpp']]], + ['hipmemset',['hipMemset',['../group__Memory.html#gac7441e74affcce4b8b69dba996c5ebc4',1,'hipMemset(void *dst, int value, size_t sizeBytes): hip_memory.cpp'],['../group__Memory.html#gac7441e74affcce4b8b69dba996c5ebc4',1,'hipMemset(void *dst, int value, size_t sizeBytes): hip_memory.cpp']]], + ['hipmemsetasync',['hipMemsetAsync',['../group__Memory.html#gae7d90e14c387e49f10db597f12915c54',1,'hipMemsetAsync(void *dst, int value, size_t sizeBytes, hipStream_t stream): hip_memory.cpp'],['../group__Memory.html#gae7d90e14c387e49f10db597f12915c54',1,'hipMemsetAsync(void *dst, int value, size_t sizeBytes, hipStream_t stream): hip_memory.cpp']]], + ['hipmodulegetfunction',['hipModuleGetFunction',['../group__Driver.html#ga9648b457bb837838cb936b417b56a65d',1,'hipModuleGetFunction(hipFunction_t *function, hipModule_t module, const char *kname): hip_module.cpp'],['../group__Driver.html#ga9648b457bb837838cb936b417b56a65d',1,'hipModuleGetFunction(hipFunction_t *hfunc, hipModule_t hmod, const char *name): hip_module.cpp']]], + ['hipmodulegetglobal',['hipModuleGetGlobal',['../group__Driver.html#ga3e425a680285f495e776f096e9632c89',1,'hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name): hip_module.cpp'],['../group__Driver.html#ga3e425a680285f495e776f096e9632c89',1,'hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name): hip_module.cpp']]], + ['hipmodulelaunchkernel',['hipModuleLaunchKernel',['../group__Driver.html#ga2e4de5937aa8171e9eda16c881ed0674',1,'hip_runtime_api.h']]], + ['hipmoduleload',['hipModuleLoad',['../group__Driver.html#ga31d806d976e91d36bd990ae3004d8760',1,'hipModuleLoad(hipModule_t *module, const char *fname): hip_module.cpp'],['../group__Driver.html#ga31d806d976e91d36bd990ae3004d8760',1,'hipModuleLoad(hipModule_t *module, const char *fname): hip_module.cpp']]], + ['hipmoduleloaddata',['hipModuleLoadData',['../group__Driver.html#gaabdbd73e952a741e861d01109c4790f3',1,'hipModuleLoadData(hipModule_t *module, const void *image): hip_module.cpp'],['../group__Driver.html#gaabdbd73e952a741e861d01109c4790f3',1,'hipModuleLoadData(hipModule_t *module, const void *image): hip_module.cpp']]], + ['hipmoduleunload',['hipModuleUnload',['../group__Driver.html#gae58e345f55bb3ec13dca80d2df88e0ed',1,'hipModuleUnload(hipModule_t module): hip_module.cpp'],['../group__Driver.html#gae58e345f55bb3ec13dca80d2df88e0ed',1,'hipModuleUnload(hipModule_t hmod): hip_module.cpp']]], + ['hippeekatlasterror',['hipPeekAtLastError',['../group__Error.html#ga1dd660bc739f7e13edd34615660f0148',1,'hipPeekAtLastError(void): hip_error.cpp'],['../group__Error.html#ga1dd660bc739f7e13edd34615660f0148',1,'hipPeekAtLastError(): hip_error.cpp']]], + ['hippointergetattributes',['hipPointerGetAttributes',['../group__Memory.html#ga3d68ba64959615d4ab84f10caa12433b',1,'hipPointerGetAttributes(hipPointerAttribute_t *attributes, void *ptr): hip_memory.cpp'],['../group__Memory.html#ga3d68ba64959615d4ab84f10caa12433b',1,'hipPointerGetAttributes(hipPointerAttribute_t *attributes, void *ptr): hip_memory.cpp']]], + ['hipprofilerstart',['hipProfilerStart',['../group__Profiler.html#ga40aa20d731f9c8f0586127d589759e1d',1,'hipProfilerStart(): hip_hcc.cpp'],['../group__Profiler.html#ga40aa20d731f9c8f0586127d589759e1d',1,'hipProfilerStart(): hip_hcc.cpp']]], + ['hipprofilerstop',['hipProfilerStop',['../group__Profiler.html#ga777994cae0f7bdef24131fc1f3ed626e',1,'hipProfilerStop(): hip_hcc.cpp'],['../group__Profiler.html#ga777994cae0f7bdef24131fc1f3ed626e',1,'hipProfilerStop(): hip_hcc.cpp']]], + ['hipruntimegetversion',['hipRuntimeGetVersion',['../group__Driver.html#gae8b7ba34d2e11e334650aa51a4dd87ee',1,'hipRuntimeGetVersion(int *runtimeVersion): hip_context.cpp'],['../group__Driver.html#gae8b7ba34d2e11e334650aa51a4dd87ee',1,'hipRuntimeGetVersion(int *runtimeVersion): hip_context.cpp']]], + ['hipsetdevice',['hipSetDevice',['../group__Device.html#ga43c1e7f15925eeb762195ccb5e063eae',1,'hipSetDevice(int deviceId): hip_device.cpp'],['../group__Device.html#ga43c1e7f15925eeb762195ccb5e063eae',1,'hipSetDevice(int deviceId): hip_device.cpp']]], + ['hipsetdeviceflags',['hipSetDeviceFlags',['../group__Device.html#ga6e54db382768827e84725632018307aa',1,'hip_runtime_api.h']]], + ['hipstreamaddcallback',['hipStreamAddCallback',['../group__Stream.html#ga3e098cd7478828b2104abb41a7bb00d3',1,'hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags): hip_stream.cpp'],['../group__Stream.html#ga3e098cd7478828b2104abb41a7bb00d3',1,'hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags): hip_stream.cpp']]], + ['hipstreamcreate',['hipStreamCreate',['../group__Stream.html#gaff5b62d6e9502d80879f7176f4d03102',1,'hipStreamCreate(hipStream_t *stream): hip_stream.cpp'],['../group__Stream.html#gaff5b62d6e9502d80879f7176f4d03102',1,'hipStreamCreate(hipStream_t *stream): hip_stream.cpp']]], + ['hipstreamcreatewithflags',['hipStreamCreateWithFlags',['../group__Stream.html#gaf2382e3cc6632332a8983a0f58e43494',1,'hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags): hip_stream.cpp'],['../group__Stream.html#gaf2382e3cc6632332a8983a0f58e43494',1,'hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags): hip_stream.cpp']]], + ['hipstreamdestroy',['hipStreamDestroy',['../group__Stream.html#ga3076a3499ed2c7821311006100bb95ec',1,'hipStreamDestroy(hipStream_t stream): hip_stream.cpp'],['../group__Stream.html#ga3076a3499ed2c7821311006100bb95ec',1,'hipStreamDestroy(hipStream_t stream): hip_stream.cpp']]], + ['hipstreamgetflags',['hipStreamGetFlags',['../group__Stream.html#ga3249555a26439591b8873f70b39bb116',1,'hipStreamGetFlags(hipStream_t stream, unsigned int *flags): hip_stream.cpp'],['../group__Stream.html#ga3249555a26439591b8873f70b39bb116',1,'hipStreamGetFlags(hipStream_t stream, unsigned int *flags): hip_stream.cpp']]], + ['hipstreamquery',['hipStreamQuery',['../group__Stream.html#ga925b39ff78d3b5fd458bd9e2cade9f4e',1,'hipStreamQuery(hipStream_t stream): hip_stream.cpp'],['../group__Stream.html#ga925b39ff78d3b5fd458bd9e2cade9f4e',1,'hipStreamQuery(hipStream_t stream): hip_stream.cpp']]], + ['hipstreamsynchronize',['hipStreamSynchronize',['../group__Stream.html#gabbfb9f573a6ebe8c478605ecb5504a74',1,'hipStreamSynchronize(hipStream_t stream): hip_stream.cpp'],['../group__Stream.html#gabbfb9f573a6ebe8c478605ecb5504a74',1,'hipStreamSynchronize(hipStream_t stream): hip_stream.cpp']]], + ['hipstreamwaitevent',['hipStreamWaitEvent',['../group__Stream.html#gacdd84c8f8ef1539c96c57c1d5bcae633',1,'hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags): hip_stream.cpp'],['../group__Stream.html#gacdd84c8f8ef1539c96c57c1d5bcae633',1,'hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags): hip_stream.cpp']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_0.html b/docs/RuntimeAPI/html/search/groups_0.html new file mode 100644 index 0000000000..2090afa206 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_0.js b/docs/RuntimeAPI/html/search/groups_0.js new file mode 100644 index 0000000000..41450aa73c --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['control',['Control',['../group__Profiler.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_1.html b/docs/RuntimeAPI/html/search/groups_1.html new file mode 100644 index 0000000000..3ad8a3ba5e --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_1.js b/docs/RuntimeAPI/html/search/groups_1.js new file mode 100644 index 0000000000..9f4a599ed0 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_1.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['device_20management',['Device Management',['../group__Device.html',1,'']]], + ['device_20memory_20access',['Device Memory Access',['../group__PeerToPeer.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_2.html b/docs/RuntimeAPI/html/search/groups_2.html new file mode 100644 index 0000000000..361832c940 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_2.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_2.js b/docs/RuntimeAPI/html/search/groups_2.js new file mode 100644 index 0000000000..c9fd17e83e --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_2.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['error_20handling',['Error Handling',['../group__Error.html',1,'']]], + ['event_20management',['Event Management',['../group__Event.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_3.html b/docs/RuntimeAPI/html/search/groups_3.html new file mode 100644 index 0000000000..f42200a1c1 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_3.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_3.js b/docs/RuntimeAPI/html/search/groups_3.js new file mode 100644 index 0000000000..04e0d8aba5 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_3.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['fence_20functions',['Fence Functions',['../group__Fence.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_4.html b/docs/RuntimeAPI/html/search/groups_4.html new file mode 100644 index 0000000000..b07fdf95ae --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_4.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_4.js b/docs/RuntimeAPI/html/search/groups_4.js new file mode 100644 index 0000000000..529e9394c3 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_4.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['global_20enum_20and_20defines',['Global enum and defines',['../group__GlobalDefs.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_5.html b/docs/RuntimeAPI/html/search/groups_5.html new file mode 100644 index 0000000000..35324c03fa --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_5.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_5.js b/docs/RuntimeAPI/html/search/groups_5.js new file mode 100644 index 0000000000..178e3f55de --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_5.js @@ -0,0 +1,6 @@ +var searchData= +[ + ['hip_20api',['HIP API',['../group__API.html',1,'']]], + ['hcc_2dspecific_20accessors',['HCC-Specific Accessors',['../group__HCC__Specific.html',1,'']]], + ['hip_20environment_20variables',['HIP Environment Variables',['../group__HIP-ENV.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_6.html b/docs/RuntimeAPI/html/search/groups_6.html new file mode 100644 index 0000000000..323140233c --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_6.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_6.js b/docs/RuntimeAPI/html/search/groups_6.js new file mode 100644 index 0000000000..f99f43b337 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_6.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['initialization_20and_20version',['Initialization and Version',['../group__Driver.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_7.html b/docs/RuntimeAPI/html/search/groups_7.html new file mode 100644 index 0000000000..27334b6a7d --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_7.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_7.js b/docs/RuntimeAPI/html/search/groups_7.js new file mode 100644 index 0000000000..0541757e52 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_7.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['management',['Management',['../group__Context.html',1,'']]], + ['memory_20management',['Memory Management',['../group__Memory.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_8.html b/docs/RuntimeAPI/html/search/groups_8.html new file mode 100644 index 0000000000..863d13654f --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_8.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_8.js b/docs/RuntimeAPI/html/search/groups_8.js new file mode 100644 index 0000000000..7cc1c54c23 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_8.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['stream_20management',['Stream Management',['../group__Stream.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/groups_9.html b/docs/RuntimeAPI/html/search/groups_9.html new file mode 100644 index 0000000000..e42b2804e7 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_9.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/groups_9.js b/docs/RuntimeAPI/html/search/groups_9.js new file mode 100644 index 0000000000..dec0aef6b9 --- /dev/null +++ b/docs/RuntimeAPI/html/search/groups_9.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['texture_20reference_20management',['Texture Reference Management',['../group__Texture.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/mag_sel.png b/docs/RuntimeAPI/html/search/mag_sel.png new file mode 100644 index 0000000000000000000000000000000000000000..81f6040a2092402b4d98f9ffa8855d12a0d4ca17 GIT binary patch literal 563 zcmV-30?hr1P)zxx&tqG15pu7)IiiXFflOc2k;dXd>%13GZAy? zRz!q0=|E6a6vV)&ZBS~G9oe0kbqyw1*gvY`{Pop2oKq#FlzgXt@Xh-7fxh>}`Fxg> z$%N%{$!4=5nM{(;=c!aG1Ofr^Do{u%Ih{^&Fc@H2)+a-?TBXrw5DW&z%Nb6mQ!L9O zl}b@6mB?f=tX3;#vl)}ggh(Vpyh(IK z(Mb0D{l{U$FsRjP;!{($+bsaaVi8T#1c0V#qEIOCYa9@UVLV`f__E81L;?WEaRA;Y zUH;rZ;vb;mk7JX|$=i3O~&If0O@oZfLg8gfIjW=dcBsz;gI=!{-r4# z4%6v$&~;q^j7Fo67yJ(NJWuX+I~I!tj^nW3?}^9bq|<3^+vapS5sgM^x7!cs(+mMT z&y%j};&~po+YO)3hoUH4E*E;e9>?R6SS&`X)p`njycAVcg{rEb41T{~Hk(bl-7eSb zmFxA2uIqo#@R?lKm50ND`~6Nfn|-b1|L6O98vt3Tx@gKz#isxO002ovPDHLkV1kyW B_l^Jn literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/search/nomatches.html b/docs/RuntimeAPI/html/search/nomatches.html new file mode 100644 index 0000000000..b1ded27e9a --- /dev/null +++ b/docs/RuntimeAPI/html/search/nomatches.html @@ -0,0 +1,12 @@ + + + + + + + +
+
No Matches
+
+ + diff --git a/docs/RuntimeAPI/html/search/pages_0.html b/docs/RuntimeAPI/html/search/pages_0.html new file mode 100644 index 0000000000..c51c834506 --- /dev/null +++ b/docs/RuntimeAPI/html/search/pages_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/pages_0.js b/docs/RuntimeAPI/html/search/pages_0.js new file mode 100644 index 0000000000..8d42650d49 --- /dev/null +++ b/docs/RuntimeAPI/html/search/pages_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['bug_20list',['Bug List',['../bug.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/pages_1.html b/docs/RuntimeAPI/html/search/pages_1.html new file mode 100644 index 0000000000..2a98fce174 --- /dev/null +++ b/docs/RuntimeAPI/html/search/pages_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/pages_1.js b/docs/RuntimeAPI/html/search/pages_1.js new file mode 100644 index 0000000000..038da5410f --- /dev/null +++ b/docs/RuntimeAPI/html/search/pages_1.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['deprecated_20list',['Deprecated List',['../deprecated.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/pages_2.html b/docs/RuntimeAPI/html/search/pages_2.html new file mode 100644 index 0000000000..0711a0b418 --- /dev/null +++ b/docs/RuntimeAPI/html/search/pages_2.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/pages_2.js b/docs/RuntimeAPI/html/search/pages_2.js new file mode 100644 index 0000000000..09f9fa0389 --- /dev/null +++ b/docs/RuntimeAPI/html/search/pages_2.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['heterogeneous_2dcomputing_20interface_20for_20portability_20_28hip_29',['Heterogeneous-computing Interface for Portability (HIP)',['../index.html',1,'']]] +]; diff --git a/docs/RuntimeAPI/html/search/related_0.html b/docs/RuntimeAPI/html/search/related_0.html new file mode 100644 index 0000000000..b6476349f4 --- /dev/null +++ b/docs/RuntimeAPI/html/search/related_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/related_0.js b/docs/RuntimeAPI/html/search/related_0.js new file mode 100644 index 0000000000..6f5dc6598c --- /dev/null +++ b/docs/RuntimeAPI/html/search/related_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['hipstreamquery',['hipStreamQuery',['../classihipStream__t.html#a46bba2e3bfbb0915a08fb608aed31858',1,'ihipStream_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/search.css b/docs/RuntimeAPI/html/search/search.css new file mode 100644 index 0000000000..4d7612ff63 --- /dev/null +++ b/docs/RuntimeAPI/html/search/search.css @@ -0,0 +1,271 @@ +/*---------------- Search Box */ + +#FSearchBox { + float: left; +} + +#MSearchBox { + white-space : nowrap; + position: absolute; + float: none; + display: inline; + margin-top: 8px; + right: 0px; + width: 170px; + z-index: 102; + background-color: white; +} + +#MSearchBox .left +{ + display:block; + position:absolute; + left:10px; + width:20px; + height:19px; + background:url('search_l.png') no-repeat; + background-position:right; +} + +#MSearchSelect { + display:block; + position:absolute; + width:20px; + height:19px; +} + +.left #MSearchSelect { + left:4px; +} + +.right #MSearchSelect { + right:5px; +} + +#MSearchField { + display:block; + position:absolute; + height:19px; + background:url('search_m.png') repeat-x; + border:none; + width:111px; + margin-left:20px; + padding-left:4px; + color: #909090; + outline: none; + font: 9pt Arial, Verdana, sans-serif; +} + +#FSearchBox #MSearchField { + margin-left:15px; +} + +#MSearchBox .right { + display:block; + position:absolute; + right:10px; + top:0px; + width:20px; + height:19px; + background:url('search_r.png') no-repeat; + background-position:left; +} + +#MSearchClose { + display: none; + position: absolute; + top: 4px; + background : none; + border: none; + margin: 0px 4px 0px 0px; + padding: 0px 0px; + outline: none; +} + +.left #MSearchClose { + left: 6px; +} + +.right #MSearchClose { + right: 2px; +} + +.MSearchBoxActive #MSearchField { + color: #000000; +} + +/*---------------- Search filter selection */ + +#MSearchSelectWindow { + display: none; + position: absolute; + left: 0; top: 0; + border: 1px solid #90A5CE; + background-color: #F9FAFC; + z-index: 1; + padding-top: 4px; + padding-bottom: 4px; + -moz-border-radius: 4px; + -webkit-border-top-left-radius: 4px; + -webkit-border-top-right-radius: 4px; + -webkit-border-bottom-left-radius: 4px; + -webkit-border-bottom-right-radius: 4px; + -webkit-box-shadow: 5px 5px 5px rgba(0, 0, 0, 0.15); +} + +.SelectItem { + font: 8pt Arial, Verdana, sans-serif; + padding-left: 2px; + padding-right: 12px; + border: 0px; +} + +span.SelectionMark { + margin-right: 4px; + font-family: monospace; + outline-style: none; + text-decoration: none; +} + +a.SelectItem { + display: block; + outline-style: none; + color: #000000; + text-decoration: none; + padding-left: 6px; + padding-right: 12px; +} + +a.SelectItem:focus, +a.SelectItem:active { + color: #000000; + outline-style: none; + text-decoration: none; +} + +a.SelectItem:hover { + color: #FFFFFF; + background-color: #3D578C; + outline-style: none; + text-decoration: none; + cursor: pointer; + display: block; +} + +/*---------------- Search results window */ + +iframe#MSearchResults { + width: 60ex; + height: 15em; +} + +#MSearchResultsWindow { + display: none; + position: absolute; + left: 0; top: 0; + border: 1px solid #000; + background-color: #EEF1F7; +} + +/* ----------------------------------- */ + + +#SRIndex { + clear:both; + padding-bottom: 15px; +} + +.SREntry { + font-size: 10pt; + padding-left: 1ex; +} + +.SRPage .SREntry { + font-size: 8pt; + padding: 1px 5px; +} + +body.SRPage { + margin: 5px 2px; +} + +.SRChildren { + padding-left: 3ex; padding-bottom: .5em +} + +.SRPage .SRChildren { + display: none; +} + +.SRSymbol { + font-weight: bold; + color: #425E97; + font-family: Arial, Verdana, sans-serif; + text-decoration: none; + outline: none; +} + +a.SRScope { + display: block; + color: #425E97; + font-family: Arial, Verdana, sans-serif; + text-decoration: none; + outline: none; +} + +a.SRSymbol:focus, a.SRSymbol:active, +a.SRScope:focus, a.SRScope:active { + text-decoration: underline; +} + +span.SRScope { + padding-left: 4px; +} + +.SRPage .SRStatus { + padding: 2px 5px; + font-size: 8pt; + font-style: italic; +} + +.SRResult { + display: none; +} + +DIV.searchresults { + margin-left: 10px; + margin-right: 10px; +} + +/*---------------- External search page results */ + +.searchresult { + background-color: #F0F3F8; +} + +.pages b { + color: white; + padding: 5px 5px 3px 5px; + background-image: url("../tab_a.png"); + background-repeat: repeat-x; + text-shadow: 0 1px 1px #000000; +} + +.pages { + line-height: 17px; + margin-left: 4px; + text-decoration: none; +} + +.hl { + font-weight: bold; +} + +#searchresults { + margin-bottom: 20px; +} + +.searchpages { + margin-top: 10px; +} + diff --git a/docs/RuntimeAPI/html/search/search.js b/docs/RuntimeAPI/html/search/search.js new file mode 100644 index 0000000000..f85f411038 --- /dev/null +++ b/docs/RuntimeAPI/html/search/search.js @@ -0,0 +1,815 @@ +// Search script generated by doxygen +// Copyright (C) 2009 by Dimitri van Heesch. + +// The code in this file is loosly based on main.js, part of Natural Docs, +// which is Copyright (C) 2003-2008 Greg Valure +// Natural Docs is licensed under the GPL. + +var indexSectionsWithContent = +{ + 0: "_abcdefghilmnprstuwxyz", + 1: "_cdfhilpstu", + 2: "h", + 3: "_h", + 4: "_achilmnprstwxyz", + 5: "dh", + 6: "h", + 7: "h", + 8: "h", + 9: "_h", + 10: "cdefghimst", + 11: "bdh" +}; + +var indexSectionNames = +{ + 0: "all", + 1: "classes", + 2: "files", + 3: "functions", + 4: "variables", + 5: "typedefs", + 6: "enums", + 7: "enumvalues", + 8: "related", + 9: "defines", + 10: "groups", + 11: "pages" +}; + +function convertToId(search) +{ + var result = ''; + for (i=0;i do a search + { + this.Search(); + } + } + + this.OnSearchSelectKey = function(evt) + { + var e = (evt) ? evt : window.event; // for IE + if (e.keyCode==40 && this.searchIndex0) // Up + { + this.searchIndex--; + this.OnSelectItem(this.searchIndex); + } + else if (e.keyCode==13 || e.keyCode==27) + { + this.OnSelectItem(this.searchIndex); + this.CloseSelectionWindow(); + this.DOMSearchField().focus(); + } + return false; + } + + // --------- Actions + + // Closes the results window. + this.CloseResultsWindow = function() + { + this.DOMPopupSearchResultsWindow().style.display = 'none'; + this.DOMSearchClose().style.display = 'none'; + this.Activate(false); + } + + this.CloseSelectionWindow = function() + { + this.DOMSearchSelectWindow().style.display = 'none'; + } + + // Performs a search. + this.Search = function() + { + this.keyTimeout = 0; + + // strip leading whitespace + var searchValue = this.DOMSearchField().value.replace(/^ +/, ""); + + var code = searchValue.toLowerCase().charCodeAt(0); + var idxChar = searchValue.substr(0, 1).toLowerCase(); + if ( 0xD800 <= code && code <= 0xDBFF && searchValue > 1) // surrogate pair + { + idxChar = searchValue.substr(0, 2); + } + + var resultsPage; + var resultsPageWithSearch; + var hasResultsPage; + + var idx = indexSectionsWithContent[this.searchIndex].indexOf(idxChar); + if (idx!=-1) + { + var hexCode=idx.toString(16); + resultsPage = this.resultsPath + '/' + indexSectionNames[this.searchIndex] + '_' + hexCode + '.html'; + resultsPageWithSearch = resultsPage+'?'+escape(searchValue); + hasResultsPage = true; + } + else // nothing available for this search term + { + resultsPage = this.resultsPath + '/nomatches.html'; + resultsPageWithSearch = resultsPage; + hasResultsPage = false; + } + + window.frames.MSearchResults.location = resultsPageWithSearch; + var domPopupSearchResultsWindow = this.DOMPopupSearchResultsWindow(); + + if (domPopupSearchResultsWindow.style.display!='block') + { + var domSearchBox = this.DOMSearchBox(); + this.DOMSearchClose().style.display = 'inline'; + if (this.insideFrame) + { + var domPopupSearchResults = this.DOMPopupSearchResults(); + domPopupSearchResultsWindow.style.position = 'relative'; + domPopupSearchResultsWindow.style.display = 'block'; + var width = document.body.clientWidth - 8; // the -8 is for IE :-( + domPopupSearchResultsWindow.style.width = width + 'px'; + domPopupSearchResults.style.width = width + 'px'; + } + else + { + var domPopupSearchResults = this.DOMPopupSearchResults(); + var left = getXPos(domSearchBox) + 150; // domSearchBox.offsetWidth; + var top = getYPos(domSearchBox) + 20; // domSearchBox.offsetHeight + 1; + domPopupSearchResultsWindow.style.display = 'block'; + left -= domPopupSearchResults.offsetWidth; + domPopupSearchResultsWindow.style.top = top + 'px'; + domPopupSearchResultsWindow.style.left = left + 'px'; + } + } + + this.lastSearchValue = searchValue; + this.lastResultsPage = resultsPage; + } + + // -------- Activation Functions + + // Activates or deactivates the search panel, resetting things to + // their default values if necessary. + this.Activate = function(isActive) + { + if (isActive || // open it + this.DOMPopupSearchResultsWindow().style.display == 'block' + ) + { + this.DOMSearchBox().className = 'MSearchBoxActive'; + + var searchField = this.DOMSearchField(); + + if (searchField.value == this.searchLabel) // clear "Search" term upon entry + { + searchField.value = ''; + this.searchActive = true; + } + } + else if (!isActive) // directly remove the panel + { + this.DOMSearchBox().className = 'MSearchBoxInactive'; + this.DOMSearchField().value = this.searchLabel; + this.searchActive = false; + this.lastSearchValue = '' + this.lastResultsPage = ''; + } + } +} + +// ----------------------------------------------------------------------- + +// The class that handles everything on the search results page. +function SearchResults(name) +{ + // The number of matches from the last run of . + this.lastMatchCount = 0; + this.lastKey = 0; + this.repeatOn = false; + + // Toggles the visibility of the passed element ID. + this.FindChildElement = function(id) + { + var parentElement = document.getElementById(id); + var element = parentElement.firstChild; + + while (element && element!=parentElement) + { + if (element.nodeName == 'DIV' && element.className == 'SRChildren') + { + return element; + } + + if (element.nodeName == 'DIV' && element.hasChildNodes()) + { + element = element.firstChild; + } + else if (element.nextSibling) + { + element = element.nextSibling; + } + else + { + do + { + element = element.parentNode; + } + while (element && element!=parentElement && !element.nextSibling); + + if (element && element!=parentElement) + { + element = element.nextSibling; + } + } + } + } + + this.Toggle = function(id) + { + var element = this.FindChildElement(id); + if (element) + { + if (element.style.display == 'block') + { + element.style.display = 'none'; + } + else + { + element.style.display = 'block'; + } + } + } + + // Searches for the passed string. If there is no parameter, + // it takes it from the URL query. + // + // Always returns true, since other documents may try to call it + // and that may or may not be possible. + this.Search = function(search) + { + if (!search) // get search word from URL + { + search = window.location.search; + search = search.substring(1); // Remove the leading '?' + search = unescape(search); + } + + search = search.replace(/^ +/, ""); // strip leading spaces + search = search.replace(/ +$/, ""); // strip trailing spaces + search = search.toLowerCase(); + search = convertToId(search); + + var resultRows = document.getElementsByTagName("div"); + var matches = 0; + + var i = 0; + while (i < resultRows.length) + { + var row = resultRows.item(i); + if (row.className == "SRResult") + { + var rowMatchName = row.id.toLowerCase(); + rowMatchName = rowMatchName.replace(/^sr\d*_/, ''); // strip 'sr123_' + + if (search.length<=rowMatchName.length && + rowMatchName.substr(0, search.length)==search) + { + row.style.display = 'block'; + matches++; + } + else + { + row.style.display = 'none'; + } + } + i++; + } + document.getElementById("Searching").style.display='none'; + if (matches == 0) // no results + { + document.getElementById("NoMatches").style.display='block'; + } + else // at least one result + { + document.getElementById("NoMatches").style.display='none'; + } + this.lastMatchCount = matches; + return true; + } + + // return the first item with index index or higher that is visible + this.NavNext = function(index) + { + var focusItem; + while (1) + { + var focusName = 'Item'+index; + focusItem = document.getElementById(focusName); + if (focusItem && focusItem.parentNode.parentNode.style.display=='block') + { + break; + } + else if (!focusItem) // last element + { + break; + } + focusItem=null; + index++; + } + return focusItem; + } + + this.NavPrev = function(index) + { + var focusItem; + while (1) + { + var focusName = 'Item'+index; + focusItem = document.getElementById(focusName); + if (focusItem && focusItem.parentNode.parentNode.style.display=='block') + { + break; + } + else if (!focusItem) // last element + { + break; + } + focusItem=null; + index--; + } + return focusItem; + } + + this.ProcessKeys = function(e) + { + if (e.type == "keydown") + { + this.repeatOn = false; + this.lastKey = e.keyCode; + } + else if (e.type == "keypress") + { + if (!this.repeatOn) + { + if (this.lastKey) this.repeatOn = true; + return false; // ignore first keypress after keydown + } + } + else if (e.type == "keyup") + { + this.lastKey = 0; + this.repeatOn = false; + } + return this.lastKey!=0; + } + + this.Nav = function(evt,itemIndex) + { + var e = (evt) ? evt : window.event; // for IE + if (e.keyCode==13) return true; + if (!this.ProcessKeys(e)) return false; + + if (this.lastKey==38) // Up + { + var newIndex = itemIndex-1; + var focusItem = this.NavPrev(newIndex); + if (focusItem) + { + var child = this.FindChildElement(focusItem.parentNode.parentNode.id); + if (child && child.style.display == 'block') // children visible + { + var n=0; + var tmpElem; + while (1) // search for last child + { + tmpElem = document.getElementById('Item'+newIndex+'_c'+n); + if (tmpElem) + { + focusItem = tmpElem; + } + else // found it! + { + break; + } + n++; + } + } + } + if (focusItem) + { + focusItem.focus(); + } + else // return focus to search field + { + parent.document.getElementById("MSearchField").focus(); + } + } + else if (this.lastKey==40) // Down + { + var newIndex = itemIndex+1; + var focusItem; + var item = document.getElementById('Item'+itemIndex); + var elem = this.FindChildElement(item.parentNode.parentNode.id); + if (elem && elem.style.display == 'block') // children visible + { + focusItem = document.getElementById('Item'+itemIndex+'_c0'); + } + if (!focusItem) focusItem = this.NavNext(newIndex); + if (focusItem) focusItem.focus(); + } + else if (this.lastKey==39) // Right + { + var item = document.getElementById('Item'+itemIndex); + var elem = this.FindChildElement(item.parentNode.parentNode.id); + if (elem) elem.style.display = 'block'; + } + else if (this.lastKey==37) // Left + { + var item = document.getElementById('Item'+itemIndex); + var elem = this.FindChildElement(item.parentNode.parentNode.id); + if (elem) elem.style.display = 'none'; + } + else if (this.lastKey==27) // Escape + { + parent.searchBox.CloseResultsWindow(); + parent.document.getElementById("MSearchField").focus(); + } + else if (this.lastKey==13) // Enter + { + return true; + } + return false; + } + + this.NavChild = function(evt,itemIndex,childIndex) + { + var e = (evt) ? evt : window.event; // for IE + if (e.keyCode==13) return true; + if (!this.ProcessKeys(e)) return false; + + if (this.lastKey==38) // Up + { + if (childIndex>0) + { + var newIndex = childIndex-1; + document.getElementById('Item'+itemIndex+'_c'+newIndex).focus(); + } + else // already at first child, jump to parent + { + document.getElementById('Item'+itemIndex).focus(); + } + } + else if (this.lastKey==40) // Down + { + var newIndex = childIndex+1; + var elem = document.getElementById('Item'+itemIndex+'_c'+newIndex); + if (!elem) // last child, jump to parent next parent + { + elem = this.NavNext(itemIndex+1); + } + if (elem) + { + elem.focus(); + } + } + else if (this.lastKey==27) // Escape + { + parent.searchBox.CloseResultsWindow(); + parent.document.getElementById("MSearchField").focus(); + } + else if (this.lastKey==13) // Enter + { + return true; + } + return false; + } +} + +function setKeyActions(elem,action) +{ + elem.setAttribute('onkeydown',action); + elem.setAttribute('onkeypress',action); + elem.setAttribute('onkeyup',action); +} + +function setClassAttr(elem,attr) +{ + elem.setAttribute('class',attr); + elem.setAttribute('className',attr); +} + +function createResults() +{ + var results = document.getElementById("SRResults"); + for (var e=0; ek7RCwB~R6VQOP#AvB$vH7i{6H{96zot$7cZT<7246EF5Np6N}+$IbiG6W zg#87A+NFaX+=_^xM1#gCtshC=E{%9^uQX_%?YwXvo{#q&MnpJ8uh(O?ZRc&~_1%^SsPxG@rfElJg-?U zm!Cz-IOn(qJP3kDp-^~qt+FGbl=5jNli^Wj_xIBG{Rc0en{!oFvyoNC7{V~T8}b>| z=jL2WIReZzX(YN(_9fV;BBD$VXQIxNasAL8ATvEu822WQ%mvv4FO#qs` BFGc_W literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/search/search_r.png b/docs/RuntimeAPI/html/search/search_r.png new file mode 100644 index 0000000000000000000000000000000000000000..97ee8b439687084201b79c6f776a41f495c6392a GIT binary patch literal 612 zcmV-q0-ODbP)PbXFRCwB?)W514K@j&X?z2*SxFI6-@HT2E2K=9X9%Pb zEK*!TBw&g(DMC;|A)uGlRkOS9vd-?zNs%bR4d$w+ox_iFnE8fvIvv7^5<(>Te12Li z7C)9srCzmK{ZcNM{YIl9j{DePFgOWiS%xG@5CnnnJa4nvY<^glbz7^|-ZY!dUkAwd z{gaTC@_>b5h~;ug#R0wRL0>o5!hxm*s0VW?8dr}O#zXTRTnrQm_Z7z1Mrnx>&p zD4qifUjzLvbVVWi?l?rUzwt^sdb~d!f_LEhsRVIXZtQ=qSxuxqm zEX#tf>$?M_Y1-LSDT)HqG?`%-%ZpY!#{N!rcNIiL;G7F0`l?)mNGTD9;f9F5Up3Kg zw}a<-JylhG&;=!>B+fZaCX+?C+kHYrP%c?X2!Zu_olK|GcS4A70HEy;vn)I0>0kLH z`jc(WIaaHc7!HS@f*^R^Znx8W=_jIl2oWJoQ*h1^$FX!>*PqR1J8k|fw}w_y}TpE>7m8DqDO<3z`OzXt$ccSejbEZCg@0000 + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/typedefs_0.js b/docs/RuntimeAPI/html/search/typedefs_0.js new file mode 100644 index 0000000000..8cfe168754 --- /dev/null +++ b/docs/RuntimeAPI/html/search/typedefs_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['dim3',['dim3',['../group__GlobalDefs.html#gacb37281795c3567d0b10a61c056d512b',1,'hip_runtime_api.h']]] +]; diff --git a/docs/RuntimeAPI/html/search/typedefs_1.html b/docs/RuntimeAPI/html/search/typedefs_1.html new file mode 100644 index 0000000000..455fe2b2b9 --- /dev/null +++ b/docs/RuntimeAPI/html/search/typedefs_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/typedefs_1.js b/docs/RuntimeAPI/html/search/typedefs_1.js new file mode 100644 index 0000000000..ab13806a51 --- /dev/null +++ b/docs/RuntimeAPI/html/search/typedefs_1.js @@ -0,0 +1,7 @@ +var searchData= +[ + ['hipfunccache_5ft',['hipFuncCache_t',['../group__GlobalDefs.html#ga7effbca2af70714feaa3330bf1a77a72',1,'hip_runtime_api.h']]], + ['hipmemcpykind',['hipMemcpyKind',['../group__GlobalDefs.html#ga0c04e67413ce030817361f02673e5c85',1,'hip_runtime_api.h']]], + ['hipsharedmemconfig',['hipSharedMemConfig',['../group__GlobalDefs.html#ga6b1ca424fa26a5fb718937d662eaee7f',1,'hip_runtime_api.h']]], + ['hipstreamcallback_5ft',['hipStreamCallback_t',['../group__Stream.html#gadd389da2e3926a8ff17658d029015dcd',1,'hip_runtime_api.h']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_0.html b/docs/RuntimeAPI/html/search/variables_0.html new file mode 100644 index 0000000000..1b8adc9b33 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_0.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_0.js b/docs/RuntimeAPI/html/search/variables_0.js new file mode 100644 index 0000000000..12df77b8fe --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_0.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['_5fcomputeunits',['_computeUnits',['../classihipDevice__t.html#a655e03136394df32571a52707aa371c5',1,'ihipDevice_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_1.html b/docs/RuntimeAPI/html/search/variables_1.html new file mode 100644 index 0000000000..78f63cd1eb --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_1.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_1.js b/docs/RuntimeAPI/html/search/variables_1.js new file mode 100644 index 0000000000..2fba32867b --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_1.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['arch',['arch',['../structhipDeviceProp__t.html#afc58158e44bef6ad26f2be401434b049',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_2.html b/docs/RuntimeAPI/html/search/variables_2.html new file mode 100644 index 0000000000..ea80d20140 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_2.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_2.js b/docs/RuntimeAPI/html/search/variables_2.js new file mode 100644 index 0000000000..82dc751418 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_2.js @@ -0,0 +1,8 @@ +var searchData= +[ + ['canmaphostmemory',['canMapHostMemory',['../structhipDeviceProp__t.html#ac2143f5448607d1a02a9e8783fcf06a1',1,'hipDeviceProp_t']]], + ['clockinstructionrate',['clockInstructionRate',['../structhipDeviceProp__t.html#a6fbf3b08a1a08ae700f1a06265f6666b',1,'hipDeviceProp_t']]], + ['clockrate',['clockRate',['../structhipDeviceProp__t.html#a1dd15bee43692b8649dfbdc1adbaaf96',1,'hipDeviceProp_t']]], + ['computemode',['computeMode',['../structhipDeviceProp__t.html#ae7d9216f8583a703359d0b9373823f5d',1,'hipDeviceProp_t']]], + ['concurrentkernels',['concurrentKernels',['../structhipDeviceProp__t.html#ad8461a28caf9c38c58cf358583b5bee3',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_3.html b/docs/RuntimeAPI/html/search/variables_3.html new file mode 100644 index 0000000000..0dca26f47b --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_3.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_3.js b/docs/RuntimeAPI/html/search/variables_3.js new file mode 100644 index 0000000000..0fba3440bc --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_3.js @@ -0,0 +1,20 @@ +var searchData= +[ + ['has3dgrid',['has3dGrid',['../structhipDeviceArch__t.html#aa5e22d295cce0d9a34ee6e7e7e378c26',1,'hipDeviceArch_t']]], + ['hasdoubles',['hasDoubles',['../structhipDeviceArch__t.html#a2d624e3d85e615b71d1182f8912893b4',1,'hipDeviceArch_t']]], + ['hasdynamicparallelism',['hasDynamicParallelism',['../structhipDeviceArch__t.html#a9114bfc718bf0648b54ff9a319a36b35',1,'hipDeviceArch_t']]], + ['hasfloatatomicadd',['hasFloatAtomicAdd',['../structhipDeviceArch__t.html#a2321d6ef74aac91c044f3289d25b2d41',1,'hipDeviceArch_t']]], + ['hasfunnelshift',['hasFunnelShift',['../structhipDeviceArch__t.html#aaec4f2d983d4602858fae8e9ddeee3ff',1,'hipDeviceArch_t']]], + ['hasglobalfloatatomicexch',['hasGlobalFloatAtomicExch',['../structhipDeviceArch__t.html#a8f213ae9a4729dff1c636ac5de0e2fa2',1,'hipDeviceArch_t']]], + ['hasglobalint32atomics',['hasGlobalInt32Atomics',['../structhipDeviceArch__t.html#a8d00c3ab98869b602c714fe7abe68e93',1,'hipDeviceArch_t']]], + ['hasglobalint64atomics',['hasGlobalInt64Atomics',['../structhipDeviceArch__t.html#ad5aa54dbab22dbcd8cf98f57a96c6636',1,'hipDeviceArch_t']]], + ['hassharedfloatatomicexch',['hasSharedFloatAtomicExch',['../structhipDeviceArch__t.html#aff005558b4edabd27b27f286ac5b2f2b',1,'hipDeviceArch_t']]], + ['hassharedint32atomics',['hasSharedInt32Atomics',['../structhipDeviceArch__t.html#a1596330b1cb9cc73f142aee11b2ab853',1,'hipDeviceArch_t']]], + ['hassharedint64atomics',['hasSharedInt64Atomics',['../structhipDeviceArch__t.html#a00c2b930fcdcad9ea7b54b449db13966',1,'hipDeviceArch_t']]], + ['hassurfacefuncs',['hasSurfaceFuncs',['../structhipDeviceArch__t.html#a9eb2462148686d4c048b69b6e09f835e',1,'hipDeviceArch_t']]], + ['hassyncthreadsext',['hasSyncThreadsExt',['../structhipDeviceArch__t.html#ade6a3b21ad5f344dcd92c52102c274ba',1,'hipDeviceArch_t']]], + ['hasthreadfencesystem',['hasThreadFenceSystem',['../structhipDeviceArch__t.html#ac2818e3b91cba8beb36741e9867bb887',1,'hipDeviceArch_t']]], + ['haswarpballot',['hasWarpBallot',['../structhipDeviceArch__t.html#af1e934a8a5106995bcc256287585564c',1,'hipDeviceArch_t']]], + ['haswarpshuffle',['hasWarpShuffle',['../structhipDeviceArch__t.html#a3d922e8fc97ca1e8ecc39600b138fa2d',1,'hipDeviceArch_t']]], + ['haswarpvote',['hasWarpVote',['../structhipDeviceArch__t.html#a35bde017352eca1d4e0eceb3bf79f274',1,'hipDeviceArch_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_4.html b/docs/RuntimeAPI/html/search/variables_4.html new file mode 100644 index 0000000000..400e8e9b41 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_4.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_4.js b/docs/RuntimeAPI/html/search/variables_4.js new file mode 100644 index 0000000000..f7bbc877f4 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_4.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['ipc_5fhandle',['ipc_handle',['../classihipIpcMemHandle__t.html#af2142ab7d9f820acbad7638428509d42',1,'ihipIpcMemHandle_t']]], + ['ismultigpuboard',['isMultiGpuBoard',['../structhipDeviceProp__t.html#a9bb19b2b0cdee8977ed63964532d639d',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_5.html b/docs/RuntimeAPI/html/search/variables_5.html new file mode 100644 index 0000000000..7f1241f94b --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_5.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_5.js b/docs/RuntimeAPI/html/search/variables_5.js new file mode 100644 index 0000000000..41a7c59602 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_5.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['l2cachesize',['l2CacheSize',['../structhipDeviceProp__t.html#a24404decccc16833973c803ced6f3a51',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_6.html b/docs/RuntimeAPI/html/search/variables_6.html new file mode 100644 index 0000000000..7536df8d5a --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_6.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_6.js b/docs/RuntimeAPI/html/search/variables_6.js new file mode 100644 index 0000000000..8fc226ebc6 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_6.js @@ -0,0 +1,13 @@ +var searchData= +[ + ['major',['major',['../structhipDeviceProp__t.html#aec9e4173c2e34cc232300c415dbd5e4f',1,'hipDeviceProp_t']]], + ['maxgridsize',['maxGridSize',['../structhipDeviceProp__t.html#ae529c23929f592120081fed31d877a55',1,'hipDeviceProp_t']]], + ['maxsharedmemorypermultiprocessor',['maxSharedMemoryPerMultiProcessor',['../structhipDeviceProp__t.html#aa1a32a7f387f6da845db7b228711fce8',1,'hipDeviceProp_t']]], + ['maxthreadsdim',['maxThreadsDim',['../structhipDeviceProp__t.html#a8ebba6fc12f80c9a9cf9b9193f0da465',1,'hipDeviceProp_t']]], + ['maxthreadsperblock',['maxThreadsPerBlock',['../structhipDeviceProp__t.html#af971cf1ca3ec1f68ad09036c0cc672e0',1,'hipDeviceProp_t']]], + ['maxthreadspermultiprocessor',['maxThreadsPerMultiProcessor',['../structhipDeviceProp__t.html#a23a39f4fd795addb3b125e9c3f6295ea',1,'hipDeviceProp_t']]], + ['memorybuswidth',['memoryBusWidth',['../structhipDeviceProp__t.html#a1d710ce8bd34daf66f892a66a2273132',1,'hipDeviceProp_t']]], + ['memoryclockrate',['memoryClockRate',['../structhipDeviceProp__t.html#a6db0ab8e7e8cc13c84d7bb7f70226d5e',1,'hipDeviceProp_t']]], + ['minor',['minor',['../structhipDeviceProp__t.html#abb51208e2509a7a1d107f0da69108938',1,'hipDeviceProp_t']]], + ['multiprocessorcount',['multiProcessorCount',['../structhipDeviceProp__t.html#add8d9d2ad52aece9fd1dbe25c18d9d57',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_7.html b/docs/RuntimeAPI/html/search/variables_7.html new file mode 100644 index 0000000000..66186a6994 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_7.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_7.js b/docs/RuntimeAPI/html/search/variables_7.js new file mode 100644 index 0000000000..124bf0ddb8 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_7.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['name',['name',['../structhipDeviceProp__t.html#a5b44bf8fa46faefcde989942b1d11a5e',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_8.html b/docs/RuntimeAPI/html/search/variables_8.html new file mode 100644 index 0000000000..aa13bf24a5 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_8.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_8.js b/docs/RuntimeAPI/html/search/variables_8.js new file mode 100644 index 0000000000..71b6a5df56 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_8.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['pcibusid',['pciBusID',['../structhipDeviceProp__t.html#a1350f64d49b717ed3a06458f7549ccb0',1,'hipDeviceProp_t']]], + ['pcideviceid',['pciDeviceID',['../structhipDeviceProp__t.html#ae6aa845dc2d540f85098ea30be35f4eb',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_9.html b/docs/RuntimeAPI/html/search/variables_9.html new file mode 100644 index 0000000000..78cc249f7d --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_9.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_9.js b/docs/RuntimeAPI/html/search/variables_9.js new file mode 100644 index 0000000000..44ba50e0b7 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_9.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['regsperblock',['regsPerBlock',['../structhipDeviceProp__t.html#a73c1c21648a901799ff6bef83c11135b',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_a.html b/docs/RuntimeAPI/html/search/variables_a.html new file mode 100644 index 0000000000..592abaa33c --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_a.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_a.js b/docs/RuntimeAPI/html/search/variables_a.js new file mode 100644 index 0000000000..a582d60c6a --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_a.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['sharedmemperblock',['sharedMemPerBlock',['../structhipDeviceProp__t.html#a3b9138678a0795c2677eddcfb1c67156',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_b.html b/docs/RuntimeAPI/html/search/variables_b.html new file mode 100644 index 0000000000..78052a1e03 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_b.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_b.js b/docs/RuntimeAPI/html/search/variables_b.js new file mode 100644 index 0000000000..b356d6e142 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_b.js @@ -0,0 +1,5 @@ +var searchData= +[ + ['totalconstmem',['totalConstMem',['../structhipDeviceProp__t.html#a29880232c56120be3455ce00d5379665',1,'hipDeviceProp_t']]], + ['totalglobalmem',['totalGlobalMem',['../structhipDeviceProp__t.html#acedd6a2d23423441e4bf51c4a1b719f9',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_c.html b/docs/RuntimeAPI/html/search/variables_c.html new file mode 100644 index 0000000000..b2260e465b --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_c.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_c.js b/docs/RuntimeAPI/html/search/variables_c.js new file mode 100644 index 0000000000..46a1400a7b --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_c.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['warpsize',['warpSize',['../structhipDeviceProp__t.html#af3357d33c004608bf05bc21a352be81b',1,'hipDeviceProp_t']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_d.html b/docs/RuntimeAPI/html/search/variables_d.html new file mode 100644 index 0000000000..f47799968f --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_d.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_d.js b/docs/RuntimeAPI/html/search/variables_d.js new file mode 100644 index 0000000000..250c203caf --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_d.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['x',['x',['../structdim3.html#ac866c05f83a28dac20a153fc65b3b16c',1,'dim3']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_e.html b/docs/RuntimeAPI/html/search/variables_e.html new file mode 100644 index 0000000000..1165006622 --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_e.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_e.js b/docs/RuntimeAPI/html/search/variables_e.js new file mode 100644 index 0000000000..133dd9dc6e --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_e.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['y',['y',['../structdim3.html#a83e60e072f7e8bdfde6ac05053cbb370',1,'dim3']]] +]; diff --git a/docs/RuntimeAPI/html/search/variables_f.html b/docs/RuntimeAPI/html/search/variables_f.html new file mode 100644 index 0000000000..537dd7c28e --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_f.html @@ -0,0 +1,26 @@ + + + + + + + + + +
+
Loading...
+
+ +
Searching...
+
No Matches
+ +
+ + diff --git a/docs/RuntimeAPI/html/search/variables_f.js b/docs/RuntimeAPI/html/search/variables_f.js new file mode 100644 index 0000000000..e8bf38b99c --- /dev/null +++ b/docs/RuntimeAPI/html/search/variables_f.js @@ -0,0 +1,4 @@ +var searchData= +[ + ['z',['z',['../structdim3.html#a866e38993ecc4e76fd47311236c16b04',1,'dim3']]] +]; diff --git a/docs/RuntimeAPI/html/structDbName-members.html b/docs/RuntimeAPI/html/structDbName-members.html new file mode 100644 index 0000000000..7ea2987d5f --- /dev/null +++ b/docs/RuntimeAPI/html/structDbName-members.html @@ -0,0 +1,103 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
DbName Member List
+
+
+ +

This is the complete list of members for DbName, including all inherited members.

+ + + +
_color (defined in DbName)DbName
_shortName (defined in DbName)DbName
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structDbName.html b/docs/RuntimeAPI/html/structDbName.html new file mode 100644 index 0000000000..5d14f3369c --- /dev/null +++ b/docs/RuntimeAPI/html/structDbName.html @@ -0,0 +1,114 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: DbName Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
DbName Struct Reference
+
+
+ + + + + + +

+Public Attributes

+const char * _color
 
+const char * _shortName
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structLockedBase-members.html b/docs/RuntimeAPI/html/structLockedBase-members.html new file mode 100644 index 0000000000..6672fa92f6 --- /dev/null +++ b/docs/RuntimeAPI/html/structLockedBase-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
LockedBase< MUTEX_TYPE > Member List
+
+
+ +

This is the complete list of members for LockedBase< MUTEX_TYPE >, including all inherited members.

+ + + + +
_mutex (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >
lock() (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >inline
unlock() (defined in LockedBase< MUTEX_TYPE >)LockedBase< MUTEX_TYPE >inline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structLockedBase.html b/docs/RuntimeAPI/html/structLockedBase.html new file mode 100644 index 0000000000..3dfbff92a4 --- /dev/null +++ b/docs/RuntimeAPI/html/structLockedBase.html @@ -0,0 +1,131 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: LockedBase< MUTEX_TYPE > Struct Template Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Public Attributes | +List of all members
+
+
LockedBase< MUTEX_TYPE > Struct Template Reference
+
+
+
+Inheritance diagram for LockedBase< MUTEX_TYPE >:
+
+
+ + +ihipCtxCriticalBase_t< MUTEX_TYPE > +ihipStreamCriticalBase_t< MUTEX_TYPE > + +
+ + + + + + +

+Public Member Functions

+void lock ()
 
+void unlock ()
 
+ + + +

+Public Attributes

+MUTEX_TYPE _mutex
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structLockedBase.png b/docs/RuntimeAPI/html/structLockedBase.png new file mode 100644 index 0000000000000000000000000000000000000000..41977d7d5872321ae5be79d6c0c76c16574a9f5c GIT binary patch literal 1146 zcmeAS@N?(olHy`uVBq!ia0y~yVB!F>12~w0q<2E?BOoOa;1lBd|Nnm=^TnI5rTvE{ z09jys;J^Xa&O7ozE=Ng_UoZnu5eQs86=KA|z{2V2;uuoF_;&8?qSXdGuHv60|NWor zFH|#8f69alo*KI!-))%KtE#&syrDmP*0Ps}m?p?6FvNs0IWC;=y+N9Z(LsoVVUh#G zgl{cL88>|IsV)00f1qvB+1itzen(VBOV=-Qy{B!TcYn$o#&4bqf7Ceq^Q42P8{hlm zEq>MS-Nxu|RaUHrUMn2Q*RSg~jLclewD8Qd*~$(lE;BJcPkU(4c)tLLdqcp!Z&shzC5HFbtrYT|)UGf^ch=*l&LX>S)v0qaJ&X{T zX7j^c)Ba?d;mV*XT&MnZJ$Nbg>w^hUN<*MVI^Ut-?t!Og)GLKqw*TZ42oYrX_KH`a zNa(IFzfevz(HXTi&N%_kJnU<&K{keB=qov+S%EgFFn4LRV0~t>m*0Z zAFl>mAr7wAJ?wWZoHeHD|1)xR=?wi!@FPS?pm_rc;3yldtq9KG%EdbD;~{Q6Wa2HamwnhA1zV4 z>L%=*qbfCp_mbk}=1&&yGt*M{yeXMi$K4`-r+@yvPqmf$k9VJwnJg+$6saBVyZ3Xu z+3Y(j>i;lCt^Rdn>7USxvX-%pKbBm6oMgkcu+(1tkm}*THU*ZIKG(Dk9q7~&IXjcR z=1G9quGL}JaxPezlis9xIR5QlmDONCu`P9)3E3gM;%50ec8T?1SzxPURce+{p|nKSDY xVG+x8@XQ|p4kpD$0Vj^@Q<)W81lFzo&%&AY?9Gdw9$-Pi;OXk;vd$@?2>>)>^>_dP literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/structProfTrigger-members.html b/docs/RuntimeAPI/html/structProfTrigger-members.html new file mode 100644 index 0000000000..7467446c8d --- /dev/null +++ b/docs/RuntimeAPI/html/structProfTrigger-members.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ProfTrigger Member List
+
+
+ +

This is the complete list of members for ProfTrigger, including all inherited members.

+ + + + + + +
add(uint64_t trigger) (defined in ProfTrigger)ProfTriggerinline
MAX_TRIGGER (defined in ProfTrigger)ProfTriggerstatic
nextTrigger() (defined in ProfTrigger)ProfTriggerinline
print(int tid) (defined in ProfTrigger)ProfTriggerinline
sort() (defined in ProfTrigger)ProfTriggerinline
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structProfTrigger.html b/docs/RuntimeAPI/html/structProfTrigger.html new file mode 100644 index 0000000000..a62d6dfb48 --- /dev/null +++ b/docs/RuntimeAPI/html/structProfTrigger.html @@ -0,0 +1,127 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ProfTrigger Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Member Functions | +Static Public Attributes | +List of all members
+
+
ProfTrigger Struct Reference
+
+
+ + + + + + + + + + +

+Public Member Functions

+void print (int tid)
 
+uint64_t nextTrigger ()
 
+void add (uint64_t trigger)
 
+void sort ()
 
+ + + +

+Static Public Attributes

+static const uint64_t MAX_TRIGGER = std::numeric_limits<uint64_t>::max()
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/struct____half-members.html b/docs/RuntimeAPI/html/struct____half-members.html new file mode 100644 index 0000000000..5eb5166825 --- /dev/null +++ b/docs/RuntimeAPI/html/struct____half-members.html @@ -0,0 +1,102 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
__half Member List
+
+
+ +

This is the complete list of members for __half, including all inherited members.

+ + +
x (defined in __half)__half
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/struct____half.html b/docs/RuntimeAPI/html/struct____half.html new file mode 100644 index 0000000000..73172bbe87 --- /dev/null +++ b/docs/RuntimeAPI/html/struct____half.html @@ -0,0 +1,111 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: __half Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
__half Struct Reference
+
+
+ + + + +

+Public Attributes

+unsigned x: 16
 
+
The documentation for this struct was generated from the following file:
    +
  • /home/rocm/hip/include/hip/hcc_detail/hip_fp16.h
    • +
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structchar3-members.html b/docs/RuntimeAPI/html/structchar3-members.html new file mode 100644 index 0000000000..84f58bd6c8 --- /dev/null +++ b/docs/RuntimeAPI/html/structchar3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
char3 Member List
+
+
+ +

This is the complete list of members for char3, including all inherited members.

+ + + + +
x (defined in char3)char3
y (defined in char3)char3
z (defined in char3)char3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structchar3.html b/docs/RuntimeAPI/html/structchar3.html new file mode 100644 index 0000000000..e6b4a355e6 --- /dev/null +++ b/docs/RuntimeAPI/html/structchar3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: char3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
char3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+signed char x
 
+signed char y
 
+signed char z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structchar4-members.html b/docs/RuntimeAPI/html/structchar4-members.html new file mode 100644 index 0000000000..37df9b2b9b --- /dev/null +++ b/docs/RuntimeAPI/html/structchar4-members.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
char4 Member List
+
+
+ +

This is the complete list of members for char4, including all inherited members.

+ + + + + + +
val (defined in char4)char4
w (defined in char4)char4
x (defined in char4)char4
y (defined in char4)char4
z (defined in char4)char4
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structchar4.html b/docs/RuntimeAPI/html/structchar4.html new file mode 100644 index 0000000000..ea4aff9f7a --- /dev/null +++ b/docs/RuntimeAPI/html/structchar4.html @@ -0,0 +1,127 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: char4 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
char4 Struct Reference
+
+
+ + + + + + + + + + + + + + + +

+Public Attributes

+union {
+   signed char   x
 
+   signed char   y
 
+   signed char   z
 
+   signed char   w
 
+   unsigned int   val
 
}; 
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structdim3-members.html b/docs/RuntimeAPI/html/structdim3-members.html new file mode 100644 index 0000000000..231fa8d0f9 --- /dev/null +++ b/docs/RuntimeAPI/html/structdim3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
dim3 Member List
+
+
+ +

This is the complete list of members for dim3, including all inherited members.

+ + + + +
xdim3
ydim3
zdim3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structdim3.html b/docs/RuntimeAPI/html/structdim3.html new file mode 100644 index 0000000000..a93484f438 --- /dev/null +++ b/docs/RuntimeAPI/html/structdim3.html @@ -0,0 +1,124 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: dim3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
dim3 Struct Reference
Global enum and defines
+
+
+ +

#include <hip_runtime_api.h>

+ + + + + + + + + + + +

+Public Attributes

+uint32_t x
 x
 
+uint32_t y
 y
 
+uint32_t z
 z
 
+

Detailed Description

+

Struct for data in 3D

+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structdouble1-members.html b/docs/RuntimeAPI/html/structdouble1-members.html new file mode 100644 index 0000000000..39e0794417 --- /dev/null +++ b/docs/RuntimeAPI/html/structdouble1-members.html @@ -0,0 +1,102 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
double1 Member List
+
+
+ +

This is the complete list of members for double1, including all inherited members.

+ + +
x (defined in double1)double1
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structdouble1.html b/docs/RuntimeAPI/html/structdouble1.html new file mode 100644 index 0000000000..2c80f23d2d --- /dev/null +++ b/docs/RuntimeAPI/html/structdouble1.html @@ -0,0 +1,111 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: double1 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
double1 Struct Reference
+
+
+ + + + +

+Public Attributes

+double x
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structdouble3-members.html b/docs/RuntimeAPI/html/structdouble3-members.html new file mode 100644 index 0000000000..5dac28d9cf --- /dev/null +++ b/docs/RuntimeAPI/html/structdouble3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
double3 Member List
+
+
+ +

This is the complete list of members for double3, including all inherited members.

+ + + + +
x (defined in double3)double3
y (defined in double3)double3
z (defined in double3)double3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structdouble3.html b/docs/RuntimeAPI/html/structdouble3.html new file mode 100644 index 0000000000..6ad945ab3b --- /dev/null +++ b/docs/RuntimeAPI/html/structdouble3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: double3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
double3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+double x
 
+double y
 
+double z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structfloat1-members.html b/docs/RuntimeAPI/html/structfloat1-members.html new file mode 100644 index 0000000000..ac66596d17 --- /dev/null +++ b/docs/RuntimeAPI/html/structfloat1-members.html @@ -0,0 +1,102 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
float1 Member List
+
+
+ +

This is the complete list of members for float1, including all inherited members.

+ + +
x (defined in float1)float1
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structfloat1.html b/docs/RuntimeAPI/html/structfloat1.html new file mode 100644 index 0000000000..33b66ec49f --- /dev/null +++ b/docs/RuntimeAPI/html/structfloat1.html @@ -0,0 +1,111 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: float1 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
float1 Struct Reference
+
+
+ + + + +

+Public Attributes

+float x
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structfloat3-members.html b/docs/RuntimeAPI/html/structfloat3-members.html new file mode 100644 index 0000000000..20edd625b0 --- /dev/null +++ b/docs/RuntimeAPI/html/structfloat3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
float3 Member List
+
+
+ +

This is the complete list of members for float3, including all inherited members.

+ + + + +
x (defined in float3)float3
y (defined in float3)float3
z (defined in float3)float3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structfloat3.html b/docs/RuntimeAPI/html/structfloat3.html new file mode 100644 index 0000000000..edcb5b0788 --- /dev/null +++ b/docs/RuntimeAPI/html/structfloat3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: float3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
float3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+float x
 
+float y
 
+float z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipArray-members.html b/docs/RuntimeAPI/html/structhipArray-members.html new file mode 100644 index 0000000000..395a5a1ada --- /dev/null +++ b/docs/RuntimeAPI/html/structhipArray-members.html @@ -0,0 +1,105 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipArray Member List
+
+
+ +

This is the complete list of members for hipArray, including all inherited members.

+ + + + + +
data (defined in hipArray)hipArray
f (defined in hipArray)hipArray
height (defined in hipArray)hipArray
width (defined in hipArray)hipArray
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipArray.html b/docs/RuntimeAPI/html/structhipArray.html new file mode 100644 index 0000000000..feba6e7a2d --- /dev/null +++ b/docs/RuntimeAPI/html/structhipArray.html @@ -0,0 +1,120 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipArray Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipArray Struct Reference
+
+
+ + + + + + + + + + +

+Public Attributes

+unsigned int width
 
+unsigned int height
 
+hipChannelFormatKind f
 
+void * data
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipChannelFormatDesc-members.html b/docs/RuntimeAPI/html/structhipChannelFormatDesc-members.html new file mode 100644 index 0000000000..276fe8b79b --- /dev/null +++ b/docs/RuntimeAPI/html/structhipChannelFormatDesc-members.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipChannelFormatDesc Member List
+
+
+ +

This is the complete list of members for hipChannelFormatDesc, including all inherited members.

+ + + + + + +
f (defined in hipChannelFormatDesc)hipChannelFormatDesc
w (defined in hipChannelFormatDesc)hipChannelFormatDesc
x (defined in hipChannelFormatDesc)hipChannelFormatDesc
y (defined in hipChannelFormatDesc)hipChannelFormatDesc
z (defined in hipChannelFormatDesc)hipChannelFormatDesc
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipChannelFormatDesc.html b/docs/RuntimeAPI/html/structhipChannelFormatDesc.html new file mode 100644 index 0000000000..98bf065336 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipChannelFormatDesc.html @@ -0,0 +1,123 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipChannelFormatDesc Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipChannelFormatDesc Struct Reference
+
+
+ + + + + + + + + + + + +

+Public Attributes

+int x
 
+int y
 
+int z
 
+int w
 
+hipChannelFormatKind f
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipDeviceArch__t-members.html b/docs/RuntimeAPI/html/structhipDeviceArch__t-members.html new file mode 100644 index 0000000000..0810a6ca90 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipDeviceArch__t-members.html @@ -0,0 +1,118 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipDeviceArch_t Member List
+
+
+ +

This is the complete list of members for hipDeviceArch_t, including all inherited members.

+ + + + + + + + + + + + + + + + + + +
has3dGridhipDeviceArch_t
hasDoubleshipDeviceArch_t
hasDynamicParallelismhipDeviceArch_t
hasFloatAtomicAddhipDeviceArch_t
hasFunnelShifthipDeviceArch_t
hasGlobalFloatAtomicExchhipDeviceArch_t
hasGlobalInt32AtomicshipDeviceArch_t
hasGlobalInt64AtomicshipDeviceArch_t
hasSharedFloatAtomicExchhipDeviceArch_t
hasSharedInt32AtomicshipDeviceArch_t
hasSharedInt64AtomicshipDeviceArch_t
hasSurfaceFuncshipDeviceArch_t
hasSyncThreadsExthipDeviceArch_t
hasThreadFenceSystemhipDeviceArch_t
hasWarpBallothipDeviceArch_t
hasWarpShufflehipDeviceArch_t
hasWarpVotehipDeviceArch_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipDeviceArch__t.html b/docs/RuntimeAPI/html/structhipDeviceArch__t.html new file mode 100644 index 0000000000..7c66d3cbc9 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipDeviceArch__t.html @@ -0,0 +1,176 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipDeviceArch_t Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipDeviceArch_t Struct Reference
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Public Attributes

+unsigned hasGlobalInt32Atomics: 1
 32-bit integer atomics for global memory.
 
+unsigned hasGlobalFloatAtomicExch: 1
 32-bit float atomic exch for global memory.
 
+unsigned hasSharedInt32Atomics: 1
 32-bit integer atomics for shared memory.
 
+unsigned hasSharedFloatAtomicExch: 1
 32-bit float atomic exch for shared memory.
 
+unsigned hasFloatAtomicAdd: 1
 32-bit float atomic add in global and shared memory.
 
+unsigned hasGlobalInt64Atomics: 1
 64-bit integer atomics for global memory.
 
+unsigned hasSharedInt64Atomics: 1
 64-bit integer atomics for shared memory.
 
+unsigned hasDoubles: 1
 Double-precision floating point.
 
+unsigned hasWarpVote: 1
 Warp vote instructions (__any, __all).
 
+unsigned hasWarpBallot: 1
 Warp ballot instructions (__ballot).
 
+unsigned hasWarpShuffle: 1
 Warp shuffle operations. (__shfl_*).
 
+unsigned hasFunnelShift: 1
 Funnel two words into one with shift&mask caps.
 
+unsigned hasThreadFenceSystem: 1
 __threadfence_system.
 
+unsigned hasSyncThreadsExt: 1
 __syncthreads_count, syncthreads_and, syncthreads_or.
 
+unsigned hasSurfaceFuncs: 1
 Surface functions.
 
+unsigned has3dGrid: 1
 Grid and group dims are 3D (rather than 2D).
 
+unsigned hasDynamicParallelism: 1
 Dynamic parallelism.
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipDeviceProp__t-members.html b/docs/RuntimeAPI/html/structhipDeviceProp__t-members.html new file mode 100644 index 0000000000..1bdc39fa96 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipDeviceProp__t-members.html @@ -0,0 +1,127 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipDeviceProp_t Member List
+
+
+ +

This is the complete list of members for hipDeviceProp_t, including all inherited members.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + +
archhipDeviceProp_t
canMapHostMemoryhipDeviceProp_t
clockInstructionRatehipDeviceProp_t
clockRatehipDeviceProp_t
computeModehipDeviceProp_t
concurrentKernelshipDeviceProp_t
isMultiGpuBoardhipDeviceProp_t
l2CacheSizehipDeviceProp_t
majorhipDeviceProp_t
maxGridSizehipDeviceProp_t
maxSharedMemoryPerMultiProcessorhipDeviceProp_t
maxThreadsDimhipDeviceProp_t
maxThreadsPerBlockhipDeviceProp_t
maxThreadsPerMultiProcessorhipDeviceProp_t
memoryBusWidthhipDeviceProp_t
memoryClockRatehipDeviceProp_t
minorhipDeviceProp_t
multiProcessorCounthipDeviceProp_t
namehipDeviceProp_t
pciBusIDhipDeviceProp_t
pciDeviceIDhipDeviceProp_t
regsPerBlockhipDeviceProp_t
sharedMemPerBlockhipDeviceProp_t
totalConstMemhipDeviceProp_t
totalGlobalMemhipDeviceProp_t
warpSizehipDeviceProp_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipDeviceProp__t.html b/docs/RuntimeAPI/html/structhipDeviceProp__t.html new file mode 100644 index 0000000000..c3ec16cdec --- /dev/null +++ b/docs/RuntimeAPI/html/structhipDeviceProp__t.html @@ -0,0 +1,216 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipDeviceProp_t Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipDeviceProp_t Struct Reference
+
+
+ +

#include <hip_runtime_api.h>

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

+Public Attributes

+char name [256]
 Device name.
 
+size_t totalGlobalMem
 Size of global memory region (in bytes).
 
+size_t sharedMemPerBlock
 Size of shared memory region (in bytes).
 
+int regsPerBlock
 Registers per block.
 
+int warpSize
 Warp size.
 
+int maxThreadsPerBlock
 Max work items per work group or workgroup max size.
 
+int maxThreadsDim [3]
 Max number of threads in each dimension (XYZ) of a block.
 
+int maxGridSize [3]
 Max grid dimensions (XYZ).
 
+int clockRate
 Max clock frequency of the multiProcessors in khz.
 
+int memoryClockRate
 Max global memory clock frequency in khz.
 
+int memoryBusWidth
 Global memory bus width in bits.
 
+size_t totalConstMem
 Size of shared memory region (in bytes).
 
+int major
 Major compute capability. On HCC, this is an approximation and features may differ from CUDA CC. See the arch feature flags for portable ways to query feature caps.
 
+int minor
 Minor compute capability. On HCC, this is an approximation and features may differ from CUDA CC. See the arch feature flags for portable ways to query feature caps.
 
+int multiProcessorCount
 Number of multi-processors (compute units).
 
+int l2CacheSize
 L2 cache size.
 
+int maxThreadsPerMultiProcessor
 Maximum resident threads per multi-processor.
 
+int computeMode
 Compute mode.
 
+int clockInstructionRate
 Frequency in khz of the timer used by the device-side "clock*" instructions. New for HIP.
 
+hipDeviceArch_t arch
 Architectural feature flags. New for HIP.
 
+int concurrentKernels
 Device can possibly execute multiple kernels concurrently.
 
+int pciBusID
 PCI Bus ID.
 
+int pciDeviceID
 PCI Device ID.
 
+size_t maxSharedMemoryPerMultiProcessor
 Maximum Shared Memory Per Multiprocessor.
 
+int isMultiGpuBoard
 1 if device is on a multi-GPU board, 0 if not.
 
+int canMapHostMemory
 Check whether HIP can map host memory.
 
+

Detailed Description

+

hipDeviceProp

+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipDoubleComplex-members.html b/docs/RuntimeAPI/html/structhipDoubleComplex-members.html new file mode 100644 index 0000000000..8af91a6348 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipDoubleComplex-members.html @@ -0,0 +1,103 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipDoubleComplex Member List
+
+
+ +

This is the complete list of members for hipDoubleComplex, including all inherited members.

+ + + +
x (defined in hipDoubleComplex)hipDoubleComplex
y (defined in hipDoubleComplex)hipDoubleComplex
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipDoubleComplex.html b/docs/RuntimeAPI/html/structhipDoubleComplex.html new file mode 100644 index 0000000000..96315b9ef2 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipDoubleComplex.html @@ -0,0 +1,114 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipDoubleComplex Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipDoubleComplex Struct Reference
+
+
+ + + + + + +

+Public Attributes

+double x
 
+double y
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipFloatComplex-members.html b/docs/RuntimeAPI/html/structhipFloatComplex-members.html new file mode 100644 index 0000000000..1cf90de8f3 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipFloatComplex-members.html @@ -0,0 +1,103 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipFloatComplex Member List
+
+
+ +

This is the complete list of members for hipFloatComplex, including all inherited members.

+ + + +
x (defined in hipFloatComplex)hipFloatComplex
y (defined in hipFloatComplex)hipFloatComplex
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipFloatComplex.html b/docs/RuntimeAPI/html/structhipFloatComplex.html new file mode 100644 index 0000000000..14bb04fc9f --- /dev/null +++ b/docs/RuntimeAPI/html/structhipFloatComplex.html @@ -0,0 +1,114 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipFloatComplex Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipFloatComplex Struct Reference
+
+
+ + + + + + +

+Public Attributes

+float x
 
+float y
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipPointerAttribute__t-members.html b/docs/RuntimeAPI/html/structhipPointerAttribute__t-members.html new file mode 100644 index 0000000000..b600b4fa3f --- /dev/null +++ b/docs/RuntimeAPI/html/structhipPointerAttribute__t-members.html @@ -0,0 +1,107 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
hipPointerAttribute_t Member List
+
+
+ +

This is the complete list of members for hipPointerAttribute_t, including all inherited members.

+ + + + + + + +
allocationFlags (defined in hipPointerAttribute_t)hipPointerAttribute_t
device (defined in hipPointerAttribute_t)hipPointerAttribute_t
devicePointer (defined in hipPointerAttribute_t)hipPointerAttribute_t
hostPointer (defined in hipPointerAttribute_t)hipPointerAttribute_t
isManaged (defined in hipPointerAttribute_t)hipPointerAttribute_t
memoryType (defined in hipPointerAttribute_t)hipPointerAttribute_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structhipPointerAttribute__t.html b/docs/RuntimeAPI/html/structhipPointerAttribute__t.html new file mode 100644 index 0000000000..6c6fb0d5b9 --- /dev/null +++ b/docs/RuntimeAPI/html/structhipPointerAttribute__t.html @@ -0,0 +1,130 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: hipPointerAttribute_t Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
hipPointerAttribute_t Struct Reference
+
+
+ +

#include <hip_runtime_api.h>

+ + + + + + + + + + + + + + +

+Public Attributes

+enum hipMemoryType memoryType
 
+int device
 
+void * devicePointer
 
+void * hostPointer
 
+int isManaged
 
+unsigned allocationFlags
 
+

Detailed Description

+

Pointer attributes

+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structihipEvent__t-members.html b/docs/RuntimeAPI/html/structihipEvent__t-members.html new file mode 100644 index 0000000000..0ed940c02e --- /dev/null +++ b/docs/RuntimeAPI/html/structihipEvent__t-members.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ihipEvent_t Member List
+
+
+ +

This is the complete list of members for ihipEvent_t, including all inherited members.

+ + + + + + +
_flags (defined in ihipEvent_t)ihipEvent_t
_marker (defined in ihipEvent_t)ihipEvent_t
_state (defined in ihipEvent_t)ihipEvent_t
_stream (defined in ihipEvent_t)ihipEvent_t
_timestamp (defined in ihipEvent_t)ihipEvent_t
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structihipEvent__t.html b/docs/RuntimeAPI/html/structihipEvent__t.html new file mode 100644 index 0000000000..7092e3dc44 --- /dev/null +++ b/docs/RuntimeAPI/html/structihipEvent__t.html @@ -0,0 +1,123 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ihipEvent_t Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
ihipEvent_t Struct Reference
+
+
+ + + + + + + + + + + + +

+Public Attributes

+hipEventStatus_t _state
 
+hipStream_t _stream
 
+unsigned _flags
 
+hc::completion_future _marker
 
+uint64_t _timestamp
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structint3-members.html b/docs/RuntimeAPI/html/structint3-members.html new file mode 100644 index 0000000000..fc82310b40 --- /dev/null +++ b/docs/RuntimeAPI/html/structint3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
int3 Member List
+
+
+ +

This is the complete list of members for int3, including all inherited members.

+ + + + +
x (defined in int3)int3
y (defined in int3)int3
z (defined in int3)int3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structint3.html b/docs/RuntimeAPI/html/structint3.html new file mode 100644 index 0000000000..57b6e81773 --- /dev/null +++ b/docs/RuntimeAPI/html/structint3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: int3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
int3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+signed int x
 
+signed int y
 
+signed int z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structlong3-members.html b/docs/RuntimeAPI/html/structlong3-members.html new file mode 100644 index 0000000000..001b1cf117 --- /dev/null +++ b/docs/RuntimeAPI/html/structlong3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
long3 Member List
+
+
+ +

This is the complete list of members for long3, including all inherited members.

+ + + + +
x (defined in long3)long3
y (defined in long3)long3
z (defined in long3)long3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structlong3.html b/docs/RuntimeAPI/html/structlong3.html new file mode 100644 index 0000000000..15f9630874 --- /dev/null +++ b/docs/RuntimeAPI/html/structlong3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: long3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
long3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+long int x
 
+long int y
 
+long int z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structlonglong3-members.html b/docs/RuntimeAPI/html/structlonglong3-members.html new file mode 100644 index 0000000000..d9b512c754 --- /dev/null +++ b/docs/RuntimeAPI/html/structlonglong3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
longlong3 Member List
+
+
+ +

This is the complete list of members for longlong3, including all inherited members.

+ + + + +
x (defined in longlong3)longlong3
y (defined in longlong3)longlong3
z (defined in longlong3)longlong3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structlonglong3.html b/docs/RuntimeAPI/html/structlonglong3.html new file mode 100644 index 0000000000..2167c4f811 --- /dev/null +++ b/docs/RuntimeAPI/html/structlonglong3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: longlong3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
longlong3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+long long int x
 
+long long int y
 
+long long int z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structshort3-members.html b/docs/RuntimeAPI/html/structshort3-members.html new file mode 100644 index 0000000000..e16bf8b5c3 --- /dev/null +++ b/docs/RuntimeAPI/html/structshort3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
short3 Member List
+
+
+ +

This is the complete list of members for short3, including all inherited members.

+ + + + +
x (defined in short3)short3
y (defined in short3)short3
z (defined in short3)short3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structshort3.html b/docs/RuntimeAPI/html/structshort3.html new file mode 100644 index 0000000000..f5075a1305 --- /dev/null +++ b/docs/RuntimeAPI/html/structshort3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: short3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
short3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+signed short x
 
+signed short y
 
+signed short z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structstruct__float-members.html b/docs/RuntimeAPI/html/structstruct__float-members.html new file mode 100644 index 0000000000..5a3aab1708 --- /dev/null +++ b/docs/RuntimeAPI/html/structstruct__float-members.html @@ -0,0 +1,103 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
struct_float Member List
+
+
+ +

This is the complete list of members for struct_float, including all inherited members.

+ + + +
f (defined in struct_float)struct_float
u (defined in struct_float)struct_float
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structstruct__float.html b/docs/RuntimeAPI/html/structstruct__float.html new file mode 100644 index 0000000000..438a7868d7 --- /dev/null +++ b/docs/RuntimeAPI/html/structstruct__float.html @@ -0,0 +1,118 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: struct_float Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
struct_float Struct Reference
+
+
+ + + + + + + + + +

+Public Attributes

+union {
+   float   f
 
+   unsigned   u
 
}; 
 
+
The documentation for this struct was generated from the following file:
    +
  • /home/rocm/hip/src/hip_fp16.cpp
    • +
+
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structtextureReference-members.html b/docs/RuntimeAPI/html/structtextureReference-members.html new file mode 100644 index 0000000000..09b0950f1f --- /dev/null +++ b/docs/RuntimeAPI/html/structtextureReference-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
textureReference Member List
+
+
+ +

This is the complete list of members for textureReference, including all inherited members.

+ + + + +
channelDesc (defined in textureReference)textureReference
filterMode (defined in textureReference)textureReference
normalized (defined in textureReference)textureReference
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structtextureReference.html b/docs/RuntimeAPI/html/structtextureReference.html new file mode 100644 index 0000000000..23cdba0ce7 --- /dev/null +++ b/docs/RuntimeAPI/html/structtextureReference.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: textureReference Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
textureReference Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+hipTextureFilterMode filterMode
 
+bool normalized
 
+hipChannelFormatDesc channelDesc
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structuchar3-members.html b/docs/RuntimeAPI/html/structuchar3-members.html new file mode 100644 index 0000000000..8a71a7aad5 --- /dev/null +++ b/docs/RuntimeAPI/html/structuchar3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
uchar3 Member List
+
+
+ +

This is the complete list of members for uchar3, including all inherited members.

+ + + + +
x (defined in uchar3)uchar3
y (defined in uchar3)uchar3
z (defined in uchar3)uchar3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structuchar3.html b/docs/RuntimeAPI/html/structuchar3.html new file mode 100644 index 0000000000..dce48635b2 --- /dev/null +++ b/docs/RuntimeAPI/html/structuchar3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: uchar3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
uchar3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+unsigned char x
 
+unsigned char y
 
+unsigned char z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structuchar4-members.html b/docs/RuntimeAPI/html/structuchar4-members.html new file mode 100644 index 0000000000..1dba273012 --- /dev/null +++ b/docs/RuntimeAPI/html/structuchar4-members.html @@ -0,0 +1,106 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
uchar4 Member List
+
+
+ +

This is the complete list of members for uchar4, including all inherited members.

+ + + + + + +
val (defined in uchar4)uchar4
w (defined in uchar4)uchar4
x (defined in uchar4)uchar4
y (defined in uchar4)uchar4
z (defined in uchar4)uchar4
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structuchar4.html b/docs/RuntimeAPI/html/structuchar4.html new file mode 100644 index 0000000000..c4a79303d6 --- /dev/null +++ b/docs/RuntimeAPI/html/structuchar4.html @@ -0,0 +1,127 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: uchar4 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
uchar4 Struct Reference
+
+
+ + + + + + + + + + + + + + + +

+Public Attributes

+union {
+   unsigned char   x
 
+   unsigned char   y
 
+   unsigned char   z
 
+   unsigned char   w
 
+   unsigned int   val
 
}; 
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structuint3-members.html b/docs/RuntimeAPI/html/structuint3-members.html new file mode 100644 index 0000000000..1d2bc9ffd4 --- /dev/null +++ b/docs/RuntimeAPI/html/structuint3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
uint3 Member List
+
+
+ +

This is the complete list of members for uint3, including all inherited members.

+ + + + +
x (defined in uint3)uint3
y (defined in uint3)uint3
z (defined in uint3)uint3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structuint3.html b/docs/RuntimeAPI/html/structuint3.html new file mode 100644 index 0000000000..efbd9f5194 --- /dev/null +++ b/docs/RuntimeAPI/html/structuint3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: uint3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
uint3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+unsigned int x
 
+unsigned int y
 
+unsigned int z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structulong3-members.html b/docs/RuntimeAPI/html/structulong3-members.html new file mode 100644 index 0000000000..59ae827ac7 --- /dev/null +++ b/docs/RuntimeAPI/html/structulong3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ulong3 Member List
+
+
+ +

This is the complete list of members for ulong3, including all inherited members.

+ + + + +
x (defined in ulong3)ulong3
y (defined in ulong3)ulong3
z (defined in ulong3)ulong3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structulong3.html b/docs/RuntimeAPI/html/structulong3.html new file mode 100644 index 0000000000..490a0233a0 --- /dev/null +++ b/docs/RuntimeAPI/html/structulong3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ulong3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
ulong3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+unsigned long x
 
+unsigned long y
 
+unsigned long z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structulonglong3-members.html b/docs/RuntimeAPI/html/structulonglong3-members.html new file mode 100644 index 0000000000..99328e4c6c --- /dev/null +++ b/docs/RuntimeAPI/html/structulonglong3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ulonglong3 Member List
+
+
+ +

This is the complete list of members for ulonglong3, including all inherited members.

+ + + + +
x (defined in ulonglong3)ulonglong3
y (defined in ulonglong3)ulonglong3
z (defined in ulonglong3)ulonglong3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structulonglong3.html b/docs/RuntimeAPI/html/structulonglong3.html new file mode 100644 index 0000000000..5906b64824 --- /dev/null +++ b/docs/RuntimeAPI/html/structulonglong3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ulonglong3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
ulonglong3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+unsigned long long int x
 
+unsigned long long int y
 
+unsigned long long int z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structushort3-members.html b/docs/RuntimeAPI/html/structushort3-members.html new file mode 100644 index 0000000000..05559529c4 --- /dev/null +++ b/docs/RuntimeAPI/html/structushort3-members.html @@ -0,0 +1,104 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: Member List + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+
ushort3 Member List
+
+
+ +

This is the complete list of members for ushort3, including all inherited members.

+ + + + +
x (defined in ushort3)ushort3
y (defined in ushort3)ushort3
z (defined in ushort3)ushort3
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/structushort3.html b/docs/RuntimeAPI/html/structushort3.html new file mode 100644 index 0000000000..2a24ed4382 --- /dev/null +++ b/docs/RuntimeAPI/html/structushort3.html @@ -0,0 +1,117 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: ushort3 Struct Reference + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ +
+
+
+Public Attributes | +List of all members
+
+
ushort3 Struct Reference
+
+
+ + + + + + + + +

+Public Attributes

+unsigned short x
 
+unsigned short y
 
+unsigned short z
 
+
The documentation for this struct was generated from the following file: +
+ +
+Generated on Mon Dec 19 2016 04:03:44 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + diff --git a/docs/RuntimeAPI/html/sync_off.png b/docs/RuntimeAPI/html/sync_off.png new file mode 100644 index 0000000000000000000000000000000000000000..3b443fc62892114406e3d399421b2a881b897acc GIT binary patch literal 853 zcmV-b1FHOqP)oT|#XixUYy%lpuf3i8{fX!o zUyDD0jOrAiT^tq>fLSOOABs-#u{dV^F$b{L9&!2=9&RmV;;8s^x&UqB$PCj4FdKbh zoB1WTskPUPu05XzFbA}=KZ-GP1fPpAfSs>6AHb12UlR%-i&uOlTpFNS7{jm@mkU1V zh`nrXr~+^lsV-s1dkZOaI|kYyVj3WBpPCY{n~yd%u%e+d=f%`N0FItMPtdgBb@py; zq@v6NVArhyTC7)ULw-Jy8y42S1~4n(3LkrW8mW(F-4oXUP3E`e#g**YyqI7h-J2zK zK{m9##m4ri!7N>CqQqCcnI3hqo1I;Yh&QLNY4T`*ptiQGozK>FF$!$+84Z`xwmeMh zJ0WT+OH$WYFALEaGj2_l+#DC3t7_S`vHpSivNeFbP6+r50cO8iu)`7i%Z4BTPh@_m3Tk!nAm^)5Bqnr%Ov|Baunj#&RPtRuK& z4RGz|D5HNrW83-#ydk}tVKJrNmyYt-sTxLGlJY5nc&Re zU4SgHNPx8~Yxwr$bsju?4q&%T1874xxzq+_%?h8_ofw~(bld=o3iC)LUNR*BY%c0y zWd_jX{Y8`l%z+ol1$@Qa?Cy!(0CVIEeYpKZ`(9{z>3$CIe;pJDQk$m3p}$>xBm4lb zKo{4S)`wdU9Ba9jJbVJ0C=SOefZe%d$8=2r={nu<_^a3~>c#t_U6dye5)JrR(_a^E f@}b6j1K9lwFJq@>o)+Ry00000NkvXXu0mjfWa5j* literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/sync_on.png b/docs/RuntimeAPI/html/sync_on.png new file mode 100644 index 0000000000000000000000000000000000000000..e08320fb64e6fa33b573005ed6d8fe294e19db76 GIT binary patch literal 845 zcmV-T1G4;yP)Y;xxyHF2B5Wzm| zOOGupOTn@c(JmBOl)e;XMNnZuiTJP>rM8<|Q`7I_))aP?*T)ow&n59{}X4$3Goat zgjs?*aasfbrokzG5cT4K=uG`E14xZl@z)F={P0Y^?$4t z>v!teRnNZym<6h{7sLyF1V0HsfEl+l6TrZpsfr1}luH~F7L}ktXu|*uVX^RG$L0`K zWs3j|0tIvVe(N%_?2{(iCPFGf#B6Hjy6o&}D$A%W%jfO8_W%ZO#-mh}EM$LMn7joJ z05dHr!5Y92g+31l<%i1(=L1a1pXX+OYnalY>31V4K}BjyRe3)9n#;-cCVRD_IG1fT zOKGeNY8q;TL@K{dj@D^scf&VCs*-Jb>8b>|`b*osv52-!A?BpbYtTQBns5EAU**$m zSnVSm(teh>tQi*S*A>#ySc=n;`BHz`DuG4&g4Kf8lLhca+zvZ7t7RflD6-i-mcK=M z!=^P$*u2)bkY5asG4gsss!Hn%u~>}kIW`vMs%lJLH+u*9<4PaV_c6U`KqWXQH%+Nu zTv41O(^ZVi@qhjQdG!fbZw&y+2o!iYymO^?ud3{P*HdoX83YV*Uu_HB=?U&W9%AU# z80}k1SS-CXTU7dcQlsm<^oYLxVSseqY6NO}dc`Nj?8vrhNuCdm@^{a3AQ_>6myOj+ z`1RsLUXF|dm|3k7s2jD(B{rzE>WI2scH8i1;=O5Cc9xB3^aJk%fQjqsu+kH#0=_5a z0nCE8@dbQa-|YIuUVvG0L_IwHMEhOj$Mj4Uq05 X8=0q~qBNan00000NkvXXu0mjfptF>5 literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/tab_a.png b/docs/RuntimeAPI/html/tab_a.png new file mode 100644 index 0000000000000000000000000000000000000000..3b725c41c5a527a3a3e40097077d0e206a681247 GIT binary patch literal 142 zcmeAS@N?(olHy`uVBq!ia0vp^j6kfy!2~3aiye;!QlXwMjv*C{Z|8b*H5dputLHD# z=<0|*y7z(Vor?d;H&?EG&cXR}?!j-Lm&u1OOI7AIF5&c)RFE;&p0MYK>*Kl@eiymD r@|NpwKX@^z+;{u_Z~trSBfrMKa%3`zocFjEXaR$#tDnm{r-UW|TZ1%4 literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/tab_b.png b/docs/RuntimeAPI/html/tab_b.png new file mode 100644 index 0000000000000000000000000000000000000000..e2b4a8638cb3496a016eaed9e16ffc12846dea18 GIT binary patch literal 169 zcmeAS@N?(olHy`uVBq!ia0vp^j6kfy!2~3aiye;!QU#tajv*C{Z}0l@H7kg?K0Lnr z!j&C6_(~HV9oQ0Pa6x{-v0AGV_E?vLn=ZI-;YrdjIl`U`uzuDWSP?o#Dmo{%SgM#oan kX~E1%D-|#H#QbHoIja2U-MgvsK&LQxy85}Sb4q9e0Efg%P5=M^ literal 0 HcmV?d00001 diff --git a/docs/RuntimeAPI/html/tabs.css b/docs/RuntimeAPI/html/tabs.css new file mode 100644 index 0000000000..9cf578f23a --- /dev/null +++ b/docs/RuntimeAPI/html/tabs.css @@ -0,0 +1,60 @@ +.tabs, .tabs2, .tabs3 { + background-image: url('tab_b.png'); + width: 100%; + z-index: 101; + font-size: 13px; + font-family: 'Lucida Grande',Geneva,Helvetica,Arial,sans-serif; +} + +.tabs2 { + font-size: 10px; +} +.tabs3 { + font-size: 9px; +} + +.tablist { + margin: 0; + padding: 0; + display: table; +} + +.tablist li { + float: left; + display: table-cell; + background-image: url('tab_b.png'); + line-height: 36px; + list-style: none; +} + +.tablist a { + display: block; + padding: 0 20px; + font-weight: bold; + background-image:url('tab_s.png'); + background-repeat:no-repeat; + background-position:right; + color: #283A5D; + text-shadow: 0px 1px 1px rgba(255, 255, 255, 0.9); + text-decoration: none; + outline: none; +} + +.tabs3 .tablist a { + padding: 0 10px; +} + +.tablist a:hover { + background-image: url('tab_h.png'); + background-repeat:repeat-x; + color: #fff; + text-shadow: 0px 1px 1px rgba(0, 0, 0, 1.0); + text-decoration: none; +} + +.tablist li.current a { + background-image: url('tab_a.png'); + background-repeat:repeat-x; + color: #fff; + text-shadow: 0px 1px 1px rgba(0, 0, 0, 1.0); +} diff --git a/docs/RuntimeAPI/html/trace__helper_8h_source.html b/docs/RuntimeAPI/html/trace__helper_8h_source.html new file mode 100644 index 0000000000..dcd5ea1eb7 --- /dev/null +++ b/docs/RuntimeAPI/html/trace__helper_8h_source.html @@ -0,0 +1,238 @@ + + + + + + +HIP: Heterogenous-computing Interface for Portability: /home/rocm/hip/src/trace_helper.h Source File + + + + + + + + + +
+
+ + + + + + +
+
HIP: Heterogenous-computing Interface for Portability +
+
+
+ + + + + + + + + +
+ +
+ + +
+
+
+
trace_helper.h
+
+
+
1 /*
+
2 Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved.
+
3 
+
4 Permission is hereby granted, free of charge, to any person obtaining a copy
+
5 of this software and associated documentation files (the "Software"), to deal
+
6 in the Software without restriction, including without limitation the rights
+
7 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+
8 copies of the Software, and to permit persons to whom the Software is
+
9 furnished to do so, subject to the following conditions:
+
10 
+
11 The above copyright notice and this permission notice shall be included in
+
12 all copies or substantial portions of the Software.
+
13 
+
14 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+
15 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+
16 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+
17 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+
18 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+
19 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+
20 THE SOFTWARE.
+
21 */
+
22 
+
23 //#pragma once
+
24 
+
25 #ifndef TRACE_HELPER_H
+
26 #define TRACE_HELPER_H
+
27 
+
28 #include <iostream>
+
29 #include <iomanip>
+
30 #include <string>
+
31 
+
32 //---
+
33 // Helper functions to convert HIP function arguments into strings.
+
34 // Handles POD data types as well as enumerations (ie hipMemcpyKind).
+
35 // The implementation uses C++11 variadic templates and template specialization.
+
36 // The hipMemcpyKind example below is a good example that shows how to implement conversion for a new HSA type.
+
37 
+
38 
+
39 // Handy macro to convert an enumeration to a stringified version of same:
+
40 #define CASE_STR(x) case x: return #x;
+
41 
+
42 
+
43 // Building block functions:
+
44 template <typename T>
+
45 inline std::string ToHexString(T v)
+
46 {
+
47  std::ostringstream ss;
+
48  ss << "0x" << std::hex << v;
+
49  return ss.str();
+
50 };
+
51 
+
52 
+
53 //---
+
54 // Template overloads for ToString to handle specific types
+
55 
+
56 // This is the default which works for most types:
+
57 template <typename T>
+
58 inline std::string ToString(T v)
+
59 {
+
60  std::ostringstream ss;
+
61  ss << v;
+
62  return ss.str();
+
63 };
+
64 
+
65 
+
66 // hipEvent_t specialization. TODO - maybe add an event ID for debug?
+
67 template <>
+
68 inline std::string ToString(hipEvent_t v)
+
69 {
+
70  std::ostringstream ss;
+
71  ss << v;
+
72  return ss.str();
+
73 };
+
74 
+
75 
+
76 
+
77 // hipStream_t
+
78 template <>
+
79 inline std::string ToString(hipStream_t v)
+
80 {
+
81  std::ostringstream ss;
+
82  if (v == NULL) {
+
83  ss << "stream:<null>";
+
84  } else {
+
85  ss << *v;
+
86  }
+
87 
+
88  return ss.str();
+
89 };
+
90 
+
91 // hipMemcpyKind specialization
+
92 template <>
+
93 inline std::string ToString(hipMemcpyKind v)
+
94 {
+
95  switch(v) {
+
96  CASE_STR(hipMemcpyHostToHost);
+
97  CASE_STR(hipMemcpyHostToDevice);
+
98  CASE_STR(hipMemcpyDeviceToHost);
+
99  CASE_STR(hipMemcpyDeviceToDevice);
+
100  CASE_STR(hipMemcpyDefault);
+
101  default : return ToHexString(v);
+
102  };
+
103 };
+
104 
+
105 
+
106 template <>
+
107 inline std::string ToString(hipError_t v)
+
108 {
+
109  return ihipErrorString(v);
+
110 };
+
111 
+
112 
+
113 // Catch empty arguments case
+
114 inline std::string ToString()
+
115 {
+
116  return ("");
+
117 }
+
118 
+
119 
+
120 //---
+
121 // C++11 variadic template - peels off first argument, converts to string, and calls itself again to peel the next arg.
+
122 // Strings are automatically separated by comma+space.
+
123 template <typename T, typename... Args>
+
124 inline std::string ToString(T first, Args... args)
+
125 {
+
126  return ToString(first) + ", " + ToString(args...) ;
+
127 }
+
128 
+
129 #endif
+
hipMemcpyHostToDevice
Host-to-Device Copy.
Definition: hip_runtime_api.h:167
+
hipMemcpyDeviceToHost
Device-to-Host Copy.
Definition: hip_runtime_api.h:168
+
hipError_t
hipError_t
Definition: hip_runtime_api.h:152
+
hipMemcpyKind
hipMemcpyKind
Definition: hip_runtime_api.h:165
+
hipMemcpyDeviceToDevice
Device-to-Device Copy.
Definition: hip_runtime_api.h:169
+
ihipEvent_t
Definition: hip_hcc.h:558
+
hipMemcpyDefault
Runtime will automatically determine copy-kind based on virtual addresses.
Definition: hip_runtime_api.h:170
+
ihipStream_t
Definition: hip_hcc.h:463
+
hipMemcpyHostToHost
Host-to-Host Copy.
Definition: hip_runtime_api.h:166
+
+ +
+Generated on Mon Dec 19 2016 04:03:43 for HIP: Heterogenous-computing Interface for Portability by   +doxygen + 1.8.6 +
+ + From bb8c51a12902943f193c6c1159933c5e5d8d17f7 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Fri, 7 Apr 2017 15:40:09 +0530 Subject: [PATCH 003/108] Merge branch 'amd-develop' into amd-master Change-Id: I53d5a8916d769c4f0fe60d2ee3b240551da80b4f (cherry picked from commit 01c523f6c9fc5a66d5f6e60d5efe78c29c9fe317) --- docs/markdown/hip_bugs.md | 188 +++++++++++++++--- docs/markdown/hip_faq.md | 9 +- docs/markdown/hip_kernel_language.md | 1 + docs/markdown/hip_porting_guide.md | 9 +- docs/markdown/hip_profiling.md | 46 +++-- hipify-clang/src/Cuda2Hip.cpp | 21 +- include/hip/hcc_detail/grid_launch_GGL.hpp | 23 ++- include/hip/hcc_detail/hip_complex.h | 53 +++-- include/hip/hcc_detail/hip_fp16.h | 2 +- include/hip/hcc_detail/hip_hcc.h | 43 +++- include/hip/hcc_detail/hip_runtime_api.h | 23 +-- include/hip/hcc_detail/hip_vector_types.h | 2 +- packaging/create_hip_samples_installer.sh | 23 --- packaging/hip_doc.txt | 15 +- samples/0_Intro/bit_extract/Makefile | 4 - samples/1_Utils/hipCommander/Makefile | 3 - samples/1_Utils/hipCommander/hipCommander.cpp | 1 - src/hip_event.cpp | 60 +++++- src/hip_hcc.cpp | 17 -- src/hip_hcc_internal.h | 32 ++- src/hip_memory.cpp | 13 +- src/hip_module.cpp | 32 ++- tests/src/deviceLib/hipDeviceMemcpy.cpp | 45 +++-- 23 files changed, 451 insertions(+), 214 deletions(-) delete mode 100755 packaging/create_hip_samples_installer.sh diff --git a/docs/markdown/hip_bugs.md b/docs/markdown/hip_bugs.md index e15c37fc54..73133843bc 100644 --- a/docs/markdown/hip_bugs.md +++ b/docs/markdown/hip_bugs.md @@ -1,49 +1,177 @@ -# HIP Bugs +# HIP Bugs -- [Errors related to undefined reference to `__hcLaunchKernel__***__grid_launch_parm**](#errors-related-to-undefined-reference-to-hclaunchkernel__grid_launch_parm) -- [Application hangs after a hipLaunchKernel call](#what-if-i-see-application-hangs-after-a-hiplaunchkernel-call) +- [Errors related to undefined reference to `__hcLaunchKernel__***__grid_launch_parm**`](#errors-related-to-undefined-reference-to-__hclaunchkernel____grid_launch_parm) - [What is the current limitation of HIP Generic Grid Launch method?](#what-is-the-current-limitation-of-hip-generic-grid-launch-method) +- [Errors related to `no matching constructor`](#errors-related-to-no-matching-constructor) +- [HIP is more restrictive in enforcing restrictions](#hip-is-more-restrictive-in-enforcing-restrictions) -### Errors related to undefined reference to `__hcLaunchKernel__***__grid_launch_parm** +### Errors related to undefined reference to `__hcLaunchKernel__***__grid_launch_parm**` Some common code practices may lead to hipcc generating a error with the form : undefined reference to `__hcLaunchKernel__ZN15vecAddNamespace6vecAddIidEEv16grid_launch_parmPT0_S3_S3_T_ -To workaround, try: -- Avoid calling hcLaunchKernel from a function with the __host__ attribute -__host__ MyFunc(…) { -hipLaunchKernel(myKernel, …) +Suggested workarounds: - Avoid use of static with kernel definition: +```c++ static __global__ MyKernel -- Avoid defining kernels in anonymous namespace +``` + +- Avoid defining kernels in anonymous namespace : +```c++ namespace { -__global__ MyKernel … -- Avoid calling member functions - -If hipLaunchKernel takes parameters that request explicitly memcpy, then it will cause application hang. -Reason is that the hipLaunchKernel macro locks the stream. -If kernel paramters are actually function calls which invoke other hip apis (i.e. memcpy) to the same stream, then deadlock occurs. - -To workaround, try: -Move the function calls so they occur outside the hipLaunchKernel macro, store results in temps, then use the tems inside the kernel. - + __global__ MyKernel +} ``` -// Example pseudo code causing system hang: -// "bottom[0]->gpu_data()" calls hipMemcpy() implicitly and using the same stream, cause deadlock condition. -hipLaunchKernel(HIP_KERNEL_NAME(LRNComputeDiff),dim3(CAFFE_GET_BLOCKS(n_threads)), dim3(CAFFE_HIP_NUM_THREADS), 0, 0, n_threads, - bottom[0]->gpu_data()); - -// Move "gpu_data()" ouside of hipLaunchKernel to avoid hang. -auto bot_gpu_data = bottom[0]->gpu_data(); -hipLaunchKernel( LRNComputeDiff, dim3(CAFFE_GET_BLOCKS(n_threads)), dim3(CAFFE_HIP_NUM_THREADS), 0, 0, n_threads, - bot_gpu_data); -``` ### What is the current limitation of HIP Generic Grid Launch method? 1. __global__ functions cannot be marked as static or put in an unnamed namespace i.e. they cannot be given internal linkage (this would clash with __attribute__((weak))); -2. using the macro based dispatch mechanism i.e. hipLaunchKernel* only works for functions that take no more than 20 arguments (this limit can be increased up to 126, and is temporary until we can enable C++14 mode and use variadic generic lambdas); no such limitation applies do dispatching directly through grid_launch. \ No newline at end of file +2. using the macro based dispatch mechanism i.e. hipLaunchKernel* only works for functions that take no more than 20 arguments (this limit can be increased up to 126, and is temporary until we can enable C++14 mode and use variadic generic lambdas); no such limitation applies do dispatching directly through grid_launch. + + +### Errors related to `no matching constructor` + +The symptom is the compiler would complain about errors like `no matching constructor` for classes/structs passed as arguments into a GPU kernel. Often, this is caused by a design limitation in HCC where array-typed member variables inside a class/struct can’t be correctly passed into GPU kernels. To mitigate this issue, a custom serializer/deserializer pair is provided. + +For example, `Foo` in the code snippets below contains an array-typed member variable `table`, which would fail the compiler if used as a kernel argument. + +``` +struct Foo { + // table is an array, which makes foo + int table[3]; +}; +``` + +An workaround is to provide a custom serializer on CPU side, and append the contents of the array as kernel arguments: + +``` + +struct Foo { + int table[3]; + + // user-provided CPU serializer + // must append the contents of the array member as kernel arguments +#ifdef __HCC__ + __attribute__((annotate(“serialize”))) + void __cxxamp_serialize(Kalmar::Serialize &s) const { + for (int i = 0; i < 3; ++i) + s.Append(sizeof(int), &table[i]); + } +#endif +}; +``` + +Then, provide a custom deserializer on GPU side, to help reconstruct the array within GPU kernels. Notice that the deserializer can not be a function template, and should have scalar-typed parameters of the number equals to the length of the array-typed member variable. For example: + +``` +struct Foo { + int table[3]; + + // user-provided GPU deserializer + // table has 3 int elements, so deserializer must have 3 int parameters. +#ifdef __HCC__ + __attribute__((annotate(“user_deserialize”))) + Foo(int x0, int x1, int x2) [[cpu]][[hc]] { + table[0] = x0; + table[1] = x1; + table[2] = x2; + } +#endif + +#ifdef __HCC__ + __attribute__((annotate(“serialize”))) + void __cxxamp_serialize(Kalmar::Serialize &s) const { + s.Append(sizeof(int), &table[0]); + s.Append(sizeof(int), &table[1]); + s.Append(sizeof(int), &table[2]); + } +#endif +}; +``` + + +Rather than create serializer functions, another workaround is to pass the member fields from the structure as simple data types. + + +### HIP is more restrictive in enforcing restrictions +The language specification for HIP and CUDA forbid calling a +`__device__` function in a `__host__` context. In practice, you may observe +differences in the strictness of this restriction, with HIP exhibiting a tighter +adherence to the specification and thus less tolerant of infringing code. The +solution is to ensure that all functions which are called in a +`__device__` context are correctly annotated to reflect it. An interesting case +where these differences emerge is shown below. This relies on a the common +[C++ Member Detector idiom][1], as it would be implemented pre C++11): + +```c++ +#include +#include + +struct aye { bool a[1]; }; +struct nay { bool a[2]; }; + +// Dual restriction is necessary in HIP if the detector is to work for +// __device__ contexts as well as __host__ ones. NVCC is less strict. +template +__host__ __device__ +const T& cref_t(); + +template +struct Has_call_operator { + // Dual restriction is necessary in HIP if the detector is to work for + // __device__ contexts as well as __host__ ones. NVCC is less strict. + template + __host__ __device__ + static + aye test( + C const *, + typename std::enable_if< + (sizeof(cref_t().operator()()) > 0)>::type* = nullptr); + static + nay test(...); + + enum { value = sizeof(test(static_cast(0))) == sizeof(aye) }; +}; + +template::value> +struct Wrapper { + template + V f() const { return T{1}; } +}; + + +template +struct Wrapper { + template + V f() const { return T{10}; } +}; + +// This specialisation will yield a compile-time error, if selected. +template +struct Wrapper {}; + +template +struct Functor; + +template<> struct Functor { + __device__ + float operator()() const { return 42.0f; } +}; + +__device__ +void this_will_not_compile_if_detector_is_not_marked_device() +{ + float f = Wrapper>().f(); +} + +__host__ +void this_will_not_compile_if_detector_is_marked_device_only() +{ + float f = Wrapper>().f(); +} +``` +[1]: https://en.wikibooks.org/wiki/More_C%2B%2B_Idioms/Member_Detector diff --git a/docs/markdown/hip_faq.md b/docs/markdown/hip_faq.md index 8ccb458103..e316d449ef 100644 --- a/docs/markdown/hip_faq.md +++ b/docs/markdown/hip_faq.md @@ -4,7 +4,7 @@ - [What APIs and features does HIP support?](#what-apis-and-features-does-hip-support) - [What is not supported?](#what-is-not-supported) - * [Run-time features](#run-time-features) + * [Runtime/Driver API features](#runtimedriver-api-features) * [Kernel language features](#kernel-language-features) - [Is HIP a drop-in replacement for CUDA?](#is-hip-a-drop-in-replacement-for-cuda) - [What specific version of CUDA does HIP support?](#what-specific-version-of-cuda-does-hip-support) @@ -23,10 +23,11 @@ - [On HCC, can I link HIP code with host code compiled with another compiler such as gcc, icc, or clang ?](#on-hcc-can-i-link-hip-code-with-host-code-compiled-with-another-compiler-such-as-gcc-icc-or-clang-) - [HIP detected my platform (hcc vs nvcc) incorrectly - what should I do?](#hip-detected-my-platform-hcc-vs-nvcc-incorrectly---what-should-i-do) - [Can I install both CUDA SDK and HCC on same machine?](#can-i-install-both-cuda-sdk-and-hcc-on-same-machine) +- [On CUDA, can I mix CUDA code with HIP code?](#on-cuda-can-i-mix-cuda-code-with-hip-code) +- [On HCC, can I use HC functionality with HIP?](#on-hcc-can-i-use-hc-functionality-with-hip) - [How do I trace HIP application flow?](#how-do-i-trace-hip-application-flow) - * [Using CodeXL markers for HIP Functions](#using-codexl-markers-for-hip-functions) - * [Using HIP_TRACE_API](#using-hip_trace_api) -- [How do I enable HIP Generic Grid Launch option?](#how-do-i-enable-hip-generic-grid-launch-option) +- [What if HIP generates error of "symbol multiply defined!" only on AMD machine?](#what-if-hip-generates-error-of-symbol-multiply-defined-only-on-amd-machine) +- [How do I disable HIP Generic Grid Launch option?](#how-do-i-disable-hip-generic-grid-launch-option) diff --git a/docs/markdown/hip_kernel_language.md b/docs/markdown/hip_kernel_language.md index 0c7f3c8d25..3cb7b17a0c 100644 --- a/docs/markdown/hip_kernel_language.md +++ b/docs/markdown/hip_kernel_language.md @@ -44,6 +44,7 @@ - [Pragma Unroll](#pragma-unroll) - [In-Line Assembly](#in-line-assembly) - [C++ Support](#c-support) +- [Kernel Compilation](#kernel-compilation) diff --git a/docs/markdown/hip_porting_guide.md b/docs/markdown/hip_porting_guide.md index 9f20d12423..72f6384f6d 100644 --- a/docs/markdown/hip_porting_guide.md +++ b/docs/markdown/hip_porting_guide.md @@ -21,6 +21,7 @@ and provides practical suggestions on how to port CUDA code and work through com * [Device-Architecture Properties](#device-architecture-properties) * [Table of Architecture Properties](#table-of-architecture-properties) - [Finding HIP](#finding-hip) +- [hipLaunchKernel](#hiplaunchkernel) - [Compiler Options](#compiler-options) - [Linking Issues](#linking-issues) * [Linking With hipcc](#linking-with-hipcc) @@ -31,9 +32,11 @@ and provides practical suggestions on how to port CUDA code and work through com * [Using a Standard C++ Compiler](#using-a-standard-c-compiler) + [cuda.h](#cudah) * [Choosing HIP File Extensions](#choosing-hip-file-extensions) - * [Workarounds](#workarounds) - + [warpSize](#warpsize) - + [Textures and Cache Control](#textures-and-cache-control) +- [Workarounds](#workarounds) + * [warpSize](#warpsize) +- [memcpyToSymbol](#memcpytosymbol) +- [threadfence_system](#threadfence_system) + * [Textures and Cache Control](#textures-and-cache-control) - [More Tips](#more-tips) * [HIPTRACE Mode](#hiptrace-mode) * [Environment Variables](#environment-variables) diff --git a/docs/markdown/hip_profiling.md b/docs/markdown/hip_profiling.md index 463c9c13b3..6e5cde700d 100644 --- a/docs/markdown/hip_profiling.md +++ b/docs/markdown/hip_profiling.md @@ -4,26 +4,32 @@ This section describes the profiling and debugging capabilities that HIP provide Profiling information can viewed in the CodeXL visualization tool or printed directly to stderr as the application runs. This document starts with some of the general capabilities of CodeXL and then describes some of the additional HIP marker and debug features. - * [CodeXL Profiling](#codexl-profiling) - * [Collecting and Viewing Traces](#collecting-and-viewing-traces) - * [Using rocm-profiler timestamp profiling](#using-rocm-profiler-timestamp-profiling) - * [Using rocm-profiler performance counter collection:](#using-rocm-profiler-performance-counter-collection) - * [Using CodeXL to view profiling results:](#using-codexl-to-view-profiling-results) - * [More information on CodeXL](#more-information-on-codexl) - * [HIP Markers](#hip-markers) - * [Profiling HIP APIs](#profiling-hip-apis) - * [Adding markers to applications](#adding-markers-to-applications) - * [Additional HIP Profiling Features](#additional-hip-profiling-features) - * [Demangling C Kernel Names](#demangling-c-kernel-names) - * [Controlling when profiling starts and ends](#controlling-when-profiling-starts-and-ends) - * [Reducing timeline trace output file size](#reducing-timeline-trace-output-file-size) - * [How to enable profiling at HIP build time](#how-to-enable-profiling-at-hip-build-time) - * [Tracing and Debug](#tracing-and-debug) - * [Tracing HIP APIs](#tracing-hip-apis) - * [Color](#color) - * [Using HIP_DB](#using-hip_db) - * [Using ltrace](#using-ltrace) - * [Chicken bits](#chicken-bits) + + +- [CodeXL Profiling](#codexl-profiling) + * [Collecting and Viewing Traces](#collecting-and-viewing-traces) + + [Using rocm-profiler timestamp profiling](#using-rocm-profiler-timestamp-profiling) + + [Using rocm-profiler performance counter collection:](#using-rocm-profiler-performance-counter-collection) + + [Using CodeXL to view profiling results:](#using-codexl-to-view-profiling-results) + + [More information on CodeXL](#more-information-on-codexl) + * [HIP Markers](#hip-markers) + + [Profiling HIP APIs](#profiling-hip-apis) + + [Adding markers to applications](#adding-markers-to-applications) + * [Additional HIP Profiling Features](#additional-hip-profiling-features) + + [Demangling C++ Kernel Names](#demangling-c-kernel-names) + + [Controlling when profiling starts and ends](#controlling-when-profiling-starts-and-ends) + + [Reducing timeline trace output file size](#reducing-timeline-trace-output-file-size) + + [How to enable profiling at HIP build time](#how-to-enable-profiling-at-hip-build-time) +- [Tracing and Debug](#tracing-and-debug) + * [Tracing HIP APIs](#tracing-hip-apis) + + [Color](#color) + * [Using HIP_DB](#using-hip_db) + * [Using ltrace](#using-ltrace) + * [Chicken bits](#chicken-bits) + * [Debugging HIP Applications](#debugging-hip-applications) + * [General Debugging Tips](#general-debugging-tips) + + ## CodeXL Profiling diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index 6c24fbf288..383af0440c 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -2343,9 +2343,6 @@ private: LangOptions DefaultLangOptions; SmallString<40> XStr; raw_svector_ostream OS(XStr); - StringRef initialParamList; - OS << "hipLaunchParm lp"; - size_t repLength = OS.str().size(); SourceLocation sl = kernelDecl->getNameInfo().getEndLoc(); SourceLocation kernelArgListStart = Lexer::findLocationAfterToken(sl, tok::l_paren, *SM, DefaultLangOptions, true); DEBUG(dbgs() << kernelArgListStart.printToString(*SM)); @@ -2355,14 +2352,12 @@ private: SourceLocation kernelArgListStart(pvdFirst->getLocStart()); SourceLocation kernelArgListEnd(pvdLast->getLocEnd()); SourceLocation stop = Lexer::getLocForEndOfToken(kernelArgListEnd, 0, *SM, DefaultLangOptions); - repLength += SM->getCharacterData(stop) - SM->getCharacterData(kernelArgListStart); - initialParamList = StringRef(SM->getCharacterData(kernelArgListStart), repLength); - OS << ", " << initialParamList; + size_t repLength = SM->getCharacterData(stop) - SM->getCharacterData(kernelArgListStart); + OS << StringRef(SM->getCharacterData(kernelArgListStart), repLength); + Replacement Rep0(*(Result.SourceManager), kernelArgListStart, repLength, OS.str()); + FullSourceLoc fullSL(sl, *(Result.SourceManager)); + insertReplacement(Rep0, fullSL); } - DEBUG(dbgs() << "initial paramlist: " << initialParamList << "\n" << "new paramlist: " << OS.str() << "\n"); - Replacement Rep0(*(Result.SourceManager), kernelArgListStart, repLength, OS.str()); - FullSourceLoc fullSL(sl, *(Result.SourceManager)); - insertReplacement(Rep0, fullSL); } bool cudaCall(const MatchFinder::MatchResult &Result) { @@ -2431,9 +2426,9 @@ private: XStr.clear(); if (calleeName.find(',') != StringRef::npos) { SmallString<128> tmpData; - calleeName = Twine("HIP_KERNEL_NAME(" + calleeName + ")").toStringRef(tmpData); + calleeName = Twine("(" + calleeName + ")").toStringRef(tmpData); } - OS << "hipLaunchKernel(" << calleeName << ","; + OS << "hipLaunchKernelGGL(" << calleeName << ","; const CallExpr *config = launchKernel->getConfig(); DEBUG(dbgs() << "Kernel config arguments:" << "\n"); SourceManager *SM = Result.SourceManager; @@ -2473,7 +2468,7 @@ private: Replacement Rep(*SM, launchKernel->getLocStart(), length, OS.str()); FullSourceLoc fullSL(launchKernel->getLocStart(), *SM); insertReplacement(Rep, fullSL); - hipCounter counter = {"hipLaunchKernel", CONV_KERN, API_RUNTIME}; + hipCounter counter = {"hipLaunchKernelGGL", CONV_KERN, API_RUNTIME}; updateCounters(counter, refName.str()); return true; } diff --git a/include/hip/hcc_detail/grid_launch_GGL.hpp b/include/hip/hcc_detail/grid_launch_GGL.hpp index 4fd7c3ff3a..8f1abbb70b 100644 --- a/include/hip/hcc_detail/grid_launch_GGL.hpp +++ b/include/hip/hcc_detail/grid_launch_GGL.hpp @@ -21,6 +21,7 @@ THE SOFTWARE. */ #pragma once +#if GENERIC_GRID_LAUNCH == 1 #include "concepts.hpp" #include "helpers.hpp" @@ -840,14 +841,16 @@ namespace hip_impl group_mem_bytes,\ stream,\ ...)\ - {\ - hipLaunchKernelGGL(\ - kernel_name,\ - num_blocks,\ - dim_blocks,\ - group_mem_bytes,\ - stream,\ - hipLaunchParm{},\ - ##__VA_ARGS__);\ - } + do {\ + hipLaunchKernelGGL(\ + kernel_name,\ + num_blocks,\ + dim_blocks,\ + group_mem_bytes,\ + stream,\ + hipLaunchParm{},\ + ##__VA_ARGS__);\ + } while(0) + } +#endif //GENERIC_GRID_LAUNCH diff --git a/include/hip/hcc_detail/hip_complex.h b/include/hip/hcc_detail/hip_complex.h index 9ff75d381a..26d73a21a8 100644 --- a/include/hip/hcc_detail/hip_complex.h +++ b/include/hip/hcc_detail/hip_complex.h @@ -23,8 +23,7 @@ THE SOFTWARE. #ifndef HIP_INCLUDE_HIP_HCC_DETAIL_HIP_COMPLEX_H #define HIP_INCLUDE_HIP_HCC_DETAIL_HIP_COMPLEX_H -#include "./hip_fp16.h" -#include "./hip_vector_types.h" +#include "hip/hcc_detail/hip_vector_types.h" #if __cplusplus #define COMPLEX_ADD_OP_OVERLOAD(type) \ @@ -177,45 +176,45 @@ COMPLEX_SCALAR_PRODUCT(hipDoubleComplex, unsigned long long) #endif -__device__ static inline float hipCrealf(hipFloatComplex z){ +__device__ __host__ static inline float hipCrealf(hipFloatComplex z){ return z.x; } -__device__ static inline float hipCimagf(hipFloatComplex z){ +__device__ __host__ static inline float hipCimagf(hipFloatComplex z){ return z.y; } -__device__ static inline hipFloatComplex make_hipFloatComplex(float a, float b){ +__device__ __host__ static inline hipFloatComplex make_hipFloatComplex(float a, float b){ hipFloatComplex z; z.x = a; z.y = b; return z; } -__device__ static inline hipFloatComplex hipConjf(hipFloatComplex z){ +__device__ __host__ static inline hipFloatComplex hipConjf(hipFloatComplex z){ hipFloatComplex ret; ret.x = z.x; ret.y = -z.y; return ret; } -__device__ static inline float hipCsqabsf(hipFloatComplex z){ +__device__ __host__ static inline float hipCsqabsf(hipFloatComplex z){ return z.x * z.x + z.y * z.y; } -__device__ static inline hipFloatComplex hipCaddf(hipFloatComplex p, hipFloatComplex q){ +__device__ __host__ static inline hipFloatComplex hipCaddf(hipFloatComplex p, hipFloatComplex q){ return make_hipFloatComplex(p.x + q.x, p.y + q.y); } -__device__ static inline hipFloatComplex hipCsubf(hipFloatComplex p, hipFloatComplex q){ +__device__ __host__ static inline hipFloatComplex hipCsubf(hipFloatComplex p, hipFloatComplex q){ return make_hipFloatComplex(p.x - q.x, p.y - q.y); } -__device__ static inline hipFloatComplex hipCmulf(hipFloatComplex p, hipFloatComplex q){ +__device__ __host__ static inline hipFloatComplex hipCmulf(hipFloatComplex p, hipFloatComplex q){ return make_hipFloatComplex(p.x * q.x - p.y * q.y, p.y * q.x + p.x * q.y); } -__device__ static inline hipFloatComplex hipCdivf(hipFloatComplex p, hipFloatComplex q){ +__device__ __host__ static inline hipFloatComplex hipCdivf(hipFloatComplex p, hipFloatComplex q){ float sqabs = hipCsqabsf(q); hipFloatComplex ret; ret.x = (p.x * q.x + p.y * q.y)/sqabs; @@ -223,51 +222,51 @@ __device__ static inline hipFloatComplex hipCdivf(hipFloatComplex p, hipFloatCom return ret; } -__device__ static inline float hipCabsf(hipFloatComplex z){ +__device__ __host__ static inline float hipCabsf(hipFloatComplex z){ return sqrtf(hipCsqabsf(z)); } -__device__ static inline double hipCreal(hipDoubleComplex z){ +__device__ __host__ static inline double hipCreal(hipDoubleComplex z){ return z.x; } -__device__ static inline double hipCimag(hipDoubleComplex z){ +__device__ __host__ static inline double hipCimag(hipDoubleComplex z){ return z.y; } -__device__ static inline hipDoubleComplex make_hipDoubleComplex(double a, double b){ +__device__ __host__ static inline hipDoubleComplex make_hipDoubleComplex(double a, double b){ hipDoubleComplex z; z.x = a; z.y = b; return z; } -__device__ static inline hipDoubleComplex hipConj(hipDoubleComplex z){ +__device__ __host__ static inline hipDoubleComplex hipConj(hipDoubleComplex z){ hipDoubleComplex ret; ret.x = z.x; ret.y = z.y; return ret; } -__device__ static inline double hipCsqabs(hipDoubleComplex z){ +__device__ __host__ static inline double hipCsqabs(hipDoubleComplex z){ return z.x * z.x + z.y * z.y; } -__device__ static inline hipDoubleComplex hipCadd(hipDoubleComplex p, hipDoubleComplex q){ +__device__ __host__ static inline hipDoubleComplex hipCadd(hipDoubleComplex p, hipDoubleComplex q){ return make_hipDoubleComplex(p.x + q.x, p.y + q.y); } -__device__ static inline hipDoubleComplex hipCsub(hipDoubleComplex p, hipDoubleComplex q){ +__device__ __host__ static inline hipDoubleComplex hipCsub(hipDoubleComplex p, hipDoubleComplex q){ return make_hipDoubleComplex(p.x - q.x, p.y - q.y); } -__device__ static inline hipDoubleComplex hipCmul(hipDoubleComplex p, hipDoubleComplex q){ +__device__ __host__ static inline hipDoubleComplex hipCmul(hipDoubleComplex p, hipDoubleComplex q){ return make_hipDoubleComplex(p.x * q.x - p.y * q.y, p.y * q.x + p.x * q.y); } -__device__ static inline hipDoubleComplex hipCdiv(hipDoubleComplex p, hipDoubleComplex q){ +__device__ __host__ static inline hipDoubleComplex hipCdiv(hipDoubleComplex p, hipDoubleComplex q){ double sqabs = hipCsqabs(q); hipDoubleComplex ret; ret.x = (p.x * q.x + p.y * q.y)/sqabs; @@ -275,28 +274,28 @@ __device__ static inline hipDoubleComplex hipCdiv(hipDoubleComplex p, hipDoubleC return ret; } -__device__ static inline double hipCabs(hipDoubleComplex z){ +__device__ __host__ static inline double hipCabs(hipDoubleComplex z){ return sqrtf(hipCsqabs(z)); } typedef hipFloatComplex hipComplex; -__device__ static inline hipComplex make_hipComplex(float x, +__device__ __host__ static inline hipComplex make_hipComplex(float x, float y){ return make_hipFloatComplex(x, y); } -__device__ static inline hipFloatComplex hipComplexDoubleToFloat +__device__ __host__ static inline hipFloatComplex hipComplexDoubleToFloat (hipDoubleComplex z){ return make_hipFloatComplex((float)z.x, (float)z.y); } -__device__ static inline hipDoubleComplex hipComplexFloatToDouble +__device__ __host__ static inline hipDoubleComplex hipComplexFloatToDouble (hipFloatComplex z){ return make_hipDoubleComplex((double)z.x, (double)z.y); } -__device__ static inline hipComplex hipCfmaf(hipComplex p, hipComplex q, hipComplex r){ +__device__ __host__ static inline hipComplex hipCfmaf(hipComplex p, hipComplex q, hipComplex r){ float real = (p.x * q.x) + r.x; float imag = (q.x * p.y) + r.y; @@ -306,7 +305,7 @@ __device__ static inline hipComplex hipCfmaf(hipComplex p, hipComplex q, hipComp return make_hipComplex(real, imag); } -__device__ static inline hipDoubleComplex hipCfma(hipDoubleComplex p, hipDoubleComplex q, hipDoubleComplex r){ +__device__ __host__ static inline hipDoubleComplex hipCfma(hipDoubleComplex p, hipDoubleComplex q, hipDoubleComplex r){ float real = (p.x * q.x) + r.x; float imag = (q.x * p.y) + r.y; diff --git a/include/hip/hcc_detail/hip_fp16.h b/include/hip/hcc_detail/hip_fp16.h index febc1b4fce..0a861b64af 100644 --- a/include/hip/hcc_detail/hip_fp16.h +++ b/include/hip/hcc_detail/hip_fp16.h @@ -23,7 +23,7 @@ THE SOFTWARE. #ifndef HIP_INCLUDE_HIP_HCC_DETAIL_HIP_FP16_H #define HIP_INCLUDE_HIP_HCC_DETAIL_HIP_FP16_H -#include "hip/hip_runtime.h" +#include "hip/hcc_detail/hip_vector_types.h" #if __clang_major__ > 3 diff --git a/include/hip/hcc_detail/hip_hcc.h b/include/hip/hcc_detail/hip_hcc.h index 645e980376..fc04917931 100644 --- a/include/hip/hcc_detail/hip_hcc.h +++ b/include/hip/hcc_detail/hip_hcc.h @@ -28,6 +28,17 @@ THE SOFTWARE. #if __cplusplus #ifdef __HCC__ #include + + +/** + *------------------------------------------------------------------------------------------------- + *------------------------------------------------------------------------------------------------- + * @defgroup HCC-specific features + * @warning These APIs provide access to special features of HCC compiler and are not available through the CUDA path. + * @{ + */ + + /** * @brief Return hc::accelerator associated with the specified deviceId * @return #hipSuccess, #hipErrorInvalidDevice @@ -45,6 +56,29 @@ hipError_t hipHccGetAcceleratorView(hipStream_t stream, hc::accelerator_view **a #endif // #ifdef __HCC__ +/** + * @brief launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra + * + * @param [in[ f Kernel to launch. + * @param [in] gridDimX X grid dimension specified in work-items + * @param [in] gridDimY Y grid dimension specified in work-items + * @param [in] gridDimZ Z grid dimension specified in work-items + * @param [in] blockDimX X block dimensions specified in work-items + * @param [in] blockDimY Y grid dimension specified in work-items + * @param [in] blockDimZ Z grid dimension specified in work-items + * @param [in] sharedMemBytes Amount of dynamic shared memory to allocate for this kernel. The kernel can access this with HIP_DYNAMIC_SHARED. + * @param [in] stream Stream where the kernel should be dispatched. May be 0, in which case th default stream is used with associated synchronization rules. + * @param [in] kernelParams + * @param [in] extra Pointer to kernel arguments. These are passed directly to the kernel and must be in the memory layout and alignment expected by the kernel. + * @param [in] startEvent If non-null, specified event will be updated to track the start time of the kernel launch. The event must be created before calling this API. + * @param [in] stopEvent If non-null, specified event will be updated to track the stop time of the kernel launch. The event must be created before calling this API. + * + * @returns hipSuccess, hipInvalidDevice, hipErrorNotInitialized, hipErrorInvalidValue + * + * @warning kernellParams argument is not yet implemented in HIP. Please use extra instead. Please refer to hip_porting_driver_api.md for sample usage. + + * HIP/ROCm actually updates the start event when the associated kernel completes. + */ hipError_t hipHccModuleLaunchKernel(hipFunction_t f, uint32_t globalWorkSizeX, uint32_t globalWorkSizeY, @@ -55,8 +89,15 @@ hipError_t hipHccModuleLaunchKernel(hipFunction_t f, size_t sharedMemBytes, hipStream_t hStream, void **kernelParams, - void **extra); + void **extra, + hipEvent_t startEvent=nullptr, + hipEvent_t stopEvent=nullptr + ); +// doxygen end HCC-specific features +/** + * @} + */ #endif // #if __cplusplus #endif // diff --git a/include/hip/hcc_detail/hip_runtime_api.h b/include/hip/hcc_detail/hip_runtime_api.h index 0daca7a53b..f9bfb5a310 100644 --- a/include/hip/hcc_detail/hip_runtime_api.h +++ b/include/hip/hcc_detail/hip_runtime_api.h @@ -1913,19 +1913,18 @@ hipError_t hipModuleLoadData(hipModule_t *module, const void *image); /** * @brief launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra * - * @param [in[ f - * @param [in] gridDimX - * @param [in] gridDimY - * @param [in] gridDimZ - * @param [in] blockDimX - * @param [in] blockDimY - * @param [in] blockDimZ - * @param [in] sharedMemBytes - * @param [in] stream - * @param [in] kernelParams - * @param [in] extraa + * @param [in[ f Kernel to launch. + * @param [in] gridDimX X grid dimension specified as multiple of blockDimX. + * @param [in] gridDimY Y grid dimension specified as multiple of blockDimY. + * @param [in] gridDimZ Z grid dimension specified as multiple of blockDimZ. + * @param [in] blockDimX X block dimensions specified in work-items + * @param [in] blockDimY Y grid dimension specified in work-items + * @param [in] blockDimZ Z grid dimension specified in work-items + * @param [in] sharedMemBytes Amount of dynamic shared memory to allocate for this kernel. The kernel can access this with HIP_DYNAMIC_SHARED. + * @param [in] stream Stream where the kernel should be dispatched. May be 0, in which case th default stream is used with associated synchronization rules. + * @param [in] kernelParams + * @param [in] extra Pointer to kernel arguments. These are passed directly to the kernel and must be in the memory layout and alignment expected by the kernel. * - * The function takes the above arguments and run the kernel in hipFunction_t f. with launch parameters specified in gridDimX, gridDimY, gridDimZ, blockDimX, blockDimY and blockDimmZ. The amount of shared memory is specificed and can be used with HIP_DYNAMIC_SHARED. The arguemt extra is used to pass in the arguments for the kernel. * @returns hipSuccess, hipInvalidDevice, hipErrorNotInitialized, hipErrorInvalidValue * * @warning kernellParams argument is not yet implemented in HIP. Please use extra instead. Please refer to hip_porting_driver_api.md for sample usage. diff --git a/include/hip/hcc_detail/hip_vector_types.h b/include/hip/hcc_detail/hip_vector_types.h index 42e1d6663c..82bd3b2d6f 100644 --- a/include/hip/hcc_detail/hip_vector_types.h +++ b/include/hip/hcc_detail/hip_vector_types.h @@ -32,7 +32,7 @@ THE SOFTWARE. #error("This version of HIP requires a newer version of HCC."); #endif -#include "host_defines.h" +#include "hip/hcc_detail/host_defines.h" #define MAKE_DEFAULT_CONSTRUCTOR_ONE_COMPONENT(type) \ __device__ __host__ type() {} \ diff --git a/packaging/create_hip_samples_installer.sh b/packaging/create_hip_samples_installer.sh deleted file mode 100755 index 91789d2524..0000000000 --- a/packaging/create_hip_samples_installer.sh +++ /dev/null @@ -1,23 +0,0 @@ -#!/bin/bash -function die { - echo "${1-Died}." >&2 - exit 1 -} - -payload=$1 -script=$2 -[ "$payload" != "" ] || [ "$script" != "" ] || die "Invalid arguments!" -tmp=__extract__$RANDOM - -printf "#!/bin/bash -samples_dir=\$1 -[ \"\$samples_dir\" != \"\" ] || read -e -p \"Enter the path to extract the HIP samples: \" samples_dir -mkdir -p \$samples_dir -PAYLOAD=\`awk '/^__PAYLOAD_BELOW__/ {print NR + 1; exit 0; }' \$0\` -tail -n+\$PAYLOAD \$0 | tar -xz -C \$samples_dir -echo \"HIP samples installed in \$samples_dir\" -exit 0 -__PAYLOAD_BELOW__\n" > "$tmp" - -cat "$tmp" "$payload" > "$script" && rm "$tmp" -chmod +x "$script" diff --git a/packaging/hip_doc.txt b/packaging/hip_doc.txt index bbcaf54ec8..d5a0c471b1 100644 --- a/packaging/hip_doc.txt +++ b/packaging/hip_doc.txt @@ -1,12 +1,19 @@ cmake_minimum_required(VERSION 2.8.3) project(hip_doc) -add_custom_target(build_doxygen ALL +find_program(DOXYGEN_EXE doxygen) +if(DOXYGEN_EXE) + add_custom_target(build_doxygen ALL COMMAND HIP_PATH=@hip_SOURCE_DIR@ doxygen @hip_SOURCE_DIR@/docs/doxygen-input/doxy.cfg) -add_custom_target(convert_md_to_html ALL + install(DIRECTORY RuntimeAPI/html DESTINATION docs/docs/RuntimeAPI) +endif() + +find_program(GRIP_EXE grip) +if(GRIP_EXE) + add_custom_target(convert_md_to_html ALL COMMAND @hip_SOURCE_DIR@/packaging/convert_md_to_html.sh @hip_SOURCE_DIR@ ${PROJECT_BINARY_DIR}/md2html) -install(DIRECTORY RuntimeAPI/html DESTINATION docs/docs/RuntimeAPI) -install(DIRECTORY md2html/ DESTINATION docs) + install(DIRECTORY md2html/ DESTINATION docs) +endif() ############################# # Packaging steps diff --git a/samples/0_Intro/bit_extract/Makefile b/samples/0_Intro/bit_extract/Makefile index 78f6a2faa8..08bca6e642 100644 --- a/samples/0_Intro/bit_extract/Makefile +++ b/samples/0_Intro/bit_extract/Makefile @@ -11,10 +11,6 @@ HIPCC=$(HIP_PATH)/bin/hipcc ifeq (${HIP_PLATFORM}, nvcc) HIPCC_FLAGS = -gencode=arch=compute_20,code=sm_20 endif -ifeq (${HIP_PLATFORM}, hcc) - HIPCC_FLAGS = -stdlib=libc++ -endif - EXE=bit_extract diff --git a/samples/1_Utils/hipCommander/Makefile b/samples/1_Utils/hipCommander/Makefile index e770c636a4..a411763b7f 100644 --- a/samples/1_Utils/hipCommander/Makefile +++ b/samples/1_Utils/hipCommander/Makefile @@ -10,9 +10,6 @@ OPT=-O3 CXXFLAGS = $(OPT) --std=c++11 HIP_PLATFORM=$(shell $(HIP_PATH)/bin/hipconfig --platform) -ifeq (${HIP_PLATFORM}, hcc) - CXXFLAGS += " -stdlib=libc++" -endif CODE_OBJECTS=nullkernel.hsaco diff --git a/samples/1_Utils/hipCommander/hipCommander.cpp b/samples/1_Utils/hipCommander/hipCommander.cpp index 0add1ce3e3..4b93180b18 100644 --- a/samples/1_Utils/hipCommander/hipCommander.cpp +++ b/samples/1_Utils/hipCommander/hipCommander.cpp @@ -11,7 +11,6 @@ #include #include #include -#include #endif #include diff --git a/src/hip_event.cpp b/src/hip_event.cpp index d44f201db5..61ac5cd3ab 100644 --- a/src/hip_event.cpp +++ b/src/hip_event.cpp @@ -30,6 +30,54 @@ THE SOFTWARE. //--- +ihipEvent_t::ihipEvent_t(unsigned flags) +{ + _state = hipEventStatusCreated; + _stream = NULL; + _flags = flags; + _timestamp = 0; + _type = hipEventTypeIndependent; +}; + + + +// Attach to an existing completion future: +void ihipEvent_t::attachToCompletionFuture(const hc::completion_future *cf, ihipEventType_t eventType) +{ + _state = hipEventStatusRecording; + _marker = *cf; + _type = eventType; +} + + + +void ihipEvent_t::setTimestamp() +{ + if (_state == hipEventStatusRecorded) { + // already recorded, done: + return; + } else { + // TODO - use completion-future functions to obtain ticks and timestamps: + hsa_signal_t *sig = static_cast (_marker.get_native_handle()); + if (sig) { + if (hsa_signal_load_acquire(*sig) == 0) { + + if ((_type == hipEventTypeIndependent) || (_type == hipEventTypeStopCommand)) { + _timestamp = _marker.get_end_tick(); + } else if (_type == hipEventTypeStartCommand) { + _timestamp = _marker.get_begin_tick(); + } else { + assert(0); // TODO - move to debug assert + _timestamp = 0; + } + + _state = hipEventStatusRecorded; + } + } + } +} + + hipError_t ihipEventCreate(hipEvent_t* event, unsigned flags) { hipError_t e = hipSuccess; @@ -37,12 +85,8 @@ hipError_t ihipEventCreate(hipEvent_t* event, unsigned flags) // TODO-IPC - support hipEventInterprocess. unsigned supportedFlags = hipEventDefault | hipEventBlockingSync | hipEventDisableTiming; if ((flags & ~supportedFlags) == 0) { - ihipEvent_t *eh = new ihipEvent_t(); + ihipEvent_t *eh = new ihipEvent_t(flags); - eh->_state = hipEventStatusCreated; - eh->_stream = NULL; - eh->_flags = flags; - eh->_timestamp = 0; *event = eh; } else { e = hipErrorInvalidValue; @@ -141,8 +185,8 @@ hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop) ihipEvent_t *start_eh = start; ihipEvent_t *stop_eh = stop; - ihipSetTs(start); - ihipSetTs(stop); + start->setTimestamp(); + stop->setTimestamp(); hipError_t status = hipSuccess; *ms = 0.0f; @@ -151,7 +195,7 @@ hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop) if ((start_eh->_state == hipEventStatusRecorded) && (stop_eh->_state == hipEventStatusRecorded)) { // Common case, we have good information for both events. - int64_t tickDiff = (stop_eh->_timestamp - start_eh->_timestamp); + int64_t tickDiff = (stop_eh->timestamp() - start_eh->timestamp()); uint64_t freqHz; hsa_system_get_info(HSA_SYSTEM_INFO_TIMESTAMP_FREQUENCY, &freqHz); diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 374840f91f..35a3e11e71 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -1641,23 +1641,6 @@ const char *ihipErrorString(hipError_t hip_error) }; -void ihipSetTs(hipEvent_t e) -{ - ihipEvent_t *eh = e; - if (eh->_state == hipEventStatusRecorded) { - // already recorded, done: - return; - } else { - // TODO - use completion-future functions to obtain ticks and timestamps: - hsa_signal_t *sig = static_cast (eh->_marker.get_native_handle()); - if (sig) { - if (hsa_signal_load_acquire(*sig) == 0) { - eh->_timestamp = eh->_marker.get_end_tick(); - eh->_state = hipEventStatusRecorded; - } - } - } -} // Returns true if copyEngineCtx can see the memory allocated on dstCtx and srcCtx. diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index 4b960e2820..9c17c6e98c 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -584,22 +584,40 @@ private: // Data //---- // Internal event structure: enum hipEventStatus_t { - hipEventStatusUnitialized = 0, // event is unutilized, must be "Created" before use. - hipEventStatusCreated = 1, - hipEventStatusRecording = 2, // event has been enqueued to record something. - hipEventStatusRecorded = 3, // event has been recorded - timestamps are valid. + hipEventStatusUnitialized = 0, // event is unutilized, must be "Created" before use. + hipEventStatusCreated = 1, + hipEventStatusRecording = 2, // event has been enqueued to record something. + hipEventStatusRecorded = 3, // event has been recorded - timestamps are valid. } ; +// TODO - rename to ihip type of some kind +enum ihipEventType_t { + hipEventTypeIndependent, + hipEventTypeStartCommand, + hipEventTypeStopCommand, +}; // internal hip event structure. -struct ihipEvent_t { - hipEventStatus_t _state; +class ihipEvent_t { +public: + ihipEvent_t(unsigned flags); + void attachToCompletionFuture(const hc::completion_future *cf, ihipEventType_t eventType); + void setTimestamp(); + uint64_t timestamp() const { return _timestamp; } ; + ihipEventType_t type() const { return _type; }; + +public: + hipEventStatus_t _state; hipStream_t _stream; // Stream where the event is recorded, or NULL if all streams. unsigned _flags; hc::completion_future _marker; + +private: + ihipEventType_t _type; uint64_t _timestamp; // store timestamp, may be set on host or by marker. +friend hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream); } ; @@ -822,8 +840,6 @@ extern hipError_t ihipDeviceSetState(); extern ihipDevice_t *ihipGetDevice(int); ihipCtx_t * ihipGetPrimaryCtx(unsigned deviceIndex); -extern void ihipSetTs(hipEvent_t e); - hipStream_t ihipSyncAndResolveStream(hipStream_t); diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 805fc9efc0..da5530349f 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -1260,10 +1260,15 @@ hipError_t hipIpcOpenMemHandle(void** devPtr, hipIpcMemHandle_t handle, unsigned ihipIpcMemHandle_t* iHandle = (ihipIpcMemHandle_t*) &handle; //Attach ipc memory - hsa_status_t hsa_status = - hsa_amd_ipc_memory_attach((hsa_amd_ipc_memory_t*)&(iHandle->ipc_handle), iHandle->psize, 1, agent, devPtr); - if(hsa_status != HSA_STATUS_SUCCESS) - hipStatus = hipErrorMapBufferObjectFailed; + auto ctx= ihipGetTlsDefaultCtx(); + { + LockedAccessor_CtxCrit_t crit(ctx->criticalData()); + // the peerCnt always stores self so make sure the trace actually + hsa_status_t hsa_status = + hsa_amd_ipc_memory_attach((hsa_amd_ipc_memory_t*)&(iHandle->ipc_handle), iHandle->psize, crit->peerCnt(), crit->peerAgents(), devPtr); + if(hsa_status != HSA_STATUS_SUCCESS) + hipStatus = hipErrorMapBufferObjectFailed; + } #else hipStatus = hipErrorRuntimeOther; #endif diff --git a/src/hip_module.cpp b/src/hip_module.cpp index 67bba5f935..b359e7a63c 100644 --- a/src/hip_module.cpp +++ b/src/hip_module.cpp @@ -364,10 +364,11 @@ hipError_t hipModuleGetFunction(hipFunction_t *hfunc, hipModule_t hmod, hipError_t ihipModuleLaunchKernel(hipFunction_t f, - uint32_t globalWorkSizeX, uint32_t globalWorkSizeY, uint32_t globalWorkSizeZ, - uint32_t localWorkSizeX, uint32_t localWorkSizeY, uint32_t localWorkSizeZ, - size_t sharedMemBytes, hipStream_t hStream, - void **kernelParams, void **extra) + uint32_t globalWorkSizeX, uint32_t globalWorkSizeY, uint32_t globalWorkSizeZ, + uint32_t localWorkSizeX, uint32_t localWorkSizeY, uint32_t localWorkSizeZ, + size_t sharedMemBytes, hipStream_t hStream, + void **kernelParams, void **extra, + hipEvent_t startEvent, hipEvent_t stopEvent) { auto ctx = ihipGetTlsDefaultCtx(); @@ -446,7 +447,20 @@ hipError_t ihipModuleLaunchKernel(hipFunction_t f, (HSA_FENCE_SCOPE_SYSTEM << HSA_PACKET_HEADER_RELEASE_FENCE_SCOPE); }; - lp.av->dispatch_hsa_kernel(&aql, config[1] /* kernarg*/, kernArgSize, nullptr/*completion_future*/); + + hc::completion_future cf; + + lp.av->dispatch_hsa_kernel(&aql, config[1] /* kernarg*/, kernArgSize, + (startEvent || stopEvent) ? &cf : nullptr); + + + if (startEvent) { + startEvent->attachToCompletionFuture(&cf, hipEventTypeStartCommand); + } + if (stopEvent) { + stopEvent->attachToCompletionFuture (&cf, hipEventTypeStopCommand); + } + if(kernelParams != NULL){ free(config[1]); @@ -470,7 +484,8 @@ hipError_t hipModuleLaunchKernel(hipFunction_t f, return ihipLogStatus(ihipModuleLaunchKernel(f, blockDimX * gridDimX, blockDimY * gridDimY, gridDimZ * blockDimZ, blockDimX, blockDimY, blockDimZ, - sharedMemBytes, hStream, kernelParams, extra)); + sharedMemBytes, hStream, kernelParams, extra, + nullptr, nullptr)); } @@ -478,7 +493,8 @@ hipError_t hipHccModuleLaunchKernel(hipFunction_t f, uint32_t globalWorkSizeX, uint32_t globalWorkSizeY, uint32_t globalWorkSizeZ, uint32_t localWorkSizeX, uint32_t localWorkSizeY, uint32_t localWorkSizeZ, size_t sharedMemBytes, hipStream_t hStream, - void **kernelParams, void **extra) + void **kernelParams, void **extra, + hipEvent_t startEvent, hipEvent_t stopEvent) { HIP_INIT_API(f, globalWorkSizeX, globalWorkSizeY, globalWorkSizeZ, localWorkSizeX, localWorkSizeY, localWorkSizeZ, @@ -486,7 +502,7 @@ hipError_t hipHccModuleLaunchKernel(hipFunction_t f, kernelParams, extra); return ihipLogStatus(ihipModuleLaunchKernel(f, globalWorkSizeX, globalWorkSizeY, globalWorkSizeZ, localWorkSizeX, localWorkSizeY, localWorkSizeZ, - sharedMemBytes, hStream, kernelParams, extra)); + sharedMemBytes, hStream, kernelParams, extra, startEvent, stopEvent)); } hipError_t hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, diff --git a/tests/src/deviceLib/hipDeviceMemcpy.cpp b/tests/src/deviceLib/hipDeviceMemcpy.cpp index 54fd02c0c2..3843c07bb9 100644 --- a/tests/src/deviceLib/hipDeviceMemcpy.cpp +++ b/tests/src/deviceLib/hipDeviceMemcpy.cpp @@ -1,18 +1,29 @@ -#include +#include #include "hip/hip_runtime.h" #include "hip/hip_runtime_api.h" +#include "../test_common.h" + #define LEN 1030 #define SIZE LEN << 2 -__global__ void cpy(hipLaunchParm lp, uint32_t *Out, uint32_t *In, uint32_t *Vald) +/* HIT_START + * BUILD: %t %s ../test_common.cpp + * RUN: %t + * HIT_END + */ + + +__global__ void cpy(hipLaunchParm lp, uint32_t *Out, uint32_t *In) { - memcpy(Out, In, SIZE, Vald); + int tx = hipThreadIdx_x; + memcpy(Out + tx, In + tx, SIZE/LEN); } __global__ void set(hipLaunchParm lp, uint32_t *ptr, uint8_t val, size_t size) { - memset(ptr, val, size); + int tx = hipThreadIdx_x; + memset(ptr + tx, val, size); } int main() @@ -24,19 +35,29 @@ int main() Val = new uint32_t; *Val = 0; for(int i=0;i Date: Thu, 13 Apr 2017 12:39:28 +0530 Subject: [PATCH 004/108] Merge branch 'amd-develop' into amd-master Change-Id: I05572d2b32f1df70b54e2efeb32c8a4d8055912d (cherry picked from commit 3a56e5c09b198ca144737902130e1c629674931c) --- CMakeLists.txt | 10 +- docs/markdown/hip_bugs.md | 81 +++++++---- include/hip/hcc_detail/grid_launch_GGL.hpp | 33 +++-- include/hip/hcc_detail/helpers.hpp | 150 ++++++++++----------- include/hip/hcc_detail/hip_complex.h | 1 + include/hip/hcc_detail/hip_vector_types.h | 38 +++++- include/hip/hcc_detail/host_defines.h | 2 +- include/hip/hcc_detail/math_functions.h | 1 + include/hip/hip_fp16.h | 2 +- packaging/hip-targets-release.cmake | 41 ++++++ packaging/hip-targets.cmake | 102 ++++++++++++++ packaging/hip_hcc.txt | 2 + src/grid_launch.cpp | 36 +++++ src/math_functions.cpp | 4 + 14 files changed, 374 insertions(+), 129 deletions(-) create mode 100644 packaging/hip-targets-release.cmake create mode 100644 packaging/hip-targets.cmake diff --git a/CMakeLists.txt b/CMakeLists.txt index 94ed2a7562..eee1a14a8a 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -142,7 +142,7 @@ add_to_config(_buildInfo COMPILE_HIP_ATP_MARKER) # Build steps ############################# # Rebuild cmake cache updates .hipInfo and .hipVersion -add_custom_target(update_build_and_version_info ALL COMMAND make rebuild_cache) +add_custom_target(update_build_and_version_info COMMAND make rebuild_cache) # Build clang hipify if enabled add_subdirectory(hipify-clang) @@ -372,14 +372,14 @@ endif() # Testing steps ############################# # Target: test -set(HIP_PATH ${CMAKE_INSTALL_PREFIX}) +set(HIP_ROOT_DIR ${CMAKE_INSTALL_PREFIX}) set(HIP_SRC_PATH ${CMAKE_CURRENT_SOURCE_DIR}) -execute_process(COMMAND "${CMAKE_COMMAND}" -E copy_directory "${HIP_SRC_PATH}/cmake" "${HIP_PATH}/cmake" RESULT_VARIABLE RUN_HIT ERROR_QUIET) +execute_process(COMMAND "${CMAKE_COMMAND}" -E copy_directory "${HIP_SRC_PATH}/cmake" "${HIP_ROOT_DIR}/cmake" RESULT_VARIABLE RUN_HIT ERROR_QUIET) if(${RUN_HIT} EQUAL 0) - execute_process(COMMAND "${CMAKE_COMMAND}" -E copy_directory "${HIP_SRC_PATH}/bin" "${HIP_PATH}/bin" RESULT_VARIABLE RUN_HIT ERROR_QUIET) + execute_process(COMMAND "${CMAKE_COMMAND}" -E copy_directory "${HIP_SRC_PATH}/bin" "${HIP_ROOT_DIR}/bin" RESULT_VARIABLE RUN_HIT ERROR_QUIET) endif() if(${RUN_HIT} EQUAL 0) - set(CMAKE_MODULE_PATH "${HIP_PATH}/cmake" ${CMAKE_MODULE_PATH}) + set(CMAKE_MODULE_PATH "${HIP_ROOT_DIR}/cmake" ${CMAKE_MODULE_PATH}) include(${HIP_SRC_PATH}/tests/hit/HIT.cmake) # Add tests diff --git a/docs/markdown/hip_bugs.md b/docs/markdown/hip_bugs.md index 73133843bc..9452fae2fd 100644 --- a/docs/markdown/hip_bugs.md +++ b/docs/markdown/hip_bugs.md @@ -1,5 +1,4 @@ -# HIP Bugs - +# HIP Bugs - [Errors related to undefined reference to `__hcLaunchKernel__***__grid_launch_parm**`](#errors-related-to-undefined-reference-to-__hclaunchkernel____grid_launch_parm) @@ -41,60 +40,86 @@ For example, `Foo` in the code snippets below contains an array-typed member var ``` struct Foo { + float _data; // table is an array, which makes foo int table[3]; }; ``` -An workaround is to provide a custom serializer on CPU side, and append the contents of the array as kernel arguments: +A workaround is to provide a custom serializer on host side which appends the contents of the array as kernel arguments, and a custome deserializaer on the device path to reconstruct the array inside the GPU kernels. +The deserializer can not be a function template, and should have scalar-typed parameters of the number equals to the length of the array-typed member variable. For example: ``` struct Foo { - int table[3]; + float _data; + int _table[3]; + - // user-provided CPU serializer - // must append the contents of the array member as kernel arguments #ifdef __HCC__ + // user-provided CPU serializer + // Append the contents of the array member as kernel arguments __attribute__((annotate(“serialize”))) void __cxxamp_serialize(Kalmar::Serialize &s) const { + s.Append(sizeof(float), &_data); for (int i = 0; i < 3; ++i) - s.Append(sizeof(int), &table[i]); + s.Append(sizeof(int), &_table[i]); } -#endif -}; -``` -Then, provide a custom deserializer on GPU side, to help reconstruct the array within GPU kernels. Notice that the deserializer can not be a function template, and should have scalar-typed parameters of the number equals to the length of the array-typed member variable. For example: - -``` -struct Foo { - int table[3]; // user-provided GPU deserializer // table has 3 int elements, so deserializer must have 3 int parameters. -#ifdef __HCC__ __attribute__((annotate(“user_deserialize”))) - Foo(int x0, int x1, int x2) [[cpu]][[hc]] { - table[0] = x0; - table[1] = x1; - table[2] = x2; + Foo(float d, int x0, int x1, int x2) [[cpu]][[hc]] { + _data = d; + _table[0] = x0; + _table[1] = x1; + _table[2] = x2; } -#endif -#ifdef __HCC__ - __attribute__((annotate(“serialize”))) - void __cxxamp_serialize(Kalmar::Serialize &s) const { - s.Append(sizeof(int), &table[0]); - s.Append(sizeof(int), &table[1]); - s.Append(sizeof(int), &table[2]); - } #endif }; ``` Rather than create serializer functions, another workaround is to pass the member fields from the structure as simple data types. +Note a class or struct can contain only one "user_deserialize" constructor. +For types which contain arrays which are based on template parameter, you can use partial template instantiation to implement one constructor per specialization. +However, an easier approach may be to create one user_deserializer which processes the maximum supported dimension. +This will take more memory in the structure and also require additional kernel arguments, but this may have little performance impact and the conversion is easier than partial template specialization. An example: + +``` +#define MAX_Dim 4 +template struct MyArray { + + T* dataPtr_; + //int size_[Dim]; // Original code with template-sized Dims + int size_[MAX_dim]; // Workaround code - allocate an array big enough for all dims so one serializer works. + + +... + +#ifdef __HCC__ + __attribute__((annotate("serialize"))) + void __cxxamp_serialize(Kalmar::Serialize &s) const { + s.Append(sizeof(float), &_dataPtr); + for (int i=0; i using is_new_grid_launch_t = typename std::conditional< - std::is_callable{}, + is_callable{}, New_grid_launch_tag, Old_grid_launch_tag>::type; } @@ -118,6 +119,7 @@ namespace hip_impl // TODO: these are workarounds, they should be removed. hc::accelerator_view lock_stream_hip_(hipStream_t&, void*&); + void print_prelaunch_trace_(const char*, dim3, dim3, int, hipStream_t); void unlock_stream_hip_( hipStream_t, void*, const char*, hc::accelerator_view*); @@ -137,7 +139,13 @@ namespace hip_impl void* lck_stream = nullptr; auto acc_v = lock_stream_hip_(stream, lck_stream); auto stream_guard = make_RAII_guard( - [](){ /* perhaps use a slimmed down ihipPrintKernelLaunch here */ }, + std::bind( + print_prelaunch_trace_, + kernel_name, + num_blocks, + dim_blocks, + group_mem_bytes, + stream), std::bind( unlock_stream_hip_, stream, lck_stream, kernel_name, &acc_v)); @@ -841,16 +849,15 @@ namespace hip_impl group_mem_bytes,\ stream,\ ...)\ - do {\ - hipLaunchKernelGGL(\ - kernel_name,\ - num_blocks,\ - dim_blocks,\ - group_mem_bytes,\ - stream,\ - hipLaunchParm{},\ - ##__VA_ARGS__);\ - } while(0) - + do {\ + hipLaunchKernelGGL(\ + kernel_name,\ + num_blocks,\ + dim_blocks,\ + group_mem_bytes,\ + stream,\ + hipLaunchParm{},\ + ##__VA_ARGS__);\ + } while(0) } #endif //GENERIC_GRID_LAUNCH diff --git a/include/hip/hcc_detail/helpers.hpp b/include/hip/hcc_detail/helpers.hpp index e5a84a4678..611929766b 100644 --- a/include/hip/hcc_detail/helpers.hpp +++ b/include/hip/hcc_detail/helpers.hpp @@ -21,6 +21,7 @@ THE SOFTWARE. */ #pragma once +#include "concepts.hpp" #include // For std::conditional, std::decay, std::enable_if, // std::false_type, std result_of and std::true_type. @@ -29,9 +30,6 @@ THE SOFTWARE. namespace std { // TODO: these should be removed as soon as possible. #if (__cplusplus < 201406L) - template - using void_t = void; - #if (__cplusplus < 201402L) template using enable_if_t = typename enable_if::type; @@ -43,88 +41,80 @@ namespace std using result_of_t = typename result_of::type; template using remove_reference_t = typename remove_reference::type; - template< - FunctionalProcedure F, - unsigned int n = 0u, - typename = void> - struct is_callable_impl : is_callable_impl {}; - - // Pointer to member function, call through non-pointer. - template - struct is_callable_impl< - F(C, Ts...), - 0u, - void_t().*declval())(declval()...))> - > : true_type { - }; - - // Pointer to member function, call through pointer. - template - struct is_callable_impl< - F(C, Ts...), - 1u, - void_t()).*declval())(declval()...))> - > : std::true_type { - }; - - // Pointer to member data, call through non-pointer, no args. - template - struct is_callable_impl< - F(C), - 2u, - void_t().*declval())> - > : true_type { - }; - - // Pointer to member data, call through pointer, no args. - template - struct is_callable_impl< - F(C), - 3u, - void_t().*declval())> - > : true_type { - }; - - // General call, n args. - template - struct is_callable_impl< - F(Ts...), - 4u, - void_t()(declval()...))> - > : true_type { - }; - - // Not callable. - template - struct is_callable_impl : false_type {}; - - template - struct is_callable : is_callable_impl {}; - #else - template - struct is_callable_impl : false_type {}; - - template - struct is_callable_impl< - F(Ts...), - void_t>> : true_type {}; - - template - struct is_callable : is_callable_impl {}; #endif - template - struct disjunction : false_type {}; - template - struct disjunction : B1 {}; - template - struct disjunction - : conditional_t> - {}; #endif } -namespace hip_impl // Only for documentation, macros ignore namespaces. +namespace hip_impl { + template + using void_t_ = void; + + #if (__cplusplus < 201402L) + template< + FunctionalProcedure F, + unsigned int n = 0u, + typename = void> + struct is_callable_impl : is_callable_impl {}; + + // Pointer to member function, call through non-pointer. + template + struct is_callable_impl< + F(C, Ts...), + 0u, + void_t_().*std::declval())( + std::declval()...))> + > : std::true_type {}; + + // Pointer to member function, call through pointer. + template + struct is_callable_impl< + F(C, Ts...), + 1u, + void_t_()).*std::declval())( + std::declval()...))> + > : std::true_type {}; + + // Pointer to member data, call through non-pointer, no args. + template + struct is_callable_impl< + F(C), + 2u, + void_t_().*std::declval())> + > : std::true_type {}; + + // Pointer to member data, call through pointer, no args. + template + struct is_callable_impl< + F(C), + 3u, + void_t_().*std::declval())> + > : std::true_type {}; + + // General call, n args. + template + struct is_callable_impl< + F(Ts...), + 4u, + void_t_()(std::declval()...))> + > : std::true_type {}; + + // Not callable. + template + struct is_callable_impl : std::false_type {}; + + template + struct is_callable : is_callable_impl {}; + #else + template + struct is_callable_impl : std::false_type {}; + + template + struct is_callable_impl< + F(Ts...), + void_t_>> : std::true_type {}; + #endif + #define count_macro_args_impl_hip_(\ _0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _13, _14, _15,\ _16, _17, _18, _19, _20, _21, _22, _23, _24, _25, _26, _27, _28, _29,\ diff --git a/include/hip/hcc_detail/hip_complex.h b/include/hip/hcc_detail/hip_complex.h index 26d73a21a8..c76d65b058 100644 --- a/include/hip/hcc_detail/hip_complex.h +++ b/include/hip/hcc_detail/hip_complex.h @@ -24,6 +24,7 @@ THE SOFTWARE. #define HIP_INCLUDE_HIP_HCC_DETAIL_HIP_COMPLEX_H #include "hip/hcc_detail/hip_vector_types.h" +#include #if __cplusplus #define COMPLEX_ADD_OP_OVERLOAD(type) \ diff --git a/include/hip/hcc_detail/hip_vector_types.h b/include/hip/hcc_detail/hip_vector_types.h index 82bd3b2d6f..35c6c23548 100644 --- a/include/hip/hcc_detail/hip_vector_types.h +++ b/include/hip/hcc_detail/hip_vector_types.h @@ -1270,6 +1270,15 @@ __device__ __host__ static inline type operator op (type& val, int) { \ #define DECLOP_1VAR_COMP(type, op) \ __device__ __host__ static inline bool operator op (type& lhs, type& rhs) { \ return lhs.x op rhs.x; \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, type& rhs) { \ + return lhs.x op rhs.x; \ +} \ +__device__ __host__ static inline bool operator op (type& lhs, const type& rhs) { \ + return lhs.x op rhs.x ; \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, const type& rhs) { \ + return lhs.x op rhs.x ; \ } #define DECLOP_1VAR_1IN_1OUT(type, op) \ @@ -1338,6 +1347,15 @@ __device__ __host__ static inline type operator op (type& val, int) { \ #define DECLOP_2VAR_COMP(type, op) \ __device__ __host__ static inline bool operator op (type& lhs, type& rhs) { \ return (lhs.x op rhs.x) && (lhs.y op rhs.y); \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y); \ +} \ +__device__ __host__ static inline bool operator op (type& lhs, const type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y); \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, const type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y); \ } #define DECLOP_2VAR_1IN_1OUT(type, op) \ @@ -1415,7 +1433,16 @@ __device__ __host__ static inline type operator op (type& val, int) { \ #define DECLOP_3VAR_COMP(type, op) \ __device__ __host__ static inline bool operator op (type& lhs, type& rhs) { \ return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z); \ -} +} \ +__device__ __host__ static inline bool operator op (const type& lhs, type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z); \ +} \ +__device__ __host__ static inline bool operator op (type& lhs, const type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z); \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, const type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z); \ +} \ #define DECLOP_3VAR_1IN_1OUT(type, op) \ __device__ __host__ static inline type operator op(type &rhs) { \ @@ -1500,6 +1527,15 @@ __device__ __host__ static inline type operator op (type& val, int) { \ #define DECLOP_4VAR_COMP(type, op) \ __device__ __host__ static inline bool operator op (type& lhs, type& rhs) { \ return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z) && (lhs.w op rhs.w); \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z) && (lhs.w op rhs.w); \ +} \ +__device__ __host__ static inline bool operator op (type& lhs, const type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z) && (lhs.w op rhs.w); \ +} \ +__device__ __host__ static inline bool operator op (const type& lhs, const type& rhs) { \ + return (lhs.x op rhs.x) && (lhs.y op rhs.y) && (lhs.z op rhs.z) && (lhs.w op rhs.w); \ } #define DECLOP_4VAR_1IN_1OUT(type, op) \ diff --git a/include/hip/hcc_detail/host_defines.h b/include/hip/hcc_detail/host_defines.h index b0a7421d18..5864cfa0e7 100644 --- a/include/hip/hcc_detail/host_defines.h +++ b/include/hip/hcc_detail/host_defines.h @@ -48,7 +48,7 @@ THE SOFTWARE. #define __global__ __attribute__((hc_grid_launch)) __attribute__((used)) #else //#warning "GGL global define reached" -#define __global__ [[hc]] __attribute__((weak)) +#define __global__ __attribute__((hc, weak)) #endif //GENERIC_GRID_LAUNCH #define __noinline__ __attribute__((noinline)) diff --git a/include/hip/hcc_detail/math_functions.h b/include/hip/hcc_detail/math_functions.h index c3b8186fd3..9faff2743a 100644 --- a/include/hip/hcc_detail/math_functions.h +++ b/include/hip/hcc_detail/math_functions.h @@ -51,6 +51,7 @@ __device__ float exp10f(float x); __device__ float exp2f(float x); __device__ float expf(float x); __device__ float expm1f(float x); +__device__ int abs(int x); __device__ float fabsf(float x); __device__ float fdimf(float x, float y); __device__ float fdividef(float x, float y); diff --git a/include/hip/hip_fp16.h b/include/hip/hip_fp16.h index 0e002d9396..95879dba50 100644 --- a/include/hip/hip_fp16.h +++ b/include/hip/hip_fp16.h @@ -20,7 +20,7 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ -#ifdef HIP_INCLUDE_HIP_HIP_FP16_H +#ifndef HIP_INCLUDE_HIP_HIP_FP16_H #define HIP_INCLUDE_HIP_HIP_FP16_H #include diff --git a/packaging/hip-targets-release.cmake b/packaging/hip-targets-release.cmake new file mode 100644 index 0000000000..ba0a5005f5 --- /dev/null +++ b/packaging/hip-targets-release.cmake @@ -0,0 +1,41 @@ +#---------------------------------------------------------------- +# Generated CMake target import file for configuration "Release". +#---------------------------------------------------------------- + +# Commands may need to know the format version. +set(CMAKE_IMPORT_FILE_VERSION 1) + +# Import target "hip::hip_hcc_static" for configuration "Release" +set_property(TARGET hip::hip_hcc_static APPEND PROPERTY IMPORTED_CONFIGURATIONS RELEASE) +set_target_properties(hip::hip_hcc_static PROPERTIES + IMPORTED_LINK_INTERFACE_LANGUAGES_RELEASE "CXX" + IMPORTED_LINK_INTERFACE_LIBRARIES_RELEASE "hc_am" + IMPORTED_LOCATION_RELEASE "/opt/rocm/hip/lib/libhip_hcc_static.a" + ) + +list(APPEND _IMPORT_CHECK_TARGETS hip::hip_hcc_static ) +list(APPEND _IMPORT_CHECK_FILES_FOR_hip::hip_hcc_static "/opt/rocm/hip/lib/libhip_hcc_static.a" ) + +# Import target "hip::hip_hcc" for configuration "Release" +set_property(TARGET hip::hip_hcc APPEND PROPERTY IMPORTED_CONFIGURATIONS RELEASE) +set_target_properties(hip::hip_hcc PROPERTIES + IMPORTED_LINK_INTERFACE_LIBRARIES_RELEASE "hcc::hccrt;hcc::hc_am" + IMPORTED_LOCATION_RELEASE "/opt/rocm/hip/lib/libhip_hcc.so" + IMPORTED_SONAME_RELEASE "libhip_hcc.so" + ) + +list(APPEND _IMPORT_CHECK_TARGETS hip::hip_hcc ) +list(APPEND _IMPORT_CHECK_FILES_FOR_hip::hip_hcc "/opt/rocm/hip/lib/libhip_hcc.so" ) + +# Import target "hip::hip_device" for configuration "Release" +set_property(TARGET hip::hip_device APPEND PROPERTY IMPORTED_CONFIGURATIONS RELEASE) +set_target_properties(hip::hip_device PROPERTIES + IMPORTED_LINK_INTERFACE_LANGUAGES_RELEASE "CXX" + IMPORTED_LOCATION_RELEASE "/opt/rocm/hip/lib/libhip_device.a" + ) + +list(APPEND _IMPORT_CHECK_TARGETS hip::hip_device ) +list(APPEND _IMPORT_CHECK_FILES_FOR_hip::hip_device "/opt/rocm/hip/lib/libhip_device.a" ) + +# Commands beyond this point should not need to know the version. +set(CMAKE_IMPORT_FILE_VERSION) diff --git a/packaging/hip-targets.cmake b/packaging/hip-targets.cmake new file mode 100644 index 0000000000..65370eec9e --- /dev/null +++ b/packaging/hip-targets.cmake @@ -0,0 +1,102 @@ +# Generated by CMake 3.5.1 + +if("${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION}" LESS 2.5) + message(FATAL_ERROR "CMake >= 2.6.0 required") +endif() +cmake_policy(PUSH) +cmake_policy(VERSION 2.6) +#---------------------------------------------------------------- +# Generated CMake target import file. +#---------------------------------------------------------------- + +# Commands may need to know the format version. +set(CMAKE_IMPORT_FILE_VERSION 1) + +# Protect against multiple inclusion, which would fail when already imported targets are added once more. +set(_targetsDefined) +set(_targetsNotDefined) +set(_expectedTargets) +foreach(_expectedTarget hip::hip_hcc_static hip::hip_hcc hip::hip_device) + list(APPEND _expectedTargets ${_expectedTarget}) + if(NOT TARGET ${_expectedTarget}) + list(APPEND _targetsNotDefined ${_expectedTarget}) + endif() + if(TARGET ${_expectedTarget}) + list(APPEND _targetsDefined ${_expectedTarget}) + endif() +endforeach() +if("${_targetsDefined}" STREQUAL "${_expectedTargets}") + set(CMAKE_IMPORT_FILE_VERSION) + cmake_policy(POP) + return() +endif() +if(NOT "${_targetsDefined}" STREQUAL "") + message(FATAL_ERROR "Some (but not all) targets in this export set were already defined.\nTargets Defined: ${_targetsDefined}\nTargets not yet defined: ${_targetsNotDefined}\n") +endif() +unset(_targetsDefined) +unset(_targetsNotDefined) +unset(_expectedTargets) + + +# The installation prefix configured by this project. +set(_IMPORT_PREFIX "/opt/rocm/hip") + +# Create imported target hip::hip_hcc_static +add_library(hip::hip_hcc_static STATIC IMPORTED) + +set_target_properties(hip::hip_hcc_static PROPERTIES + INTERFACE_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include;/opt/rocm/hsa/include" + INTERFACE_SYSTEM_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include;/opt/rocm/hsa/include" +) + +# Create imported target hip::hip_hcc +add_library(hip::hip_hcc SHARED IMPORTED) + +set_target_properties(hip::hip_hcc PROPERTIES + INTERFACE_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include;/opt/rocm/hsa/include" + INTERFACE_SYSTEM_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include;/opt/rocm/hsa/include" +) + +# Create imported target hip::hip_device +add_library(hip::hip_device STATIC IMPORTED) + +set_target_properties(hip::hip_device PROPERTIES + INTERFACE_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include;/opt/rocm/hsa/include" + INTERFACE_SYSTEM_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include;/opt/rocm/hsa/include" +) + +# Load information for each installed configuration. +get_filename_component(_DIR "${CMAKE_CURRENT_LIST_FILE}" PATH) +file(GLOB CONFIG_FILES "${_DIR}/hip-targets-*.cmake") +foreach(f ${CONFIG_FILES}) + include(${f}) +endforeach() + +# Cleanup temporary variables. +set(_IMPORT_PREFIX) + +# Loop over all imported files and verify that they actually exist +foreach(target ${_IMPORT_CHECK_TARGETS} ) + foreach(file ${_IMPORT_CHECK_FILES_FOR_${target}} ) + if(NOT EXISTS "${file}" ) + message(FATAL_ERROR "The imported target \"${target}\" references the file + \"${file}\" +but this file does not exist. Possible reasons include: +* The file was deleted, renamed, or moved to another location. +* An install or uninstall procedure did not complete successfully. +* The installation package was faulty and contained + \"${CMAKE_CURRENT_LIST_FILE}\" +but not all the files it references. +") + endif() + endforeach() + unset(_IMPORT_CHECK_FILES_FOR_${target}) +endforeach() +unset(_IMPORT_CHECK_TARGETS) + +# This file does not depend on other imported targets which have +# been exported from the same project but in a separate export set. + +# Commands beyond this point should not need to know the version. +set(CMAKE_IMPORT_FILE_VERSION) +cmake_policy(POP) diff --git a/packaging/hip_hcc.txt b/packaging/hip_hcc.txt index 7dd65033fd..7118c32eb9 100644 --- a/packaging/hip_hcc.txt +++ b/packaging/hip_hcc.txt @@ -6,6 +6,8 @@ install(FILES @PROJECT_BINARY_DIR@/libhip_hcc_static.a DESTINATION lib) install(FILES @PROJECT_BINARY_DIR@/libhip_device.a DESTINATION lib) install(FILES @PROJECT_BINARY_DIR@/.hipInfo DESTINATION lib) install(FILES @hip_SOURCE_DIR@/src/hip_hc.ll @hip_SOURCE_DIR@/src/hip_hc_gfx803.ll DESTINATION lib) +install(FILES @PROJECT_BINARY_DIR@/hip-config.cmake @PROJECT_BINARY_DIR@/hip-config-version.cmake DESTINATION lib/cmake/hip) +install(FILES @hip_SOURCE_DIR@/packaging/hip-targets.cmake @hip_SOURCE_DIR@/packaging/hip-targets-release.cmake DESTINATION lib/cmake/hip) ############################# # Packaging steps diff --git a/src/grid_launch.cpp b/src/grid_launch.cpp index 7739995600..cac01df7dc 100644 --- a/src/grid_launch.cpp +++ b/src/grid_launch.cpp @@ -27,6 +27,9 @@ THE SOFTWARE. #include "hc.hpp" #include "trace_helper.h" +#include +#include + namespace hip_impl { hc::accelerator_view lock_stream_hip_( @@ -42,6 +45,39 @@ namespace hip_impl return (*static_cast(locked_stream))->_av; } + void print_prelaunch_trace_( + const char* kernel_name, + dim3 num_blocks, + dim3 dim_blocks, + int group_mem_bytes, + hipStream_t stream) + { + if ((HIP_TRACE_API & (1 << TRACE_CMD)) || + HIP_PROFILE_API || + (COMPILE_HIP_DB && HIP_TRACE_API)) { + std::stringstream os; + os << tls_tidInfo.tid() << "." << tls_tidInfo.apiSeqNum() + << " hipLaunchKernel '" << kernel_name << "'" + << " gridDim:" << num_blocks + << " groupDim:" << dim_blocks + << " sharedMem:+" << group_mem_bytes + << " " << *stream; + + if (HIP_PROFILE_API == 0x1) { + std::string shortAtpString("hipLaunchKernel:"); + shortAtpString += kernel_name; + MARKER_BEGIN(shortAtpString.c_str(), "HIP"); + } else if (HIP_PROFILE_API == 0x2) { + MARKER_BEGIN(os.str().c_str(), "HIP"); + } + + if (COMPILE_HIP_DB && HIP_TRACE_API) { + std::cerr << API_COLOR << os.str() << API_COLOR_END + << std::endl; + } + } + } + void unlock_stream_hip_( hipStream_t stream, void* locked_stream, diff --git a/src/math_functions.cpp b/src/math_functions.cpp index 92cc8689fc..3472216309 100644 --- a/src/math_functions.cpp +++ b/src/math_functions.cpp @@ -114,6 +114,10 @@ __device__ float expm1f(float x) { return hc::precise_math::expm1f(x); } +__device__ int abs(int x) +{ + return x >= 0 ? x : -x; // TODO - optimize with OCML +} __device__ float fabsf(float x) { return hc::precise_math::fabsf(x); From 8ff755b6ccf3c5e4b47cc84c4ff45978eacc4410 Mon Sep 17 00:00:00 2001 From: James Edwards Date: Mon, 17 Apr 2017 23:58:34 -0500 Subject: [PATCH 005/108] Fix RPM HIP packages from specifying /opt Change-Id: Iec3c3b81eef4c8888d425eefc80b12488a8d20a1 (cherry picked from commit 7699aeb3d37576ea3fdfb0eb9f2333386f9415a6) --- packaging/hip_base.txt | 1 + packaging/hip_doc.txt | 1 + packaging/hip_hcc.txt | 1 + packaging/hip_nvcc.txt | 1 + packaging/hip_samples.txt | 1 + 5 files changed, 5 insertions(+) diff --git a/packaging/hip_base.txt b/packaging/hip_base.txt index a208bc3463..836a82657b 100644 --- a/packaging/hip_base.txt +++ b/packaging/hip_base.txt @@ -33,5 +33,6 @@ set(CPACK_RPM_POST_INSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/postinst") set(CPACK_RPM_PRE_UNINSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/prerm") set(CPACK_RPM_PACKAGE_AUTOREQPROV " no") set(CPACK_RPM_PACKAGE_REQUIRES "perl >= 5.0") +set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt") set(CPACK_SOURCE_GENERATOR "TGZ") include(CPack) diff --git a/packaging/hip_doc.txt b/packaging/hip_doc.txt index d5a0c471b1..6f602c84cf 100644 --- a/packaging/hip_doc.txt +++ b/packaging/hip_doc.txt @@ -36,5 +36,6 @@ set(CPACK_BINARY_RPM "ON") set(CPACK_RPM_PACKAGE_ARCHITECTURE "x86_64") set(CPACK_RPM_PACKAGE_AUTOREQPROV " no") set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}") +set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt") set(CPACK_SOURCE_GENERATOR "TGZ") include(CPack) diff --git a/packaging/hip_hcc.txt b/packaging/hip_hcc.txt index 7118c32eb9..b0808aa0bc 100644 --- a/packaging/hip_hcc.txt +++ b/packaging/hip_hcc.txt @@ -46,5 +46,6 @@ if(@COMPILE_HIP_ATP_MARKER@) else() set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}, ${HCC_PACKAGE_NAME} = @HCC_PACKAGE_VERSION@") endif() +set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt") set(CPACK_SOURCE_GENERATOR "TGZ") include(CPack) diff --git a/packaging/hip_nvcc.txt b/packaging/hip_nvcc.txt index ea4943f282..0d7c357623 100644 --- a/packaging/hip_nvcc.txt +++ b/packaging/hip_nvcc.txt @@ -25,5 +25,6 @@ set(CPACK_RPM_PACKAGE_ARCHITECTURE "x86_64") #set(CPACK_RPM_PRE_UNINSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/prerm") set(CPACK_RPM_PACKAGE_AUTOREQPROV " no") set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}, cuda >= 7.5") +set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt") set(CPACK_SOURCE_GENERATOR "TGZ") include(CPack) diff --git a/packaging/hip_samples.txt b/packaging/hip_samples.txt index f289f2a8e5..6d34a6fd40 100644 --- a/packaging/hip_samples.txt +++ b/packaging/hip_samples.txt @@ -24,5 +24,6 @@ set(CPACK_BINARY_RPM "ON") set(CPACK_RPM_PACKAGE_ARCHITECTURE "x86_64") set(CPACK_RPM_PACKAGE_AUTOREQPROV " no") set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}") +set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt") set(CPACK_SOURCE_GENERATOR "TGZ") include(CPack) From 0e9a3371e742b69f933d35a7c15c232e57cfb12a Mon Sep 17 00:00:00 2001 From: James Edwards Date: Fri, 21 Apr 2017 22:34:26 -0500 Subject: [PATCH 006/108] Properly link hip cmake file into top level lib directory. Change-Id: I2113a86ca6985f34fd0cfb091abdbce0f632cfc2 (cherry picked from commit 36353a560f214fbcde9b762141112bc2a8fac6eb) --- packaging/hip_hcc.postinst | 15 ++++++++++----- packaging/hip_hcc.prerm | 10 +++++++--- 2 files changed, 17 insertions(+), 8 deletions(-) diff --git a/packaging/hip_hcc.postinst b/packaging/hip_hcc.postinst index 14179db767..e7d53b742b 100755 --- a/packaging/hip_hcc.postinst +++ b/packaging/hip_hcc.postinst @@ -8,17 +8,22 @@ popd () { } ROCMDIR=/opt/rocm -HIPDIR=$ROCMDIR/hip - -# Soft-link to libraries -HIPLIBFILES=$HIPDIR/lib/* ROCMLIBDIR=$ROCMDIR/lib +HIPDIR=$ROCMDIR/hip +HIPLIBDIR=$ROCMDIR/hip/lib + +# Soft-link to library files +HIPLIBFILES=$(ls -aF $HIPLIBDIR | grep -v [-/$]) mkdir -p $ROCMLIBDIR +mkdir -p $ROCMLIBDIR/cmake pushd $ROCMLIBDIR for f in $HIPLIBFILES do ln -s $f $(basename $f) done - ln -s $HIPDIR/lib/.hipInfo .hipInfo +# Make the hip cmake directory link. +pushd cmake +ln -s $HIPLIBDIR/cmake/hip hip +popd popd diff --git a/packaging/hip_hcc.prerm b/packaging/hip_hcc.prerm index dda313a3a4..ee64aea632 100755 --- a/packaging/hip_hcc.prerm +++ b/packaging/hip_hcc.prerm @@ -9,17 +9,21 @@ popd () { } ROCMDIR=/opt/rocm +ROCMLIBDIR=$ROCMDIR/lib HIPDIR=$ROCMDIR/hip +HIPLIBDIR=$ROCMDIR/hip/lib # Remove soft-links to libraries -HIPLIBFILES=$HIPDIR/lib/* -ROCMLIBDIR=$ROCMDIR/lib +HIPLIBFILES=$(ls -aF $HIPLIBDIR | grep -v [-/$]) pushd $ROCMLIBDIR for f in $HIPLIBFILES do rm $(basename $f) done -rm .hipInfo +pushd cmake +unlink hip +popd +rmdir --ignore-fail-on-non-empty cmake popd rmdir --ignore-fail-on-non-empty $ROCMLIBDIR From 090398d52226d32a7af05cb5822b3bad03d52a9d Mon Sep 17 00:00:00 2001 From: James Edwards Date: Sat, 22 Apr 2017 15:54:14 -0500 Subject: [PATCH 007/108] Specify full path of hip libraries in link file. Change-Id: I49b788f3489e7abff6b11006ff97fdfca4e5942c (cherry picked from commit 4869bf5a7cd02ad0bce8057eebbbe098ac30963b) --- packaging/hip_hcc.postinst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packaging/hip_hcc.postinst b/packaging/hip_hcc.postinst index e7d53b742b..c7f9c3184c 100755 --- a/packaging/hip_hcc.postinst +++ b/packaging/hip_hcc.postinst @@ -19,7 +19,7 @@ mkdir -p $ROCMLIBDIR/cmake pushd $ROCMLIBDIR for f in $HIPLIBFILES do - ln -s $f $(basename $f) + ln -s $HIPLIBDIR/$f $(basename $f) done # Make the hip cmake directory link. pushd cmake From 5433d17e2efce82d5f95b69759f0db2a18ee3dfc Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Mon, 24 Apr 2017 08:50:43 +0530 Subject: [PATCH 008/108] Merge branch 'amd-develop' into amd-master Change-Id: I312fb9d1181733ef5160d1e993e2ae57ced0f6b3 (cherry picked from commit 88fb807af081f31314031d4549e98d9b621cfc41) --- RELEASE.md | 46 + cmake/FindHIP.cmake | 8 +- ...A_Driver_API_functions_supported_by_HIP.md | 499 ++++++++ docs/markdown/hip_bugs.md | 19 + hipify-clang/README.md | 29 +- hipify-clang/src/Cuda2Hip.cpp | 1044 +++++++++++------ include/hip/nvcc_detail/hip_runtime.h | 4 + include/hip/nvcc_detail/hip_runtime_api.h | 4 +- samples/0_Intro/square/Makefile | 1 + samples/0_Intro/square/square.hipref.cpp | 2 +- 10 files changed, 1259 insertions(+), 397 deletions(-) create mode 100644 docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md diff --git a/RELEASE.md b/RELEASE.md index 34eab60833..21fd8da7bb 100644 --- a/RELEASE.md +++ b/RELEASE.md @@ -13,6 +13,52 @@ Upcoming: ## Revision History: +=================================================================================================== +Release: 1.0.17102 +Date: 2017.03.07 +- Lots of improvements to hipify-clang. +- Added HIP package config for cmake. +- Several bug fixes and documentation updates. + + +=================================================================================================== +Release: 1.0.17066 +Date: 2017.02.11 +- Improved support for math device functions. +- Added several half math device functions. +- Enabled support for CUDA 8.0 in hipify-clang. +- Lots of bug fixes and documentation updates. + + +=================================================================================================== +Release: 1.0.17015 +Date: 2017.01.06 +- Several improvements to the hipify-clang infrastructure. +- Refactored module and function APIs. +- HIP now defaults to linking against the shared runtime library. +- Documentation updates. + + +=================================================================================================== +Release: 1.0.16502 +Date: 2016.12.13 +- Added several fast math and packaged math instrincs +- Improved debug and profiler documentation +- Support for building and linking to HIP shared library +- Several improvements to hipify-clang +- Several bug fixes + + +=================================================================================================== +Release: 1.0.16461 +Date: 2016.11.14 +- Significant changes to the HIP Profiling APIs. Refer to the documentation for details +- Improvements to P2P support +- New API: hipDeviceGetByPCIBusId +- Several bug fixes in NV path +- hipModuleLaunch now works for multi-dim kernels + + =================================================================================================== Release:1.0 Date: 2016.11.8 diff --git a/cmake/FindHIP.cmake b/cmake/FindHIP.cmake index 0001436fee..5a5813ba0d 100644 --- a/cmake/FindHIP.cmake +++ b/cmake/FindHIP.cmake @@ -514,7 +514,9 @@ macro(HIP_ADD_EXECUTABLE hip_target) # Separate the sources from the options HIP_GET_SOURCES_AND_OPTIONS(_sources _cmake_options _hipcc_options _hcc_options _nvcc_options ${ARGN}) HIP_PREPARE_TARGET_COMMANDS(${hip_target} OBJ _generated_files _source_files ${_sources} HIPCC_OPTIONS ${_hipcc_options} HCC_OPTIONS ${_hcc_options} NVCC_OPTIONS ${_nvcc_options}) - list(REMOVE_ITEM _sources ${_source_files}) + if(_source_files) + list(REMOVE_ITEM _sources ${_source_files}) + endif() if("x${HCC_HOME}" STREQUAL "x") set(HCC_HOME "/opt/rocm/hcc") endif() @@ -530,7 +532,9 @@ macro(HIP_ADD_LIBRARY hip_target) # Separate the sources from the options HIP_GET_SOURCES_AND_OPTIONS(_sources _cmake_options _hipcc_options _hcc_options _nvcc_options ${ARGN}) HIP_PREPARE_TARGET_COMMANDS(${hip_target} OBJ _generated_files _source_files ${_sources} ${_cmake_options} HIPCC_OPTIONS ${_hipcc_options} HCC_OPTIONS ${_hcc_options} NVCC_OPTIONS ${_nvcc_options}) - list(REMOVE_ITEM _sources ${_source_files}) + if(_source_files) + list(REMOVE_ITEM _sources ${_source_files}) + endif() add_library(${hip_target} ${_cmake_options} ${_generated_files} ${_sources}) set_target_properties(${hip_target} PROPERTIES LINKER_LANGUAGE ${HIP_C_OR_CXX}) endmacro() diff --git a/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md new file mode 100644 index 0000000000..3434d29a70 --- /dev/null +++ b/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md @@ -0,0 +1,499 @@ +# CUDA Driver API functions supported by HIP + +## **1. Data types used by CUDA driver** + +| **type** | **CUDA** | **HIP** | **CUDA description** | +|-------------:|---------------------------------------------------------------|------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------| +| struct | `CUDA_ARRAY3D_DESCRIPTOR` | | | +| struct | `CUDA_ARRAY_DESCRIPTOR` | | | +| struct | `CUDA_MEMCPY2D` | | | +| struct | `CUDA_MEMCPY3D` | | | +| struct | `CUDA_MEMCPY3D_PEER` | | | +| struct | `CUDA_POINTER_ATTRIBUTE_P2P_TOKENS` | | | +| struct | `CUDA_RESOURCE_DESC` | | | +| struct | `CUDA_RESOURCE_VIEW_DESC` | | | +| struct | `CUdevprop` | `hipDeviceProp_t` | | +| struct | `CUipcEventHandle` | | | +| struct | `CUipcMemHandle` | | | +| enum |***`CUaddress_mode`*** | | Texture reference addressing modes | +| 0 |*`CU_TR_ADDRESS_MODE_WRAP`* | | Wrapping address mode | +| 1 |*`CU_TR_ADDRESS_MODE_CLAMP`* | | Clamp to edge address mode | +| 2 |*`CU_TR_ADDRESS_MODE_MIRROR`* | | Mirror address mode | +| 3 |*`CU_TR_ADDRESS_MODE_BORDER`* | | Border address mode | +| enum |***`CUarray_cubemap_face`*** | | Array indices for cube faces | +| 0x00 |*`CU_CUBEMAP_FACE_POSITIVE_X`* | | Positive X face of cubemap | +| 0x01 |*`CU_CUBEMAP_FACE_NEGATIVE_X`* | | Negative X face of cubemap | +| 0x02 |*`CU_CUBEMAP_FACE_POSITIVE_Y`* | | Positive Y face of cubemap | +| 0x03 |*`CU_CUBEMAP_FACE_NEGATIVE_Y`* | | Negative Y face of cubemap | +| 0x04 |*`CU_CUBEMAP_FACE_POSITIVE_Z`* | | Positive Z face of cubemap | +| 0x05 |*`CU_CUBEMAP_FACE_NEGATIVE_Z`* | | Negative Z face of cubemap | +| enum |***`CUarray_format`*** | | Array formats | +| 0x01 |*`CU_AD_FORMAT_UNSIGNED_INT8`* | | Unsigned 8-bit integers | +| 0x02 |*`CU_AD_FORMAT_UNSIGNED_INT16`* | | Unsigned 16-bit integers | +| 0x03 |*`CU_AD_FORMAT_UNSIGNED_INT32`* | | Unsigned 32-bit integers | +| 0x08 |*`CU_AD_FORMAT_SIGNED_INT8`* | | Signed 8-bit integers | +| 0x09 |*`CU_AD_FORMAT_SIGNED_INT16`* | | Signed 16-bit integers | +| 0x0a |*`CU_AD_FORMAT_SIGNED_INT32`* | | Signed 32-bit integers | +| 0x10 |*`CU_AD_FORMAT_HALF`* | | 16-bit floating point | +| 0x20 |*`CU_AD_FORMAT_FLOAT`* | | 32-bit floating point | +| enum |***`CUctx_flags`*** | | Context creation flags | +| 0x00 |*`CU_CTX_SCHED_AUTO`* | | Automatic scheduling | +| 0x01 |*`CU_CTX_SCHED_SPIN`* | | Set spin as default scheduling | +| 0x02 |*`CU_CTX_SCHED_YIELD`* | | Set yield as default scheduling | +| 0x04 |*`CU_CTX_SCHED_BLOCKING_SYNC`* | | Set blocking synchronization as default scheduling | +| 0x04 |*`CU_CTX_BLOCKING_SYNC`* | | Set blocking synchronization as default scheduling Deprecated. This flag was deprecated as of CUDA 4.0 and was replaced with CU_CTX_SCHED_BLOCKING_SYNC.| +| 0x07 |*`CU_CTX_SCHED_MASK`* | | | +| 0x08 |*`CU_CTX_MAP_HOST`* | | Support mapped pinned allocations | +| 0x10 |*`CU_CTX_LMEM_RESIZE_TO_MAX`* | | Keep local memory allocation after launch | +| 0x1f |*`CU_CTX_FLAGS_MASK`* | | | +| enum |***`CUdevice_attribute`*** | | Device properties | +| 1 |*`CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK`* |*`hipDeviceAttributeMaxThreadsPerBlock`* | Maximum number of threads per block | +| 2 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X`* |*`hipDeviceAttributeMaxBlockDimX`* | Maximum block dimension X | +| 3 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y`* |*`hipDeviceAttributeMaxBlockDimY`* | Maximum block dimension Y | +| 4 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z`* |*`hipDeviceAttributeMaxBlockDimZ`* | Maximum block dimension Z | +| 5 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X`* |*`hipDeviceAttributeMaxGridDimX`* | Maximum grid dimension X | +| 6 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y`* |*`hipDeviceAttributeMaxGridDimY`* | Maximum grid dimension Y | +| 7 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z`* |*`hipDeviceAttributeMaxGridDimZ`* | Maximum grid dimension Y | +| 8 |*`CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | Maximum shared memory available per block in bytes | +| 8 |*`CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | Deprecated, use CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK | +| 9 |*`CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY`* |*`hipDeviceAttributeTotalConstantMemory`* | Memory available on device for __constant__ variables in a CUDA C kernel in bytes | +| 10 |*`CU_DEVICE_ATTRIBUTE_WARP_SIZE`* |*`hipDeviceAttributeWarpSize`* | Warp size in threads | +| 11 |*`CU_DEVICE_ATTRIBUTE_MAX_PITCH`* | | Maximum pitch in bytes allowed by memory copies | +| 12 |*`CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | Maximum number of 32-bit registers available per block | +| 12 |*`CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | Deprecated, use CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK | +| 13 |*`CU_DEVICE_ATTRIBUTE_CLOCK_RATE`* |*`hipDeviceAttributeClockRate`* | Typical clock frequency in kilohertz | +| 14 |*`CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT`* | | Alignment requirement for textures | +| 15 |*`CU_DEVICE_ATTRIBUTE_GPU_OVERLAP`* | | Device can possibly copy memory and execute a kernel concurrently. Deprecated. Use instead CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT| +| 16 |*`CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT`* |*`hipDeviceAttributeMultiprocessorCount`* | Number of multiprocessors on device | +| 17 |*`CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT`* | | Specifies whether there is a run time limit on kernels | +| 18 |*`CU_DEVICE_ATTRIBUTE_INTEGRATED`* | | Device is integrated with host memory | +| 19 |*`CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY`* | | Device can map host memory into CUDA address space | +| 20 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_MODE`* |*`hipDeviceAttributeComputeMode`* | Compute mode (See CUcomputemode for details) | +| 21 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH`* | | Maximum 1D texture width | +| 22 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH`* | | Maximum 2D texture width | +| 23 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT`* | | Maximum 2D texture height | +| 24 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH`* | | Maximum 3D texture width | +| 25 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT`* | | Maximum 3D texture height | +| 26 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH`* | | Maximum 3D texture depth | +| 27 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH`* | | Maximum 2D layered texture width | +| 28 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT`* | | Maximum 2D layered texture height | +| 29 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS`* | | Maximum layers in a 2D layered texture | +| 27 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH`* | | Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH | +| 28 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT`* | | Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT | +| 29 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES`* | | Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS | +| 30 |*`CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT`* | | Alignment requirement for surfaces | +| 31 |*`CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS`* |*`hipDeviceAttributeConcurrentKernels`* | Device can possibly execute multiple kernels concurrently | +| 32 |*`CU_DEVICE_ATTRIBUTE_ECC_ENABLED`* | | Device has ECC support enabled | +| 33 |*`CU_DEVICE_ATTRIBUTE_PCI_BUS_ID`* |*`hipDeviceAttributePciBusId`* | PCI bus ID of the device | +| 34 |*`CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID`* |*`hipDeviceAttributePciDeviceId`* | PCI device ID of the device | +| 35 |*`CU_DEVICE_ATTRIBUTE_TCC_DRIVER`* | | Device is using TCC driver model | +| 36 |*`CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE`* |*`hipDeviceAttributeMemoryClockRate`* | Peak memory clock frequency in kilohertz | +| 37 |*`CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH`* |*`hipDeviceAttributeMemoryBusWidth`* | Global memory bus width in bits | +| 38 |*`CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE`* |*`hipDeviceAttributeL2CacheSize`* | Size of L2 cache in bytes | +| 39 |*`CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR`* |*`hipDeviceAttributeMaxThreadsPerMultiProcessor`* | Maximum resident threads per multiprocessor | +| 40 |*`CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT`* | | Number of asynchronous engines | +| 41 |*`CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING`* | | Device shares a unified address space with the host | +| 42 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH`* | | Maximum 1D layered texture width | +| 43 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS`* | | Maximum layers in a 1D layered texture | +| 44 |*`CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER`* | | Deprecated, do not use | +| 45 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH`* | | Maximum 2D texture width if CUDA_ARRAY3D_TEXTURE_GATHER is set | +| 46 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT`* | | Maximum 2D texture height if CUDA_ARRAY3D_TEXTURE_GATHER is set | +| 47 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE`* | | Alternate maximum 3D texture width | +| 48 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE`* | | Alternate maximum 3D texture height | +| 49 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE`* | | Alternate maximum 3D texture depth | +| 50 |*`CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID`* | | PCI domain ID of the device | +| 51 |*`CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT`* | | Pitch alignment requirement for textures | +| 52 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH`* | | Maximum cubemap texture width/height | +| 53 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH`* | | Maximum cubemap layered texture width/height | +| 54 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS`* | | Maximum layers in a cubemap layered texture | +| 55 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH`* | | Maximum 1D surface width | +| 56 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH`* | | Maximum 2D surface width | +| 57 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT`* | | Maximum 2D surface height | +| 58 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH`* | | Maximum 3D surface width | +| 59 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT`* | | Maximum 3D surface height | +| 60 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH`* | | Maximum 3D surface depth | +| 61 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH`* | | Maximum 1D layered surface width | +| 62 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS`* | | Maximum layers in a 1D layered surface | +| 63 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH`* | | Maximum 2D layered surface width | +| 64 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT`* | | Maximum 2D layered surface height | +| 65 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS`* | | Maximum layers in a 2D layered surface | +| 66 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH`* | | Maximum cubemap surface width | +| 67 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH`* | | Maximum cubemap layered surface width | +| 68 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS`* | | Maximum layers in a cubemap layered surface | +| 69 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH`* | | Maximum 1D linear texture width | +| 70 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH`* | | Maximum 2D linear texture width | +| 71 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT`* | | Maximum 2D linear texture height | +| 72 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH`* | | Maximum 2D linear texture pitch in bytes | +| 73 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH`* | | Maximum mipmapped 2D texture width | +| 74 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT`* | | Maximum mipmapped 2D texture height | +| 75 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR`* |*`hipDeviceAttributeComputeCapabilityMajor`* | Major compute capability version number | +| 76 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR`* |*`hipDeviceAttributeComputeCapabilityMinor`* | Minor compute capability version number | +| 77 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH`* | | Maximum mipmapped 1D texture width | +| 78 |*`CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED`* | | Device supports stream priorities | +| 79 |*`CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED`* | | Device supports caching globals in L1 | +| 80 |*`CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED`* | | Device supports caching locals in L1 | +| 81 |*`CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR`* |*`hipDeviceAttributeMaxSharedMemoryPerMultiprocessor`* | Maximum shared memory available per multiprocessor in bytes | +| 82 |*`CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR`* | | Maximum number of 32-bit registers available per multiprocessor | +| 83 |*`CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY`* |*`hipDeviceAttributeManagedMemory`* | Device can allocate managed memory on this system | +| 84 |*`CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD`* | | Device is on a multi-GPU board | +| 85 |*`CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID`* | | Unique id for a group of devices on the same multi-GPU board | +| 86 |*`CU_DEVICE_ATTRIBUTE_MAX`* | | | +| enum |***`CUevent_flags`*** | | Event creation flags | +| 0x00 |*`CU_EVENT_DEFAULT`* |*`hipEventDefault`* | Default event flag | +| 0x01 |*`CU_EVENT_BLOCKING_SYNC`* |*`hipEventBlockingSync`* | Event uses blocking synchronization | +| 0x02 |*`CU_EVENT_DISABLE_TIMING`* |*`hipEventDisableTiming`* | Event will not record timing data | +| 0x04 |*`CU_EVENT_INTERPROCESS`* |*`hipEventInterprocess`* | Event is suitable for interprocess use. CU_EVENT_DISABLE_TIMING must be set | +| enum |***`CUfilter_mode`*** |***`hipTextureFilterMode`*** | Texture reference filtering modes | +| 0 |*`CU_TR_FILTER_MODE_POINT`* |*`hipFilterModePoint`* | Point filter mode | +| 1 |*`CU_TR_FILTER_MODE_LINEAR`* |*`hipFilterModeLinear`* | Linear filter mode | +| enum |***`CUfunc_cache`*** |***`hipFuncCache`*** | Function cache configurations | +| 0x00 |*`CU_FUNC_CACHE_PREFER_NONE`* |*`hipFuncCachePreferNone`* | no preference for shared memory or L1 (default) | +| 0x01 |*`CU_FUNC_CACHE_PREFER_SHARED`* |*`hipFuncCachePreferShared`* | prefer larger shared memory and smaller L1 cache | +| 0x02 |*`CU_FUNC_CACHE_PREFER_L1`* |*`hipFuncCachePreferL1`* | prefer larger L1 cache and smaller shared memory | +| 0x03 |*`CU_FUNC_CACHE_PREFER_EQUAL`* |*`hipFuncCachePreferEqual`* | prefer equal sized L1 cache and shared memory | +| enum |***`CUfunction_attribute`*** | | Function properties | +| 0 |*`CU_FUNC_ATTRIBUTE_MAX_THREADS_PER_BLOCK`* | | The maximum number of threads per block, beyond which a launch of the function would fail. This number depends on both the function and the device on which the function is currently loaded. | +| 1 |*`CU_FUNC_ATTRIBUTE_SHARED_SIZE_BYTES`* | | The size in bytes of statically-allocated shared memory required by this function. This does not include dynamically-allocated shared memory requested by the user at runtime. | +| 2 |*`CU_FUNC_ATTRIBUTE_CONST_SIZE_BYTES`* | | The size in bytes of user-allocated constant memory required by this function. | +| 3 |*`CU_FUNC_ATTRIBUTE_LOCAL_SIZE_BYTES`* | | The size in bytes of local memory used by each thread of this function. | +| 4 |*`CU_FUNC_ATTRIBUTE_NUM_REGS`* | | The number of registers used by each thread of this function. | +| 5 |*`CU_FUNC_ATTRIBUTE_PTX_VERSION`* | | The PTX virtual architecture version for which the function was compiled. This value is the major PTX version * 10 + the minor PTX version, so a PTX version 1.3 function would return the value 13. Note that this may return the undefined value of 0 for cubins compiled prior to CUDA 3.0. | +| 6 |*`CU_FUNC_ATTRIBUTE_BINARY_VERSION`* | | The binary architecture version for which the function was compiled. This value is the major binary version * 10 + the minor binary version, so a binary version 1.3 function would return the value 13. Note that this will return a value of 10 for legacy cubins that do not have a properly-encoded binary architecture version. | +| 7 |*`CU_FUNC_ATTRIBUTE_CACHE_MODE_CA`* | | The attribute to indicate whether the function has been compiled with user specified option "-Xptxas --dlcm=ca" set. | +| 8 |*`CU_FUNC_ATTRIBUTE_MAX`* | | | +| enum |***`CUgraphicsMapResourceFlags`*** | | Flags for mapping and unmapping interop resources | +| 0x00 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE`* | | | +| 0x01 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY`* | | | +| 0x02 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_WRITE_DISCARD`* | | | +| enum |***`CUgraphicsRegisterFlags`*** | | Flags to register a graphics resource | +| 0x00 |*`CU_GRAPHICS_REGISTER_FLAGS_NONE`* | | | +| 0x01 |*`CU_GRAPHICS_REGISTER_FLAGS_READ_ONLY`* | | | +| 0x02 |*`CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD`* | | | +| 0x04 |*`CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST`* | | | +| 0x08 |*`CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER`* | | | +| enum |***`CUipcMem_flags`*** | | CUDA Ipc Mem Flags | +| 0x1 |*`CU_IPC_MEM_LAZY_ENABLE_PEER_ACCESS`* |*`hipIpcMemLazyEnablePeerAccess`* | Automatically enable peer access between remote devices as needed | +| enum |***`CUjit_cacheMode`*** | | Caching modes for dlcm | +| 0 |*`CU_JIT_CACHE_OPTION_NONE`* | | Compile with no -dlcm flag specified | +| |*`CU_JIT_CACHE_OPTION_CG`* | | Compile with L1 cache disabled | +| |*`CU_JIT_CACHE_OPTION_CA`* | | Compile with L1 cache enabled | +| enum |***`CUjit_fallback`*** | | Cubin matching fallback strategies | +| 0 |*`CU_PREFER_PTX`* | | Prefer to compile ptx if exact binary match not found | +| |*`CU_PREFER_BINARY`* | | Prefer to fall back to compatible binary code if exact match not found | +| enum |***`CUjit_option`*** | | Online compiler and linker options | +| 0 |*`CU_JIT_MAX_REGISTERS`* | | Max number of registers that a thread may use. Option type: unsigned int Applies to: compiler only. | +| |*`CU_JIT_THREADS_PER_BLOCK`* | | IN: Specifies minimum number of threads per block to target compilation for OUT: Returns the number of threads the compiler actually targeted. This restricts the resource utilization fo the compiler (e.g. max registers) such that a block with the given number of threads should be able to launch based on register limitations. Note, this option does not currently take into account any other resource limitations, such as shared memory utilization. Cannot be combined with CU_JIT_TARGET. Option type: unsigned int Applies to: compiler only. | +| |*`CU_JIT_WALL_TIME`* | | Overwrites the option value with the total wall clock time, in milliseconds, spent in the compiler and linker Option type: float Applies to: compiler and linker. | +| |*`CU_JIT_INFO_LOG_BUFFER`* | | Pointer to a buffer in which to print any log messages that are informational in nature (the buffer size is specified via option CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES) Option type: char * Applies to: compiler and linker. | +| |*`CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES`* | | IN: Log buffer size in bytes. Log messages will be capped at this size (including null terminator) OUT: Amount of log buffer filled with messages Option type: unsigned int Applies to: compiler and linker. | +| |*`CU_JIT_OPTIMIZATION_LEVEL`* | | Level of optimizations to apply to generated code (0 - 4), with 4 being the default and highest level of optimizations. Option type: unsigned int Applies to: compiler only. | +| |*`CU_JIT_TARGET_FROM_CUCONTEXT`* | | No option value required. Determines the target based on the current attached context (default) Option type: No option value needed Applies to: compiler and linker. | +| |*`CU_JIT_TARGET`* | | Target is chosen based on supplied CUjit_target. Cannot be combined with CU_JIT_THREADS_PER_BLOCK. Option type: unsigned int for enumerated type CUjit_target Applies to: compiler and linker. | +| |*`CU_JIT_FALLBACK_STRATEGY`* | | Specifies choice of fallback strategy if matching cubin is not found. Choice is based on supplied CUjit_fallback. This option cannot be used with cuLink* APIs as the linker requires exact matches. Option type: unsigned int for enumerated type CUjit_fallback Applies to: compiler only. | +| |*`CU_JIT_GENERATE_DEBUG_INFO`* | | Specifies whether to create debug information in output (-g) (0: false, default) Option type: int Applies to: compiler and linker. | +| |*`CU_JIT_LOG_VERBOSE`* | | Generate verbose log messages (0: false, default) Option type: int Applies to: compiler and linker. | +| |*`CU_JIT_GENERATE_LINE_INFO`* | | Generate line number information (-lineinfo) (0: false, default) Option type: int Applies to: compiler only. | +| |*`CU_JIT_CACHE_MODE`* | | Specifies whether to enable caching explicitly (-dlcm) Choice is based on supplied CUjit_cacheMode_enum. Option type: unsigned int for enumerated type CUjit_cacheMode_enum Applies to: compiler only. | +| |*`CU_JIT_NUM_OPTIONS`* | | | +| enum |***`CUjit_target`*** | | Online compilation targets | +| 10 |*`CU_TARGET_COMPUTE_10`* | | Compute device class 1.0. | +| 11 |*`CU_TARGET_COMPUTE_11`* | | Compute device class 1.1. | +| 12 |*`CU_TARGET_COMPUTE_12`* | | Compute device class 1.2. | +| 13 |*`CU_TARGET_COMPUTE_13`* | | Compute device class 1.3. | +| 20 |*`CU_TARGET_COMPUTE_20`* | | Compute device class 2.0. | +| 21 |*`CU_TARGET_COMPUTE_21`* | | Compute device class 2.1. | +| 30 |*`CU_TARGET_COMPUTE_30`* | | Compute device class 3.0. | +| 32 |*`CU_TARGET_COMPUTE_32`* | | Compute device class 3.2. | +| 35 |*`CU_TARGET_COMPUTE_35`* | | Compute device class 3.5. | +| 37 |*`CU_TARGET_COMPUTE_37`* | | Compute device class 3.7. | +| 50 |*`CU_TARGET_COMPUTE_50`* | | Compute device class 5.0. | +| 52 |*`CU_TARGET_COMPUTE_52`* | | Compute device class 5.2. | +| enum |***`CUjitInputType`*** | | Device code formats | +| 0 |*`CU_JIT_INPUT_CUBIN`* | | Compiled device-class-specific device code Applicable options: none. | +| |*`CU_JIT_INPUT_PTX`* | | PTX source code Applicable options: PTX compiler options. | +| |*`CU_JIT_INPUT_FATBINARY`* | | Bundle of multiple cubins and/or PTX of some device code Applicable options: PTX compiler options, CU_JIT_FALLBACK_STRATEGY. | +| |*`CU_JIT_INPUT_OBJECT`* | | Host object with embedded device code Applicable options: PTX compiler options, CU_JIT_FALLBACK_STRATEGY. | +| |*`CU_JIT_INPUT_LIBRARY`* | | Archive of host objects with embedded device code Applicable options: PTX compiler options, CU_JIT_FALLBACK_STRATEGY. | +| |*`CU_JIT_NUM_INPUT_TYPES`* | | | +| enum |***`CUlimit`*** |***`hipLimit_t`*** | Limits | +| 0x00 |*`CU_LIMIT_STACK_SIZE`* | | GPU thread stack size. | +| 0x01 |*`CU_LIMIT_PRINTF_FIFO_SIZE`* | | GPU printf FIFO size. | +| 0x02 |*`CU_LIMIT_MALLOC_HEAP_SIZE`* |*`hipLimitMallocHeapSize`* | GPU malloc heap size. | +| 0x03 |*`CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH`* | | GPU device runtime launch synchronize depth. | +| 0x04 |*`CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT`* | | GPU device runtime pending launch count. | +| |*`CU_LIMIT_MAX`* | | | +| enum |***`CUmemAttach_flags`*** | | CUDA Mem Attach Flags | +| 0x1 |*`CU_MEM_ATTACH_GLOBAL`* | | Memory can be accessed by any stream on any device. | +| 0x2 |*`CU_MEM_ATTACH_HOST`* | | Memory cannot be accessed by any stream on any device. | +| 0x4 |*`CU_MEM_ATTACH_SINGLE`* | | Memory can only be accessed by a single stream on the associated device. | +| enum |***`CUmemorytype`*** | | Memory types | +| 0x01 |*`CU_MEMORYTYPE_HOST`* | | Host memory | +| 0x02 |*`CU_MEMORYTYPE_DEVICE`* | | Device memory | +| 0x03 |*`CU_MEMORYTYPE_ARRAY`* | | Array memory | +| 0x04 |*`CU_MEMORYTYPE_UNIFIED`* | | Unified device or host memory | +| enum |***`CUoccupancy_flags`*** | | Occupancy calculator flag | +| 0x00 |*`CU_OCCUPANCY_DEFAULT`* | | Default behavior | +| 0x01 |*`CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE`* | | Assume global caching is enabled and cannot be automatically turned off | +| enum |***`CUpointer_attribute`*** | | Pointer information | +| 1 |*`CU_POINTER_ATTRIBUTE_CONTEXT`* | | The CUcontext on which a pointer was allocated or registered | +| 2 |*`CU_POINTER_ATTRIBUTE_MEMORY_TYPE`* | | The CUmemorytype describing the physical location of a pointer | +| 3 |*`CU_POINTER_ATTRIBUTE_DEVICE_POINTER`* | | The address at which a pointer's memory may be accessed on the device | +| 4 |*`CU_POINTER_ATTRIBUTE_HOST_POINTER`* | | The address at which a pointer's memory may be accessed on the host | +| 5 |*`CU_POINTER_ATTRIBUTE_P2P_TOKENS`* | | A pair of tokens for use with the nv-p2p.h Linux kernel interface | +| 6 |*`CU_POINTER_ATTRIBUTE_SYNC_MEMOPS`* | | Synchronize every synchronous memory operation initiated on this region | +| 7 |*`CU_POINTER_ATTRIBUTE_BUFFER_ID`* | | A process-wide unique ID for an allocated memory region | +| 8 |*`CU_POINTER_ATTRIBUTE_IS_MANAGED`* | | Indicates if the pointer points to managed memory | +| enum |***`CUmemorytype`*** | | Resource types | +| 0x00 |*`CU_RESOURCE_TYPE_ARRAY`* | | Array resoure | +| 0x01 |*`CU_RESOURCE_TYPE_MIPMAPPED_ARRAY`* | | Mipmapped array resource | +| 0x02 |*`CU_RESOURCE_TYPE_LINEAR`* | | Linear resource | +| 0x03 |*`CU_RESOURCE_TYPE_PITCH2D`* | | Pitch 2D resource | +| enum |***`CUresourceViewFormat`*** | | Resource view format | +| 0x00 |*`CU_RES_VIEW_FORMAT_NONE`* | | No resource view format (use underlying resource format) | +| 0x01 |*`CU_RES_VIEW_FORMAT_UINT_1X8`* | | 1 channel unsigned 8-bit integers | +| 0x02 |*`CU_RES_VIEW_FORMAT_UINT_2X8`* | | 2 channel unsigned 8-bit integers | +| 0x03 |*`CU_RES_VIEW_FORMAT_UINT_4X8`* | | 4 channel unsigned 8-bit integers | +| 0x04 |*`CU_RES_VIEW_FORMAT_SINT_1X8`* | | 1 channel signed 8-bit integers | +| 0x05 |*`CU_RES_VIEW_FORMAT_SINT_2X8`* | | 2 channel signed 8-bit integers | +| 0x06 |*`CU_RES_VIEW_FORMAT_SINT_4X8`* | | 4 channel signed 8-bit integers | +| 0x07 |*`CU_RES_VIEW_FORMAT_UINT_1X16`* | | 1 channel unsigned 16-bit integers | +| 0x08 |*`CU_RES_VIEW_FORMAT_UINT_2X16`* | | 2 channel unsigned 16-bit integers | +| 0x09 |*`CU_RES_VIEW_FORMAT_UINT_4X16`* | | 4 channel unsigned 16-bit integers | +| 0x0a |*`CU_RES_VIEW_FORMAT_SINT_1X16`* | | 1 channel signed 16-bit integers | +| 0x0b |*`CU_RES_VIEW_FORMAT_SINT_2X16`* | | 2 channel signed 16-bit integers | +| 0x0c |*`CU_RES_VIEW_FORMAT_SINT_4X16`* | | 4 channel signed 16-bit integers | +| 0x0d |*`CU_RES_VIEW_FORMAT_UINT_1X32`* | | 1 channel unsigned 32-bit integers | +| 0x0e |*`CU_RES_VIEW_FORMAT_UINT_2X32`* | | 2 channel unsigned 32-bit integers | +| 0x0f |*`CU_RES_VIEW_FORMAT_UINT_4X32`* | | 4 channel unsigned 32-bit integers | +| 0x10 |*`CU_RES_VIEW_FORMAT_SINT_1X32`* | | 1 channel signed 32-bit integers | +| 0x11 |*`CU_RES_VIEW_FORMAT_SINT_2X32`* | | 2 channel signed 32-bit integers | +| 0x12 |*`CU_RES_VIEW_FORMAT_SINT_4X32`* | | 4 channel signed 32-bit integers | +| 0x13 |*`CU_RES_VIEW_FORMAT_FLOAT_1X16`* | | 1 channel 16-bit floating point | +| 0x14 |*`CU_RES_VIEW_FORMAT_FLOAT_2X16`* | | 2 channel 16-bit floating point | +| 0x15 |*`CU_RES_VIEW_FORMAT_FLOAT_4X16`* | | 4 channel 16-bit floating point | +| 0x16 |*`CU_RES_VIEW_FORMAT_FLOAT_1X32`* | | 1 channel 32-bit floating point | +| 0x17 |*`CU_RES_VIEW_FORMAT_FLOAT_2X32`* | | 2 channel 32-bit floating point | +| 0x18 |*`CU_RES_VIEW_FORMAT_FLOAT_4X32`* | | 4 channel 32-bit floating point | +| 0x19 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC1`* | | Block compressed 1 | +| 0x1a |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC3`* | | Block compressed 2 | +| 0x1b |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC3`* | | Block compressed 3 | +| 0x1c |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC4`* | | Block compressed 4 unsigned | +| 0x1d |*`CU_RES_VIEW_FORMAT_SIGNED_BC4`* | | Block compressed 4 signed | +| 0x1e |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC5`* | | Block compressed 5 unsigned | +| 0x1f |*`CU_RES_VIEW_FORMAT_SIGNED_BC5`* | | Block compressed 5 signed | +| 0x20 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC6H`* | | Block compressed 6 unsigned half-float | +| 0x21 |*`CU_RES_VIEW_FORMAT_SIGNED_BC6H`* | | Block compressed 6 signed half-float | +| 0x22 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC7`* | | Block compressed 7 | +| enum |***`CUresult`*** |***`hipError_t`*** | Error codes | +| 0 |*`CUDA_SUCCESS`* |*`hipSuccess`* | The API call returned with no errors. In the case of query calls, this can also mean that the operation being queried is complete (see cuEventQuery() and cuStreamQuery()). | +| 1 |*`CUDA_ERROR_INVALID_VALUE`* |*`hipErrorInvalidValue`* | This indicates that one or more of the parameters passed to the API call is not within an acceptable range of values. | +| 2 |*`CUDA_ERROR_OUT_OF_MEMORY`* |*`hipErrorMemoryAllocation`* | The API call failed because it was unable to allocate enough memory to perform the requested operation. | +| 3 |*`CUDA_ERROR_NOT_INITIALIZED`* |*`hipErrorNotInitialized`* | This indicates that the CUDA driver has not been initialized with cuInit() or that initialization has failed. | +| 4 |*`CUDA_ERROR_DEINITIALIZED`* |*`hipErrorDeinitialized`* | This indicates that the CUDA driver is in the process of shutting down. | +| 5 |*`CUDA_ERROR_PROFILER_DISABLED`* |*`hipErrorProfilerDisabled`* | This indicates profiler is not initialized for this run. This can happen when the application is running with external profiling tools like visual profiler. | +| 6 |*`CUDA_ERROR_PROFILER_NOT_INITIALIZED`* |*`hipErrorProfilerNotInitialized`* | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to attempt to enable/disable the profiling via cuProfilerStart or cuProfilerStop without initialization. | +| 7 |*`CUDA_ERROR_PROFILER_ALREADY_STARTED`* |*`hipErrorProfilerAlreadyStarted`* | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to call cuProfilerStart() when profiling is already enabled. | +| 8 |*`CUDA_ERROR_PROFILER_ALREADY_STOPPED`* |*`hipErrorProfilerAlreadyStopped`* | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to call cuProfilerStop() when profiling is already disabled. | +| 100 |*`CUDA_ERROR_NO_DEVICE`* |*`hipErrorNoDevice`* | This indicates that no CUDA-capable devices were detected by the installed CUDA driver. | +| 101 |*`CUDA_ERROR_INVALID_DEVICE`* |*`hipErrorInvalidDevice`* | This indicates that the device ordinal supplied by the user does not correspond to a valid CUDA device. | +| 200 |*`CUDA_ERROR_INVALID_IMAGE`* |*`hipErrorInvalidImage`* | This indicates that the device kernel image is invalid. This can also indicate an invalid CUDA module. | +| 201 |*`CUDA_ERROR_INVALID_CONTEXT`* |*`hipErrorInvalidContext`* | This most frequently indicates that there is no context bound to the current thread. This can also be returned if the context passed to an API call is not a valid handle (such as a context that has had cuCtxDestroy() invoked on it). This can also be returned if a user mixes different API versions (i.e. 3010 context with 3020 API calls). See cuCtxGetApiVersion() for more details. | +| 202 |*`CUDA_ERROR_CONTEXT_ALREADY_CURRENT`* |*`hipErrorContextAlreadyCurrent`* | This indicated that the context being supplied as a parameter to the API call was already the active context. Deprecated This error return is deprecated as of CUDA 3.2. It is no longer an error to attempt to push the active context via cuCtxPushCurrent(). | +| 205 |*`CUDA_ERROR_MAP_FAILED`* |*`hipErrorMapFailed`* | This indicates that a map or register operation has failed. | +| 206 |*`CUDA_ERROR_UNMAP_FAILED`* |*`hipErrorUnmapFailed`* | This indicates that an unmap or unregister operation has failed. | +| 207 |*`CUDA_ERROR_ARRAY_IS_MAPPED`* |*`hipErrorArrayIsMapped`* | This indicates that the specified array is currently mapped and thus cannot be destroyed. | +| 208 |*`CUDA_ERROR_ALREADY_MAPPED`* |*`hipErrorAlreadyMapped`* | This indicates that the resource is already mapped. | +| 209 |*`CUDA_ERROR_NO_BINARY_FOR_GPU`* |*`hipErrorNoBinaryForGpu* | This indicates that there is no kernel image available that is suitable for the device. This can occur when a user specifies code generation options for a particular CUDA source file that do not include the corresponding device configuration. | +| 210 |*`CUDA_ERROR_ALREADY_ACQUIRED`* |*`hipErrorAlreadyAcquired* | This indicates that a resource has already been acquired. | +| 211 |*`CUDA_ERROR_NOT_MAPPED`* |*`hipErrorNotMapped`* | This indicates that a resource is not mapped. | +| 212 |*`CUDA_ERROR_NOT_MAPPED_AS_ARRAY`* |*`hipErrorNotMappedAsArray`* | This indicates that a mapped resource is not available for access as an array. | +| 213 |*`CUDA_ERROR_NOT_MAPPED_AS_POINTER`* |*`hipErrorNotMappedAsPointer`* | This indicates that a mapped resource is not available for access as a pointer. | +| 214 |*`CUDA_ERROR_ECC_UNCORRECTABLE`* |*`hipErrorECCNotCorrectable`* | This indicates that an uncorrectable ECC error was detected during execution. | +| 215 |*`CUDA_ERROR_UNSUPPORTED_LIMIT`* |*`hipErrorUnsupportedLimit`* | This indicates that the CUlimit passed to the API call is not supported by the active device. | +| 216 |*`CUDA_ERROR_CONTEXT_ALREADY_IN_USE`* |*`hipErrorContextAlreadyInUse`* | This indicates that the CUcontext passed to the API call can only be bound to a single CPU thread at a time but is already bound to a CPU thread. | +| 217 |*`CUDA_ERROR_PEER_ACCESS_UNSUPPORTED`* |*`hipErrorPeerAccessUnsupported`* | This indicates that peer access is not supported across the given devices. | +| 218 |*`CUDA_ERROR_INVALID_PTX`* |*`hipErrorInvalidKernelFile`* | This indicates that a PTX JIT compilation failed. | +| 219 |*`CUDA_ERROR_INVALID_GRAPHICS_CONTEXT`* |*`hipErrorInvalidGraphicsContext`* | This indicates an error with OpenGL or DirectX context. | +| 300 |*`CUDA_ERROR_INVALID_SOURCE`* |*`hipErrorInvalidSource`* | This indicates that the device kernel source is invalid. | +| 301 |*`CUDA_ERROR_FILE_NOT_FOUND`* |*`hipErrorFileNotFound`* | This indicates that the file specified was not found. | +| 302 |*`CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND`* |*`hipErrorSharedObjectSymbolNotFound`* | This indicates that a link to a shared object failed to resolve. | +| 303 |*`CUDA_ERROR_SHARED_OBJECT_INIT_FAILED`* |*`hipErrorSharedObjectInitFailed`* | This indicates that initialization of a shared object failed. | +| 304 |*`CUDA_ERROR_OPERATING_SYSTEM`* |*`hipErrorOperatingSystem`* | This indicates that an OS call failed. | +| 400 |*`CUDA_ERROR_INVALID_HANDLE`* |*`hipErrorInvalidResourceHandle`* | This indicates that a resource handle passed to the API call was not valid. Resource handles are opaque types like CUstream and CUevent. | +| 500 |*`CUDA_ERROR_NOT_FOUND`* |*`hipErrorNotFound`* | This indicates that a named symbol was not found. Examples of symbols are global/constant variable names, texture names, and surface names. | +| 600 |*`CUDA_ERROR_NOT_READY`* |*`hipErrorNotReady`* | This indicates that asynchronous operations issued previously have not completed yet. This result is not actually an error, but must be indicated differently than CUDA_SUCCESS (which indicates completion). Calls that may return this value include cuEventQuery() and cuStreamQuery(). | +| 700 |*`CUDA_ERROR_ILLEGAL_ADDRESS`* |*`hipErrorIllegalAddress`* | While executing a kernel, the device encountered a load or store instruction on an invalid memory address. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | + + +## **2. Error Handling** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **3. Initialization** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **4. Version Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **5. Device Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + +## **6. Device Management [DEPRECATED]** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **7. Primary Context Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **8. Context Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **9. Context Management [DEPRECATED]** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **10. Module Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **11. Memory Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **12. Unified Addressing** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **13. Stream Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **14. Event Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **15. Execution Control** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **16. Execution Control [DEPRECATED]** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **17. Occupancy** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **18. Texture Reference Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **19. Texture Reference Management [DEPRECATED]** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **20. Surface Reference Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **21. Texture Object Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **22. Surface Object Management** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **23. Peer Context Memory Access** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **24. Graphics Interoperability** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **25. Profiler Control** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **26. OpenGL Interoperability** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **27. Direct3D 9 Interoperability** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **28. Direct3D 10 Interoperability** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **29. Direct3D 11 Interoperability** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + + +## **30. VDPAU Interoperability** + +| **CUDA** | **HIP** | **CUDA description** | +|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| + diff --git a/docs/markdown/hip_bugs.md b/docs/markdown/hip_bugs.md index 9452fae2fd..abb31d80e8 100644 --- a/docs/markdown/hip_bugs.md +++ b/docs/markdown/hip_bugs.md @@ -11,7 +11,13 @@ ### Errors related to undefined reference to `__hcLaunchKernel__***__grid_launch_parm**` Some common code practices may lead to hipcc generating a error with the form : +``` undefined reference to `__hcLaunchKernel__ZN15vecAddNamespace6vecAddIidEEv16grid_launch_parmPT0_S3_S3_T_ +``` +Or: +``` +error: weak declaration cannot have internal linkage +``` Suggested workarounds: - Avoid use of static with kernel definition: @@ -26,6 +32,19 @@ namespace { } ``` +### Can't find kernels inside dynamic linked library + +HCC requires use of the "-Bdynamic" flag when creating a dynamic library which contains kernels. The dynamic flag causes the symbols to be created with a signature which allows HCC to discover and load the kernels in the dynamic library. This flag is often not set by default and must be added to the link step of the library. If not done, HCC will be unable to find the kernels defined in the library, and will emit a message such as: + +``` +HSADevice::CreateKernel(): Unable to create kernel" +``` + +To correct, add the following flag to hcc or hipcc: +``` +$ hipcc -Wl,-Bsymbolic ... +``` + ### What is the current limitation of HIP Generic Grid Launch method? 1. __global__ functions cannot be marked as static or put in an unnamed namespace i.e. they cannot be given internal linkage (this would clash with __attribute__((weak))); diff --git a/hipify-clang/README.md b/hipify-clang/README.md index 850dfb3ffa..c0d74dbe48 100644 --- a/hipify-clang/README.md +++ b/hipify-clang/README.md @@ -13,24 +13,31 @@ `hipify-clang` is a clang-based tool which can automate the translation of CUDA source code into portable HIP C++. The tool can automatically add extra HIP arguments (notably the "hipLaunchParm" required at the beginning of every HIP kernel call). -`hipify-clang` has some additional dependencies explained below and can be built as a separate make step. The instructions below are specifically for **Ubuntu 14.04** +`hipify-clang` has some additional dependencies explained below and can be built as a separate make step. The instructions below are specifically for **Ubuntu 14.04** and **Ubuntu 16.04**. ### Build and install - Download and unpack clang+llvm 3.8 binary package preqrequisite. + +**Ubuntu 14.04**: ```shell wget http://llvm.org/releases/3.8.0/clang+llvm-3.8.0-x86_64-linux-gnu-ubuntu-14.04.tar.xz tar xvfJ clang+llvm-3.8.0-x86_64-linux-gnu-ubuntu-14.04.tar.xz ``` +**Ubuntu 16.04**: +```shell +wget http://llvm.org/releases/3.8.0/clang+llvm-3.8.0-x86_64-linux-gnu-ubuntu-16.04.tar.xz +tar xvfJ clang+llvm-3.8.0-x86_64-linux-gnu-ubuntu-16.04.tar.xz +``` - Enable build of hipify-clang and specify path to LLVM. -Note HIPIFY_CLANG_LLVM_DIR must be a full absolute path to the location extracted above. Here's an example assuming we extract the clang 3.8 package into ~/HIP/clang+llvm-3.8.0-x86_64-linux-gnu-ubuntu-14.04/ +Note HIPIFY_CLANG_LLVM_DIR must be a full absolute path to the location extracted above. Here's an example assuming we extract the clang 3.8 package into ~/HIP/clang+llvm-3.8.0/ ```shell cd HIP mkdir build cd build -cmake -DHIPIFY_CLANG_LLVM_DIR=~/HIP/clang+llvm-3.8.0-x86_64-linux-gnu-ubuntu-14.04/ -DCMAKE_BUILD_TYPE=Release .. +cmake -DHIPIFY_CLANG_LLVM_DIR=~/HIP/clang+llvm-3.8.0/ -DCMAKE_BUILD_TYPE=Release .. make make install ``` @@ -41,13 +48,20 @@ make install In the case when `hipify-clang` doesn't find cuda headers, it reports various errors about unknown keywords (e.g. '\__global\__'), API function names (e.g. 'cudaMalloc'), syntax (e.g. 'foo<<<1,n>>>(...)'), etc. -To install CUDA headers, download the "deb(network)" variant of the target installer from https://developer.nvidia.com/cuda-downloads. The commands below show how to download and install a recent version from http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1404/x86_64/cuda-repo-ubuntu1404_7.5-18_amd64.deb. +To install CUDA headers, download the "deb(network)" variant of the target installer. + +**Ubuntu 14.04**: ```shell wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1404/x86_64/cuda-repo-ubuntu1404_7.5-18_amd64.deb sudo dpkg -i cuda-repo-ubuntu1404_7.5-18_amd64.deb sudo apt-get update && sudo apt-get install cuda-minimal-build-7-5 cuda-curand-dev-7-5 ``` - +**Ubuntu 16.04**: +```shell +wget http://archive.ubuntu.com/ubuntu/pool/multiverse/n/nvidia-cuda-toolkit/nvidia-cuda-toolkit_7.5.18-0ubuntu1_amd64.deb +sudo dpkg -i nvidia-cuda-toolkit_7.5.18-0ubuntu1_amd64.deb +sudo apt-get update && sudo apt-get install cuda-minimal-build-7-5 cuda-curand-dev-7-5 +``` To set additional options like Language Selection (only "-x cuda" is supported), Preprocessor Definition (-D), Include Path (-I), etc., options delimiter "--" should be used before them, for instance: ```shell @@ -58,10 +72,11 @@ Delimiter "--" is used to separate hipify-clang options (before the delimiter) f Option "-x clang" is also worth specifying in order to convert source CUDA files with extensions other than standard extensions (*.cu, *.cuh). -#### Disclaimer +## Disclaimer The information contained herein is for informational purposes only, and is subject to change without notice. While every precaution has been taken in the preparation of this document, it may contain technical inaccuracies, omissions and typographical errors, and AMD is under no obligation to update or otherwise correct this information. Advanced Micro Devices, Inc. makes no representations or warranties with respect to the accuracy or completeness of the contents of this document, and assumes no liability of any kind, including the implied warranties of noninfringement, merchantability or fitness for particular purposes, with respect to the operation or use of AMD hardware, software or other products described herein. No license, including implied or arising by estoppel, to any intellectual property rights is granted by this document. Terms and limitations applicable to the purchase or use of AMD's products are as set forth in a signed agreement between the parties or in AMD's Standard Terms and Conditions of Sale. AMD, the AMD Arrow logo, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other product names used in this publication are for identification purposes only and may be trademarks of their respective companies. -Copyright (c) 2014-2016 Advanced Micro Devices, Inc. All rights reserved. +Copyright (c) 2014-2017 Advanced Micro Devices, Inc. All rights reserved. + diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index 383af0440c..5a2940322e 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -81,6 +81,7 @@ enum ConvTypes { CONV_GL, CONV_GRAPHICS, CONV_SURFACE, + CONV_JIT, CONV_OTHER, CONV_INCLUDE, CONV_INCLUDE_CUDA_MAIN_H, @@ -94,7 +95,7 @@ const char *counterNames[CONV_LAST] = { "driver", "dev", "mem", "kern", "coord_func", "math_func", "special_func", "stream", "event", "occupancy", "ctx", "module", "cache", "exec", "err", "def", "tex", "gl", - "graphics", "surface", "other", "include", "include_cuda_main_header", + "graphics", "surface", "jit", "other", "include", "include_cuda_main_header", "type", "literal", "numeric_literal"}; enum ApiTypes { @@ -190,24 +191,23 @@ struct cuda2hipMap { // Error codes and return types cuda2hipRename["CUresult"] = {"hipError_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["cudaError_enum"] = {"hipError_t", CONV_TYPE, API_DRIVER}; cuda2hipRename["cudaError_t"] = {"hipError_t", CONV_TYPE, API_RUNTIME}; cuda2hipRename["cudaError"] = {"hipError_t", CONV_TYPE, API_RUNTIME}; - // CUDA Driver API error code only - cuda2hipRename["CUDA_ERROR_INVALID_CONTEXT"] = {"hipErrorInvalidContext", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_CURRENT"] = {"hipErrorContextAlreadyCurrent", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_MAP_FAILED"] = {"hipErrorMapFailed", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_UNMAP_FAILED"] = {"hipErrorUnmapFailed", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_ARRAY_IS_MAPPED"] = {"hipErrorArrayIsMapped", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_ALREADY_MAPPED"] = {"hipErrorAlreadyMapped", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_ALREADY_ACQUIRED"] = {"hipErrorAlreadyAcquired", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_NOT_MAPPED"] = {"hipErrorNotMapped", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_ARRAY"] = {"hipErrorNotMappedAsArray", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_POINTER"] = {"hipErrorNotMappedAsPointer", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_IN_USE"] = {"hipErrorContextAlreadyInUse", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_INVALID_SOURCE"] = {"hipErrorInvalidSource", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_FILE_NOT_FOUND"] = {"hipErrorFileNotFound", CONV_ERR, API_DRIVER}; - cuda2hipRename["CUDA_ERROR_NOT_FOUND"] = {"hipErrorNotFound", CONV_ERR, API_DRIVER}; + // CUDA Driver API error codes only + cuda2hipRename["CUDA_ERROR_INVALID_CONTEXT"] = {"hipErrorInvalidContext", CONV_ERR, API_DRIVER}; // 201 + cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_CURRENT"] = {"hipErrorContextAlreadyCurrent", CONV_ERR, API_DRIVER}; // 202 + cuda2hipRename["CUDA_ERROR_ARRAY_IS_MAPPED"] = {"hipErrorArrayIsMapped", CONV_ERR, API_DRIVER}; // 207 + cuda2hipRename["CUDA_ERROR_ALREADY_MAPPED"] = {"hipErrorAlreadyMapped", CONV_ERR, API_DRIVER}; // 208 + cuda2hipRename["CUDA_ERROR_ALREADY_ACQUIRED"] = {"hipErrorAlreadyAcquired", CONV_ERR, API_DRIVER}; // 210 + cuda2hipRename["CUDA_ERROR_NOT_MAPPED"] = {"hipErrorNotMapped", CONV_ERR, API_DRIVER}; // 211 + cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_ARRAY"] = {"hipErrorNotMappedAsArray", CONV_ERR, API_DRIVER}; // 212 + cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_POINTER"] = {"hipErrorNotMappedAsPointer", CONV_ERR, API_DRIVER}; // 213 + cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_IN_USE"] = {"hipErrorContextAlreadyInUse", CONV_ERR, API_DRIVER}; // 216 + cuda2hipRename["CUDA_ERROR_INVALID_SOURCE"] = {"hipErrorInvalidSource", CONV_ERR, API_DRIVER}; // 300 + cuda2hipRename["CUDA_ERROR_FILE_NOT_FOUND"] = {"hipErrorFileNotFound", CONV_ERR, API_DRIVER}; // 301 + cuda2hipRename["CUDA_ERROR_NOT_FOUND"] = {"hipErrorNotFound", CONV_ERR, API_DRIVER}; // 500 // CUDA RT API error code only cuda2hipRename["cudaErrorMissingConfiguration"] = {"hipErrorMissingConfiguration", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 1 @@ -216,8 +216,6 @@ struct cuda2hipMap { cuda2hipRename["cudaErrorInvalidConfiguration"] = {"hipErrorInvalidConfiguration", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 9 cuda2hipRename["cudaErrorInvalidPitchValue"] = {"hipErrorInvalidPitchValue", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 12 cuda2hipRename["cudaErrorInvalidSymbol"] = {"hipErrorInvalidSymbol", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 13 - cuda2hipRename["cudaErrorMapBufferObjectFailed"] = {"hipErrorMapBufferObjectFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 14 - cuda2hipRename["cudaErrorUnmapBufferObjectFailed"] = {"hipErrorUnmapBufferObjectFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 15 cuda2hipRename["cudaErrorInvalidHostPointer"] = {"hipErrorInvalidHostPointer", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 16 cuda2hipRename["cudaErrorInvalidDevicePointer"] = {"hipErrorInvalidDevicePointer", CONV_ERR, API_RUNTIME}; // 17 cuda2hipRename["cudaErrorInvalidTexture"] = {"hipErrorInvalidTexture", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 18 @@ -262,17 +260,96 @@ struct cuda2hipMap { // Deprecated as of CUDA 4.1 cuda2hipRename["cudaErrorApiFailureBase"] = {"hipErrorApiFailureBase", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 10000 - cuda2hipRename["CUDA_SUCCESS"] = {"hipSuccess", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaSuccess"] = {"hipSuccess", CONV_ERR, API_RUNTIME}; // 0 - cuda2hipRename["CUDA_ERROR_OUT_OF_MEMORY"] = {"hipErrorMemoryAllocation", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorMemoryAllocation"] = {"hipErrorMemoryAllocation", CONV_ERR, API_RUNTIME}; // 2 - cuda2hipRename["CUDA_ERROR_NOT_INITIALIZED"] = {"hipErrorNotInitialized", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInitializationError"] = {"hipErrorInitializationError", CONV_ERR, API_RUNTIME}; // 3 + cuda2hipRename["CUDA_SUCCESS"] = {"hipSuccess", CONV_ERR, API_DRIVER}; // 0 + cuda2hipRename["cudaSuccess"] = {"hipSuccess", CONV_ERR, API_RUNTIME}; // 0 - cuda2hipRename["CUDA_ERROR_LAUNCH_FAILED"] = {"hipErrorLaunchFailure", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorLaunchFailure"] = {"hipErrorLaunchFailure", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 4 + cuda2hipRename["CUDA_ERROR_INVALID_VALUE"] = {"hipErrorInvalidValue", CONV_ERR, API_DRIVER}; // 1 + cuda2hipRename["cudaErrorInvalidValue"] = {"hipErrorInvalidValue", CONV_ERR, API_RUNTIME}; // 11 + + cuda2hipRename["CUDA_ERROR_OUT_OF_MEMORY"] = {"hipErrorMemoryAllocation", CONV_ERR, API_DRIVER}; // 2 + cuda2hipRename["cudaErrorMemoryAllocation"] = {"hipErrorMemoryAllocation", CONV_ERR, API_RUNTIME}; // 2 + + cuda2hipRename["CUDA_ERROR_NOT_INITIALIZED"] = {"hipErrorNotInitialized", CONV_ERR, API_DRIVER}; // 3 + cuda2hipRename["cudaErrorInitializationError"] = {"hipErrorInitializationError", CONV_ERR, API_RUNTIME}; // 3 + + cuda2hipRename["CUDA_ERROR_DEINITIALIZED"] = {"hipErrorDeinitialized", CONV_ERR, API_DRIVER}; // 4 + // TODO: double check, that these errors match + cuda2hipRename["cudaErrorCudartUnloading"] = {"hipErrorDeinitialized", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 29 + + cuda2hipRename["CUDA_ERROR_PROFILER_DISABLED"] = {"hipErrorProfilerDisabled", CONV_ERR, API_DRIVER}; // 5 + cuda2hipRename["cudaErrorProfilerDisabled"] = {"hipErrorProfilerDisabled", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 55 + + cuda2hipRename["CUDA_ERROR_PROFILER_NOT_INITIALIZED"] = {"hipErrorProfilerNotInitialized", CONV_ERR, API_DRIVER}; // 6 + // Deprecated as of CUDA 5.0 + cuda2hipRename["cudaErrorProfilerNotInitialized"] = {"hipErrorProfilerNotInitialized", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 56 + + cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STARTED"] = {"hipErrorProfilerAlreadyStarted", CONV_ERR, API_DRIVER}; // 7 + // Deprecated as of CUDA 5.0 + cuda2hipRename["cudaErrorProfilerAlreadyStarted"] = {"hipErrorProfilerAlreadyStarted", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 57 + + cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STOPPED"] = {"hipErrorProfilerAlreadyStopped", CONV_ERR, API_DRIVER}; // 8 + // Deprecated as of CUDA 5.0 + cuda2hipRename["cudaErrorProfilerAlreadyStopped"] = {"hipErrorProfilerAlreadyStopped", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 58 + + cuda2hipRename["CUDA_ERROR_NO_DEVICE"] = {"hipErrorNoDevice", CONV_ERR, API_DRIVER}; // 100 + cuda2hipRename["cudaErrorNoDevice"] = {"hipErrorNoDevice", CONV_ERR, API_RUNTIME}; // 38 + + cuda2hipRename["CUDA_ERROR_INVALID_DEVICE"] = {"hipErrorInvalidDevice", CONV_ERR, API_DRIVER}; // 101 + cuda2hipRename["cudaErrorInvalidDevice"] = {"hipErrorInvalidDevice", CONV_ERR, API_RUNTIME}; // 10 + + cuda2hipRename["CUDA_ERROR_INVALID_IMAGE"] = {"hipErrorInvalidImage", CONV_ERR, API_DRIVER}; // 200 + cuda2hipRename["cudaErrorInvalidKernelImage"] = {"hipErrorInvalidImage", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 47 + + cuda2hipRename["CUDA_ERROR_MAP_FAILED"] = {"hipErrorMapFailed", CONV_ERR, API_DRIVER}; // 205 + // TODO: double check, that these errors match + cuda2hipRename["cudaErrorMapBufferObjectFailed"] = {"hipErrorMapFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 14 + + cuda2hipRename["CUDA_ERROR_UNMAP_FAILED"] = {"hipErrorUnmapFailed", CONV_ERR, API_DRIVER}; // 206 + // TODO: double check, that these errors match + cuda2hipRename["cudaErrorUnmapBufferObjectFailed"] = {"hipErrorUnmapFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 15 + + cuda2hipRename["CUDA_ERROR_NO_BINARY_FOR_GPU"] = {"hipErrorNoBinaryForGpu", CONV_ERR, API_DRIVER}; // 209 + cuda2hipRename["cudaErrorNoKernelImageForDevice"] = {"hipErrorNoBinaryForGpu", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 48 + + cuda2hipRename["CUDA_ERROR_ECC_UNCORRECTABLE"] = {"hipErrorECCNotCorrectable", CONV_ERR, API_DRIVER}; // 214 + cuda2hipRename["cudaErrorECCUncorrectable"] = {"hipErrorECCNotCorrectable", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 39 + + cuda2hipRename["CUDA_ERROR_UNSUPPORTED_LIMIT"] = {"hipErrorUnsupportedLimit", CONV_ERR, API_DRIVER}; // 215 + cuda2hipRename["cudaErrorUnsupportedLimit"] = {"hipErrorUnsupportedLimit", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 42 + + cuda2hipRename["CUDA_ERROR_PEER_ACCESS_UNSUPPORTED"] = {"hipErrorPeerAccessUnsupported", CONV_ERR, API_DRIVER}; // 217 + cuda2hipRename["cudaErrorPeerAccessUnsupported"] = {"hipErrorPeerAccessUnsupported", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 64 + + cuda2hipRename["CUDA_ERROR_INVALID_PTX"] = {"hipErrorInvalidKernelFile", CONV_ERR, API_DRIVER}; // 218 + cuda2hipRename["cudaErrorInvalidPtx"] = {"hipErrorInvalidKernelFile", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 78 + + cuda2hipRename["CUDA_ERROR_INVALID_GRAPHICS_CONTEXT"] = {"hipErrorInvalidGraphicsContext", CONV_ERR, API_DRIVER}; // 219 + cuda2hipRename["cudaErrorInvalidGraphicsContext"] = {"hipErrorInvalidGraphicsContext", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 79 + + cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND"] = {"hipErrorSharedObjectSymbolNotFound", CONV_ERR, API_DRIVER}; // 302 + cuda2hipRename["cudaErrorSharedObjectSymbolNotFound"] = {"hipErrorSharedObjectSymbolNotFound", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 40 + + cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_INIT_FAILED"] = {"hipErrorSharedObjectInitFailed", CONV_ERR, API_DRIVER}; // 303 + cuda2hipRename["cudaErrorSharedObjectInitFailed"] = {"hipErrorSharedObjectInitFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 41 + + cuda2hipRename["CUDA_ERROR_OPERATING_SYSTEM"] = {"hipErrorOperatingSystem", CONV_ERR, API_DRIVER}; // 304 + cuda2hipRename["cudaErrorOperatingSystem"] = {"hipErrorOperatingSystem", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 63 + + cuda2hipRename["CUDA_ERROR_INVALID_HANDLE"] = {"hipErrorInvalidResourceHandle", CONV_ERR, API_DRIVER}; // 400 + cuda2hipRename["cudaErrorInvalidResourceHandle"] = {"hipErrorInvalidResourceHandle", CONV_ERR, API_RUNTIME}; // 33 + + cuda2hipRename["CUDA_ERROR_NOT_READY"] = {"hipErrorNotReady", CONV_ERR, API_DRIVER}; // 600 + cuda2hipRename["cudaErrorNotReady"] = {"hipErrorNotReady", CONV_ERR, API_RUNTIME}; // 34 + + cuda2hipRename["CUDA_ERROR_ILLEGAL_ADDRESS"] = {"hipErrorIllegalAddress", CONV_ERR, API_DRIVER}; // 700 + cuda2hipRename["cudaErrorIllegalAddress"] = {"hipErrorIllegalAddress", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 77 + + ///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + + cuda2hipRename["CUDA_ERROR_LAUNCH_FAILED"] = {"hipErrorLaunchFailure", CONV_ERR, API_DRIVER}; // 719 + cuda2hipRename["cudaErrorLaunchFailure"] = {"hipErrorLaunchFailure", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 4 cuda2hipRename["CUDA_ERROR_LAUNCH_TIMEOUT"] = {"hipErrorLaunchTimeOut", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorLaunchTimeout"] = {"hipErrorLaunchTimeOut", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 6 @@ -280,94 +357,81 @@ struct cuda2hipMap { cuda2hipRename["CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES"] = {"hipErrorLaunchOutOfResources", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorLaunchOutOfResources"] = {"hipErrorLaunchOutOfResources", CONV_ERR, API_RUNTIME}; // 7 - cuda2hipRename["CUDA_ERROR_INVALID_DEVICE"] = {"hipErrorInvalidDevice", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInvalidDevice"] = {"hipErrorInvalidDevice", CONV_ERR, API_RUNTIME}; // 10 - - cuda2hipRename["CUDA_ERROR_INVALID_VALUE"] = {"hipErrorInvalidValue", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInvalidValue"] = {"hipErrorInvalidValue", CONV_ERR, API_RUNTIME}; // 11 - - cuda2hipRename["CUDA_ERROR_DEINITIALIZED"] = {"hipErrorDeinitialized", CONV_ERR, API_DRIVER}; - // TODO: double check, that this error matches to hipErrorDeinitialized - cuda2hipRename["cudaErrorCudartUnloading"] = {"hipErrorDeinitialized", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 29 - cuda2hipRename["CUDA_ERROR_UNKNOWN"] = {"hipErrorUnknown", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorUnknown"] = {"hipErrorUnknown", CONV_ERR, API_RUNTIME}; // 30 - cuda2hipRename["CUDA_ERROR_INVALID_HANDLE"] = {"hipErrorInvalidResourceHandle", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInvalidResourceHandle"] = {"hipErrorInvalidResourceHandle", CONV_ERR, API_RUNTIME}; // 33 - // cuda2hipRename["CUDA_ERROR_NOT_INITIALIZED"] = {"hipErrorInitializationError", CONV_ERR, API_DRIVER}; // cuda2hipRename["cudaErrorInitializationError"] = {"hipErrorInitializationError", CONV_ERR, API_RUNTIME}; - cuda2hipRename["CUDA_ERROR_NOT_READY"] = {"hipErrorNotReady", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorNotReady"] = {"hipErrorNotReady", CONV_ERR, API_RUNTIME}; // 34 - - cuda2hipRename["CUDA_ERROR_NO_DEVICE"] = {"hipErrorNoDevice", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorNoDevice"] = {"hipErrorNoDevice", CONV_ERR, API_RUNTIME}; // 38 - - cuda2hipRename["CUDA_ERROR_ECC_UNCORRECTABLE"] = {"hipErrorECCNotCorrectable", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorECCUncorrectable"] = {"hipErrorECCNotCorrectable", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 39 - - cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND"] = {"hipErrorSharedObjectSymbolNotFound", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorSharedObjectSymbolNotFound"] = {"hipErrorSharedObjectSymbolNotFound", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 40 - - cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_INIT_FAILED"] = {"hipErrorSharedObjectInitFailed", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorSharedObjectInitFailed"] = {"hipErrorSharedObjectInitFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 41 - - cuda2hipRename["CUDA_ERROR_UNSUPPORTED_LIMIT"] = {"hipErrorUnsupportedLimit", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorUnsupportedLimit"] = {"hipErrorUnsupportedLimit", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 42 - - cuda2hipRename["CUDA_ERROR_INVALID_IMAGE"] = {"hipErrorInvalidImage", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInvalidKernelImage"] = {"hipErrorInvalidImage", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 47 - - cuda2hipRename["CUDA_ERROR_NO_BINARY_FOR_GPU"] = {"hipErrorNoBinaryForGpu", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorNoKernelImageForDevice"] = {"hipErrorNoBinaryForGpu", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 48 - cuda2hipRename["CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED"] = {"hipErrorPeerAccessAlreadyEnabled", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorPeerAccessAlreadyEnabled"] = {"hipErrorPeerAccessAlreadyEnabled", CONV_ERR, API_RUNTIME}; // 50 cuda2hipRename["CUDA_ERROR_PEER_ACCESS_NOT_ENABLED"] = {"hipErrorPeerAccessNotEnabled", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorPeerAccessNotEnabled"] = {"hipErrorPeerAccessNotEnabled", CONV_ERR, API_RUNTIME}; // 51 - cuda2hipRename["CUDA_ERROR_PROFILER_DISABLED"] = {"hipErrorProfilerDisabled", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorProfilerDisabled"] = {"hipErrorProfilerDisabled", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 55 - - cuda2hipRename["CUDA_ERROR_PROFILER_NOT_INITIALIZED"] = {"hipErrorProfilerNotInitialized", CONV_ERR, API_DRIVER}; - // Deprecated as of CUDA 5.0 - cuda2hipRename["cudaErrorProfilerNotInitialized"] = {"hipErrorProfilerNotInitialized", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 56 - - cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STARTED"] = {"hipErrorProfilerAlreadyStarted", CONV_ERR, API_DRIVER}; - // Deprecated as of CUDA 5.0 - cuda2hipRename["cudaErrorProfilerAlreadyStarted"] = {"hipErrorProfilerAlreadyStarted", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 57 - - cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STOPPED"] = {"hipErrorProfilerAlreadyStopped", CONV_ERR, API_DRIVER}; - // Deprecated as of CUDA 5.0 - cuda2hipRename["cudaErrorProfilerAlreadyStopped"] = {"hipErrorProfilerAlreadyStopped", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 58 - cuda2hipRename["CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED"] = {"hipErrorHostMemoryAlreadyRegistered", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorHostMemoryAlreadyRegistered"] = {"hipErrorHostMemoryAlreadyRegistered", CONV_ERR, API_RUNTIME}; // 61 cuda2hipRename["CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED"] = {"hipErrorHostMemoryNotRegistered", CONV_ERR, API_DRIVER}; cuda2hipRename["cudaErrorHostMemoryNotRegistered"] = {"hipErrorHostMemoryNotRegistered", CONV_ERR, API_RUNTIME}; // 62 - cuda2hipRename["CUDA_ERROR_OPERATING_SYSTEM"] = {"hipErrorOperatingSystem", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorOperatingSystem"] = {"hipErrorOperatingSystem", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 63 - - cuda2hipRename["CUDA_ERROR_PEER_ACCESS_UNSUPPORTED"] = {"hipErrorPeerAccessUnsupported", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorPeerAccessUnsupported"] = {"hipErrorPeerAccessUnsupported", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 64 - - cuda2hipRename["CUDA_ERROR_ILLEGAL_ADDRESS"] = {"hipErrorIllegalAddress", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorIllegalAddress"] = {"hipErrorIllegalAddress", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 77 - - cuda2hipRename["CUDA_ERROR_INVALID_PTX"] = {"hipErrorInvalidKernelFile", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInvalidPtx"] = {"hipErrorInvalidKernelFile", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 78 - - cuda2hipRename["CUDA_ERROR_INVALID_GRAPHICS_CONTEXT"] = {"hipErrorInvalidGraphicsContext", CONV_ERR, API_DRIVER}; - cuda2hipRename["cudaErrorInvalidGraphicsContext"] = {"hipErrorInvalidGraphicsContext", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 79 - - - ///////////////////////////// CUDA DRIVER API ///////////////////////////// + // enums + cuda2hipRename["CUDA_ARRAY3D_DESCRIPTOR"] = {"HIP_ARRAY3D_DESCRIPTOR", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_ARRAY_DESCRIPTOR"] = {"HIP_ARRAY_DESCRIPTOR", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_MEMCPY2D"] = {"HIP_MEMCPY2D", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_MEMCPY3D"] = {"HIP_MEMCPY3D", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_MEMCPY3D_PEER"] = {"HIP_MEMCPY3D_PEER", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_POINTER_ATTRIBUTE_P2P_TOKENS"] = {"HIP_POINTER_ATTRIBUTE_P2P_TOKENS", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_RESOURCE_DESC"] = {"HIP_RESOURCE_DESC", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUDA_RESOURCE_VIEW_DESC"] = {"HIP_RESOURCE_VIEW_DESC", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + + cuda2hipRename["CUipcEventHandle"] = {"hipIpcEventHandle", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUipcMemHandle"] = {"hipIpcMemHandle", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + + + + cuda2hipRename["CUaddress_mode"] = {"hipAddress_mode", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TR_ADDRESS_MODE_WRAP"] = {"HIP_TR_ADDRESS_MODE_WRAP", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0 + cuda2hipRename["CU_TR_ADDRESS_MODE_CLAMP"] = {"HIP_TR_ADDRESS_MODE_CLAMP", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 1 + cuda2hipRename["CU_TR_ADDRESS_MODE_MIRROR"] = {"HIP_TR_ADDRESS_MODE_MIRROR", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 2 + cuda2hipRename["CU_TR_ADDRESS_MODE_BORDER"] = {"HIP_TR_ADDRESS_MODE_BORDER", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 3 + + cuda2hipRename["CUarray_cubemap_face"] = {"hipArray_cubemap_face", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_CUBEMAP_FACE_POSITIVE_X"] = {"HIP_CUBEMAP_FACE_POSITIVE_X", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 + cuda2hipRename["CU_CUBEMAP_FACE_NEGATIVE_X"] = {"HIP_CUBEMAP_FACE_NEGATIVE_X", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_CUBEMAP_FACE_POSITIVE_Y"] = {"HIP_CUBEMAP_FACE_POSITIVE_Y", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 + cuda2hipRename["CU_CUBEMAP_FACE_NEGATIVE_Y"] = {"HIP_CUBEMAP_FACE_NEGATIVE_Y", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 + cuda2hipRename["CU_CUBEMAP_FACE_POSITIVE_Z"] = {"HIP_CUBEMAP_FACE_POSITIVE_Z", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 + cuda2hipRename["CU_CUBEMAP_FACE_NEGATIVE_Z"] = {"HIP_CUBEMAP_FACE_NEGATIVE_Z", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x05 + + cuda2hipRename["CUarray_format"] = {"hipArray_format", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_AD_FORMAT_UNSIGNED_INT8"] = {"HIP_AD_FORMAT_UNSIGNED_INT8", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_AD_FORMAT_UNSIGNED_INT16"] = {"HIP_AD_FORMAT_UNSIGNED_INT16", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 + cuda2hipRename["CU_AD_FORMAT_UNSIGNED_INT32"] = {"HIP_AD_FORMAT_UNSIGNED_INT32", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 + cuda2hipRename["CU_AD_FORMAT_SIGNED_INT8"] = {"HIP_AD_FORMAT_SIGNED_INT8", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x08 + cuda2hipRename["CU_AD_FORMAT_SIGNED_INT16"] = {"HIP_AD_FORMAT_SIGNED_INT16", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x09 + cuda2hipRename["CU_AD_FORMAT_SIGNED_INT32"] = {"HIP_AD_FORMAT_SIGNED_INT32", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x0a + cuda2hipRename["CU_AD_FORMAT_HALF"] = {"HIP_AD_FORMAT_HALF", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 + cuda2hipRename["CU_AD_FORMAT_FLOAT"] = {"HIP_AD_FORMAT_FLOAT", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x20 + // Compute mode + cuda2hipRename["CUcomputemode"] = {"hipComputemode", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_RUNTIME ANALOGUE (cudaComputeMode) + cuda2hipRename["CU_COMPUTEMODE_DEFAULT"] = {"hipComputeModeDefault", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0 + cuda2hipRename["CU_COMPUTEMODE_EXCLUSIVE"] = {"hipComputeModeExclusive", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 1 + cuda2hipRename["CU_COMPUTEMODE_PROHIBITED"] = {"hipComputeModeProhibited", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 2 + cuda2hipRename["CU_COMPUTEMODE_EXCLUSIVE_PROCESS"] = {"hipComputeModeExclusiveProcess", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 3 + // Context flags + cuda2hipRename["CUctx_flags"] = {"hipCctx_flags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_CTX_SCHED_AUTO"] = {"HIP_CTX_SCHED_AUTO", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 + cuda2hipRename["CU_CTX_SCHED_SPIN"] = {"HIP_CTX_SCHED_SPIN", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_CTX_SCHED_YIELD"] = {"HIP_CTX_SCHED_YIELD", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 + cuda2hipRename["CU_CTX_SCHED_BLOCKING_SYNC"] = {"HIP_CTX_SCHED_BLOCKING_SYNC", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 + cuda2hipRename["CU_CTX_BLOCKING_SYNC"] = {"HIP_CTX_BLOCKING_SYNC", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 + cuda2hipRename["CU_CTX_SCHED_MASK"] = {"HIP_CTX_SCHED_MASK", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x07 + cuda2hipRename["CU_CTX_MAP_HOST"] = {"HIP_CTX_MAP_HOST", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x08 + cuda2hipRename["CU_CTX_LMEM_RESIZE_TO_MAX"] = {"HIP_CTX_LMEM_RESIZE_TO_MAX", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 + cuda2hipRename["CU_CTX_FLAGS_MASK"] = {"HIP_CTX_FLAGS_MASK", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x1f + // Defines cuda2hipRename["CU_LAUNCH_PARAM_BUFFER_POINTER"] = {"HIP_LAUNCH_PARAM_BUFFER_POINTER", CONV_DEV, API_DRIVER}; cuda2hipRename["CU_LAUNCH_PARAM_BUFFER_SIZE"] = {"HIP_LAUNCH_PARAM_BUFFER_SIZE", CONV_DEV, API_DRIVER}; @@ -375,104 +439,108 @@ struct cuda2hipMap { // Types // NOTE: CUdevice might be changed to typedef int in the future. - cuda2hipRename["CUdevice"] = {"hipDevice_t", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUdevice_attribute_enum"] = {"hipDeviceAttribute_t", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUdevice_attribute"] = {"hipDeviceAttribute_t", CONV_TYPE, API_DRIVER}; - - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X"] = {"hipDeviceAttributeMaxBlockDimX", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y"] = {"hipDeviceAttributeMaxBlockDimY", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X"] = {"hipDeviceAttributeMaxGridDimX", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y"] = {"hipDeviceAttributeMaxGridDimY", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z"] = {"hipDeviceAttributeMaxGridDimZ", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY"] = {"hipDeviceAttributeTotalConstantMemory", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_WARP_SIZE"] = {"hipDeviceAttributeWarpSize", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CLOCK_RATE"] = {"hipDeviceAttributeClockRate", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE"] = {"hipDeviceAttributeMemoryClockRate", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH"] = {"hipDeviceAttributeMemoryBusWidth", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_MODE"] = {"hipDeviceAttributeComputeMode", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE"] = {"hipDeviceAttributeL2CacheSize", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS"] = {"hipDeviceAttributeConcurrentKernels", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_BUS_ID"] = {"hipDeviceAttributePciBusId", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID"] = {"hipDeviceAttributePciDeviceId", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_DEV, API_DRIVER}; + cuda2hipRename["CUdevice"] = {"hipDevice_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUdevice_attribute_enum"] = {"hipDeviceAttribute_t", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaDeviceAttr) + cuda2hipRename["CUdevice_attribute"] = {"hipDeviceAttribute_t", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaDeviceAttr) // unsupported yet by HIP - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_PITCH"] = {"hipDeviceAttributeMaxPitch", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT"] = {"hipDeviceAttributeTextureAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (cudaDevAttrMaxThreadsPerBlock = 1) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X"] = {"hipDeviceAttributeMaxBlockDimX", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimX = 2) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y"] = {"hipDeviceAttributeMaxBlockDimY", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimY = 3) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 4 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimZ = 4) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X"] = {"hipDeviceAttributeMaxGridDimX", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 5 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimX =5) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y"] = {"hipDeviceAttributeMaxGridDimY", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 6 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimY = 6) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z"] = {"hipDeviceAttributeMaxGridDimZ", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 7 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimZ - 7) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 8 // API_Runtime ANALOGUE (cudaDevAttrMaxSharedMemoryPerBlock = 8) + // Deprecated, use CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK + cuda2hipRename["CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 8 + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY"] = {"hipDeviceAttributeTotalConstantMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 9 // API_Runtime ANALOGUE (cudaDevAttrTotalConstantMemory = 9) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_WARP_SIZE"] = {"hipDeviceAttributeWarpSize", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 10 // API_Runtime ANALOGUE (cudaDevAttrWarpSize = 10) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_PITCH"] = {"hipDeviceAttributeMaxPitch", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 11 // API_Runtime ANALOGUE (cudaDevAttrMaxPitch = 11) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 12 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerBlock = 12) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 12 + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CLOCK_RATE"] = {"hipDeviceAttributeClockRate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 13 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerBlock = 13) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT"] = {"hipDeviceAttributeTextureAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 14 // API_Runtime ANALOGUE (cudaDevAttrTextureAlignment = 14) // Deprecated. Use instead CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT - cuda2hipRename["CU_DEVICE_ATTRIBUTE_GPU_OVERLAP"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT"] = {"hipDeviceAttributeMultiprocessorCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT"] = {"hipDeviceAttributeKernelExecTimeout", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_INTEGRATED"] = {"hipDeviceAttributeIntegrated", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY"] = {"hipDeviceAttributeCanMapHostMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT"] = {"hipDeviceAttributeSurfaceAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_ECC_ENABLED"] = {"hipDeviceAttributeEccEnabled", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TCC_DRIVER"] = {"hipDeviceAttributeTccDriver", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING"] = {"hipDeviceAttributeUnifiedAddressing", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID"] = {"hipDeviceAttributePciDomainId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY"] = {"hipDeviceAttributeManagedMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID"] = {"hipDeviceAttributeMultiGpuBoardGroupId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX"] = {"hipDeviceAttributeMax", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_DEVICE_ATTRIBUTE_GPU_OVERLAP"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 15 // API_Runtime ANALOGUE (cudaDevAttrGpuOverlap = 15) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT"] = {"hipDeviceAttributeMultiprocessorCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 16 // API_Runtime ANALOGUE (cudaDevAttrMultiProcessorCount = 16) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT"] = {"hipDeviceAttributeKernelExecTimeout", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 17 // API_Runtime ANALOGUE (cudaDevAttrKernelExecTimeout = 17) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_INTEGRATED"] = {"hipDeviceAttributeIntegrated", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 18 // API_Runtime ANALOGUE (cudaDevAttrIntegrated = 18) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY"] = {"hipDeviceAttributeCanMapHostMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 19 // API_Runtime ANALOGUE (cudaDevAttrCanMapHostMemory = 19) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_MODE"] = {"hipDeviceAttributeComputeMode", CONV_DEV, API_DRIVER}; // 20 // API_Runtime ANALOGUE (cudaDevAttrComputeMode = 20) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 21 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DWidth = 21) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 22 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DWidth = 22) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 23 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DHeight = 23) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 24 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DWidth = 24) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 25 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DHeight = 25) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 26 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DDepth = 26) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 27 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredWidth = 27) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 28 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredHeight = 28) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 29 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredLayers = 29) + // Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 27 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredWidth = 27) + // Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 28 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredHeight = 28) + // Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 29 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredLayers = 29) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT"] = {"hipDeviceAttributeSurfaceAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 30 // API_Runtime ANALOGUE (cudaDevAttrSurfaceAlignment = 30) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS"] = {"hipDeviceAttributeConcurrentKernels", CONV_DEV, API_DRIVER}; // 31 // API_Runtime ANALOGUE (cudaDevAttrConcurrentKernels = 31) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_ECC_ENABLED"] = {"hipDeviceAttributeEccEnabled", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 32 // API_Runtime ANALOGUE (cudaDevAttrEccEnabled = 32) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_BUS_ID"] = {"hipDeviceAttributePciBusId", CONV_DEV, API_DRIVER}; // 33 // API_Runtime ANALOGUE (cudaDevAttrPciBusId = 33) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID"] = {"hipDeviceAttributePciDeviceId", CONV_DEV, API_DRIVER}; // 34 // API_Runtime ANALOGUE (cudaDevAttrPciDeviceId = 34) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TCC_DRIVER"] = {"hipDeviceAttributeTccDriver", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 35 // API_Runtime ANALOGUE (cudaDevAttrTccDriver = 35) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE"] = {"hipDeviceAttributeMemoryClockRate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 36 // API_Runtime ANALOGUE (cudaDevAttrMemoryClockRate = 36) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH"] = {"hipDeviceAttributeMemoryBusWidth", CONV_DEV, API_DRIVER}; // 37 // API_Runtime ANALOGUE (cudaDevAttrGlobalMemoryBusWidth = 37) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE"] = {"hipDeviceAttributeL2CacheSize", CONV_DEV, API_DRIVER}; // 38 // API_Runtime ANALOGUE (cudaDevAttrL2CacheSize = 38) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_DEV, API_DRIVER}; // 39 // API_Runtime ANALOGUE (cudaDevAttrMaxThreadsPerMultiProcessor = 39) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 40 // API_Runtime ANALOGUE (cudaDevAttrAsyncEngineCount = 40) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING"] = {"hipDeviceAttributeUnifiedAddressing", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 41 // API_Runtime ANALOGUE (cudaDevAttrUnifiedAddressing = 41) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 42 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLayeredWidth = 42) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 43 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLayeredLayers = 43) // deprecated, do not use - // cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER"] = {"hipDeviceAttributeCanTex2DGather", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER"] = {"hipDeviceAttributeCanTex2DGather", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 44 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 45 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DGatherWidth = 45) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 46 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DGatherHeight = 46) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 47 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DWidthAlt = 47) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 48 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DHeightAlt = 48) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 49 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DDepthAlt = 49) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID"] = {"hipDeviceAttributePciDomainId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 50 // API_Runtime ANALOGUE (cudaDevAttrPciDomainId = 50) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 51 // API_Runtime ANALOGUE (cudaDevAttrTexturePitchAlignment = 51) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 52 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapWidth = 52) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 53 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapLayeredWidth = 53) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 54 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapLayeredLayers = 54) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 55 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DWidth = 55) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 56 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DWidth = 56) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 57 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DHeight = 57) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 58 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DWidth = 58) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 59 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DHeight = 59) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 60 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DDepth = 60) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 61 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DLayeredWidth = 61) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 62 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DLayeredLayers = 62) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 63 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredWidth = 63) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 64 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredHeight = 64) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 65 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredLayers = 65) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 66 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapWidth = 66) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 67 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapLayeredWidth = 67) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 68 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapLayeredLayers = 68) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 69 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLinearWidth = 69) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 70 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearWidth = 70) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 71 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearHeight = 71) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 72 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearPitch = 72) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 73 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DMipmappedWidth = 73) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 74 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DMipmappedHeight = 74) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_DEV, API_DRIVER}; // 75 // API_Runtime ANALOGUE (cudaDevAttrComputeCapabilityMajor = 75) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_DEV, API_DRIVER}; // 76 // API_Runtime ANALOGUE (cudaDevAttrComputeCapabilityMinor = 76) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 77 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DMipmappedWidth = 77) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 78 // API_Runtime ANALOGUE (cudaDevAttrStreamPrioritiesSupported = 78) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 79 // API_Runtime ANALOGUE (cudaDevAttrGlobalL1CacheSupported = 79) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 80 // API_Runtime ANALOGUE (cudaDevAttrLocalL1CacheSupported = 80) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_DEV, API_DRIVER}; // 81 // API_Runtime ANALOGUE (cudaDevAttrMaxSharedMemoryPerMultiprocessor = 81) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 82 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerMultiprocessor = 82) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY"] = {"hipDeviceAttributeManagedMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 83 // API_Runtime ANALOGUE (cudaDevAttrManagedMemory = 83) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_DEV, API_DRIVER}; // 84 // API_Runtime ANALOGUE (cudaDevAttrIsMultiGpuBoard = 84) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID"] = {"hipDeviceAttributeMultiGpuBoardGroupId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 85 // API_Runtime ANALOGUE (cudaDevAttrMultiGpuBoardGroupID = 85) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX"] = {"hipDeviceAttributeMax", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 86 // API_Runtime ANALOGUE (no) + // unsupported yet by HIP [CUDA 8.0.44] cuda2hipRename["CU_DEVICE_ATTRIBUTE_HOST_NATIVE_ATOMIC_SUPPORTED"] = {"hipDeviceAttributeHostNativeAtomicSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["CU_DEVICE_ATTRIBUTE_SINGLE_TO_DOUBLE_PRECISION_PERF_RATIO"] = {"hipDeviceAttributeSingleToDoublePrecisionPerfRatio", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; @@ -481,52 +549,232 @@ struct cuda2hipMap { cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_PREEMPTION_SUPPORTED"] = {"hipDeviceAttributeComputePreemptionSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_USE_HOST_POINTER_FOR_REGISTERED_MEM"] = {"hipDeviceAttributeCanUseHostPointerForRegisteredMem", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CUdevprop_st"] = {"hipDeviceProp_t", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUdevprop"] = {"hipDeviceProp_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUdevprop_st"] = {"hipDeviceProp_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUdevprop"] = {"hipDeviceProp_t", CONV_TYPE, API_DRIVER}; // TODO: Analogues enum is needed in HIP. Couldn't map enum to struct hipPointerAttribute_t. // TODO: Do for Pointer Attributes the same as for Device Attributes. - // cuda2hipRename["CUpointer_attribute_enum"] = {"hipPointerAttribute_t", CONV_TYPE, API_DRIVER}; - // cuda2hipRename["CUpointer_attribute"] = {"hipPointerAttribute_t", CONV_TYPE, API_DRIVER}; + // cuda2hipRename["CUpointer_attribute_enum"] = {"hipPointerAttribute", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + // cuda2hipRename["CUpointer_attribute"] = {"hipPointerAttribute", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_CONTEXT"] = {"hipPointerAttributeContext", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_MEMORY_TYPE"] = {"hipPointerAttributeMemoryType", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_DEVICE_POINTER"] = {"hipPointerAttributeDevicePointer", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_HOST_POINTER"] = {"hipPointerAttributeHostPointer", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 4 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_P2P_TOKENS"] = {"hipPointerAttributeP2pTokens", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 5 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_SYNC_MEMOPS"] = {"hipPointerAttributeSyncMemops", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 6 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_BUFFER_ID"] = {"hipPointerAttributeBufferId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 7 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_IS_MANAGED"] = {"hipPointerAttributeIsManaged", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 8 // API_Runtime ANALOGUE (no) + // pointer to CUfunc_st - cuda2hipRename["CUfunction"] = {"hipFunction_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUfunction"] = {"hipFunction_t", CONV_TYPE, API_DRIVER}; // TODO: in HIP ihipModuleSymbol_t should be declared in hip_runtime_api.h, not in hcc_detail/hip_runtime_api.h, as it's analogue CUfunc_st is declared also in cuda.h // ToDO: examples are needed with CUfunc_st - // cuda2hipRename["CUfunc_st"] = {"ihipModuleSymbol_t", CONV_TYPE, API_DRIVER}; + // cuda2hipRename["CUfunc_st"] = {"ihipModuleSymbol_t", CONV_TYPE, API_DRIVER}; // unsupported yet by HIP - cuda2hipRename["CUfunction_attribute_enum"] = {"hipFuncAttribute_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CUfunction_attribute"] = {"hipFuncAttribute_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUfunction_attribute"] = {"hipFuncAttribute_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CUfunction_attribute_enum"] = {"hipFuncAttribute_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_MAX_THREADS_PER_BLOCK"] = {"hipFuncAttributeMaxThreadsPerBlocks", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_SHARED_SIZE_BYTES"] = {"hipFuncAttributeSharedSizeBytes", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_CONST_SIZE_BYTES"] = {"hipFuncAttributeConstSizeBytes", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_LOCAL_SIZE_BYTES"] = {"hipFuncAttributeLocalSizeBytes", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_NUM_REGS"] = {"hipFuncAttributeNumRegs", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_PTX_VERSION"] = {"hipFuncAttributePtxVersion", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_BINARY_VERSION"] = {"hipFuncAttributeBinaryVersion", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_CACHE_MODE_CA"] = {"hipFuncAttributeCacheModeCA", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_FUNC_ATTRIBUTE_MAX"] = {"hipFuncAttributeMax", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CUfunc_cache_enum"] = {"hipFuncCache", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUfunc_cache"] = {"hipFuncCache", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CU_FUNC_CACHE_PREFER_NONE"] = {"hipFuncCachePreferNone", CONV_CACHE, API_DRIVER}; - cuda2hipRename["CU_FUNC_CACHE_PREFER_SHARED"] = {"hipFuncCachePreferShared", CONV_CACHE, API_DRIVER}; - cuda2hipRename["CU_FUNC_CACHE_PREFER_L1"] = {"hipFuncCachePreferL1", CONV_CACHE, API_DRIVER}; - cuda2hipRename["CU_FUNC_CACHE_PREFER_EQUAL"] = {"hipFuncCachePreferEqual", CONV_CACHE, API_DRIVER}; + // enum CUgraphicsMapResourceFlags/CUgraphicsMapResourceFlags_enum + cuda2hipRename["CUgraphicsMapResourceFlags"] = {"hipGraphicsMapFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsMapFlags) + cuda2hipRename["CUgraphicsMapResourceFlags_enum"] = {"hipGraphicsMapFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsMapFlags) + cuda2hipRename["CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE"] = {"hipGraphicsMapFlagsNone", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaGraphicsMapFlagsNone = 0) + cuda2hipRename["CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY"] = {"hipGraphicsMapFlagsReadOnly", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaGraphicsMapFlagsReadOnly = 1) + cuda2hipRename["CU_GRAPHICS_MAP_RESOURCE_FLAGS_WRITE_DISCARD"] = {"hipGraphicsMapFlagsWriteDiscard", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaGraphicsMapFlagsWriteDiscard = 2) - cuda2hipRename["CUsharedconfig_enum"] = {"hipSharedMemConfig", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUsharedconfig"] = {"hipSharedMemConfig", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CU_SHARED_MEM_CONFIG_DEFAULT_BANK_SIZE"] = {"hipSharedMemBankSizeDefault", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_SHARED_MEM_CONFIG_FOUR_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeFourByte", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_SHARED_MEM_CONFIG_EIGHT_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeEightByte", CONV_DEV, API_DRIVER}; + // enum CUgraphicsRegisterFlags/CUgraphicsRegisterFlags_enum + cuda2hipRename["CUgraphicsRegisterFlags"] = {"hipGraphicsRegisterFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsRegisterFlags) + cuda2hipRename["CUgraphicsRegisterFlags_enum"] = {"hipGraphicsRegisterFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsRegisterFlags) + cuda2hipRename["CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE"] = {"hipGraphicsRegisterFlagsNone", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaGraphicsRegisterFlagsNone = 0) + cuda2hipRename["CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY"] = {"hipGraphicsRegisterFlagsReadOnly", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaGraphicsRegisterFlagsReadOnly = 1) + cuda2hipRename["CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD"] = {"hipGraphicsRegisterFlagsWriteDiscard", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaGraphicsRegisterFlagsWriteDiscard = 2) + cuda2hipRename["CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST"] = {"hipGraphicsRegisterFlagsSurfaceLoadStore", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 // API_Runtime ANALOGUE (cudaGraphicsRegisterFlagsSurfaceLoadStore = 4) + cuda2hipRename["CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER"] = {"hipGraphicsRegisterFlagsTextureGather", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x08 // API_Runtime ANALOGUE (cudaGraphicsRegisterFlagsTextureGather = 8) - cuda2hipRename["CUcontext"] = {"hipCtx_t", CONV_TYPE, API_DRIVER}; + // enum CUoccupancy_flags/CUoccupancy_flags_enum + cuda2hipRename["CUoccupancy_flags"] = {"hipOccupancyFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUoccupancy_flags_enum"] = {"hipOccupancyFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_OCCUPANCY_DEFAULT"] = {"hipOccupancyDefault", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaOccupancyDefault = 0x0) + cuda2hipRename["CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE"] = {"hipOccupancyDisableCachingOverride", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaOccupancyDisableCachingOverride = 0x1) + + + + cuda2hipRename["CUfunc_cache_enum"] = {"hipFuncCache", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaFuncCache) + cuda2hipRename["CUfunc_cache"] = {"hipFuncCache", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaFuncCache) + cuda2hipRename["CU_FUNC_CACHE_PREFER_NONE"] = {"hipFuncCachePreferNone", CONV_CACHE, API_DRIVER}; // 0x00 // API_Runtime ANALOGUE (cudaFilterModePoint = 0) + cuda2hipRename["CU_FUNC_CACHE_PREFER_SHARED"] = {"hipFuncCachePreferShared", CONV_CACHE, API_DRIVER}; // 0x01 // API_Runtime ANALOGUE (cudaFuncCachePreferShared = 1) + cuda2hipRename["CU_FUNC_CACHE_PREFER_L1"] = {"hipFuncCachePreferL1", CONV_CACHE, API_DRIVER}; // 0x02 // API_Runtime ANALOGUE (cudaFuncCachePreferL1 = 2) + cuda2hipRename["CU_FUNC_CACHE_PREFER_EQUAL"] = {"hipFuncCachePreferEqual", CONV_CACHE, API_DRIVER}; // 0x03 // API_Runtime ANALOGUE (cudaFuncCachePreferEqual = 3) + + // enum CUipcMem_flags/CUipcMem_flags_enum + cuda2hipRename["CUipcMem_flags"] = {"hipIpcMemFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUipcMem_flags_enum"] = {"hipIpcMemFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_IPC_MEM_LAZY_ENABLE_PEER_ACCESS"] = {"hipIpcMemLazyEnablePeerAccess", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x1 // API_Runtime ANALOGUE (cudaIpcMemLazyEnablePeerAccess = 0x01) + + // enum CUipcMem_flags/CUipcMem_flags_enum + cuda2hipRename["CUipcMem_flags"] = {"hipIpcMemFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + + // JIT + // enum CUjit_cacheMode/CUjit_cacheMode_enum + cuda2hipRename["CUjit_cacheMode"] = {"hipJitCacheMode", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUjit_cacheMode_enum"] = {"hipJitCacheMode", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_CACHE_OPTION_NONE"] = {"hipJitCacheModeOptionNone", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_CACHE_OPTION_CG"] = {"hipJitCacheModeOptionCG", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_CACHE_OPTION_CA"] = {"hipJitCacheModeOptionCA", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + // enum CUjit_fallback/CUjit_fallback_enum + cuda2hipRename["CUjit_fallback"] = {"hipJitFallback", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUjit_fallback_enum"] = {"hipJitFallback", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_PREFER_PTX"] = {"hipJitFallbackPreferPtx", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_PREFER_BINARY"] = {"hipJitFallbackPreferBinary", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + // enum CUjit_option/CUjit_option_enum + cuda2hipRename["CUjit_option"] = {"hipJitOption", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUjit_option_enum"] = {"hipJitOption", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_MAX_REGISTERS"] = {"hipJitOptionMaxRegisters", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_THREADS_PER_BLOCK"] = {"hipJitOptionThreadsPerBlock", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_WALL_TIME"] = {"hipJitOptionWallTime", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INFO_LOG_BUFFER"] = {"hipJitOptionInfoLogBuffer", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES"] = {"hipJitOptionInfoLogBufferSizeBytes", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_ERROR_LOG_BUFFER"] = {"hipJitOptionErrorLogBuffer", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_ERROR_LOG_BUFFER_SIZE_BYTES"] = {"hipJitOptionErrorLogBufferSizeBytes", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_OPTIMIZATION_LEVEL"] = {"hipJitOptionOptimizationLevel", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_TARGET_FROM_CUCONTEXT"] = {"hipJitOptionTargetFromContext", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_TARGET"] = {"hipJitOptionTarget", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_FALLBACK_STRATEGY"] = {"hipJitOptionFallbackStrategy", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_GENERATE_DEBUG_INFO"] = {"hipJitOptionGenerateDebugInfo", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_LOG_VERBOSE"] = {"hipJitOptionLogVerbose", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_GENERATE_LINE_INFO"] = {"hipJitOptionLogVerbose", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_CACHE_MODE"] = {"hipJitOptionCacheMode", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_NUM_OPTIONS"] = {"hipJitOptionNumOptions", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + // enum CUjit_target/CUjit_target_enum + cuda2hipRename["CUjit_target"] = {"hipJitTarget", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUjit_target_enum"] = {"hipJitTarget", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_10"] = {"hipJitTargetCompute10", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_11"] = {"hipJitTargetCompute11", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_12"] = {"hipJitTargetCompute12", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_13"] = {"hipJitTargetCompute13", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_20"] = {"hipJitTargetCompute20", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_21"] = {"hipJitTargetCompute21", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_30"] = {"hipJitTargetCompute30", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_32"] = {"hipJitTargetCompute32", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_35"] = {"hipJitTargetCompute35", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_37"] = {"hipJitTargetCompute37", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_50"] = {"hipJitTargetCompute50", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_TARGET_COMPUTE_52"] = {"hipJitTargetCompute52", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + // enum CUjitInputType/CUjitInputType_enum + cuda2hipRename["CUjitInputType"] = {"hipJitInputType", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUjitInputType_enum"] = {"hipJitInputType", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INPUT_CUBIN"] = {"hipJitInputTypeBin", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INPUT_PTX"] = {"hipJitInputTypePtx", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INPUT_FATBINARY"] = {"hipJitInputTypeFatBinary", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INPUT_OBJECT"] = {"hipJitInputTypeObject", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_INPUT_LIBRARY"] = {"hipJitInputTypeLibrary", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_JIT_NUM_INPUT_TYPES"] = {"hipJitInputTypeNumInputTypes", CONV_JIT, API_DRIVER, HIP_UNSUPPORTED}; + + // Limits + cuda2hipRename["CUlimit"] = {"hipLimit_t", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaLimit) + cuda2hipRename["CUlimit_enum"] = {"hipLimit_t", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaLimit) + cuda2hipRename["CU_LIMIT_STACK_SIZE"] = {"hipLimitStackSize", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaLimitStackSize = 0x00) + cuda2hipRename["CU_LIMIT_PRINTF_FIFO_SIZE"] = {"hipLimitPrintfFifoSize", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaLimitPrintfFifoSize = 0x01) + cuda2hipRename["CU_LIMIT_MALLOC_HEAP_SIZE"] = {"hipLimitMallocHeapSize", CONV_TYPE, API_DRIVER}; // 0x02 // API_Runtime ANALOGUE (cudaLimitMallocHeapSize = 0x02) + cuda2hipRename["CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH"] = {"hipLimitDevRuntimeSyncDepth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaLimitDevRuntimeSyncDepth = 0x03) + cuda2hipRename["CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT"] = {"hipLimitDevRuntimePendingLaunchCount", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 // API_Runtime ANALOGUE (cudaLimitDevRuntimePendingLaunchCount = 0x04) + cuda2hipRename["CU_LIMIT_STACK_SIZE"] = {"hipLimitStackSize", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + + // enum CUmemAttach_flags/CUmemAttach_flags_enum + cuda2hipRename["CUmemAttach_flags"] = {"hipMemAttachFlags_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CUmemAttach_flags_enum"] = {"hipMemAttachFlags_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_MEM_ATTACH_GLOBAL"] = {"hipMemAttachGlobal", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x1 // API_Runtime ANALOGUE (#define cudaMemAttachGlobal 0x01) + cuda2hipRename["CU_MEM_ATTACH_HOST"] = {"hipMemAttachHost", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x2 // API_Runtime ANALOGUE (#define cudaMemAttachHost 0x02) + cuda2hipRename["CU_MEM_ATTACH_SINGLE"] = {"hipMemAttachSingle", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x4 // API_Runtime ANALOGUE (#define cudaMemAttachSingle 0x04) + + // enum CUmemorytype/CUmemorytype_enum + cuda2hipRename["CUmemorytype"] = {"hipMemType_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no - cudaMemoryType is not an analogue) + cuda2hipRename["CUmemorytype_enum"] = {"hipMemType_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no - cudaMemoryType is not an analogue) + cuda2hipRename["CU_MEMORYTYPE_HOST"] = {"hipMemTypeHost", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_MEMORYTYPE_DEVICE"] = {"hipMemTypeDevice", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_MEMORYTYPE_ARRAY"] = {"hipMemTypeArray", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_MEMORYTYPE_UNIFIED"] = {"hipMemTypeUnified", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 // API_Runtime ANALOGUE (no) + + // enum CUresourcetype + cuda2hipRename["CUresourcetype"] = {"hipResourceType", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaResourceType) + cuda2hipRename["CUresourcetype_enum"] = {"hipResourceType", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaResourceType) + cuda2hipRename["CU_RESOURCE_TYPE_ARRAY"] = {"hipResourceTypeArray", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaResourceTypeArray = 0x00) + cuda2hipRename["CU_RESOURCE_TYPE_MIPMAPPED_ARRAY"] = {"hipResourceTypeMipmappedArray", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaResourceTypeMipmappedArray = 0x01) + cuda2hipRename["CU_RESOURCE_TYPE_LINEAR"] = {"hipResourceTypeLinear", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaResourceTypeLinear = 0x02) + cuda2hipRename["CU_RESOURCE_TYPE_PITCH2D"] = {"hipResourceTypePitch2D", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaResourceTypePitch2D = 0x03) + + // enum CUresourceViewFormat/CUresourceViewFormat_enum + cuda2hipRename["CUresourceViewFormat"] = {"hipResourceViewFormat", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaResourceViewFormat) + cuda2hipRename["CUresourceViewFormat_enum"] = {"hipResourceViewFormat", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaResourceViewFormat) + cuda2hipRename["CU_RES_VIEW_FORMAT_NONE"] = {"hipResViewFormatNone", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaResViewFormatNone = 0x00) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_1X8"] = {"hipResViewFormatUnsignedChar1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedChar1 = 0x01) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_2X8"] = {"hipResViewFormatUnsignedChar2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedChar2 = 0x02) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_4X8"] = {"hipResViewFormatUnsignedChar4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedChar4 = 0x03) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_1X8"] = {"hipResViewFormatSignedChar1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 // API_Runtime ANALOGUE (cudaResViewFormatSignedChar1 = 0x04) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_2X8"] = {"hipResViewFormatSignedChar2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x05 // API_Runtime ANALOGUE (cudaResViewFormatSignedChar2 = 0x05) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_4X8"] = {"hipResViewFormatSignedChar4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x06 // API_Runtime ANALOGUE (cudaResViewFormatSignedChar4 = 0x06) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_1X16"] = {"hipResViewFormatUnsignedShort1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x07 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedShort1 = 0x07) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_2X16"] = {"hipResViewFormatUnsignedShort2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x08 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedShort2 = 0x08) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_4X16"] = {"hipResViewFormatUnsignedShort4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x09 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedShort4 = 0x09) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_1X16"] = {"hipResViewFormatSignedShort1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x0a // API_Runtime ANALOGUE (cudaResViewFormatSignedShort1 = 0x0a) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_2X16"] = {"hipResViewFormatSignedShort2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x0b // API_Runtime ANALOGUE (cudaResViewFormatSignedShort2 = 0x0b) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_4X16"] = {"hipResViewFormatSignedShort4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x0c // API_Runtime ANALOGUE (cudaResViewFormatSignedShort4 = 0x0c) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_1X32"] = {"hipResViewFormatUnsignedInt1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x0d // API_Runtime ANALOGUE (cudaResViewFormatUnsignedInt1 = 0x0d) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_2X32"] = {"hipResViewFormatUnsignedInt2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x0e // API_Runtime ANALOGUE (cudaResViewFormatUnsignedInt2 = 0x0e) + cuda2hipRename["CU_RES_VIEW_FORMAT_UINT_4X32"] = {"hipResViewFormatUnsignedInt4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x0f // API_Runtime ANALOGUE (cudaResViewFormatUnsignedInt4 = 0x0f) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_1X32"] = {"hipResViewFormatSignedInt1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 // API_Runtime ANALOGUE (cudaResViewFormatSignedInt1 = 0x10) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_2X32"] = {"hipResViewFormatSignedInt2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x11 // API_Runtime ANALOGUE (cudaResViewFormatSignedInt2 = 0x11) + cuda2hipRename["CU_RES_VIEW_FORMAT_SINT_4X32"] = {"hipResViewFormatSignedInt4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x12 // API_Runtime ANALOGUE (cudaResViewFormatSignedInt4 = 0x12) + cuda2hipRename["CU_RES_VIEW_FORMAT_FLOAT_1X16"] = {"hipResViewFormatHalf1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x13 // API_Runtime ANALOGUE (cudaResViewFormatHalf1 = 0x13) + cuda2hipRename["CU_RES_VIEW_FORMAT_FLOAT_2X16"] = {"hipResViewFormatHalf2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x14 // API_Runtime ANALOGUE (cudaResViewFormatHalf2 = 0x14) + cuda2hipRename["CU_RES_VIEW_FORMAT_FLOAT_4X16"] = {"hipResViewFormatHalf4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x15 // API_Runtime ANALOGUE (cudaResViewFormatHalf4 = 0x15) + cuda2hipRename["CU_RES_VIEW_FORMAT_FLOAT_1X32"] = {"hipResViewFormatFloat1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x16 // API_Runtime ANALOGUE (cudaResViewFormatFloat1 = 0x16) + cuda2hipRename["CU_RES_VIEW_FORMAT_FLOAT_2X32"] = {"hipResViewFormatFloat2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x17 // API_Runtime ANALOGUE (cudaResViewFormatFloat2 = 0x17) + cuda2hipRename["CU_RES_VIEW_FORMAT_FLOAT_4X32"] = {"hipResViewFormatFloat4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x18 // API_Runtime ANALOGUE (cudaResViewFormatFloat4 = 0x18) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC1"] = {"hipResViewFormatUnsignedBlockCompressed1", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x19 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed1 = 0x19) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC2"] = {"hipResViewFormatUnsignedBlockCompressed2", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x1a // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed2 = 0x1a) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC3"] = {"hipResViewFormatUnsignedBlockCompressed3", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x1b // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed3 = 0x1b) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC4"] = {"hipResViewFormatUnsignedBlockCompressed4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x1c // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed4 = 0x1c) + cuda2hipRename["CU_RES_VIEW_FORMAT_SIGNED_BC4"] = {"hipResViewFormatSignedBlockCompressed4", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x1d // API_Runtime ANALOGUE (cudaResViewFormatSignedBlockCompressed4 = 0x1d) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC5"] = {"hipResViewFormatUnsignedBlockCompressed5", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x1e // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed5 = 0x1e) + cuda2hipRename["CU_RES_VIEW_FORMAT_SIGNED_BC5"] = {"hipResViewFormatSignedBlockCompressed5", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x1f // API_Runtime ANALOGUE (cudaResViewFormatSignedBlockCompressed5 = 0x1f) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC6H"] = {"hipResViewFormatUnsignedBlockCompressed6H", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x20 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed6H = 0x20) + cuda2hipRename["CU_RES_VIEW_FORMAT_SIGNED_BC6H"] = {"hipResViewFormatSignedBlockCompressed6H", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x21 // API_Runtime ANALOGUE (cudaResViewFormatSignedBlockCompressed6H = 0x21) + cuda2hipRename["CU_RES_VIEW_FORMAT_UNSIGNED_BC7"] = {"hipResViewFormatUnsignedBlockCompressed7", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x22 // API_Runtime ANALOGUE (cudaResViewFormatUnsignedBlockCompressed7 = 0x22) + + + + cuda2hipRename["CUsharedconfig_enum"] = {"hipSharedMemConfig", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUsharedconfig"] = {"hipSharedMemConfig", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CU_SHARED_MEM_CONFIG_DEFAULT_BANK_SIZE"] = {"hipSharedMemBankSizeDefault", CONV_DEV, API_DRIVER}; + cuda2hipRename["CU_SHARED_MEM_CONFIG_FOUR_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeFourByte", CONV_DEV, API_DRIVER}; + cuda2hipRename["CU_SHARED_MEM_CONFIG_EIGHT_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeEightByte", CONV_DEV, API_DRIVER}; + + cuda2hipRename["CUcontext"] = {"hipCtx_t", CONV_TYPE, API_DRIVER}; // TODO: // cuda2hipRename["CUctx_st"] = {"XXXX", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUmodule"] = {"hipModule_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUmodule"] = {"hipModule_t", CONV_TYPE, API_DRIVER}; // TODO: // cuda2hipRename["CUmod_st"] = {"XXXX", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CUstream"] = {"hipStream_t", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CUstream"] = {"hipStream_t", CONV_TYPE, API_DRIVER}; // TODO: // cuda2hipRename["CUstream_st"] = {"XXXX", CONV_TYPE, API_DRIVER}; // Stream Flags - cuda2hipRename["CU_STREAM_DEFAULT"] = {"hipStreamDefault", CONV_STREAM, API_DRIVER}; - cuda2hipRename["CU_STREAM_NON_BLOCKING"] = {"hipStreamNonBlocking", CONV_STREAM, API_DRIVER}; + cuda2hipRename["CU_STREAM_DEFAULT"] = {"hipStreamDefault", CONV_STREAM, API_DRIVER}; + cuda2hipRename["CU_STREAM_NON_BLOCKING"] = {"hipStreamNonBlocking", CONV_STREAM, API_DRIVER}; // Init - cuda2hipRename["cuInit"] = {"hipInit", CONV_DRIVER, API_DRIVER}; + cuda2hipRename["cuInit"] = {"hipInit", CONV_DRIVER, API_DRIVER}; // Driver cuda2hipRename["cuDriverGetVersion"] = {"hipDriverGetVersion", CONV_DRIVER, API_DRIVER}; @@ -568,14 +816,17 @@ struct cuda2hipMap { // Events // pointer to CUevent_st cuda2hipRename["CUevent"] = {"hipEvent_t", CONV_TYPE, API_DRIVER}; - // ToDO: - // cuda2hipRename["CUevent_st"] = {"XXXX", CONV_TYPE, API_DRIVER}; + // ToDo: + // cuda2hipRename["CUevent_st"] = {"XXXX", CONV_TYPE, API_DRIVER}; // Event Flags + cuda2hipRename["CUevent_flags"] = {"hipEventFlags", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; + // ToDo: + // cuda2hipRename["CUevent_flags_enum"] = {"hipEventFlags", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["CU_EVENT_DEFAULT"] = {"hipEventDefault", CONV_EVENT, API_DRIVER}; cuda2hipRename["CU_EVENT_BLOCKING_SYNC"] = {"hipEventBlockingSync", CONV_EVENT, API_DRIVER}; cuda2hipRename["CU_EVENT_DISABLE_TIMING"] = {"hipEventDisableTiming", CONV_EVENT, API_DRIVER}; cuda2hipRename["CU_EVENT_INTERPROCESS"] = {"hipEventInterprocess", CONV_EVENT, API_DRIVER}; - + // Event functions cuda2hipRename["cuEventCreate"] = {"hipEventCreate", CONV_EVENT, API_DRIVER}; cuda2hipRename["cuEventDestroy_v2"] = {"hipEventDestroy", CONV_EVENT, API_DRIVER}; cuda2hipRename["cuEventElapsedTime"] = {"hipEventElapsedTime", CONV_EVENT, API_DRIVER}; @@ -627,7 +878,7 @@ struct cuda2hipMap { cuda2hipRename["cuMemsetD16_v2"] = {"hipMemsetD16", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemsetD16Async"] = {"hipMemsetD16Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemsetD2D16_v2"] = {"hipMemsetD2D16", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D16Async"] = {"hipMemsetD2D16Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D16Async"] = {"hipMemsetD2D16Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemsetD32_v2"] = {"hipMemset", CONV_MEM, API_DRIVER}; cuda2hipRename["cuMemsetD32Async"] = {"hipMemsetAsync", CONV_MEM, API_DRIVER}; @@ -639,6 +890,14 @@ struct cuda2hipMap { cuda2hipRename["cuMemHostRegister_v2"] = {"hipHostRegister", CONV_MEM, API_DRIVER}; cuda2hipRename["cuMemHostUnregister"] = {"hipHostUnregister", CONV_MEM, API_DRIVER}; + // Texture Reference Mngmnt + // Texture reference filtering modes + cuda2hipRename["CUfilter_mode"] = {"hipTextureFilterMode", CONV_TEX, API_DRIVER}; // API_Runtime ANALOGUE (cudaTextureFilterMode) + // ToDo: + // cuda2hipRename["CUfilter_mode"] = {"CUfilter_mode_enum", CONV_TEX, API_DRIVER}; // API_Runtime ANALOGUE (cudaTextureFilterMode) + cuda2hipRename["CU_TR_FILTER_MODE_POINT"] = {"hipFilterModePoint", CONV_TEX, API_DRIVER}; // 0 // API_Runtime ANALOGUE (cudaFilterModePoint = 0) + cuda2hipRename["CU_TR_FILTER_MODE_LINEAR"] = {"hipFilterModeLinear", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (cudaFilterModeLinear = 1) + // Profiler // unsupported yet by HIP cuda2hipRename["cuProfilerInitialize"] = {"hipProfilerInitialize", CONV_OTHER, API_DRIVER, HIP_UNSUPPORTED}; @@ -676,6 +935,14 @@ struct cuda2hipMap { cuda2hipRename["MINOR_VERSION"] = {"hipLibraryMinorVersion", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["PATCH_LEVEL"] = {"hipLibraryPatchVersion", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; + // defines + cuda2hipRename["cudaMemAttachGlobal"] = {"hipMemAttachGlobal", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_MEM_ATTACH_GLOBAL = 0x1) + cuda2hipRename["cudaMemAttachHost"] = {"hipMemAttachHost", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_MEM_ATTACH_HOST = 0x2) + cuda2hipRename["cudaMemAttachSingle"] = {"hipMemAttachSingle", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x04 // API_Driver ANALOGUE (CU_MEM_ATTACH_SINGLE = 0x4) + + cuda2hipRename["cudaOccupancyDefault"] = {"hipOccupancyDefault", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_OCCUPANCY_DEFAULT = 0x0) + cuda2hipRename["cudaOccupancyDisableCachingOverride"] = {"hipOccupancyDisableCachingOverride", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE = 0x1) + // Error API cuda2hipRename["cudaGetLastError"] = {"hipGetLastError", CONV_ERR, API_RUNTIME}; cuda2hipRename["cudaPeekAtLastError"] = {"hipPeekAtLastError", CONV_ERR, API_RUNTIME}; @@ -766,7 +1033,7 @@ struct cuda2hipMap { cuda2hipRename["cudaHostAlloc"] = {"hipHostMalloc", CONV_MEM, API_RUNTIME}; // Memory types - cuda2hipRename["cudaMemoryType"] = {"hipMemoryType", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaMemoryType"] = {"hipMemoryType", CONV_MEM, API_RUNTIME}; // API_Driver ANALOGUE (no - CUmemorytype is not an analogue) cuda2hipRename["cudaMemoryTypeHost"] = {"hipMemoryTypeHost", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMemoryTypeDevice"] = {"hipMemoryTypeDevice", CONV_MEM, API_RUNTIME}; @@ -838,7 +1105,6 @@ struct cuda2hipMap { cuda2hipRename["cudaEventBlockingSync"] = {"hipEventBlockingSync", CONV_EVENT, API_RUNTIME}; cuda2hipRename["cudaEventDisableTiming"] = {"hipEventDisableTiming", CONV_EVENT, API_RUNTIME}; cuda2hipRename["cudaEventInterprocess"] = {"hipEventInterprocess", CONV_EVENT, API_RUNTIME}; - // Streams cuda2hipRename["cudaStream_t"] = {"hipStream_t", CONV_TYPE, API_RUNTIME}; cuda2hipRename["cudaStreamCreate"] = {"hipStreamCreate", CONV_STREAM, API_RUNTIME}; @@ -874,93 +1140,94 @@ struct cuda2hipMap { // Attributes cuda2hipRename["cudaDeviceGetAttribute"] = {"hipDeviceGetAttribute", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceAttr"] = {"hipDeviceAttribute_t", CONV_TYPE, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxThreadsPerBlock"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxBlockDimX"] = {"hipDeviceAttributeMaxBlockDimX", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxBlockDimY"] = {"hipDeviceAttributeMaxBlockDimY", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxBlockDimZ"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxGridDimX"] = {"hipDeviceAttributeMaxGridDimX", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxGridDimY"] = {"hipDeviceAttributeMaxGridDimY", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxGridDimZ"] = {"hipDeviceAttributeMaxGridDimZ", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxSharedMemoryPerBlock"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrTotalConstantMemory"] = {"hipDeviceAttributeTotalConstantMemory", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrWarpSize"] = {"hipDeviceAttributeWarpSize", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxRegistersPerBlock"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrClockRate"] = {"hipDeviceAttributeClockRate", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMemoryClockRate"] = {"hipDeviceAttributeMemoryClockRate", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrGlobalMemoryBusWidth"] = {"hipDeviceAttributeMemoryBusWidth", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMultiProcessorCount"] = {"hipDeviceAttributeMultiprocessorCount", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrComputeMode"] = {"hipDeviceAttributeComputeMode", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrL2CacheSize"] = {"hipDeviceAttributeL2CacheSize", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxThreadsPerMultiProcessor"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrComputeCapabilityMajor"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrComputeCapabilityMinor"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrConcurrentKernels"] = {"hipDeviceAttributeConcurrentKernels", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrPciBusId"] = {"hipDeviceAttributePciBusId", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrPciDeviceId"] = {"hipDeviceAttributePciDeviceId", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrMaxSharedMemoryPerMultiprocessor"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDevAttrIsMultiGpuBoard"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_DEV, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaDevAttrMaxPitch"] = {"hipDeviceAttributeMaxPitch", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrTextureAlignment"] = {"hipDeviceAttributeTextureAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaDeviceAttr"] = {"hipDeviceAttribute_t", CONV_TYPE, API_RUNTIME}; // API_DRIVER ANALOGUE (CUdevice_attribute) + cuda2hipRename["cudaDevAttrMaxThreadsPerBlock"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_DEV, API_RUNTIME}; // 1 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK = 1) + cuda2hipRename["cudaDevAttrMaxBlockDimX"] = {"hipDeviceAttributeMaxBlockDimX", CONV_DEV, API_RUNTIME}; // 2 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X = 2) + cuda2hipRename["cudaDevAttrMaxBlockDimY"] = {"hipDeviceAttributeMaxBlockDimY", CONV_DEV, API_RUNTIME}; // 3 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y = 3) + cuda2hipRename["cudaDevAttrMaxBlockDimZ"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_DEV, API_RUNTIME}; // 4 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z = 4) + cuda2hipRename["cudaDevAttrMaxGridDimX"] = {"hipDeviceAttributeMaxGridDimX", CONV_DEV, API_RUNTIME}; // 5 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 5) + cuda2hipRename["cudaDevAttrMaxGridDimY"] = {"hipDeviceAttributeMaxGridDimY", CONV_DEV, API_RUNTIME}; // 6 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 6) + cuda2hipRename["cudaDevAttrMaxGridDimZ"] = {"hipDeviceAttributeMaxGridDimZ", CONV_DEV, API_RUNTIME}; // 7 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 7) + cuda2hipRename["cudaDevAttrMaxSharedMemoryPerBlock"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_RUNTIME}; // 8 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK = 8) + cuda2hipRename["cudaDevAttrTotalConstantMemory"] = {"hipDeviceAttributeTotalConstantMemory", CONV_DEV, API_RUNTIME}; // 9 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY =9) + cuda2hipRename["cudaDevAttrWarpSize"] = {"hipDeviceAttributeWarpSize", CONV_DEV, API_RUNTIME}; // 10 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_WARP_SIZE = 10) + cuda2hipRename["cudaDevAttrMaxPitch"] = {"hipDeviceAttributeMaxPitch", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 11 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_PITCH = 11) + cuda2hipRename["cudaDevAttrMaxRegistersPerBlock"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_RUNTIME}; // 12 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK = 12) + cuda2hipRename["cudaDevAttrClockRate"] = {"hipDeviceAttributeClockRate", CONV_DEV, API_RUNTIME}; // 13 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CLOCK_RATE = 13) + cuda2hipRename["cudaDevAttrTextureAlignment"] = {"hipDeviceAttributeTextureAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 14 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT = 14) // Is not deprecated as CUDA Driver's API analogue CU_DEVICE_ATTRIBUTE_GPU_OVERLAP - cuda2hipRename["cudaDevAttrGpuOverlap"] = {"hipDeviceAttributeGpuOverlap", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrKernelExecTimeout"] = {"hipDeviceAttributeKernelExecTimeout", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrIntegrated"] = {"hipDeviceAttributeIntegrated", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrCanMapHostMemory"] = {"hipDeviceAttributeCanMapHostMemory", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture1DWidth"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DWidth"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DHeight"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture3DWidth"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture3DHeight"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture3DDepth"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DLayeredWidth"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DLayeredHeight"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DLayeredLayers"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrSurfaceAlignment"] = {"hipDeviceAttributeSurfaceAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrEccEnabled"] = {"hipDeviceAttributeEccEnabled", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrTccDriver"] = {"hipDeviceAttributeTccDriver", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrAsyncEngineCount"] = {"hipDevAttrAsyncEngineCount", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrUnifiedAddressing"] = {"hipDeviceAttributeUnifiedAddressing", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture1DLayeredWidth"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture1DLayeredLayers"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DGatherWidth"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DGatherHeight"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture3DWidthAlt"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture3DHeightAlt"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture3DDepthAlt"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrPciDomainId"] = {"hipDeviceAttributePciDomainId", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrTexturePitchAlignment"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTextureCubemapWidth"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredWidth"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredLayers"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface1DWidth"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface2DWidth"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface2DHeight"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface3DWidth"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface3DHeight"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface3DDepth"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface1DLayeredWidth"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface1DLayeredLayers"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface2DLayeredWidth"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface2DLayeredHeight"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurface2DLayeredLayers"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurfaceCubemapWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredLayers"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture1DLinearWidth"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DLinearWidth"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DLinearHeight"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DLinearPitch"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedHeight"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxTexture1DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrStreamPrioritiesSupported"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrGlobalL1CacheSupported"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrLocalL1CacheSupported"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMaxRegistersPerMultiprocessor"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrManagedMemory"] = {"hipDeviceAttributeManagedMemory", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDevAttrMultiGpuBoardGroupID"] = {"hipDeviceAttributeMultiGpuBoardGroupID", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaDevAttrGpuOverlap"] = {"hipDeviceAttributeGpuOverlap", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 15 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GPU_OVERLAP = 15) + cuda2hipRename["cudaDevAttrMultiProcessorCount"] = {"hipDeviceAttributeMultiprocessorCount", CONV_DEV, API_RUNTIME}; // 16 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT = 16) + cuda2hipRename["cudaDevAttrKernelExecTimeout"] = {"hipDeviceAttributeKernelExecTimeout", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 17 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT = 17) + cuda2hipRename["cudaDevAttrIntegrated"] = {"hipDeviceAttributeIntegrated", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 18 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_INTEGRATED = 18) + cuda2hipRename["cudaDevAttrCanMapHostMemory"] = {"hipDeviceAttributeCanMapHostMemory", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 19 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY = 19) + cuda2hipRename["cudaDevAttrComputeMode"] = {"hipDeviceAttributeComputeMode", CONV_DEV, API_RUNTIME}; // 20 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_MODE = 20) + cuda2hipRename["cudaDevAttrMaxTexture1DWidth"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 21 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH = 21) + cuda2hipRename["cudaDevAttrMaxTexture2DWidth"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 22 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH = 22) + cuda2hipRename["cudaDevAttrMaxTexture2DHeight"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 23 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT = 23) + cuda2hipRename["cudaDevAttrMaxTexture3DWidth"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 24 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH = 24) + cuda2hipRename["cudaDevAttrMaxTexture3DHeight"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 25 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT = 25) + cuda2hipRename["cudaDevAttrMaxTexture3DDepth"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 26 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH = 26) + cuda2hipRename["cudaDevAttrMaxTexture2DLayeredWidth"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 27 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH = 27) + cuda2hipRename["cudaDevAttrMaxTexture2DLayeredHeight"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 28 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT = 28) + cuda2hipRename["cudaDevAttrMaxTexture2DLayeredLayers"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 29 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS = 29) + cuda2hipRename["cudaDevAttrSurfaceAlignment"] = {"hipDeviceAttributeSurfaceAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 30 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT = 30) + cuda2hipRename["cudaDevAttrConcurrentKernels"] = {"hipDeviceAttributeConcurrentKernels", CONV_DEV, API_RUNTIME}; // 31 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS = 31) + cuda2hipRename["cudaDevAttrEccEnabled"] = {"hipDeviceAttributeEccEnabled", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 32 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_ECC_ENABLED = 32) + cuda2hipRename["cudaDevAttrPciBusId"] = {"hipDeviceAttributePciBusId", CONV_DEV, API_RUNTIME}; // 33 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_BUS_ID = 33) + cuda2hipRename["cudaDevAttrPciDeviceId"] = {"hipDeviceAttributePciDeviceId", CONV_DEV, API_RUNTIME}; // 34 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID = 34) + cuda2hipRename["cudaDevAttrTccDriver"] = {"hipDeviceAttributeTccDriver", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 35 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TCC_DRIVER = 35) + cuda2hipRename["cudaDevAttrMemoryClockRate"] = {"hipDeviceAttributeMemoryClockRate", CONV_DEV, API_RUNTIME}; // 36 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE = 36) + cuda2hipRename["cudaDevAttrGlobalMemoryBusWidth"] = {"hipDeviceAttributeMemoryBusWidth", CONV_DEV, API_RUNTIME}; // 37 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH = 37) + cuda2hipRename["cudaDevAttrL2CacheSize"] = {"hipDeviceAttributeL2CacheSize", CONV_DEV, API_RUNTIME}; // 38 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE = 38) + cuda2hipRename["cudaDevAttrMaxThreadsPerMultiProcessor"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_DEV, API_RUNTIME}; // 39 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR = 39) + cuda2hipRename["cudaDevAttrAsyncEngineCount"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 40 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT = 40) + cuda2hipRename["cudaDevAttrUnifiedAddressing"] = {"hipDeviceAttributeUnifiedAddressing", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 41 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING = 41) + cuda2hipRename["cudaDevAttrMaxTexture1DLayeredWidth"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 42 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH = 42) + cuda2hipRename["cudaDevAttrMaxTexture1DLayeredLayers"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 43 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS = 43) + // 44 - no + cuda2hipRename["cudaDevAttrMaxTexture2DGatherWidth"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 45 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH = 45) + cuda2hipRename["cudaDevAttrMaxTexture2DGatherHeight"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 46 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT = 46) + cuda2hipRename["cudaDevAttrMaxTexture3DWidthAlt"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 47 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE = 47) + cuda2hipRename["cudaDevAttrMaxTexture3DHeightAlt"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 48 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE = 48) + cuda2hipRename["cudaDevAttrMaxTexture3DDepthAlt"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 49 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE = 49) + cuda2hipRename["cudaDevAttrPciDomainId"] = {"hipDeviceAttributePciDomainId", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 50 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID = 50) + cuda2hipRename["cudaDevAttrTexturePitchAlignment"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 51 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT = 51) + cuda2hipRename["cudaDevAttrMaxTextureCubemapWidth"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 52 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH = 52) + cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredWidth"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 53 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH = 53) + cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredLayers"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 54 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS = 54) + cuda2hipRename["cudaDevAttrMaxSurface1DWidth"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 55 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH = 55) + cuda2hipRename["cudaDevAttrMaxSurface2DWidth"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 56 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH = 56) + cuda2hipRename["cudaDevAttrMaxSurface2DHeight"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 57 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT = 57) + cuda2hipRename["cudaDevAttrMaxSurface3DWidth"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 58 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH = 58) + cuda2hipRename["cudaDevAttrMaxSurface3DHeight"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 59 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT = 59) + cuda2hipRename["cudaDevAttrMaxSurface3DDepth"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 60 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH = 60) + cuda2hipRename["cudaDevAttrMaxSurface1DLayeredWidth"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 61 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH = 61) + cuda2hipRename["cudaDevAttrMaxSurface1DLayeredLayers"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 62 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS = 62) + cuda2hipRename["cudaDevAttrMaxSurface2DLayeredWidth"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 63 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH = 63) + cuda2hipRename["cudaDevAttrMaxSurface2DLayeredHeight"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 64 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT = 64) + cuda2hipRename["cudaDevAttrMaxSurface2DLayeredLayers"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 65 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS = 65) + cuda2hipRename["cudaDevAttrMaxSurfaceCubemapWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 66 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH = 66) + cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 67 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH = 67) + cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredLayers"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 68 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS = 68) + cuda2hipRename["cudaDevAttrMaxTexture1DLinearWidth"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 69 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH = 69) + cuda2hipRename["cudaDevAttrMaxTexture2DLinearWidth"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 70 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH = 70) + cuda2hipRename["cudaDevAttrMaxTexture2DLinearHeight"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 71 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT = 71) + cuda2hipRename["cudaDevAttrMaxTexture2DLinearPitch"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 72 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH = 72) + cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 73 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH = 73) + cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedHeight"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 74 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT = 74) + cuda2hipRename["cudaDevAttrComputeCapabilityMajor"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_DEV, API_RUNTIME}; // 75 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR = 75) + cuda2hipRename["cudaDevAttrComputeCapabilityMinor"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_DEV, API_RUNTIME}; // 76 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR = 76) + cuda2hipRename["cudaDevAttrMaxTexture1DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 77 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH = 77) + cuda2hipRename["cudaDevAttrStreamPrioritiesSupported"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 78 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED = 78) + cuda2hipRename["cudaDevAttrGlobalL1CacheSupported"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 79 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED = 79) + cuda2hipRename["cudaDevAttrLocalL1CacheSupported"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 80 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED = 80) + cuda2hipRename["cudaDevAttrMaxSharedMemoryPerMultiprocessor"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_DEV, API_RUNTIME}; // 81 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR = 81) + cuda2hipRename["cudaDevAttrMaxRegistersPerMultiprocessor"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 82 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR = 82) + cuda2hipRename["cudaDevAttrManagedMemory"] = {"hipDeviceAttributeManagedMemory", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 83 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY = 83) + cuda2hipRename["cudaDevAttrIsMultiGpuBoard"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_DEV, API_RUNTIME}; // 84 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD = 84) + cuda2hipRename["cudaDevAttrMultiGpuBoardGroupID"] = {"hipDeviceAttributeMultiGpuBoardGroupID", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 85 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID = 85) + // unsupported yet by HIP [CUDA 8.0.44] cuda2hipRename["cudaDevAttrHostNativeAtomicSupported"] = {"hipDeviceAttributeHostNativeAtomicSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaDevAttrSingleToDoublePrecisionPerfRatio"] = {"hipDeviceAttributeSingleToDoublePrecisionPerfRatio", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; @@ -970,10 +1237,11 @@ struct cuda2hipMap { cuda2hipRename["cudaDevAttrCanUseHostPointerForRegisteredMem"] = {"hipDeviceAttributeCanUseHostPointerForRegisteredMem", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // Pointer Attributes - cuda2hipRename["cudaPointerAttributes"] = {"hipPointerAttribute_t", CONV_TYPE, API_RUNTIME}; - cuda2hipRename["cudaPointerGetAttributes"] = {"hipPointerGetAttributes", CONV_MEM, API_RUNTIME}; + // struct cudaPointerAttributes + cuda2hipRename["cudaPointerAttributes"] = {"hipPointerAttribute_t", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaPointerGetAttributes"] = {"hipPointerGetAttributes", CONV_MEM, API_RUNTIME}; - cuda2hipRename["cudaHostGetDevicePointer"] = {"hipHostGetDevicePointer", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaHostGetDevicePointer"] = {"hipHostGetDevicePointer", CONV_MEM, API_RUNTIME}; // Device cuda2hipRename["cudaDeviceProp"] = {"hipDeviceProp_t", CONV_TYPE, API_RUNTIME}; @@ -985,11 +1253,11 @@ struct cuda2hipMap { cuda2hipRename["cudaSetValidDevices"] = {"hipSetValidDevices", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // Compute mode - cuda2hipRename["cudaComputeMode"] = {"hipComputeMode", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaComputeModeDefault"] = {"hipComputeModeDefault", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaComputeModeExclusive"] = {"hipComputeModeExclusive", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaComputeModeProhibited"] = {"hipComputeModeProhibited", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaComputeModeExclusiveProcess"] = {"hipComputeModeExclusiveProcess", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaComputeMode"] = {"hipComputeMode", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (CUcomputemode) + cuda2hipRename["cudaComputeModeDefault"] = {"hipComputeModeDefault", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 0 + cuda2hipRename["cudaComputeModeExclusive"] = {"hipComputeModeExclusive", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 1 + cuda2hipRename["cudaComputeModeProhibited"] = {"hipComputeModeProhibited", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 2 + cuda2hipRename["cudaComputeModeExclusiveProcess"] = {"hipComputeModeExclusiveProcess", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 3 // Device Flags // unsupported yet by HIP @@ -1020,11 +1288,11 @@ struct cuda2hipMap { // Execution control // CUDA function cache configurations - cuda2hipRename["cudaFuncCache"] = {"hipFuncCache_t", CONV_CACHE, API_RUNTIME}; - cuda2hipRename["cudaFuncCachePreferNone"] = {"hipFuncCachePreferNone", CONV_CACHE, API_RUNTIME}; - cuda2hipRename["cudaFuncCachePreferShared"] = {"hipFuncCachePreferShared", CONV_CACHE, API_RUNTIME}; - cuda2hipRename["cudaFuncCachePreferL1"] = {"hipFuncCachePreferL1", CONV_CACHE, API_RUNTIME}; - cuda2hipRename["cudaFuncCachePreferEqual"] = {"hipFuncCachePreferEqual", CONV_CACHE, API_RUNTIME}; + cuda2hipRename["cudaFuncCache"] = {"hipFuncCache_t", CONV_CACHE, API_RUNTIME}; // API_Driver ANALOGUE (CUfunc_cache) + cuda2hipRename["cudaFuncCachePreferNone"] = {"hipFuncCachePreferNone", CONV_CACHE, API_RUNTIME}; // 0 // API_Driver ANALOGUE (CU_FUNC_CACHE_PREFER_NONE = 0x00) + cuda2hipRename["cudaFuncCachePreferShared"] = {"hipFuncCachePreferShared", CONV_CACHE, API_RUNTIME}; // 1 // API_Driver ANALOGUE (CU_FUNC_CACHE_PREFER_SHARED = 0x01) + cuda2hipRename["cudaFuncCachePreferL1"] = {"hipFuncCachePreferL1", CONV_CACHE, API_RUNTIME}; // 2 // API_Driver ANALOGUE (CU_FUNC_CACHE_PREFER_L1 = 0x02) + cuda2hipRename["cudaFuncCachePreferEqual"] = {"hipFuncCachePreferEqual", CONV_CACHE, API_RUNTIME}; // 3 // API_Driver ANALOGUE (CU_FUNC_CACHE_PREFER_EQUAL = 0x03) // Execution control functions // unsupported yet by HIP @@ -1062,7 +1330,9 @@ struct cuda2hipMap { cuda2hipRename["cudaDeviceEnablePeerAccess"] = {"hipDeviceEnablePeerAccess", CONV_DEV, API_RUNTIME}; cuda2hipRename["cudaMemcpyPeerAsync"] = {"hipMemcpyPeerAsync", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMemcpyPeer"] = {"hipMemcpyPeer", CONV_MEM, API_RUNTIME}; - cuda2hipRename["cudaIpcMemLazyEnablePeerAccess"] = {"hipIpcMemLazyEnablePeerAccess", CONV_ERR, API_RUNTIME}; + + // #define cudaIpcMemLazyEnablePeerAccess 0x01 + cuda2hipRename["cudaIpcMemLazyEnablePeerAccess"] = {"hipIpcMemLazyEnablePeerAccess", CONV_TYPE, API_RUNTIME}; // 0x01 // API_Driver ANALOGUE (CU_IPC_MEM_LAZY_ENABLE_PEER_ACCESS = 0x1) // Shared memory cuda2hipRename["cudaDeviceSetSharedMemConfig"] = {"hipDeviceSetSharedMemConfig", CONV_DEV, API_RUNTIME}; @@ -1078,14 +1348,12 @@ struct cuda2hipMap { cuda2hipRename["cudaSharedMemBankSizeEightByte"] = {"hipSharedMemBankSizeEightByte", CONV_DEV, API_RUNTIME}; // Limits - cuda2hipRename["cudaLimit"] = {"hipLimit_t", CONV_DEV, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaLimitStackSize"] = {"hipLimitStackSize", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaLimitPrintfFifoSize"] = {"hipLimitPrintfFifoSize", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaLimitMallocHeapSize"] = {"hipLimitMallocHeapSize", CONV_DEV, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaLimitDevRuntimeSyncDepth"] = {"hipLimitPrintfFifoSize", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaLimitDevRuntimePendingLaunchCount"] = {"hipLimitMallocHeapSize", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaLimit"] = {"hipLimit_t", CONV_TYPE, API_RUNTIME}; // API_Driver ANALOGUE (CUlimit) + cuda2hipRename["cudaLimitStackSize"] = {"hipLimitStackSize", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_LIMIT_STACK_SIZE = 0x00) + cuda2hipRename["cudaLimitPrintfFifoSize"] = {"hipLimitPrintfFifoSize", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_LIMIT_PRINTF_FIFO_SIZE = 0x01) + cuda2hipRename["cudaLimitMallocHeapSize"] = {"hipLimitMallocHeapSize", CONV_TYPE, API_RUNTIME}; // 0x02 // API_Driver ANALOGUE (CU_LIMIT_MALLOC_HEAP_SIZE = 0x02) + cuda2hipRename["cudaLimitDevRuntimeSyncDepth"] = {"hipLimitDevRuntimeSyncDepth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH = 0x03) + cuda2hipRename["cudaLimitDevRuntimePendingLaunchCount"] = {"hipLimitDevRuntimePendingLaunchCount", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x04 // API_Driver ANALOGUE (CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT = 0x04) cuda2hipRename["cudaDeviceGetLimit"] = {"hipDeviceGetLimit", CONV_DEV, API_RUNTIME}; @@ -1108,10 +1376,9 @@ struct cuda2hipMap { // unsupported yet by HIP cuda2hipRename["cudaReadModeNormalizedFloat"] = {"hipReadModeNormalizedFloat", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaTextureFilterMode"] = {"hipTextureFilterMode", CONV_TEX, API_RUNTIME}; - cuda2hipRename["cudaFilterModePoint"] = {"hipFilterModePoint", CONV_TEX, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaFilterModeLinear"] = {"hipFilterModeLinear", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaTextureFilterMode"] = {"hipTextureFilterMode", CONV_TEX, API_RUNTIME}; // API_DRIVER ANALOGUE (CUfilter_mode) + cuda2hipRename["cudaFilterModePoint"] = {"hipFilterModePoint", CONV_TEX, API_RUNTIME}; // 0 // API_DRIVER ANALOGUE (CU_TR_FILTER_MODE_POINT = 0) + cuda2hipRename["cudaFilterModeLinear"] = {"hipFilterModeLinear", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_DRIVER ANALOGUE (CU_TR_FILTER_MODE_POINT = 1) cuda2hipRename["cudaBindTexture"] = {"hipBindTexture", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaUnbindTexture"] = {"hipUnbindTexture", CONV_TEX, API_RUNTIME}; @@ -1131,7 +1398,7 @@ struct cuda2hipMap { cuda2hipRename["cudaChannelFormatDesc"] = {"hipChannelFormatDesc", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaCreateChannelDesc"] = {"hipCreateChannelDesc", CONV_TEX, API_RUNTIME}; // unsupported yet by HIP - cuda2hipRename["cudaGetChannelDesc"] = {"hipGetChannelDesc", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaGetChannelDesc"] = {"hipGetChannelDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // Texture Object Management // structs @@ -1139,49 +1406,52 @@ struct cuda2hipMap { cuda2hipRename["cudaResourceDesc"] = {"hipResourceDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaResourceViewDesc"] = {"hipResourceViewDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaTextureDesc"] = {"hipTextureDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - // enums - // unsupported yet by HIP - cuda2hipRename["cudaResourceType"] = {"hipResourceType", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResourceTypeArray"] = {"hipResourceTypeArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResourceTypeMipmappedArray"] = {"hipResourceTypeMipmappedArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResourceTypeLinear"] = {"hipResourceTypeLinear", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResourceTypePitch2D"] = {"hipResourceTypePitch2D", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResourceViewFormat"] = {"hipResourceViewFormat", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatNone"] = {"hipResViewFormatNone", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedChar1"] = {"hipResViewFormatUnsignedChar1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedChar2"] = {"hipResViewFormatUnsignedChar2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedChar4"] = {"hipResViewFormatUnsignedChar4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedChar1"] = {"hipResViewFormatSignedChar1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedChar2"] = {"hipResViewFormatSignedChar2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedChar4"] = {"hipResViewFormatSignedChar4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedShort1"] = {"hipResViewFormatUnsignedShort1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedShort2"] = {"hipResViewFormatUnsignedShort2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedShort4"] = {"hipResViewFormatUnsignedShort4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedShort1"] = {"hipResViewFormatSignedShort1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedShort2"] = {"hipResViewFormatSignedShort2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedShort4"] = {"hipResViewFormatSignedShort4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedInt1"] = {"hipResViewFormatUnsignedInt1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedInt2"] = {"hipResViewFormatUnsignedInt2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedInt4"] = {"hipResViewFormatUnsignedInt4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedInt1"] = {"hipResViewFormatSignedInt1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedInt2"] = {"hipResViewFormatSignedInt2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedInt4"] = {"hipResViewFormatSignedInt4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatHalf1"] = {"hipResViewFormatHalf1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatHalf2"] = {"hipResViewFormatHalf2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatHalf4"] = {"hipResViewFormatHalf4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatFloat1"] = {"hipResViewFormatFloat1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatFloat2"] = {"hipResViewFormatFloat2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatFloat4"] = {"hipResViewFormatFloat4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed1"] = {"hipResViewFormatUnsignedBlockCompressed1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed2"] = {"hipResViewFormatUnsignedBlockCompressed2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed3"] = {"hipResViewFormatUnsignedBlockCompressed3", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed4"] = {"hipResViewFormatUnsignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedBlockCompressed4"] = {"hipResViewFormatSignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed5"] = {"hipResViewFormatUnsignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedBlockCompressed5"] = {"hipResViewFormatSignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatSignedBlockCompressed6H"] = {"hipResViewFormatSignedBlockCompressed6H", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed7"] = {"hipResViewFormatUnsignedBlockCompressed7", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + // enums + // enum cudaResourceType + cuda2hipRename["cudaResourceType"] = {"hipResourceType", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUresourcetype) + cuda2hipRename["cudaResourceTypeArray"] = {"hipResourceTypeArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_ARRAY = 0x00) + cuda2hipRename["cudaResourceTypeMipmappedArray"] = {"hipResourceTypeMipmappedArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_MIPMAPPED_ARRAY = 0x01) + cuda2hipRename["cudaResourceTypeLinear"] = {"hipResourceTypeLinear", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_LINEAR = 0x02) + cuda2hipRename["cudaResourceTypePitch2D"] = {"hipResourceTypePitch2D", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_PITCH2D = 0x03) + + + cuda2hipRename["cudaResourceViewFormat"] = {"hipResourceViewFormat", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUresourceViewFormat) + cuda2hipRename["cudaResViewFormatNone"] = {"hipResViewFormatNone", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_NONE = 0x00) + cuda2hipRename["cudaResViewFormatUnsignedChar1"] = {"hipResViewFormatUnsignedChar1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X8 = 0x01) + cuda2hipRename["cudaResViewFormatUnsignedChar2"] = {"hipResViewFormatUnsignedChar2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X8 = 0x02) + cuda2hipRename["cudaResViewFormatUnsignedChar4"] = {"hipResViewFormatUnsignedChar4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X8 = 0x03) + cuda2hipRename["cudaResViewFormatSignedChar1"] = {"hipResViewFormatSignedChar1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x04 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X8 = 0x04) + cuda2hipRename["cudaResViewFormatSignedChar2"] = {"hipResViewFormatSignedChar2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x05 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X8 = 0x05) + cuda2hipRename["cudaResViewFormatSignedChar4"] = {"hipResViewFormatSignedChar4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x06 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X8 = 0x06) + cuda2hipRename["cudaResViewFormatUnsignedShort1"] = {"hipResViewFormatUnsignedShort1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x07 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X16 = 0x07) + cuda2hipRename["cudaResViewFormatUnsignedShort2"] = {"hipResViewFormatUnsignedShort2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x08 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X16 = 0x08) + cuda2hipRename["cudaResViewFormatUnsignedShort4"] = {"hipResViewFormatUnsignedShort4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x09 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X16 = 0x09) + cuda2hipRename["cudaResViewFormatSignedShort1"] = {"hipResViewFormatSignedShort1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0a // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X16 = 0x0a) + cuda2hipRename["cudaResViewFormatSignedShort2"] = {"hipResViewFormatSignedShort2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0b // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X16 = 0x0b) + cuda2hipRename["cudaResViewFormatSignedShort4"] = {"hipResViewFormatSignedShort4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0c // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X16 = 0x0c) + cuda2hipRename["cudaResViewFormatUnsignedInt1"] = {"hipResViewFormatUnsignedInt1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X32 = 0x0d) + cuda2hipRename["cudaResViewFormatUnsignedInt2"] = {"hipResViewFormatUnsignedInt2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X32 = 0x0e) + cuda2hipRename["cudaResViewFormatUnsignedInt4"] = {"hipResViewFormatUnsignedInt4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X32 = 0x0f) + cuda2hipRename["cudaResViewFormatSignedInt1"] = {"hipResViewFormatSignedInt1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x10 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X32 = 0x10) + cuda2hipRename["cudaResViewFormatSignedInt2"] = {"hipResViewFormatSignedInt2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x11 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X32 = 0x11) + cuda2hipRename["cudaResViewFormatSignedInt4"] = {"hipResViewFormatSignedInt4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x12 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X32 = 0x12) + cuda2hipRename["cudaResViewFormatHalf1"] = {"hipResViewFormatHalf1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x13 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_1X16 = 0x13) + cuda2hipRename["cudaResViewFormatHalf2"] = {"hipResViewFormatHalf2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x14 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_2X16 = 0x14) + cuda2hipRename["cudaResViewFormatHalf4"] = {"hipResViewFormatHalf4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x15 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_4X16 = 0x15) + cuda2hipRename["cudaResViewFormatFloat1"] = {"hipResViewFormatFloat1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x16 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_1X32 = 0x16) + cuda2hipRename["cudaResViewFormatFloat2"] = {"hipResViewFormatFloat2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x17 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_2X32 = 0x17) + cuda2hipRename["cudaResViewFormatFloat4"] = {"hipResViewFormatFloat4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x18 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_4X32 = 0x18) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed1"] = {"hipResViewFormatUnsignedBlockCompressed1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x19 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC1 = 0x19) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed2"] = {"hipResViewFormatUnsignedBlockCompressed2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1a // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC2 = 0x1a) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed3"] = {"hipResViewFormatUnsignedBlockCompressed3", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1b // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC3 = 0x1b) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed4"] = {"hipResViewFormatUnsignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1c // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC4 = 0x1c) + cuda2hipRename["cudaResViewFormatSignedBlockCompressed4"] = {"hipResViewFormatSignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC4 = 0x1d) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed5"] = {"hipResViewFormatUnsignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC5 = 0x1e) + cuda2hipRename["cudaResViewFormatSignedBlockCompressed5"] = {"hipResViewFormatSignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC5 = 0x1f) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed6H"] = {"hipResViewFormatUnsignedBlockCompressed6H", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x20 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC6H = 0x20) + cuda2hipRename["cudaResViewFormatSignedBlockCompressed6H"] = {"hipResViewFormatSignedBlockCompressed6H", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x21 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC6H = 0x21) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed7"] = {"hipResViewFormatUnsignedBlockCompressed7", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x22 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC7 = 0x22) cuda2hipRename["cudaTextureAddressMode"] = {"hipTextureAddressMode", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaAddressModeWrap"] = {"hipAddressModeWrap", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; @@ -1255,17 +1525,19 @@ struct cuda2hipMap { cuda2hipRename["cudaGraphicsCubeFacePositiveZ"] = {"hipGraphicsCubeFacePositiveZ", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaGraphicsCubeFaceNegativeZ"] = {"hipGraphicsCubeFaceNegativeZ", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsMapFlags"] = {"hipGraphicsMapFlags", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsMapFlagsNone"] = {"hipGraphicsMapFlagsNone", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsMapFlagsReadOnly"] = {"hipGraphicsMapFlagsReadOnly", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsMapFlagsWriteDiscard"] = {"hipGraphicsMapFlagsWriteDiscard", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; + // enum cudaGraphicsMapFlags + cuda2hipRename["cudaGraphicsMapFlags"] = {"hipGraphicsMapFlags", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUgraphicsMapResourceFlags) + cuda2hipRename["cudaGraphicsMapFlagsNone"] = {"hipGraphicsMapFlagsNone", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_Driver ANALOGUE (CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE = 0x00) + cuda2hipRename["cudaGraphicsMapFlagsReadOnly"] = {"hipGraphicsMapFlagsReadOnly", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY = 0x01) + cuda2hipRename["cudaGraphicsMapFlagsWriteDiscard"] = {"hipGraphicsMapFlagsWriteDiscard", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_GRAPHICS_MAP_RESOURCE_FLAGS_WRITE_DISCARD = 0x02) - cuda2hipRename["cudaGraphicsRegisterFlags"] = {"hipGraphicsRegisterFlags", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsRegisterFlagsNone"] = {"hipGraphicsRegisterFlagsNone", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsRegisterFlagsReadOnly"] = {"hipGraphicsRegisterFlagsReadOnly", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsRegisterFlagsWriteDiscard"] = {"hipGraphicsRegisterFlagsWriteDiscard", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsRegisterFlagsSurfaceLoadStore"] = {"hipGraphicsRegisterFlagsSurfaceLoadStore", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsRegisterFlagsTextureGather"] = {"hipGraphicsRegisterFlagsTextureGather", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; + // enum cudaGraphicsRegisterFlags + cuda2hipRename["cudaGraphicsRegisterFlags"] = {"hipGraphicsRegisterFlags", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUgraphicsRegisterFlags) + cuda2hipRename["cudaGraphicsRegisterFlagsNone"] = {"hipGraphicsRegisterFlagsNone", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_Driver ANALOGUE (CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE = 0x00) + cuda2hipRename["cudaGraphicsRegisterFlagsReadOnly"] = {"hipGraphicsRegisterFlagsReadOnly", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY = 0x01) + cuda2hipRename["cudaGraphicsRegisterFlagsWriteDiscard"] = {"hipGraphicsRegisterFlagsWriteDiscard", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD = 0x02) + cuda2hipRename["cudaGraphicsRegisterFlagsSurfaceLoadStore"] = {"hipGraphicsRegisterFlagsSurfaceLoadStore", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 4 // API_Driver ANALOGUE (CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST = 0x04) + cuda2hipRename["cudaGraphicsRegisterFlagsTextureGather"] = {"hipGraphicsRegisterFlagsTextureGather", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 8 // API_Driver ANALOGUE (CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER = 0x08) //---------------------------------------BLAS-------------------------------------// // Blas types diff --git a/include/hip/nvcc_detail/hip_runtime.h b/include/hip/nvcc_detail/hip_runtime.h index b4fa13f48c..80da388007 100644 --- a/include/hip/nvcc_detail/hip_runtime.h +++ b/include/hip/nvcc_detail/hip_runtime.h @@ -36,6 +36,10 @@ do {\ kernelName<<>>(0, ##__VA_ARGS__);\ } while(0) +#define hipLaunchKernelGGL(kernelName, numblocks, numthreads, memperblock, streamId, ...) \ +do {\ +kernelName<<>>(__VA_ARGS__);\ +} while(0) #define hipReadModeElementType cudaReadModeElementType diff --git a/include/hip/nvcc_detail/hip_runtime_api.h b/include/hip/nvcc_detail/hip_runtime_api.h index 7e881df3ab..0cc40f32af 100644 --- a/include/hip/nvcc_detail/hip_runtime_api.h +++ b/include/hip/nvcc_detail/hip_runtime_api.h @@ -948,4 +948,6 @@ inline static hipChannelFormatDesc hipCreateChannelDesc() { return cudaCreateChannelDesc(); } -#endif +#endif //__CUDACC__ + +#endif //HIP_INCLUDE_HIP_NVCC_DETAIL_HIP_RUNTIME_API_H diff --git a/samples/0_Intro/square/Makefile b/samples/0_Intro/square/Makefile index 1e8cdba080..aa48cc5864 100644 --- a/samples/0_Intro/square/Makefile +++ b/samples/0_Intro/square/Makefile @@ -15,5 +15,6 @@ square.hip.out: square.hipref.cpp + clean: rm -f *.o *.out diff --git a/samples/0_Intro/square/square.hipref.cpp b/samples/0_Intro/square/square.hipref.cpp index 963ab63260..e694bfb8a4 100644 --- a/samples/0_Intro/square/square.hipref.cpp +++ b/samples/0_Intro/square/square.hipref.cpp @@ -83,7 +83,7 @@ int main(int argc, char *argv[]) const unsigned threadsPerBlock = 256; printf ("info: launch 'vector_square' kernel\n"); - hipLaunchKernel(vector_square, dim3(blocks), dim3(threadsPerBlock), 0, nullptr, C_d, A_d, N); + hipLaunchKernel(vector_square, dim3(blocks), dim3(threadsPerBlock), 0, 0, C_d, A_d, N); printf ("info: copy Device2Host\n"); CHECK ( hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost)); From 662f4f81393280187ab8ca6abf16d169939c415e Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Wed, 26 Apr 2017 23:58:44 +0530 Subject: [PATCH 009/108] Merge branch 'amd-develop' into amd-master Change-Id: I51545bb66e9c9ff39bed86e2a8621e49a0e3f1c1 (cherry picked from commit cbb34eaf766e5e8374b1d514154be1d102b4115c) --- .vimrc | 1 - include/hip/hcc_detail/hip_runtime_api.h | 6 +- include/hip/nvcc_detail/hip_complex.h | 6 +- src/hip_hcc.cpp | 82 +++- src/hip_memory.cpp | 26 +- .../src/runtimeApi/memory/hipHostRegister.cpp | 143 ++++--- tests/src/runtimeApi/memory/hipMemcpy.cpp | 367 ++++++++++++++---- .../runtimeApi/memory/hipMemoryAllocate.cpp | 10 + 8 files changed, 500 insertions(+), 141 deletions(-) delete mode 100644 .vimrc diff --git a/.vimrc b/.vimrc deleted file mode 100644 index 019afa57e6..0000000000 --- a/.vimrc +++ /dev/null @@ -1 +0,0 @@ -:set makeprg=make\ -C\ build.hcc-LC.db diff --git a/include/hip/hcc_detail/hip_runtime_api.h b/include/hip/hcc_detail/hip_runtime_api.h index f9bfb5a310..7a99ff0810 100644 --- a/include/hip/hcc_detail/hip_runtime_api.h +++ b/include/hip/hcc_detail/hip_runtime_api.h @@ -853,7 +853,7 @@ hipError_t hipEventQuery(hipEvent_t event) ; * * @see hipGetDeviceCount, hipGetDevice, hipSetDevice, hipChooseDevice */ -hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void* ptr); +hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void* ptr); /** * @brief Allocate memory on the default accelerator @@ -863,7 +863,7 @@ hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void* ptr) * * If size is 0, no memory is allocated, *ptr returns nullptr, and hipSuccess is returned. * - * @return #hipSuccess + * @return #hipSuccess, #hipErrorMemoryAllocation, #hipErrorInvalidValue (bad context, null *ptr) * * @see hipMallocPitch, hipFree, hipMallocArray, hipFreeArray, hipMalloc3D, hipMalloc3DArray, hipHostFree, hipHostMalloc */ @@ -1922,7 +1922,7 @@ hipError_t hipModuleLoadData(hipModule_t *module, const void *image); * @param [in] blockDimZ Z grid dimension specified in work-items * @param [in] sharedMemBytes Amount of dynamic shared memory to allocate for this kernel. The kernel can access this with HIP_DYNAMIC_SHARED. * @param [in] stream Stream where the kernel should be dispatched. May be 0, in which case th default stream is used with associated synchronization rules. - * @param [in] kernelParams + * @param [in] kernelParams * @param [in] extra Pointer to kernel arguments. These are passed directly to the kernel and must be in the memory layout and alignment expected by the kernel. * * @returns hipSuccess, hipInvalidDevice, hipErrorNotInitialized, hipErrorInvalidValue diff --git a/include/hip/nvcc_detail/hip_complex.h b/include/hip/nvcc_detail/hip_complex.h index 84afb13e50..20cb24460c 100644 --- a/include/hip/nvcc_detail/hip_complex.h +++ b/include/hip/nvcc_detail/hip_complex.h @@ -64,7 +64,7 @@ __device__ __host__ static inline hipFloatComplex hipCdivf(hipFloatComplex p, hi } __device__ __host__ static inline float hipCabsf(hipFloatComplex z){ - return cuCabsf(p, q); + return cuCabsf(z); } typedef cuDoubleComplex hipDoubleComplex; @@ -85,7 +85,7 @@ __device__ __host__ static inline hipDoubleComplex hipConj(hipDoubleComplex z){ return cuConj(z); } -__device__ __host__ static inline hipDoubleComplex hipCsqabs(hipDoubleComplex z){ +__device__ __host__ static inline double hipCsqabs(hipDoubleComplex z){ return cuCabs(z) * cuCabs(z); } @@ -123,7 +123,7 @@ __device__ __host__ static inline hipComplex hipCfmaf(hipComplex p, hipComplex q return cuCfmaf(p, q, r); } -__device__ __host__ static inline hipDoubleComplex hipCfma(hipComplex p, hipComplex q, hipComplex r){ +__device__ __host__ static inline hipDoubleComplex hipCfma(hipDoubleComplex p, hipDoubleComplex q, hipDoubleComplex r){ return cuCfma(p, q, r); } diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 35a3e11e71..71d947488d 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -1765,20 +1765,24 @@ void ihipStream_t::resolveHcMemcpyDirection(unsigned hipMemKind, if (HIP_FORCE_P2P_HOST & 0x1) { *forceUnpinnedCopy = true; - tprintf (DB_COPY, "P2P. Copy engine (dev:%d agent=0x%lx) can see src and dst but HIP_FORCE_P2P_HOST=0, forcing copy through staging buffers.\n", - (*copyDevice)->getDeviceNum(), (*copyDevice)->getDevice()->_hsaAgent.handle); + tprintf (DB_COPY, "Copy engine (dev:%d agent=0x%lx) can see src and dst but HIP_FORCE_P2P_HOST=0, forcing copy through staging buffers.\n", + *copyDevice ? (*copyDevice)->getDeviceNum() : -1, + *copyDevice ? (*copyDevice)->getDevice()->_hsaAgent.handle : 0x0); } else { - tprintf (DB_COPY, "P2P. Copy engine (dev:%d agent=0x%lx) can see src and dst.\n", - (*copyDevice)->getDeviceNum(), (*copyDevice)->getDevice()->_hsaAgent.handle); + tprintf (DB_COPY, "Copy engine (dev:%d agent=0x%lx) can see src and dst.\n", + *copyDevice ? (*copyDevice)->getDeviceNum() : -1, + *copyDevice ? (*copyDevice)->getDevice()->_hsaAgent.handle : 0x0); } } else { *forceUnpinnedCopy = true; tprintf (DB_COPY, "P2P: Copy engine(dev:%d agent=0x%lx) cannot see both host and device pointers - forcing copy with unpinned engine.\n", - (*copyDevice)->getDeviceNum(), (*copyDevice)->getDevice()->_hsaAgent.handle); + *copyDevice ? (*copyDevice)->getDeviceNum() : -1, + *copyDevice ? (*copyDevice)->getDevice()->_hsaAgent.handle : 0x0); if (HIP_FAIL_SOC & 0x2) { fprintf (stderr, "HIP_FAIL_SOC: P2P: copy engine(dev:%d agent=0x%lx) cannot see both host and device pointers - forcing copy with unpinned engine.\n", - (*copyDevice)->getDeviceNum(), (*copyDevice)->getDevice()->_hsaAgent.handle); + *copyDevice ? (*copyDevice)->getDeviceNum() : -1, + *copyDevice ? (*copyDevice)->getDevice()->_hsaAgent.handle : 0x0); throw ihipException(hipErrorRuntimeOther); } } @@ -1794,6 +1798,62 @@ void printPointerInfo(unsigned dbFlag, const char *tag, const void *ptr, const h } +// the pointer-info as returned by HC refers to the allocation +// This routine modifies the pointer-info so it appears to refer to the specific ptr and sizeBytes. +// TODO -remove this when HCC uses HSA pointer info functions directly. +void tailorPtrInfo(hc::AmPointerInfo *ptrInfo, const void * ptr, size_t sizeBytes) +{ + const char *ptrc = static_cast (ptr); + if (ptrInfo->_sizeBytes == 0) { + // invalid ptrInfo, don't modify + return; + } else if (ptrInfo->_isInDeviceMem) { + assert (ptrInfo->_devicePointer != nullptr); + std::ptrdiff_t diff = ptrc - static_cast (ptrInfo->_devicePointer); + + //TODO : assert-> runtime assert that only appears in debug mode + assert (diff >= 0); + assert (diff <= ptrInfo->_sizeBytes); + + ptrInfo->_devicePointer = const_cast (ptr); + + if (ptrInfo->_hostPointer != nullptr) { + ptrInfo->_hostPointer = static_cast(ptrInfo->_hostPointer) + diff; + } + + } else { + + assert (ptrInfo->_hostPointer != nullptr); + std::ptrdiff_t diff = ptrc - static_cast (ptrInfo->_hostPointer); + + //TODO : assert-> runtime assert that only appears in debug mode + assert (diff >= 0); + assert (diff <= ptrInfo->_sizeBytes); + + ptrInfo->_hostPointer = const_cast(ptr); + + if (ptrInfo->_devicePointer != nullptr) { + ptrInfo->_devicePointer = static_cast(ptrInfo->_devicePointer) + diff; + } + } + + assert (sizeBytes <= ptrInfo->_sizeBytes); + ptrInfo->_sizeBytes = sizeBytes; +}; + + +bool getTailoredPtrInfo(hc::AmPointerInfo *ptrInfo, const void * ptr, size_t sizeBytes) +{ + bool tracked = (hc::am_memtracker_getinfo(ptrInfo, ptr) == AM_SUCCESS); + + if (tracked) { + tailorPtrInfo(ptrInfo, ptr, sizeBytes); + } + + return tracked; +}; + + // TODO : For registered and host memory, if the portable flag is set, we need to recognize that and perform appropriate copy operation. // What can happen now is that Portable memory is mapped into multiple devices but Peer access is not enabled. i // The peer detection logic doesn't see that the memory is already mapped and so tries to use an unpinned copy algorithm. If this is PinInPlace, then an error can occur. @@ -1812,8 +1872,8 @@ void ihipStream_t::locked_copySync(void* dst, const void* src, size_t sizeBytes, hc::accelerator acc; hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); - bool dstTracked = (hc::am_memtracker_getinfo(&dstPtrInfo, dst) == AM_SUCCESS); - bool srcTracked = (hc::am_memtracker_getinfo(&srcPtrInfo, src) == AM_SUCCESS); + bool dstTracked = getTailoredPtrInfo(&dstPtrInfo, dst, sizeBytes); + bool srcTracked = getTailoredPtrInfo(&srcPtrInfo, src, sizeBytes); // Some code in HCC and in printPointerInfo uses _sizeBytes==0 as an indication ptr is not valid, so check it here: @@ -1873,6 +1933,7 @@ void ihipStream_t::lockedSymbolCopySync(hc::accelerator &acc, void* dst, void* s void ihipStream_t::lockedSymbolCopyAsync(hc::accelerator &acc, void* dst, void* src, size_t sizeBytes, size_t offset, unsigned kind) { + // TODO - review - this looks broken , should not be adding pointers to tracker dynamically: if(kind == hipMemcpyHostToDevice) { hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); bool srcTracked = (hc::am_memtracker_getinfo(&srcPtrInfo, src) == AM_SUCCESS); @@ -1899,6 +1960,7 @@ void ihipStream_t::lockedSymbolCopyAsync(hc::accelerator &acc, void* dst, void* } } + void ihipStream_t::locked_copyAsync(void* dst, const void* src, size_t sizeBytes, unsigned kind) { @@ -1926,8 +1988,8 @@ void ihipStream_t::locked_copyAsync(void* dst, const void* src, size_t sizeBytes hc::accelerator acc; hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); - bool dstTracked = (hc::am_memtracker_getinfo(&dstPtrInfo, dst) == AM_SUCCESS); - bool srcTracked = (hc::am_memtracker_getinfo(&srcPtrInfo, src) == AM_SUCCESS); + bool dstTracked = getTailoredPtrInfo(&dstPtrInfo, dst, sizeBytes); + bool srcTracked = getTailoredPtrInfo(&srcPtrInfo, src, sizeBytes); hc::hcCommandKind hcCopyDir; diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index da5530349f..821f64bc76 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -133,7 +133,7 @@ void * allocAndSharePtr(const char *msg, size_t sizeBytes, ihipCtx_t *ctx, unsig //_appAllocationFlags : These are flags provided by the user when allocation is performed. They are returned to user in hipHostGetFlags and other APIs. // TODO - add more info here when available. // -hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void* ptr) +hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void* ptr) { HIP_INIT_API(attributes, ptr); @@ -149,10 +149,10 @@ hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void* ptr) attributes->devicePointer = amPointerInfo._devicePointer; attributes->isManaged = 0; if(attributes->memoryType == hipMemoryTypeHost){ - attributes->hostPointer = ptr; + attributes->hostPointer = (void*)ptr; } if(attributes->memoryType == hipMemoryTypeDevice){ - attributes->devicePointer = ptr; + attributes->devicePointer = (void*)ptr; } attributes->allocationFlags = amPointerInfo._appAllocationFlags; attributes->device = amPointerInfo._appId; @@ -207,22 +207,26 @@ hipError_t hipMalloc(void** ptr, size_t sizeBytes) HIP_INIT_API(ptr, sizeBytes); HIP_SET_DEVICE(); hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); // return NULL pointer when malloc size is 0 if (sizeBytes == 0) { *ptr = NULL; - return ihipLogStatus(hipSuccess); - } + hip_status = hipSuccess; - auto ctx = ihipGetTlsDefaultCtx(); + } else if ((ctx==nullptr) || (ptr == nullptr)) { + hip_status = hipErrorInvalidValue; - if (ctx) { + } else { auto device = ctx->getWriteableDevice(); *ptr = hip_internal::allocAndSharePtr("device_mem", sizeBytes, ctx, 0/*amFlags*/, 0/*hipFlags*/); - } else { - hip_status = hipErrorMemoryAllocation; - } + if(sizeBytes && (*ptr == NULL)){ + hip_status = hipErrorMemoryAllocation; + } + + } return ihipLogStatus(hip_status); @@ -1268,7 +1272,7 @@ hipError_t hipIpcOpenMemHandle(void** devPtr, hipIpcMemHandle_t handle, unsigned hsa_amd_ipc_memory_attach((hsa_amd_ipc_memory_t*)&(iHandle->ipc_handle), iHandle->psize, crit->peerCnt(), crit->peerAgents(), devPtr); if(hsa_status != HSA_STATUS_SUCCESS) hipStatus = hipErrorMapBufferObjectFailed; - } + } #else hipStatus = hipErrorRuntimeOther; #endif diff --git a/tests/src/runtimeApi/memory/hipHostRegister.cpp b/tests/src/runtimeApi/memory/hipHostRegister.cpp index 1a1319c500..8cf0979261 100644 --- a/tests/src/runtimeApi/memory/hipHostRegister.cpp +++ b/tests/src/runtimeApi/memory/hipHostRegister.cpp @@ -19,87 +19,142 @@ THE SOFTWARE. /* HIT_START * BUILD: %t %s ../../test_common.cpp - * RUN: %t + * RUN: %t --tests 0x1 + * RUN: %t --tests 0x2 + * RUN: %t --tests 0x4 * HIT_END */ +// TODO - bug if run both back-to-back, once fixed should just need one command line + #include"test_common.h" #include __global__ void Inc(hipLaunchParm lp, float *Ad){ -int tx = hipThreadIdx_x + hipBlockIdx_x * hipBlockDim_x; -Ad[tx] = Ad[tx] + float(1); + int tx = hipThreadIdx_x + hipBlockIdx_x * hipBlockDim_x; + Ad[tx] = Ad[tx] + float(1); } -int main(){ - float *A, **Ad; - int num_devices; - HIPCHECK(hipGetDeviceCount(&num_devices)); - Ad = new float*[num_devices]; - const size_t size = N * sizeof(float); - A = (float*)malloc(size); - HIPCHECK(hipHostRegister(A, size, 0)); + +template +void doMemCopy(size_t numElements, int offset, T *A, T *Bh, T *Bd, bool internalRegister) +{ + A = A + offset; + numElements -= offset; + + size_t sizeBytes = numElements * sizeof(T); + + if (internalRegister) { + HIPCHECK(hipHostRegister(A, sizeBytes, 0)); + } - for(int i=0;iOFFSETS_TO_TRY); + + if (p_tests & 0x2) { + for (size_t i=0; i +class DeviceMemory +{ +public: + DeviceMemory(size_t numElements); + ~DeviceMemory(); + + T *A_d() const { return _A_d + _offset; }; + T *B_d() const { return _B_d + _offset; }; + T *C_d() const { return _C_d + _offset; }; + T *C_dd() const { return _C_dd + _offset; }; + + size_t maxNumElements() const { return _maxNumElements; }; + + + void offset(int offset) { _offset = offset; }; + int offset() const { return _offset; }; + +private: + T * _A_d; + T* _B_d; + T* _C_d; + T* _C_dd; + + + size_t _maxNumElements; + int _offset; +}; + +template +DeviceMemory::DeviceMemory(size_t numElements) + : _maxNumElements(numElements), + _offset(0) +{ + T ** np = nullptr; + HipTest::initArrays (&_A_d, &_B_d, &_C_d, np, np, np, numElements, 0); + + + size_t sizeElements = numElements * sizeof(T); + + + HIPCHECK ( hipMalloc(&_C_dd, sizeElements) ); +} + + +template +DeviceMemory::~DeviceMemory () +{ + T * np = nullptr; + HipTest::freeArrays (_A_d, _B_d, _C_d, np, np, np, 0); + + HIPCHECK (hipFree(_C_dd)); + + _C_dd = NULL; +}; + + + +//------- +template +class HostMemory +{ +public: + HostMemory(size_t numElements, bool usePinnedHost); + void reset(size_t numElements, bool full=false) ; + ~HostMemory(); + + + T *A_h() const { return _A_h + _offset; }; + T *B_h() const { return _B_h + _offset; }; + T *C_h() const { return _C_h + _offset; }; + + + + size_t maxNumElements() const { return _maxNumElements; }; + + void offset(int offset) { _offset = offset; }; + int offset() const { return _offset; }; +public: + + // Host arrays, secondary copy + T * A_hh; + T* B_hh; + + bool _usePinnedHost; +private: + size_t _maxNumElements; + + int _offset; + + // Host arrays + T * _A_h; + T* _B_h; + T* _C_h; +}; + +template +HostMemory::HostMemory(size_t numElements, bool usePinnedHost) + : _maxNumElements(numElements), + _usePinnedHost(usePinnedHost), + _offset(0) +{ + T ** np = nullptr; + HipTest::initArrays (np, np, np, &_A_h, &_B_h, &_C_h, numElements, usePinnedHost); + + A_hh = NULL; + B_hh = NULL; + + + size_t sizeElements = numElements * sizeof(T); + + if (usePinnedHost) { + HIPCHECK ( hipHostMalloc((void**)&A_hh, sizeElements, hipHostMallocDefault) ); + HIPCHECK ( hipHostMalloc((void**)&B_hh, sizeElements, hipHostMallocDefault) ); + } else { + A_hh = (T*)malloc(sizeElements); + B_hh = (T*)malloc(sizeElements); + } + +} + + +template +void +HostMemory::reset(size_t numElements, bool full) +{ + // Initialize the host data: + for (size_t i=0; i +HostMemory::~HostMemory () +{ + HipTest::freeArraysForHost (_A_h, _B_h, _C_h, _usePinnedHost); + + if (_usePinnedHost) { + HIPCHECK (hipHostFree(A_hh)); + HIPCHECK (hipHostFree(B_hh)); + + } else { + free(A_hh); + free(B_hh); + } + T *A_hh = NULL; + T *B_hh = NULL; + +}; @@ -52,71 +210,57 @@ void printSep() // IN: useMemkindDefault : If true, use memkinddefault (runtime figures out direction). if false, use explicit memcpy direction. // template -void memcpytest2(size_t numElements, bool usePinnedHost, bool useHostToHost, bool useDeviceToDevice, bool useMemkindDefault) +void memcpytest2(DeviceMemory *dmem, HostMemory *hmem, size_t numElements, bool useHostToHost, bool useDeviceToDevice, bool useMemkindDefault) { size_t sizeElements = numElements * sizeof(T); - printf ("test: %s<%s> size=%lu (%6.2fMB) usePinnedHost:%d, useHostToHost:%d, useDeviceToDevice:%d, useMemkindDefault:%d\n", + printf ("test: %s<%s> size=%lu (%6.2fMB) usePinnedHost:%d, useHostToHost:%d, useDeviceToDevice:%d, useMemkindDefault:%d, offsets:dev:%+d host:+%d\n", __func__, TYPENAME(T), sizeElements, sizeElements/1024.0/1024.0, - usePinnedHost, useHostToHost, useDeviceToDevice, useMemkindDefault); + hmem->_usePinnedHost, useHostToHost, useDeviceToDevice, useMemkindDefault, + dmem->offset(), hmem->offset() + ); - T *A_d, *B_d, *C_d; - T *A_h, *B_h, *C_h; - - - HipTest::initArrays (&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, numElements, usePinnedHost); + hmem->reset(numElements); unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, numElements); - T *A_hh = NULL; - T *B_hh = NULL; - T *C_dd = NULL; + assert (numElements <= dmem->maxNumElements()); + assert (numElements <= hmem->maxNumElements()); if (useHostToHost) { - if (usePinnedHost) { - HIPCHECK ( hipHostMalloc((void**)&A_hh, sizeElements, hipHostMallocDefault) ); - HIPCHECK ( hipHostMalloc((void**)&B_hh, sizeElements, hipHostMallocDefault) ); - } else { - A_hh = (T*)malloc(sizeElements); - B_hh = (T*)malloc(sizeElements); - } - - // Do some extra host-to-host copies here to mix things up: - HIPCHECK ( hipMemcpy(A_hh, A_h, sizeElements, useMemkindDefault? hipMemcpyDefault : hipMemcpyHostToHost)); - HIPCHECK ( hipMemcpy(B_hh, B_h, sizeElements, useMemkindDefault? hipMemcpyDefault : hipMemcpyHostToHost)); + HIPCHECK ( hipMemcpy(hmem->A_hh, hmem->A_h(), sizeElements, useMemkindDefault? hipMemcpyDefault : hipMemcpyHostToHost)); + HIPCHECK ( hipMemcpy(hmem->B_hh, hmem->B_h(), sizeElements, useMemkindDefault? hipMemcpyDefault : hipMemcpyHostToHost)); - HIPCHECK ( hipMemcpy(A_d, A_hh, sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); - HIPCHECK ( hipMemcpy(B_d, B_hh, sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); + HIPCHECK ( hipMemcpy(dmem->A_d(), hmem->A_hh, sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); + HIPCHECK ( hipMemcpy(dmem->B_d(), hmem->B_hh, sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); } else { - HIPCHECK ( hipMemcpy(A_d, A_h, sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); - HIPCHECK ( hipMemcpy(B_d, B_h, sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); + HIPCHECK ( hipMemcpy(dmem->A_d(), hmem->A_h(), sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); + HIPCHECK ( hipMemcpy(dmem->B_d(), hmem->B_h(), sizeElements, useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice)); } - hipLaunchKernel(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock), 0, 0, A_d, B_d, C_d, numElements); + hipLaunchKernel(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock), 0, 0, dmem->A_d(), dmem->B_d(), dmem->C_d(), numElements); if (useDeviceToDevice) { - HIPCHECK ( hipMalloc(&C_dd, sizeElements) ); + // Do an extra device-to-device copy here to mix things up: + HIPCHECK ( hipMemcpy(dmem->C_dd(), dmem->C_d(), sizeElements, useMemkindDefault? hipMemcpyDefault : hipMemcpyDeviceToDevice)); - // Do an extra device-to-device copies here to mix things up: - HIPCHECK ( hipMemcpy(C_dd, C_d, sizeElements, useMemkindDefault? hipMemcpyDefault : hipMemcpyDeviceToDevice)); + //Destroy the original dmem->C_d(): + HIPCHECK ( hipMemset(dmem->C_d(), 0x5A, sizeElements)); - //Destroy the original C_d: - HIPCHECK ( hipMemset(C_d, 0x5A, sizeElements)); - - HIPCHECK ( hipMemcpy(C_h, C_dd, sizeElements, useMemkindDefault? hipMemcpyDefault:hipMemcpyDeviceToHost)); + HIPCHECK ( hipMemcpy(hmem->C_h(), dmem->C_dd(), sizeElements, useMemkindDefault? hipMemcpyDefault:hipMemcpyDeviceToHost)); } else { - HIPCHECK ( hipMemcpy(C_h, C_d, sizeElements, useMemkindDefault? hipMemcpyDefault:hipMemcpyDeviceToHost)); + HIPCHECK ( hipMemcpy(hmem->C_h(), dmem->C_d(), sizeElements, useMemkindDefault? hipMemcpyDefault:hipMemcpyDeviceToHost)); } HIPCHECK ( hipDeviceSynchronize() ); - HipTest::checkVectorADD(A_h, B_h, C_h, numElements); + HipTest::checkVectorADD(hmem->A_h(), hmem->B_h(), hmem->C_h(), numElements); + - HipTest::freeArrays (A_d, B_d, C_d, A_h, B_h, C_h, usePinnedHost); printf (" %s success\n", __func__); } @@ -129,11 +273,15 @@ void memcpytest2_for_type(size_t numElements) { printSep(); + DeviceMemory memD(numElements); + HostMemory memU(numElements, 0/*usePinnedHost*/); + HostMemory memP(numElements, 1/*usePinnedHost*/); + for (int usePinnedHost =0; usePinnedHost<=1; usePinnedHost++) { for (int useHostToHost =0; useHostToHost<=1; useHostToHost++) { // TODO for (int useDeviceToDevice =0; useDeviceToDevice<=1; useDeviceToDevice++) { for (int useMemkindDefault =0; useMemkindDefault<=1; useMemkindDefault++) { - memcpytest2(numElements, usePinnedHost, useHostToHost, useDeviceToDevice, useMemkindDefault); + memcpytest2(&memD, usePinnedHost ? &memP : &memU, numElements, useHostToHost, useDeviceToDevice, useMemkindDefault); } } } @@ -144,7 +292,7 @@ void memcpytest2_for_type(size_t numElements) //--- //Try many different sizes to memory copy. template -void memcpytest2_sizes(size_t maxElem=0, size_t offset=0) +void memcpytest2_sizes(size_t maxElem=0) { printSep(); printf ("test: %s<%s>\n", __func__, TYPENAME(T)); @@ -159,14 +307,68 @@ void memcpytest2_sizes(size_t maxElem=0, size_t offset=0) maxElem = free/sizeof(T)/5; } - printf (" device#%d: hipMemGetInfo: free=%zu (%4.2fMB) total=%zu (%4.2fMB) maxSize=%6.1fMB offset=%lu\n", - deviceId, free, (float)(free/1024.0/1024.0), total, (float)(total/1024.0/1024.0), maxElem*sizeof(T)/1024.0/1024.0, offset); + printf (" device#%d: hipMemGetInfo: free=%zu (%4.2fMB) total=%zu (%4.2fMB) maxSize=%6.1fMB\n", + deviceId, free, (float)(free/1024.0/1024.0), total, (float)(total/1024.0/1024.0), maxElem*sizeof(T)/1024.0/1024.0); + HIPCHECK ( hipDeviceReset() ); + DeviceMemory memD(maxElem); + HostMemory memU(maxElem, 0/*usePinnedHost*/); + HostMemory memP(maxElem, 1/*usePinnedHost*/); - for (size_t elem=64; elem+offset<=maxElem; elem*=2) { - HIPCHECK ( hipDeviceReset() ); - memcpytest2(elem+offset, 0, 1, 1, 0); // unpinned host - HIPCHECK ( hipDeviceReset() ); - memcpytest2(elem+offset, 1, 1, 1, 0); // pinned host + for (size_t elem=1; elem<=maxElem; elem*=2) { + memcpytest2(&memD, &memU, elem, 1, 1, 0); // unpinned host + memcpytest2(&memD, &memP, elem, 1, 1, 0); // pinned host + } +} + + +//--- +//Try many different sizes to memory copy. +template +void memcpytest2_offsets(size_t maxElem, bool devOffsets, bool hostOffsets) +{ + printSep(); + printf ("test: %s<%s>\n", __func__, TYPENAME(T)); + + int deviceId; + HIPCHECK(hipGetDevice(&deviceId)); + + size_t free, total; + HIPCHECK(hipMemGetInfo(&free, &total)); + + + printf (" device#%d: hipMemGetInfo: free=%zu (%4.2fMB) total=%zu (%4.2fMB) maxSize=%6.1fMB\n", + deviceId, free, (float)(free/1024.0/1024.0), total, (float)(total/1024.0/1024.0), maxElem*sizeof(T)/1024.0/1024.0); + HIPCHECK ( hipDeviceReset() ); + DeviceMemory memD(maxElem); + HostMemory memU(maxElem, 0/*usePinnedHost*/); + HostMemory memP(maxElem, 1/*usePinnedHost*/); + + size_t elem = maxElem / 2; + + for (int offset=0; offset < 512; offset++) { + assert (elem + offset < maxElem); + if (devOffsets) { + memD.offset(offset); + } + if (hostOffsets) { + memU.offset(offset); + memP.offset(offset); + } + memcpytest2(&memD, &memU, elem, 1, 1, 0); // unpinned host + memcpytest2(&memD, &memP, elem, 1, 1, 0); // pinned host + } + + for (int offset=512; offset < elem; offset*=2) { + assert (elem + offset < maxElem); + if (devOffsets) { + memD.offset(offset); + } + if (hostOffsets) { + memU.offset(offset); + memP.offset(offset); + } + memcpytest2(&memD, &memU, elem, 1, 1, 0); // unpinned host + memcpytest2(&memD, &memP, elem, 1, 1, 0); // pinned host } } @@ -178,13 +380,17 @@ void multiThread_1(bool serialize, bool usePinnedHost) { printSep(); printf ("test: %s<%s> serialize=%d usePinnedHost=%d\n", __func__, TYPENAME(T), serialize, usePinnedHost); - std::thread t1 (memcpytest2,N, usePinnedHost,0,0,0); + DeviceMemory memD(N); + HostMemory mem1(N, usePinnedHost); + HostMemory mem2(N, usePinnedHost); + + std::thread t1 (memcpytest2, &memD, &mem1, N, 0,0,0); if (serialize) { t1.join(); } - std::thread t2 (memcpytest2,N, usePinnedHost,0,0,0); + std::thread t2 (memcpytest2,&memD, &mem2, N, 0,0,0); if (serialize) { t2.join(); } @@ -218,37 +424,39 @@ int main(int argc, char *argv[]) if (p_tests & 0x2) { - // Some tests around the 64MB boundary which have historically shown issues: - printf ("\n\n=== tests&0x2 (64MB boundary)\n"); -#if 0 + // Some tests around the 64KB boundary which have historically shown issues: + printf ("\n\n=== tests&0x2 (64KB boundary)\n"); + size_t maxElem = 32*1024*1024; + DeviceMemory memD(maxElem); + HostMemory memU(maxElem, 0/*usePinnedHost*/); + HostMemory memP(maxElem, 0/*usePinnedHost*/); // These all pass: - memcpytest2(15*1024*1024, 1, 0, 0, 0); - memcpytest2(16*1024*1024, 1, 0, 0, 0); - memcpytest2(16*1024*1024+16*1024, 1, 0, 0, 0); -#endif + memcpytest2(&memD, &memP, 15*1024*1024, 0, 0, 0); + memcpytest2(&memD, &memP, 16*1024*1024, 0, 0, 0); + memcpytest2(&memD, &memP, 16*1024*1024+16*1024, 0, 0, 0); + // Just over 64MB: - memcpytest2(16*1024*1024+512*1024, 1, 0, 0, 0); - memcpytest2(17*1024*1024+1024, 1, 0, 0, 0); - memcpytest2(32*1024*1024, 1, 0, 0, 0); - memcpytest2(32*1024*1024, 0, 0, 0, 0); - memcpytest2(32*1024*1024, 1, 1, 1, 0); - memcpytest2(32*1024*1024, 1, 1, 1, 0); + memcpytest2(&memD, &memP, 16*1024*1024+512*1024, 0, 0, 0); + memcpytest2(&memD, &memP, 17*1024*1024+1024, 0, 0, 0); + memcpytest2(&memD, &memP, 32*1024*1024, 0, 0, 0); + memcpytest2(&memD, &memU, 32*1024*1024, 0, 0, 0); + memcpytest2(&memD, &memP, 32*1024*1024, 1, 1, 0); + memcpytest2(&memD, &memP, 32*1024*1024, 1, 1, 0); + + } + if (p_tests & 0x4) { - printf ("\n\n=== tests&4 (test sizes and offsets)\n"); + printf ("\n\n=== tests&4 (test sizes)\n"); HIPCHECK ( hipDeviceReset() ); + memcpytest2_sizes(0); printSep(); - memcpytest2_sizes(0,0); - printSep(); - memcpytest2_sizes(0,64); - printSep(); - memcpytest2_sizes(1024*1024, 13); - printSep(); - memcpytest2_sizes(1024*1024, 50); } + + if (p_tests & 0x8) { printf ("\n\n=== tests&8\n"); HIPCHECK ( hipDeviceReset() ); @@ -270,6 +478,27 @@ int main(int argc, char *argv[]) } + if (p_tests & 0x10) { + printf ("\n\n=== tests&0x10 (test device offsets)\n"); + HIPCHECK ( hipDeviceReset() ); + size_t maxSize = 256*1024; + memcpytest2_offsets (maxSize, true, false); + memcpytest2_offsets (maxSize, true, false); + memcpytest2_offsets(maxSize, true, false); + } + + + if (p_tests & 0x20) { + printf ("\n\n=== tests&0x10 (test device offsets)\n"); + HIPCHECK ( hipDeviceReset() ); + size_t maxSize = 256*1024; + memcpytest2_offsets (maxSize, false, true); + memcpytest2_offsets (maxSize, false, true); + memcpytest2_offsets(maxSize, false, true); + } + + + passed(); } diff --git a/tests/src/runtimeApi/memory/hipMemoryAllocate.cpp b/tests/src/runtimeApi/memory/hipMemoryAllocate.cpp index 1f7599491a..0a256d6362 100644 --- a/tests/src/runtimeApi/memory/hipMemoryAllocate.cpp +++ b/tests/src/runtimeApi/memory/hipMemoryAllocate.cpp @@ -56,5 +56,15 @@ int main(){ HIPCHECK_API(hipFree(NULL) , hipSuccess); HIPCHECK_API(hipHostFree(NULL) , hipSuccess); + + { + // Some negative testing - request a too-big allocation and verify it fails: + // Someday when we support virtual memory may need to refactor these: + size_t tooBig = 128LL*1024*1024*1024*1024; // 128 TB; + void *p; + HIPCHECK_API ( hipMalloc(&p, tooBig), hipErrorMemoryAllocation ); + HIPCHECK_API ( hipHostMalloc(&p, tooBig), hipErrorMemoryAllocation ); + } + passed(); } From f5b335b147c3fc3cdc337fd6b7c632cd7b19e219 Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Fri, 16 Jun 2017 16:33:28 -0500 Subject: [PATCH 010/108] Adding initial set Jenkinsfile and dockerfiles For continuous integration enablement --- Jenkinsfile | 319 +++++++++++++++++++++++++++ docker/dockerfile-build-ubuntu-16.04 | 26 +++ docker/dockerfile-hip-ubuntu-16.04 | 21 ++ 3 files changed, 366 insertions(+) create mode 100644 Jenkinsfile create mode 100644 docker/dockerfile-build-ubuntu-16.04 create mode 100644 docker/dockerfile-hip-ubuntu-16.04 diff --git a/Jenkinsfile b/Jenkinsfile new file mode 100644 index 0000000000..16354285ec --- /dev/null +++ b/Jenkinsfile @@ -0,0 +1,319 @@ +#!/usr/bin/env groovy + +// Generated from snippet generator 'properties; set job properties' +properties([buildDiscarder(logRotator( + artifactDaysToKeepStr: '', + artifactNumToKeepStr: '', + daysToKeepStr: '', + numToKeepStr: '10')), + disableConcurrentBuilds(), + [$class: 'CopyArtifactPermissionProperty', projectNames: '*'] + ]) + +//////////////////////////////////////////////////////////////////////// +// -- AUXILLARY HELPER FUNCTIONS + +//////////////////////////////////////////////////////////////////////// +// Construct the relative path of the build directory +String build_directory_rel( String build_config ) +{ + if( build_config.equalsIgnoreCase( 'release' ) ) + { + return "build/release" + } + else + { + return "build/debug" + } +} + +//////////////////////////////////////////////////////////////////////// +// -- BUILD RELATED FUNCTIONS + +//////////////////////////////////////////////////////////////////////// +// Checkout source code, source dependencies and update version number numbers +// Returns a relative path to the directory where the source exists in the workspace +String checkout_and_version( String platform ) +{ + String source_dir_rel = "src" + String source_hip_rel = "${source_dir_rel}/hip" + + stage("${platform} clone") + { + dir( "${source_hip_rel}" ) + { + // checkout hip + checkout([ + $class: 'GitSCM', + branches: scm.branches, + doGenerateSubmoduleConfigurations: scm.doGenerateSubmoduleConfigurations, + extensions: scm.extensions + [[$class: 'CleanCheckout']], + userRemoteConfigs: scm.userRemoteConfigs + ]) + } + } + + return source_hip_rel +} + + +//////////////////////////////////////////////////////////////////////// +// This creates the docker image that we use to build the project in +// The docker images contains all dependencies, including OS platform, to build +def docker_build_image( String platform, String source_hip_rel, String from_image ) +{ + String project = "hip" + String build_type_name = "build-ubuntu-16.04" + String dockerfile_name = "dockerfile-${build_type_name}" + String build_image_name = "${build_type_name}" + def build_image = null + + stage("${platform} build image") + { + dir("${source_hip_rel}") + { + def user_uid = sh( script: 'id -u', returnStdout: true ).trim() + + // Docker 17.05 introduced the ability to use ARG values in FROM statements + // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 + //build_image = docker.build( "${project}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) + + // JENKINS-44836 workaround + sh "docker build -t ${project}/${build_image_name}:latest --pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." + build_image = docker.image( "${project}/${build_image_name}:latest" ) + } + } + + return build_image +} + +//////////////////////////////////////////////////////////////////////// +// This encapsulates the cmake configure, build and package commands +// Leverages docker containers to encapsulate the build in a fixed environment +def docker_build_inside_image( def build_image, String inside_args, String platform, String optional_configure, String build_config, String source_hip_rel, String build_dir_rel ) +{ + String source_hip_abs = pwd() + "/" + source_hip_rel + + build_image.inside( inside_args ) + { + stage("${platform} make ${build_config}") + { + // The rm command needs to run as sudo because the test steps below create files owned by root + sh """#!/usr/bin/env bash + set -x + sudo rm -rf ${build_dir_rel} + mkdir -p ${build_dir_rel} + cd ${build_dir_rel} + cmake -DCMAKE_BUILD_TYPE=${build_config} -DCMAKE_INSTALL_PREFIX=staging ${optional_configure} ${source_hip_abs} + make -j\$(nproc) + """ + } + + // Cap the maximum amount of testing, in case of hangs + timeout(time: 1, unit: 'HOURS') + { + stage("${platform} unit testing") + { + sh """#!/usr/bin/env bash + set -x + cd ${build_dir_rel} + make install -j\$(nproc) + make build_tests -j\$(nproc) + make test + """ + // If unit tests output a junit or xunit file in the future, jenkins can parse that file + // to display test results on the dashboard + // junit "${build_dir_rel}/*.xml" + } + } + + // Only create packages from hcc based builds + if( platform.toLowerCase( ).startsWith( 'hcc-' ) ) + { + stage("${platform} packaging") + { + sh """#!/usr/bin/env bash + set -x + cd ${build_dir_rel} + make package + """ + + // No matter the base platform, all packages have the same name + // Only upload 1 set of packages, so we don't have a race condition uploading packages + // I arbitrarily pick hcc-1.6 as the most stable + if( platform.toLowerCase( ).startsWith( 'hcc-1.6' ) ) + { + archiveArtifacts artifacts: "${build_dir_rel}/*.deb", fingerprint: true + archiveArtifacts artifacts: "${build_dir_rel}/*.rpm", fingerprint: true + } + } + } + } + + return void +} + +//////////////////////////////////////////////////////////////////////// +// This builds a fresh docker image FROM a clean base image, with no build dependencies included +// Uploads the new docker image to internal artifactory +def docker_upload_artifactory( String hcc_ver, String from_image, String source_hip_rel, String build_dir_rel ) +{ + def hip_install_image = null + String image_name = "hip-${hcc_ver}-ubuntu-16.04" + String artifactory_org = env.JOB_NAME.toLowerCase( ) + + stage( 'artifactory' ) + { + println "artifactory_org: ${artifactory_org}" + + // We copy the docker files into the bin directory where the .deb lives so that it's a clean build everytime + sh "cp -r ${source_hip_rel}/docker/* ${build_dir_rel}" + + // Docker 17.05 introduced the ability to use ARG values in FROM statements + // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 + // hip_install_image = docker.build( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}", "--pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" ) + + // JENKINS-44836 workaround + sh "docker build -t ${artifactory_org}/${image_name}:${env.BUILD_NUMBER} --pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" + hip_install_image = docker.image( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}" ) + + // The connection to artifactory can fail sometimes, but this should not be treated as a build fail + try + { + // Don't push pull requests to artifactory, these tend to accumulate over time + if( env.BRANCH_NAME.toLowerCase( ).startsWith( 'pr-' ) ) + { + println 'Pull Request (PR-xxx) detected; NOT pushing to artifactory' + } + else + { + docker.withRegistry('http://compute-artifactory:5001', 'artifactory-cred' ) + { + hip_install_image.push( "${env.BUILD_NUMBER}" ) + hip_install_image.push( 'latest' ) + } + } + } + catch( err ) + { + currentBuild.result = 'SUCCESS' + } + + // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" + } +} + +//////////////////////////////////////////////////////////////////////// +// -- MAIN +// Following this line is the start of MAIN of this Jenkinsfile +String build_config = 'Release' + +parallel hcc_ctu: +{ + node('docker && rocm && gfx803') + { + String hcc_ver = 'hcc-ctu' + String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/clang_tot_upgrade/hcc-lc-ubuntu-16.04:latest' + String inside_args = '--device=/dev/kfd' + + // Checkout source code, dependencies and version files + String source_hip_rel = checkout_and_version( hcc_ver ) + + // Create/reuse a docker image that represents the hip build environment + def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) + + // Print system information for the log + hip_build_image.inside( inside_args ) + { + sh """#!/usr/bin/env bash + set -x + /opt/rocm/bin/rocm_agent_enumerator -t ALL + /opt/rocm/bin/hcc --version + """ + } + + // Conctruct a binary directory path based on build config + String build_hip_rel = build_directory_rel( build_config ); + + // Build hip inside of the build environment + docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) + + // After a successful build, upload a docker image of the results + docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + } +}, +hcc_1_6: +{ + node('docker && rocm && gfx803') + { + String hcc_ver = 'hcc-1.6' + String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/roc-1.6.x/hcc-lc-ubuntu-16.04:latest' + String inside_args = '--device=/dev/kfd' + + // Checkout source code, dependencies and version files + String source_hip_rel = checkout_and_version( hcc_ver ) + + // Create/reuse a docker image that represents the hip build environment + def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) + + // Print system information for the log + hip_build_image.inside( inside_args ) + { + sh """#!/usr/bin/env bash + set -x + /opt/rocm/bin/rocm_agent_enumerator -t ALL + /opt/rocm/bin/hcc --version + """ + } + + // Conctruct a binary directory path based on build config + String build_hip_rel = build_directory_rel( build_config ); + + // Build hip inside of the build environment + docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) + + // After a successful build, upload a docker image of the results + docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + } +}, +nvcc: +{ + node('docker && cuda') + { + //////////////////////////////////////////////////////////////////////// + // Block of string constants customizing behavior for cuda + String nvcc_ver = 'nvcc-8.0' + String from_image = 'nvidia/cuda:8.0-devel' + + // This unfortunately hardcodes the driver version nvidia_driver_375.66 in the volume mount. Research if a way + // exists to get volume driver to customize the volume names to leave out driver version + String inside_args = '''--device=/dev/nvidiactl --device=/dev/nvidia0 --device=/dev/nvidia-uvm --device=/dev/nvidia-uvm-tools + --volume-driver=nvidia-docker --volume=nvidia_driver_375.66:/usr/local/nvidia:ro'''; + + // Checkout source code, dependencies and version files + String source_hip_rel = checkout_and_version( nvcc_ver ) + + // We pull public nvidia images + def hip_build_image = docker_build_image( nvcc_ver, source_hip_rel, from_image ) + + // Print system information for the log + hip_build_image.inside( inside_args ) + { + sh """#!/usr/bin/env bash + set -x + nvidia-smi + nvcc --version + """ + } + + // Conctruct a binary directory path based on build config + String build_hip_rel = build_directory_rel( build_config ); + + // Build hip inside of the build environment + docker_build_inside_image( hip_build_image, inside_args, nvcc_ver, "-DHIP_NVCC_FLAGS=--Wno-deprecated-gpu-targets", build_config, source_hip_rel, build_hip_rel ) + + // Not pushing an Nvidia based HiP to artifactory at this time + } +} \ No newline at end of file diff --git a/docker/dockerfile-build-ubuntu-16.04 b/docker/dockerfile-build-ubuntu-16.04 new file mode 100644 index 0000000000..0c2e4fa26d --- /dev/null +++ b/docker/dockerfile-build-ubuntu-16.04 @@ -0,0 +1,26 @@ +# Parameters related to building hip +ARG base_image + +FROM ${base_image} +MAINTAINER Kent Knox + +ARG user_uid + +# Install Packages +RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \ + sudo \ + build-essential \ + cmake \ + git \ + libelf-dev \ + rpm \ + && \ + apt-get clean && \ + rm -rf /var/lib/apt/lists/* + +# docker pipeline runs containers with uid 1000 +# create a jenkins user with this specific uid so it can use sudo priviledges +# Grant any member of sudo group password-less sudo privileges +RUN useradd --create-home -u ${user_uid} -G sudo --shell /bin/bash jenkins && \ + mkdir -p /etc/sudoers.d/ && \ + echo '%sudo ALL=(ALL) NOPASSWD:ALL' | tee /etc/sudoers.d/sudo-nopasswd diff --git a/docker/dockerfile-hip-ubuntu-16.04 b/docker/dockerfile-hip-ubuntu-16.04 new file mode 100644 index 0000000000..cecf5911b8 --- /dev/null +++ b/docker/dockerfile-hip-ubuntu-16.04 @@ -0,0 +1,21 @@ +# Parameters related to building hip +ARG base_image + +FROM ${base_image} +MAINTAINER Kent Knox + +# Copy the debian package of hip into the container from host +COPY *.deb /tmp/ + +# Install the debian package +RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends -y curl \ + && curl -sL http://packages.amd.com/rocm/apt/debian/rocm.gpg.key | apt-key add - \ + && echo deb [arch=amd64] http://packages.amd.com/rocm/apt/debian/ xenial main | tee /etc/apt/sources.list.d/rocm.list \ + && apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends --allow-unauthenticated -y \ + /tmp/hip_base-*.deb \ + /tmp/hip_hcc-*.deb \ + /tmp/hip_doc-*.deb \ + /tmp/hip_samples-* \ + && rm -f /tmp/*.deb \ + && apt-get clean \ + && rm -rf /var/lib/apt/lists/* \ No newline at end of file From f4ea3b76c84ac059a62476797ab3afbdb1a63c2f Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Wed, 28 Jun 2017 23:48:27 +0530 Subject: [PATCH 011/108] Fixed hipDeviceGetPCIBusId for HIP/HCC Change-Id: I3688fa2476e1baada2d3c5fc3735cec3f15a1e21 --- include/hip/hip_runtime_api.h | 1 + src/hip_device.cpp | 13 ++++--------- src/hip_hcc.cpp | 2 +- 3 files changed, 6 insertions(+), 10 deletions(-) diff --git a/include/hip/hip_runtime_api.h b/include/hip/hip_runtime_api.h index dc163d5c25..0cdace0e99 100644 --- a/include/hip/hip_runtime_api.h +++ b/include/hip/hip_runtime_api.h @@ -102,6 +102,7 @@ typedef struct hipDeviceProp_t { int clockInstructionRate; ///< Frequency in khz of the timer used by the device-side "clock*" instructions. New for HIP. hipDeviceArch_t arch; ///< Architectural feature flags. New for HIP. int concurrentKernels; ///< Device can possibly execute multiple kernels concurrently. + int pciDomainID; ///< PCI Domain ID int pciBusID; ///< PCI Bus ID. int pciDeviceID; ///< PCI Device ID. size_t maxSharedMemoryPerMultiProcessor; ///< Maximum Shared Memory Per Multiprocessor. diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 2bb9970d35..62518a1ba7 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -376,15 +376,10 @@ hipError_t hipDeviceGetPCIBusId (char *pciBusId,int len, int device) e = hipErrorInvalidDevice; } else { if((pciBusId != nullptr) && (len > 0)) { - int tempPciBusId = 0; - e = ihipDeviceGetAttribute( &tempPciBusId, hipDeviceAttributePciBusId, device); - if( e == hipSuccess) { - std::string tempPciStr = std::to_string(tempPciBusId); - if( len < tempPciStr.length()){ - e = hipErrorInvalidValue; - } else { - memcpy( pciBusId , tempPciStr.c_str() , tempPciStr.length() ); - } + auto deviceHandle = ihipGetDevice(device); + int retVal = snprintf(pciBusId,len, "%04x:%02x:%02x.0",deviceHandle->_props.pciDomainID,deviceHandle->_props.pciBusID,deviceHandle->_props.pciDeviceID); + if( retVal > 0 && retVal < len) { + e = hipSuccess; } } } diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index be591f2f04..f7f617615b 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -799,7 +799,7 @@ hipError_t ihipDevice_t::initProperties(hipDeviceProp_t* prop) DeviceErrorCheck(err); // BDFID is 16bit uint: [8bit - BusID | 5bit - Device ID | 3bit - Function/DomainID] - // prop->pciDomainID = bdf_id & 0x7; + prop->pciDomainID = bdf_id & 0x7; prop->pciDeviceID = (bdf_id>>3) & 0x1F; prop->pciBusID = (bdf_id>>8) & 0xFF; From d6db525f51082100afc61235f864f888223d99c3 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Thu, 29 Jun 2017 00:13:02 +0530 Subject: [PATCH 012/108] Fixed hipDeviceGetPCIBusId for HIP/NVCC Change-Id: I662efa148257a710f09002850b41d57bef00dff3 --- include/hip/nvcc_detail/hip_runtime_api.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/hip/nvcc_detail/hip_runtime_api.h b/include/hip/nvcc_detail/hip_runtime_api.h index b1011aac6c..ae4d99d787 100644 --- a/include/hip/nvcc_detail/hip_runtime_api.h +++ b/include/hip/nvcc_detail/hip_runtime_api.h @@ -873,7 +873,7 @@ inline static hipError_t hipDeviceGetName(char *name,int len,hipDevice_t device) inline static hipError_t hipDeviceGetPCIBusId(char* pciBusId,int len,hipDevice_t device) { - return hipCUResultTohipError(cuDeviceGetPCIBusId(pciBusId,len,device)); + return hipCUDAErrorTohipError(cudaDeviceGetPCIBusId(pciBusId,len,device)); } inline static hipError_t hipDeviceGetByPCIBusId(int* device, const int *pciBusId) From 3e610ab0fe7857ef073afb4b453c2f7606c819b1 Mon Sep 17 00:00:00 2001 From: Aditya Atluri Date: Thu, 29 Jun 2017 12:01:40 -0500 Subject: [PATCH 013/108] automate gcnarch detection Change-Id: Ibbad22db136f7f5e2be84c82e9169298a144cc77 --- samples/1_Utils/hipInfo/hipInfo.cpp | 1 + src/hip_hcc.cpp | 16 +--------------- 2 files changed, 2 insertions(+), 15 deletions(-) diff --git a/samples/1_Utils/hipInfo/hipInfo.cpp b/samples/1_Utils/hipInfo/hipInfo.cpp index cf4660eae7..0401745efd 100644 --- a/samples/1_Utils/hipInfo/hipInfo.cpp +++ b/samples/1_Utils/hipInfo/hipInfo.cpp @@ -129,6 +129,7 @@ void printDeviceProp (int deviceId) cout << setw(w1) << "arch.hasSurfaceFuncs: " << props.arch.hasSurfaceFuncs << endl; cout << setw(w1) << "arch.has3dGrid: " << props.arch.has3dGrid << endl; cout << setw(w1) << "arch.hasDynamicParallelism: " << props.arch.hasDynamicParallelism << endl; + cout << setw(w1) << "gcnArch: " << props.gcnArch << endl; int deviceCnt; hipGetDeviceCount(&deviceCnt); diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index f7f617615b..9c54c8917f 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -737,21 +737,7 @@ hipError_t ihipDevice_t::initProperties(hipDeviceProp_t* prop) char archName[256]; err = hsa_agent_get_info(_hsaAgent, HSA_AGENT_INFO_NAME, &archName); - if(strcmp(archName,"gfx701")==0){ - prop->gcnArch = 701; - } - if(strcmp(archName,"gfx801")==0){ - prop->gcnArch = 801; - } - if(strcmp(archName,"gfx802")==0){ - prop->gcnArch = 802; - } - if(strcmp(archName,"gfx803")==0){ - prop->gcnArch = 803; - } - if(strcmp(archName,"gfx900")==0){ - prop->gcnArch = 900; - } + prop->gcnArch = atoi(archName+3); DeviceErrorCheck(err); From 65cb024b61ab43a3ac75a3922ca8765e80e32bcb Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Fri, 30 Jun 2017 10:11:41 +0530 Subject: [PATCH 014/108] Fixed hipDeviceGetByPCIBusId Change-Id: Ia36bb9425671ef7659541c9aeedae4098456a31b --- include/hip/nvcc_detail/hip_runtime_api.h | 4 ++-- src/hip_device.cpp | 28 +++++++++++++++-------- 2 files changed, 20 insertions(+), 12 deletions(-) diff --git a/include/hip/nvcc_detail/hip_runtime_api.h b/include/hip/nvcc_detail/hip_runtime_api.h index ae4d99d787..502d205ac4 100644 --- a/include/hip/nvcc_detail/hip_runtime_api.h +++ b/include/hip/nvcc_detail/hip_runtime_api.h @@ -876,9 +876,9 @@ inline static hipError_t hipDeviceGetPCIBusId(char* pciBusId,int len,hipDevice_t return hipCUDAErrorTohipError(cudaDeviceGetPCIBusId(pciBusId,len,device)); } -inline static hipError_t hipDeviceGetByPCIBusId(int* device, const int *pciBusId) +inline static hipError_t hipDeviceGetByPCIBusId(int* device, const char *pciBusId) { - return hipCUDAErrorTohipError(cudaDeviceGetByPCIBusId(device,(char*)pciBusId)); + return hipCUDAErrorTohipError(cudaDeviceGetByPCIBusId(device, pciBusId)); } inline static hipError_t hipDeviceGetLimit(size_t *pValue, hipLimit_t limit) diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 62518a1ba7..66c6c7db5f 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -399,17 +399,25 @@ hipError_t hipDeviceGetByPCIBusId (int* device, const int* pciBusId ) { HIP_INIT_API(device,pciBusId); hipDeviceProp_t tempProp; - int deviceCount; + int deviceCount = 0 ; hipError_t e = hipErrorInvalidValue; - ihipGetDeviceCount( &deviceCount ); - *device = 0; - for (int i = 0; i< deviceCount; i++) { - ihipGetDeviceProperties( &tempProp, i ); - if(tempProp.pciBusID == *pciBusId) { - *device =i; - e = hipSuccess; - break; - } + if((device != nullptr) && (pciBusId != nullptr)) { + int pciBusID = -1; + int pciDeviceID = -1; + int pciDomainID = -1; + int len = 0; + len = sscanf (pciBusId,"%04x:%02x:%02x",&pciDomainID,&pciBusID,&pciDeviceID); + if(len == 3) { + ihipGetDeviceCount( &deviceCount ); + for (int i = 0; i< deviceCount; i++) { + ihipGetDeviceProperties( &tempProp, i ); + if(tempProp.pciBusID == pciBusID) { + *device = i; + e = hipSuccess; + break; + } + } + } } return ihipLogStatus(e); } From 9c229f75301dbb2f8a46df806b6313c00adcbdab Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Fri, 30 Jun 2017 08:23:25 -0500 Subject: [PATCH 015/108] Refactor hip_hcc. Don't require setting __HIP_PLATFORM_HCC__ ; Instead check if compiler is HCC. The functions here use HCC-specific functionality so only work with HCC. ANd they work regardless of the __HIP_PLATFORM_HCC__ setting so don't require that. Also remove the "hcc_detail" extra hierarchy level for hip_hcc.h as it is not necessary. Remove hcc_detail/hip_hcc.h. --- include/hip/hcc_detail/grid_launch_GGL.hpp | 2 +- include/hip/hcc_detail/hip_hcc.h | 103 --------------------- include/hip/hip_hcc.h | 83 ++++++++++++++++- 3 files changed, 80 insertions(+), 108 deletions(-) delete mode 100644 include/hip/hcc_detail/hip_hcc.h diff --git a/include/hip/hcc_detail/grid_launch_GGL.hpp b/include/hip/hcc_detail/grid_launch_GGL.hpp index eac48b595e..4c632f9d68 100644 --- a/include/hip/hcc_detail/grid_launch_GGL.hpp +++ b/include/hip/hcc_detail/grid_launch_GGL.hpp @@ -28,7 +28,7 @@ THE SOFTWARE. #include "helpers.hpp" #include "hc.hpp" -#include "hip_hcc.h" +#include "hip/hip_hcc.h" #include "hip_runtime.h" #include diff --git a/include/hip/hcc_detail/hip_hcc.h b/include/hip/hcc_detail/hip_hcc.h deleted file mode 100644 index fc04917931..0000000000 --- a/include/hip/hcc_detail/hip_hcc.h +++ /dev/null @@ -1,103 +0,0 @@ -/* -Copyright (c) 2015 - present Advanced Micro Devices, Inc. All rights reserved. - -Permission is hereby granted, free of charge, to any person obtaining a copy -of this software and associated documentation files (the "Software"), to deal -in the Software without restriction, including without limitation the rights -to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -copies of the Software, and to permit persons to whom the Software is -furnished to do so, subject to the following conditions: - -The above copyright notice and this permission notice shall be included in -all copies or substantial portions of the Software. - -THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -THE SOFTWARE. -*/ - -#ifndef HIP_INCLUDE_HIP_HCC_DETAIL_HIP_HCC_H -#define HIP_INCLUDE_HIP_HCC_DETAIL_HIP_HCC_H - -#include "hip/hip_runtime_api.h" - -#if __cplusplus -#ifdef __HCC__ -#include - - -/** - *------------------------------------------------------------------------------------------------- - *------------------------------------------------------------------------------------------------- - * @defgroup HCC-specific features - * @warning These APIs provide access to special features of HCC compiler and are not available through the CUDA path. - * @{ - */ - - -/** - * @brief Return hc::accelerator associated with the specified deviceId - * @return #hipSuccess, #hipErrorInvalidDevice - */ -hipError_t hipHccGetAccelerator(int deviceId, hc::accelerator *acc); - -/** - * @brief Return hc::accelerator_view associated with the specified stream - * - * If stream is 0, the accelerator_view for the default stream is returned. - * @return #hipSuccess - */ -hipError_t hipHccGetAcceleratorView(hipStream_t stream, hc::accelerator_view **av); - - -#endif // #ifdef __HCC__ - -/** - * @brief launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra - * - * @param [in[ f Kernel to launch. - * @param [in] gridDimX X grid dimension specified in work-items - * @param [in] gridDimY Y grid dimension specified in work-items - * @param [in] gridDimZ Z grid dimension specified in work-items - * @param [in] blockDimX X block dimensions specified in work-items - * @param [in] blockDimY Y grid dimension specified in work-items - * @param [in] blockDimZ Z grid dimension specified in work-items - * @param [in] sharedMemBytes Amount of dynamic shared memory to allocate for this kernel. The kernel can access this with HIP_DYNAMIC_SHARED. - * @param [in] stream Stream where the kernel should be dispatched. May be 0, in which case th default stream is used with associated synchronization rules. - * @param [in] kernelParams - * @param [in] extra Pointer to kernel arguments. These are passed directly to the kernel and must be in the memory layout and alignment expected by the kernel. - * @param [in] startEvent If non-null, specified event will be updated to track the start time of the kernel launch. The event must be created before calling this API. - * @param [in] stopEvent If non-null, specified event will be updated to track the stop time of the kernel launch. The event must be created before calling this API. - * - * @returns hipSuccess, hipInvalidDevice, hipErrorNotInitialized, hipErrorInvalidValue - * - * @warning kernellParams argument is not yet implemented in HIP. Please use extra instead. Please refer to hip_porting_driver_api.md for sample usage. - - * HIP/ROCm actually updates the start event when the associated kernel completes. - */ -hipError_t hipHccModuleLaunchKernel(hipFunction_t f, - uint32_t globalWorkSizeX, - uint32_t globalWorkSizeY, - uint32_t globalWorkSizeZ, - uint32_t localWorkSizeX, - uint32_t localWorkSizeY, - uint32_t localWorkSizeZ, - size_t sharedMemBytes, - hipStream_t hStream, - void **kernelParams, - void **extra, - hipEvent_t startEvent=nullptr, - hipEvent_t stopEvent=nullptr - ); - -// doxygen end HCC-specific features -/** - * @} - */ -#endif // #if __cplusplus - -#endif // diff --git a/include/hip/hip_hcc.h b/include/hip/hip_hcc.h index 3407a311bd..582d544eb9 100644 --- a/include/hip/hip_hcc.h +++ b/include/hip/hip_hcc.h @@ -23,8 +23,83 @@ THE SOFTWARE. #ifndef HIP_INCLUDE_HIP_HIP_HCC_H #define HIP_INCLUDE_HIP_HIP_HCC_H -#if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__) -#include "hip/hcc_detail/hip_hcc.h" -#endif +#ifdef __HCC__ -#endif +#include "hip/hip_runtime_api.h" + +// Forward declarations: +namespace hc { + class accelerator; + class accelerator_view; +}; + + +/** + *------------------------------------------------------------------------------------------------- + *------------------------------------------------------------------------------------------------- + * @defgroup HCC-specific features + * @warning These APIs provide access to special features of HCC compiler and are not available through the CUDA path. + * @{ + */ + + +/** + * @brief Return hc::accelerator associated with the specified deviceId + * @return #hipSuccess, #hipErrorInvalidDevice + */ +hipError_t hipHccGetAccelerator(int deviceId, hc::accelerator *acc); + +/** + * @brief Return hc::accelerator_view associated with the specified stream + * + * If stream is 0, the accelerator_view for the default stream is returned. + * @return #hipSuccess + */ +hipError_t hipHccGetAcceleratorView(hipStream_t stream, hc::accelerator_view **av); + + + +/** + * @brief launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelparams or extra + * + * @param [in[ f Kernel to launch. + * @param [in] gridDimX X grid dimension specified in work-items + * @param [in] gridDimY Y grid dimension specified in work-items + * @param [in] gridDimZ Z grid dimension specified in work-items + * @param [in] blockDimX X block dimensions specified in work-items + * @param [in] blockDimY Y grid dimension specified in work-items + * @param [in] blockDimZ Z grid dimension specified in work-items + * @param [in] sharedMemBytes Amount of dynamic shared memory to allocate for this kernel. The kernel can access this with HIP_DYNAMIC_SHARED. + * @param [in] stream Stream where the kernel should be dispatched. May be 0, in which case th default stream is used with associated synchronization rules. + * @param [in] kernelParams + * @param [in] extra Pointer to kernel arguments. These are passed directly to the kernel and must be in the memory layout and alignment expected by the kernel. + * @param [in] startEvent If non-null, specified event will be updated to track the start time of the kernel launch. The event must be created before calling this API. + * @param [in] stopEvent If non-null, specified event will be updated to track the stop time of the kernel launch. The event must be created before calling this API. + * + * @returns hipSuccess, hipInvalidDevice, hipErrorNotInitialized, hipErrorInvalidValue + * + * @warning kernellParams argument is not yet implemented in HIP. Please use extra instead. Please refer to hip_porting_driver_api.md for sample usage. + + * HIP/ROCm actually updates the start event when the associated kernel completes. + */ +hipError_t hipHccModuleLaunchKernel(hipFunction_t f, + uint32_t globalWorkSizeX, + uint32_t globalWorkSizeY, + uint32_t globalWorkSizeZ, + uint32_t localWorkSizeX, + uint32_t localWorkSizeY, + uint32_t localWorkSizeZ, + size_t sharedMemBytes, + hipStream_t hStream, + void **kernelParams, + void **extra, + hipEvent_t startEvent=nullptr, + hipEvent_t stopEvent=nullptr + ); + +// doxygen end HCC-specific features +/** + * @} + */ +#endif // #ifdef __HCC__ +#endif // #ifdef HIP_INCLUDE_HIP_HIP_HCC_H From 626b5191e8cc56386ad6d0c83559af6f514ed5a0 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Fri, 30 Jun 2017 19:01:14 -0500 Subject: [PATCH 016/108] Set default HIP_SYNC_NULL_STREAM=1. --- src/hip_hcc.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 9c54c8917f..39466aa834 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -92,7 +92,7 @@ int HIP_SYNC_HOST_ALLOC = 1; // Chicken bit to sync on host to implement null stream. // If 0, null stream synchronization is performed on the GPU -int HIP_SYNC_NULL_STREAM = 0; +int HIP_SYNC_NULL_STREAM = 1; // HIP needs to change some behavior based on HCC_OPT_FLUSH : // TODO - set this to 1 From 34fcaa23bbc011499524fe25e63dd5544161dd20 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Sat, 1 Jul 2017 07:56:30 +0530 Subject: [PATCH 017/108] Fixed build error in hipDeviceGetByPCIBusId Change-Id: I33be362910c4c5dca7435fb9c41f427b54f0e9a7 --- include/hip/hcc_detail/hip_runtime_api.h | 2 +- src/hip_device.cpp | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/include/hip/hcc_detail/hip_runtime_api.h b/include/hip/hcc_detail/hip_runtime_api.h index 724bf09b21..2036aa5008 100644 --- a/include/hip/hcc_detail/hip_runtime_api.h +++ b/include/hip/hcc_detail/hip_runtime_api.h @@ -1867,7 +1867,7 @@ hipError_t hipDeviceGetPCIBusId (char *pciBusId,int len,int device); * * @returns #hipSuccess, #hipErrorInavlidDevice, #hipErrorInvalidValue */ -hipError_t hipDeviceGetByPCIBusId ( int* device,const int* pciBusId ); +hipError_t hipDeviceGetByPCIBusId ( int* device,const char* pciBusId ); /** diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 66c6c7db5f..9086cd8012 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -395,7 +395,7 @@ hipError_t hipDeviceTotalMem (size_t *bytes,hipDevice_t device) return ihipLogStatus(e); } -hipError_t hipDeviceGetByPCIBusId (int* device, const int* pciBusId ) +hipError_t hipDeviceGetByPCIBusId (int* device, const char* pciBusId ) { HIP_INIT_API(device,pciBusId); hipDeviceProp_t tempProp; From 40d0a203f8b7a0719e6d7c3f704b4bec9594b900 Mon Sep 17 00:00:00 2001 From: "Sun, Peng" Date: Sat, 1 Jul 2017 09:43:31 -0500 Subject: [PATCH 018/108] fix typo in hipcc Change-Id: I5f5fb3d8f1e5e8104dca45dc213c13ae9d98a583 --- bin/hipcc | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/bin/hipcc b/bin/hipcc index bcd3e3a591..4a12b9e541 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -47,7 +47,7 @@ sub parse_config_file { } $verbose = $ENV{'HIPCC_VERBOSE'} // 0; -# Verbose: 0x1=commands, 0x2=paths, 0x4=hippc args +# Verbose: 0x1=commands, 0x2=paths, 0x4=hipcc args $HIP_PATH=$ENV{'HIP_PATH'} // dirname (dirname $0); # use parent directory of hipcc From 0459f3db72706af10996a178df76650b5fb2d45c Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Mon, 3 Jul 2017 15:05:30 -0500 Subject: [PATCH 019/108] Add hipdbPrintMem - wrapper for hcc memory tracker. --- CMakeLists.txt | 1 + include/hip/hcc_detail/hip_db.h | 22 ++++++++++++++++++++++ src/hip_db.cpp | 12 ++++++++++++ 3 files changed, 35 insertions(+) create mode 100644 include/hip/hcc_detail/hip_db.h create mode 100644 src/hip_db.cpp diff --git a/CMakeLists.txt b/CMakeLists.txt index b3ea5a3ca3..7d4039b4b4 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -179,6 +179,7 @@ if(HIP_PLATFORM STREQUAL "hcc") src/hip_peer.cpp src/hip_stream.cpp src/hip_module.cpp + src/hip_db.cpp src/grid_launch.cpp src/env.cpp) diff --git a/include/hip/hcc_detail/hip_db.h b/include/hip/hcc_detail/hip_db.h new file mode 100644 index 0000000000..eb5c3c0ac8 --- /dev/null +++ b/include/hip/hcc_detail/hip_db.h @@ -0,0 +1,22 @@ +/** + * @defgroup HipDb HCC-specific debug facilities + * @{ + */ + + +/** + * @brief * Print memory tracker information for this pointer. + * + * HIP maintains a table for all memory allocations performed by the application. + * If targetAddress is 0, the entire table is printed to stderr. + * If targetAddress is non-null, this routine will perform some forensic analysis + * to find the pointer + */ +void hipdbPrintMem(void *targetAddress); + + + +// doxygen end HipDb +/** + * @} + */ diff --git a/src/hip_db.cpp b/src/hip_db.cpp new file mode 100644 index 0000000000..21fa677fe6 --- /dev/null +++ b/src/hip_db.cpp @@ -0,0 +1,12 @@ + +#include "hcc/hc_am.hpp" + + + + +void hipdbPrintMem(void *targetAddress) +{ + hc::am_memtracker_print(targetAddress); +}; + + From db959e648bc8bacc1d993b264afa7bb65bda3463 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Wed, 5 Jul 2017 11:44:44 +0530 Subject: [PATCH 020/108] GPUOpen-ProfessionalCompute-Tools -> ROCm-Developer-Tools Change-Id: I9f5b29dd1097385acecb0c672770d8adca2fdcf7 --- INSTALL.md | 4 ++-- README.md | 8 ++++---- RELEASE.md | 4 ++-- packaging/convert_md_to_html.sh | 4 ++-- samples/2_Cookbook/0_MatrixTranspose/Readme.md | 18 +++++++++--------- samples/2_Cookbook/10_inline_asm/Readme.md | 18 +++++++++--------- samples/2_Cookbook/1_hipEvent/Readme.md | 18 +++++++++--------- samples/2_Cookbook/2_Profiler/Readme.md | 16 ++++++++-------- samples/2_Cookbook/3_shared_memory/Readme.md | 18 +++++++++--------- samples/2_Cookbook/4_shfl/Readme.md | 18 +++++++++--------- samples/2_Cookbook/5_2dshfl/Readme.md | 18 +++++++++--------- samples/2_Cookbook/6_dynamic_shared/Readme.md | 18 +++++++++--------- samples/2_Cookbook/7_streams/Readme.md | 18 +++++++++--------- samples/2_Cookbook/9_unroll/Readme.md | 18 +++++++++--------- 14 files changed, 99 insertions(+), 99 deletions(-) diff --git a/INSTALL.md b/INSTALL.md index ef584dafa0..dc9ae41b9c 100644 --- a/INSTALL.md +++ b/INSTALL.md @@ -51,14 +51,14 @@ Run hipconfig (instructions below assume default installation path) : /opt/rocm/bin/hipconfig --full ``` -Compile and run the [square sample](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/tree/master/samples/0_Intro/square). +Compile and run the [square sample](https://github.com/ROCm-Developer-Tools/HIP/tree/master/samples/0_Intro/square). # Building HIP from source HIP source code is available and the project can be built from source on the HCC platform. 1. Follow the above steps to install and validate the binary packages. -2. Download HIP source code (from the [GitHub repot](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP).) +2. Download HIP source code (from the [GitHub repot](https://github.com/ROCm-Developer-Tools/HIP).) 3. Build and install HIP (This is the simple version assuming default paths ; see below for additional options.) ``` cd HIP diff --git a/README.md b/README.md index 565fd6a36d..568b105049 100644 --- a/README.md +++ b/README.md @@ -29,7 +29,7 @@ HIP releases are typically of two types. The tag naming convention is different - [Installation](INSTALL.md) - [HIP FAQ](docs/markdown/hip_faq.md) - [HIP Kernel Language](docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) - [HIP Porting Guide](docs/markdown/hip_porting_guide.md) - [HIP Porting Driver Guide](docs/markdown/hip_porting_driver_api.md) - [HIP Profiling ](docs/markdown/hip_profiling.md) @@ -122,8 +122,8 @@ make ## More Examples -The GitHub repository [HIP-Examples](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP-Examples.git) contains a hipified version of the popular Rodinia benchmark suite. -The README with the procedures and tips the team used during this porting effort is here: [Rodinia Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP-Examples/blob/master/rodinia_3.0/hip/README.hip_porting) +The GitHub repository [HIP-Examples](https://github.com/ROCm-Developer-Tools/HIP-Examples.git) contains a hipified version of the popular Rodinia benchmark suite. +The README with the procedures and tips the team used during this porting effort is here: [Rodinia Porting Guide](https://github.com/ROCm-Developer-Tools/HIP-Examples/blob/master/rodinia_3.0/hip/README.hip_porting) ## Tour of the HIP Directories * **include**: @@ -141,6 +141,6 @@ The README with the procedures and tips the team used during this porting effort * **doc**: Documentation - markdown and doxygen info ## Reporting an issue -Use the [GitHub issue tracker] (https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/issues). +Use the [GitHub issue tracker](https://github.com/ROCm-Developer-Tools/HIP/issues). If reporting a bug, include the output of "hipconfig --full" and samples/1_hipInfo/hipInfo (if possible). diff --git a/RELEASE.md b/RELEASE.md index a1e580b7b0..d426f8563c 100644 --- a/RELEASE.md +++ b/RELEASE.md @@ -1,6 +1,6 @@ # Release notes -We have attempted to document known bugs and limitations - in particular the [HIP Kernel Language](docs/markdown/hip_kernel_language.md) document uses the phrase "Under Development", and the [HIP Runtime API bug list](http://gpuopen-professionalcompute-tools.github.io/HIP/bug.html) lists known bugs. +We have attempted to document known bugs and limitations - in particular the [HIP Kernel Language](docs/markdown/hip_kernel_language.md) document uses the phrase "Under Development", and the [HIP Runtime API bug list](http://rocm-developer-tools.github.io/HIP/bug.html) lists known bugs. =================================================================================================== @@ -113,7 +113,7 @@ Date: 2016.06.06 - Add cross-linking support between G++ and HCC, in particular for interfaces that use standard C++ libraries (ie std::vectors, std::strings). HIPCC now uses libstdc++ by default on the HCC compilation path. -- More samples including gpu-burn, SHOC, nbody, rtm. See [HIP-Examples](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP-Examples) +- More samples including gpu-burn, SHOC, nbody, rtm. See [HIP-Examples](https://github.com/ROCm-Developer-Tools/HIP-Examples) =================================================================================================== diff --git a/packaging/convert_md_to_html.sh b/packaging/convert_md_to_html.sh index b2c868cded..e6442a9ef2 100755 --- a/packaging/convert_md_to_html.sh +++ b/packaging/convert_md_to_html.sh @@ -54,8 +54,8 @@ popd # replace github.io links pushd $html_destdir -sed -i "s?http://gpuopen-professionalcompute-tools.github.io/HIP?docs/RuntimeAPI/html/index.html?g" README.html -sed -i "s?http://gpuopen-professionalcompute-tools.github.io/HIP?docs/RuntimeAPI/html/?g" RELEASE.html +sed -i "s?http://rocm-developer-tools.github.io/HIP?docs/RuntimeAPI/html/index.html?g" README.html +sed -i "s?http://rocm-developer-tools.github.io/HIP?docs/RuntimeAPI/html/?g" RELEASE.html popd exit 0 diff --git a/samples/2_Cookbook/0_MatrixTranspose/Readme.md b/samples/2_Cookbook/0_MatrixTranspose/Readme.md index 5e9483b595..ab5dbdc958 100644 --- a/samples/2_Cookbook/0_MatrixTranspose/Readme.md +++ b/samples/2_Cookbook/0_MatrixTranspose/Readme.md @@ -7,7 +7,7 @@ This tutorial shows how to get write simple HIP application. We will write the s HIP is a C++ runtime API and kernel language that allows developers to create portable applications that can run on AMD and other GPU’s. Our goal was to rise above the lowest-common-denominator paths and deliver a solution that allows you, the developer, to use essential hardware features and maximize your application’s performance on GPU hardware. ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -90,11 +90,11 @@ Use the make command and execute it using ./exe Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [hipify-clang](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/hipify-clang/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [hipify-clang](https://github.com/ROCm-Developer-Tools/HIP/hipify-clang/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/10_inline_asm/Readme.md b/samples/2_Cookbook/10_inline_asm/Readme.md index 8c98547220..0e64fe9c6e 100644 --- a/samples/2_Cookbook/10_inline_asm/Readme.md +++ b/samples/2_Cookbook/10_inline_asm/Readme.md @@ -15,7 +15,7 @@ For more information: [User Guide for AMDGPU Back-end](llvm.org/docs/AMDGPUUsage.html) ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -37,11 +37,11 @@ Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/1_hipEvent/Readme.md b/samples/2_Cookbook/1_hipEvent/Readme.md index e3ec8ad780..ea4f3a67e9 100644 --- a/samples/2_Cookbook/1_hipEvent/Readme.md +++ b/samples/2_Cookbook/1_hipEvent/Readme.md @@ -7,7 +7,7 @@ This tutorial is follow-up of the previous one where we learn how to write our f Memory transfer and kernel execution are the most important parameter in parallel computing (specially HPC and machine learning). Memory bottlenecks is the main problem why we are not able to get the highest performance, therefore obtaining the memory transfer timing and kernel execution timing plays key role in application optimization. ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -64,11 +64,11 @@ Use the make command and execute it using ./exe Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [hipify-clang](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/hipify-clang/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [hipify-clang](https://github.com/ROCm-Developer-Tools/HIP/hipify-clang/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/2_Profiler/Readme.md b/samples/2_Cookbook/2_Profiler/Readme.md index 92a8be228e..4059e42193 100644 --- a/samples/2_Cookbook/2_Profiler/Readme.md +++ b/samples/2_Cookbook/2_Profiler/Readme.md @@ -37,11 +37,11 @@ You can also print the HIP function strings to stderr using HIP_TRACE_API enviro Note this trace mode uses colors. "less -r" can handle raw control characters and will display the debug output in proper colors. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [hipify-clang](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/hipify-clang/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [hipify-clang](https://github.com/ROCm-Developer-Tools/HIP/hipify-clang/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/3_shared_memory/Readme.md b/samples/2_Cookbook/3_shared_memory/Readme.md index 6b9393397c..8b9e102ec9 100644 --- a/samples/2_Cookbook/3_shared_memory/Readme.md +++ b/samples/2_Cookbook/3_shared_memory/Readme.md @@ -7,7 +7,7 @@ Earlier we learned how to write our first hip program, in which we compute Matri As we mentioned earlier that Memory bottlenecks is the main problem why we are not able to get the highest performance, therefore minimizing the latency for memory access plays prominent role in application optimization. In this tutorial, we'll learn how to use static shared memory and will explain the dynamic one latter. ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -32,11 +32,11 @@ Use the make command and execute it using ./exe Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/4_shfl/Readme.md b/samples/2_Cookbook/4_shfl/Readme.md index da62901851..923d2f3837 100644 --- a/samples/2_Cookbook/4_shfl/Readme.md +++ b/samples/2_Cookbook/4_shfl/Readme.md @@ -15,7 +15,7 @@ Let's talk about Warp first. The kernel code is executed in groups of fixed numb ` float __shfl_xor (float var, int laneMask, int width=warpSize); ` ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -41,11 +41,11 @@ Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia please make sure you have a 3.0 or higher compute capable device in order to use warp shfl operations and add `-gencode arch=compute=30, code=sm_30` nvcc flag in the Makefile while using this application. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/5_2dshfl/Readme.md b/samples/2_Cookbook/5_2dshfl/Readme.md index fba114152a..8efff49d8e 100644 --- a/samples/2_Cookbook/5_2dshfl/Readme.md +++ b/samples/2_Cookbook/5_2dshfl/Readme.md @@ -15,7 +15,7 @@ Let's talk about Warp first. The kernel code is executed in groups of fixed numb ` float __shfl_xor (float var, int laneMask, int width=warpSize); ` ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -41,11 +41,11 @@ Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia please make sure you have a 3.0 or higher compute capable device in order to use warp shfl operations and add `-gencode arch=compute=30, code=sm_30` nvcc flag in the Makefile while using this application. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/6_dynamic_shared/Readme.md b/samples/2_Cookbook/6_dynamic_shared/Readme.md index a10fd56a95..15ea299a9c 100644 --- a/samples/2_Cookbook/6_dynamic_shared/Readme.md +++ b/samples/2_Cookbook/6_dynamic_shared/Readme.md @@ -7,7 +7,7 @@ Earlier we learned how to use static shared memory. In this tutorial, we'll expl As we mentioned earlier that Memory bottlenecks is the main problem why we are not able to get the highest performance, therefore minimizing the latency for memory access plays prominent role in application optimization. In this tutorial, we'll learn how to use dynamic shared memory. ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -37,11 +37,11 @@ Use the make command and execute it using ./exe Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/7_streams/Readme.md b/samples/2_Cookbook/7_streams/Readme.md index a75149925e..ca295d3f49 100644 --- a/samples/2_Cookbook/7_streams/Readme.md +++ b/samples/2_Cookbook/7_streams/Readme.md @@ -7,7 +7,7 @@ In all Earlier tutorial we used single stream, In this tutorial, we'll explain h The various instances of kernel to be executed on device in exact launch order defined by Host are called streams. We can launch multiple streams on a single device. We will learn how to learn two streams which can we scaled with ease. ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -47,11 +47,11 @@ Use the make command and execute it using ./exe Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) diff --git a/samples/2_Cookbook/9_unroll/Readme.md b/samples/2_Cookbook/9_unroll/Readme.md index 3c2635c0eb..194eb9c7ff 100644 --- a/samples/2_Cookbook/9_unroll/Readme.md +++ b/samples/2_Cookbook/9_unroll/Readme.md @@ -8,7 +8,7 @@ Loop unrolling optimization hints can be specified with #pragma unroll and #prag Specifying #pragma unroll without a parameter directs the loop unroller to attempt to fully unroll the loop if the trip count is known at compile time and attempt to partially unroll the loop if the trip count is not known at compile time. ## Requirement: -For hardware requirement and software installation [Installation](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/INSTALL.md) +For hardware requirement and software installation [Installation](https://github.com/ROCm-Developer-Tools/HIP/INSTALL.md) ## prerequiste knowledge: @@ -38,11 +38,11 @@ Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia please make sure you have a 3.0 or higher compute capable device in order to use warp shfl operations and add `-gencode arch=compute=30, code=sm_30` nvcc flag in the Makefile while using this application. ## More Info: -- [HIP FAQ](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_faq.md) -- [HIP Kernel Language](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_kernel_language.md) -- [HIP Runtime API (Doxygen)](http://gpuopen-professionalcompute-tools.github.io/HIP) -- [HIP Porting Guide](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_porting_guide.md) -- [HIP Terminology](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) -- [clang-hipify](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/clang-hipify/README.md) -- [Developer/CONTRIBUTING Info](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/CONTRIBUTING.md) -- [Release Notes](https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/RELEASE.md) +- [HIP FAQ](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_faq.md) +- [HIP Kernel Language](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_kernel_language.md) +- [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) +- [HIP Porting Guide](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_porting_guide.md) +- [HIP Terminology](https://github.com/ROCm-Developer-Tools/HIP/docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) +- [clang-hipify](https://github.com/ROCm-Developer-Tools/HIP/clang-hipify/README.md) +- [Developer/CONTRIBUTING Info](https://github.com/ROCm-Developer-Tools/HIP/CONTRIBUTING.md) +- [Release Notes](https://github.com/ROCm-Developer-Tools/HIP/RELEASE.md) From 85cd86c3bdf30a331a3dc8e0295f16a6f4502698 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Wed, 5 Jul 2017 11:45:18 +0530 Subject: [PATCH 021/108] ignore build directory Change-Id: I227ac377b6eed3aee1cf2121de00c86f34c12405 --- .gitignore | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.gitignore b/.gitignore index 22cd23f2c6..3dfb1afef5 100644 --- a/.gitignore +++ b/.gitignore @@ -8,7 +8,7 @@ hip-amdinternal HIP-Examples lib packages - +build bin/hipInfo bin/hipBusBandwidth bin/hipDispatchLatency From 17a58bdca5dd4674862a9175f1137d3d24dbb42b Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Fri, 7 Jul 2017 10:36:05 +0530 Subject: [PATCH 022/108] [ci] Use new repo server --- docker/dockerfile-hip-ubuntu-16.04 | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/docker/dockerfile-hip-ubuntu-16.04 b/docker/dockerfile-hip-ubuntu-16.04 index cecf5911b8..054dea76ce 100644 --- a/docker/dockerfile-hip-ubuntu-16.04 +++ b/docker/dockerfile-hip-ubuntu-16.04 @@ -9,8 +9,8 @@ COPY *.deb /tmp/ # Install the debian package RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends -y curl \ - && curl -sL http://packages.amd.com/rocm/apt/debian/rocm.gpg.key | apt-key add - \ - && echo deb [arch=amd64] http://packages.amd.com/rocm/apt/debian/ xenial main | tee /etc/apt/sources.list.d/rocm.list \ + && curl -sL http://repo.radeon.com/rocm/apt/debian/rocm.gpg.key | apt-key add - \ + && echo deb [arch=amd64] http://repo.radeon.com/rocm/apt/debian/ xenial main | tee /etc/apt/sources.list.d/rocm.list \ && apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends --allow-unauthenticated -y \ /tmp/hip_base-*.deb \ /tmp/hip_hcc-*.deb \ @@ -18,4 +18,4 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-instal /tmp/hip_samples-* \ && rm -f /tmp/*.deb \ && apt-get clean \ - && rm -rf /var/lib/apt/lists/* \ No newline at end of file + && rm -rf /var/lib/apt/lists/* From 30ee797dd4e29340245e46e31ecafe39a99785bb Mon Sep 17 00:00:00 2001 From: "Wen-Heng (Jack) Chung" Date: Fri, 7 Jul 2017 12:46:29 -0500 Subject: [PATCH 023/108] Remove explicit address spaces declarations --- include/hip/hcc_detail/hip_runtime.h | 8 -------- src/device_util.cpp | 4 ++-- tests/src/deviceLib/hipTestDeviceSymbol.cpp | 8 -------- 3 files changed, 2 insertions(+), 18 deletions(-) diff --git a/include/hip/hcc_detail/hip_runtime.h b/include/hip/hcc_detail/hip_runtime.h index da3b7ba50e..5536ded5e2 100644 --- a/include/hip/hcc_detail/hip_runtime.h +++ b/include/hip/hcc_detail/hip_runtime.h @@ -50,14 +50,6 @@ THE SOFTWARE. #include -#if USE_PROMOTE_FREE_HCC == 1 -#define ADDRESS_SPACE_1 -#define ADDRESS_SPACE_3 -#else -#define ADDRESS_SPACE_1 __attribute__((address_space(1))) -#define ADDRESS_SPACE_3 __attribute__((address_space(3))) -#endif - //--- // Remainder of this file only compiles with HCC #if defined __HCC__ diff --git a/src/device_util.cpp b/src/device_util.cpp index 1efda02933..cd27ada4f5 100644 --- a/src/device_util.cpp +++ b/src/device_util.cpp @@ -35,8 +35,8 @@ THE SOFTWARE. This is the best place to put them because the device global variables need to be initialized at the start. */ -__device__ ADDRESS_SPACE_1 char gpuHeap[SIZE_OF_HEAP]; -__device__ ADDRESS_SPACE_1 uint32_t gpuFlags[NUM_PAGES]; +__device__ char gpuHeap[SIZE_OF_HEAP]; +__device__ uint32_t gpuFlags[NUM_PAGES]; __device__ void *__hip_hc_malloc(size_t size) { diff --git a/tests/src/deviceLib/hipTestDeviceSymbol.cpp b/tests/src/deviceLib/hipTestDeviceSymbol.cpp index c2ffb5ce7d..2ee1e02dc0 100644 --- a/tests/src/deviceLib/hipTestDeviceSymbol.cpp +++ b/tests/src/deviceLib/hipTestDeviceSymbol.cpp @@ -31,16 +31,8 @@ THE SOFTWARE. #define NUM 1024 #define SIZE 1024*4 -// TODO - collapse: -#ifdef __HIP_PLATFORM_HCC__ -__device__ ADDRESS_SPACE_1 int globalIn[NUM]; -__device__ ADDRESS_SPACE_1 int globalOut[NUM]; -#endif - -#ifdef __HIP_PLATFORM_NVCC__ __device__ int globalIn[NUM]; __device__ int globalOut[NUM]; -#endif __global__ void Assign(hipLaunchParm lp, int* Out) { From 7ba09c4137a52e33c1b073565f25b5adaa43e7aa Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Tue, 11 Jul 2017 11:21:53 +0530 Subject: [PATCH 024/108] [ci] Pass IP for repo.radeon.com as docker buildargs --- Jenkinsfile | 6 +++--- docker/dockerfile-hip-ubuntu-16.04 | 5 +++-- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 16354285ec..62135e3cde 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -76,10 +76,10 @@ def docker_build_image( String platform, String source_hip_rel, String from_imag // Docker 17.05 introduced the ability to use ARG values in FROM statements // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 - //build_image = docker.build( "${project}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) + //build_image = docker.build( "${project}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg REPO_RADEON=10.255.8.5 --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) // JENKINS-44836 workaround - sh "docker build -t ${project}/${build_image_name}:latest --pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." + sh "docker build -t ${project}/${build_image_name}:latest --pull -f docker/${dockerfile_name} --build-arg REPO_RADEON=10.255.8.5 --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." build_image = docker.image( "${project}/${build_image_name}:latest" ) } } @@ -316,4 +316,4 @@ nvcc: // Not pushing an Nvidia based HiP to artifactory at this time } -} \ No newline at end of file +} diff --git a/docker/dockerfile-hip-ubuntu-16.04 b/docker/dockerfile-hip-ubuntu-16.04 index 054dea76ce..1272bf30af 100644 --- a/docker/dockerfile-hip-ubuntu-16.04 +++ b/docker/dockerfile-hip-ubuntu-16.04 @@ -1,5 +1,6 @@ # Parameters related to building hip ARG base_image +ARG REPO_RADEON=$REPO_RADEON FROM ${base_image} MAINTAINER Kent Knox @@ -9,8 +10,8 @@ COPY *.deb /tmp/ # Install the debian package RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends -y curl \ - && curl -sL http://repo.radeon.com/rocm/apt/debian/rocm.gpg.key | apt-key add - \ - && echo deb [arch=amd64] http://repo.radeon.com/rocm/apt/debian/ xenial main | tee /etc/apt/sources.list.d/rocm.list \ + && curl -sL http://$REPO_RADEON/rocm/apt/debian/rocm.gpg.key | apt-key add - \ + && echo deb [arch=amd64] http://$REPO_RADEON/rocm/apt/debian/ xenial main | tee /etc/apt/sources.list.d/rocm.list \ && apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends --allow-unauthenticated -y \ /tmp/hip_base-*.deb \ /tmp/hip_hcc-*.deb \ From 35ebfda9d43c56982370a8f36e0f7fc15bb73a0a Mon Sep 17 00:00:00 2001 From: "Wen-Heng (Jack) Chung" Date: Wed, 12 Jul 2017 16:08:20 +0000 Subject: [PATCH 025/108] Fix hip_db header inclusion path Change-Id: I8a95798301de73887dde3316dc593a1847002712 --- src/hip_db.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hip_db.cpp b/src/hip_db.cpp index 21fa677fe6..7565d70422 100644 --- a/src/hip_db.cpp +++ b/src/hip_db.cpp @@ -1,5 +1,5 @@ -#include "hcc/hc_am.hpp" +#include From 17555eabc0a6885931161c8060651ba236e159dc Mon Sep 17 00:00:00 2001 From: "Wen-Heng (Jack) Chung" Date: Wed, 12 Jul 2017 16:08:36 +0000 Subject: [PATCH 026/108] Temporarily disables HCC_OPT_FLUSH Change-Id: I290791e58dd52ab3823f6c3315e42b0d386e9d64 --- src/hip_hcc.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 39466aa834..8aaa7af02b 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -96,7 +96,7 @@ int HIP_SYNC_NULL_STREAM = 1; // HIP needs to change some behavior based on HCC_OPT_FLUSH : // TODO - set this to 1 -int HCC_OPT_FLUSH = 1; +int HCC_OPT_FLUSH = 0; From 7066ff4478ee0060dfdd3686f9a63061369d90e4 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Fri, 14 Jul 2017 11:31:40 +0530 Subject: [PATCH 027/108] remove unsupported erf(c)inv related host functionality Change-Id: I665c33616359a0124b5552076359d2f8faa54930 --- src/device_util.cpp | 102 ----------------------------------------- src/device_util.h | 6 --- src/math_functions.cpp | 20 -------- 3 files changed, 128 deletions(-) diff --git a/src/device_util.cpp b/src/device_util.cpp index cd27ada4f5..8bba36c9c7 100644 --- a/src/device_util.cpp +++ b/src/device_util.cpp @@ -1267,108 +1267,6 @@ __device__ void __threadfence_system(void){ // no-op } -float __hip_host_erfinvf(float x) -{ - float ret; - int sign; - if (x < -1 || x > 1){ - return NAN; - } - if (x == 0){ - return 0; - } - if (x > 0){ - sign = 1; - } else { - sign = -1; - x = -x; - } - if (x <= 0.7) { - float x1 = x * x; - float x2 = std::fma(__hip_erfinva3, x1, __hip_erfinva2); - float x3 = std::fma(x2, x1, __hip_erfinva1); - float x4 = x * std::fma(x3, x1, __hip_erfinva0); - - float r1 = std::fma(__hip_erfinvb4, x1, __hip_erfinvb3); - float r2 = std::fma(r1, x1, __hip_erfinvb2); - float r3 = std::fma(r2, x1, __hip_erfinvb1); - ret = x4 / std::fma(r3, x1, __hip_erfinvb0); - } else { - float x1 = std::sqrt(-std::log((1 - x) / 2)); - float x2 = std::fma(__hip_erfinvc3, x1, __hip_erfinvc2); - float x3 = std::fma(x2, x1, __hip_erfinvc1); - float x4 = std::fma(x3, x1, __hip_erfinvc0); - - float r1 = std::fma(__hip_erfinvd2, x1, __hip_erfinvd1); - ret = x4 / std::fma(r1, x1, __hip_erfinvd0); - } - - ret = ret * sign; - x = x * sign; - - ret -= (std::erf(ret) - x) / (2 / std::sqrt(HIP_PI) * std::exp(-ret * ret)); - ret -= (std::erf(ret) - x) / (2 / std::sqrt(HIP_PI) * std::exp(-ret * ret)); - - return ret; - -} - -double __hip_host_erfinv(double x) -{ - double ret; - int sign; - if (x < -1 || x > 1){ - return NAN; - } - if (x == 0){ - return 0; - } - if (x > 0){ - sign = 1; - } else { - sign = -1; - x = -x; - } - if (x <= 0.7) { - double x1 = x * x; - double x2 = std::fma(__hip_erfinva3, x1, __hip_erfinva2); - double x3 = std::fma(x2, x1, __hip_erfinva1); - double x4 = x * std::fma(x3, x1, __hip_erfinva0); - - double r1 = std::fma(__hip_erfinvb4, x1, __hip_erfinvb3); - double r2 = std::fma(r1, x1, __hip_erfinvb2); - double r3 = std::fma(r2, x1, __hip_erfinvb1); - ret = x4 / std::fma(r3, x1, __hip_erfinvb0); - } else { - double x1 = std::sqrt(-std::log((1 - x) / 2)); - double x2 = std::fma(__hip_erfinvc3, x1, __hip_erfinvc2); - double x3 = std::fma(x2, x1, __hip_erfinvc1); - double x4 = std::fma(x3, x1, __hip_erfinvc0); - - double r1 = std::fma(__hip_erfinvd2, x1, __hip_erfinvd1); - ret = x4 / std::fma(r1, x1, __hip_erfinvd0); - } - - ret = ret * sign; - x = x * sign; - - ret -= (std::erf(ret) - x) / (2 / std::sqrt(HIP_PI) * std::exp(-ret * ret)); - ret -= (std::erf(ret) - x) / (2 / std::sqrt(HIP_PI) * std::exp(-ret * ret)); - - return ret; - -} - -float __hip_host_erfcinvf(float y) -{ - return __hip_host_erfinvf(1 - y); -} - -double __hip_host_erfcinv(double y) -{ - return __hip_host_erfinv(1 - y); -} - double __hip_host_j0(double x) { double ret, a = std::fabs(x); diff --git a/src/device_util.h b/src/device_util.h index ad8b2607dd..f942fdf5e5 100644 --- a/src/device_util.h +++ b/src/device_util.h @@ -128,12 +128,6 @@ __device__ double __hip_fast_dsqrt_ru(double x); __device__ double __hip_fast_dsqrt_rz(double x); __device__ void __threadfence_system(void); -float __hip_host_erfinvf(float x); -double __hip_host_erfinv(double x); - -float __hip_host_erfcinvf(float y); -double __hip_host_erfcinv(double y); - float __hip_host_j0f(float x); double __hip_host_j0(double x); diff --git a/src/math_functions.cpp b/src/math_functions.cpp index f66f0a4312..80ccece1a3 100644 --- a/src/math_functions.cpp +++ b/src/math_functions.cpp @@ -820,16 +820,6 @@ __host__ float modff(float x, float *iptr) return std::modf(x, iptr); } -__host__ float erfcinvf(float y) -{ - return __hip_host_erfcinvf(y); -} - -__host__ double erfcinv(double y) -{ - return __hip_host_erfcinv(y); -} - __host__ double fdivide(double x, double y) { return x/y; @@ -937,16 +927,6 @@ __host__ void sincospi(double x, double *sptr, double *cptr) *cptr = std::cos(HIP_PI*x); } -//__host__ float normcdfinvf(float x) -//{ -// return std::sqrt(2) * erfinvf(2*x-1); -//} - -//__host__ double normcdfinv(double x) -//{ -// return std::sqrt(2) * erfinv(2*x-1); -//} - __host__ float nextafterf(float x, float y) { return std::nextafter(x, y); From ed82afad2ea07c5daa99c543b3daa077331c4619 Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Fri, 14 Jul 2017 17:28:16 -0500 Subject: [PATCH 028/108] [ci] Fixes to remove repo.radeon.com build-args HiP docker images should inherit the repo information from the hcc image, so no need to duplicate in HiP again --- Jenkinsfile | 9 +++++---- docker/dockerfile-build-ubuntu-16.04 | 2 +- docker/dockerfile-hip-ubuntu-16.04 | 3 --- 3 files changed, 6 insertions(+), 8 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 62135e3cde..250a8949a3 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -76,10 +76,10 @@ def docker_build_image( String platform, String source_hip_rel, String from_imag // Docker 17.05 introduced the ability to use ARG values in FROM statements // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 - //build_image = docker.build( "${project}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg REPO_RADEON=10.255.8.5 --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) + //build_image = docker.build( "${project}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) - // JENKINS-44836 workaround - sh "docker build -t ${project}/${build_image_name}:latest --pull -f docker/${dockerfile_name} --build-arg REPO_RADEON=10.255.8.5 --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." + // JENKINS-44836 workaround by using a bash script instead of docker.build() + sh "docker build -t ${project}/${build_image_name}:latest --pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." build_image = docker.image( "${project}/${build_image_name}:latest" ) } } @@ -173,7 +173,8 @@ def docker_upload_artifactory( String hcc_ver, String from_image, String source_ // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 // hip_install_image = docker.build( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}", "--pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" ) - // JENKINS-44836 workaround + // The --build-arg REPO_RADEON= is a temporary fix to get around a DNS issue with our build machines + // JENKINS-44836 workaround by using a bash script instead of docker.build() sh "docker build -t ${artifactory_org}/${image_name}:${env.BUILD_NUMBER} --pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" hip_install_image = docker.image( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}" ) diff --git a/docker/dockerfile-build-ubuntu-16.04 b/docker/dockerfile-build-ubuntu-16.04 index 0c2e4fa26d..031bf72437 100644 --- a/docker/dockerfile-build-ubuntu-16.04 +++ b/docker/dockerfile-build-ubuntu-16.04 @@ -18,7 +18,7 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-ins apt-get clean && \ rm -rf /var/lib/apt/lists/* -# docker pipeline runs containers with uid 1000 +# docker pipeline runs containers with particular uid # create a jenkins user with this specific uid so it can use sudo priviledges # Grant any member of sudo group password-less sudo privileges RUN useradd --create-home -u ${user_uid} -G sudo --shell /bin/bash jenkins && \ diff --git a/docker/dockerfile-hip-ubuntu-16.04 b/docker/dockerfile-hip-ubuntu-16.04 index 1272bf30af..0852ae38c5 100644 --- a/docker/dockerfile-hip-ubuntu-16.04 +++ b/docker/dockerfile-hip-ubuntu-16.04 @@ -1,6 +1,5 @@ # Parameters related to building hip ARG base_image -ARG REPO_RADEON=$REPO_RADEON FROM ${base_image} MAINTAINER Kent Knox @@ -10,8 +9,6 @@ COPY *.deb /tmp/ # Install the debian package RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends -y curl \ - && curl -sL http://$REPO_RADEON/rocm/apt/debian/rocm.gpg.key | apt-key add - \ - && echo deb [arch=amd64] http://$REPO_RADEON/rocm/apt/debian/ xenial main | tee /etc/apt/sources.list.d/rocm.list \ && apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends --allow-unauthenticated -y \ /tmp/hip_base-*.deb \ /tmp/hip_hcc-*.deb \ From e671cec36782edc1901066adf4454d3c2dd48afe Mon Sep 17 00:00:00 2001 From: Weixing Zhang Date: Mon, 17 Jul 2017 15:16:12 -0400 Subject: [PATCH 029/108] HIP Texture Support --- CMakeLists.txt | 1 + bin/hipify-perl | 7 +- include/hip/hcc_detail/driver_types.h | 218 +- include/hip/hcc_detail/hip_runtime.h | 2 +- include/hip/hcc_detail/hip_runtime_api.h | 235 +- include/hip/hcc_detail/hip_texture.h | 107 - include/hip/hcc_detail/hip_texture_types.h | 83 + include/hip/hcc_detail/texture_functions.h | 6771 +++++++++++++++++ include/hip/hcc_detail/texture_types.h | 85 +- .../{hip_texture.h => hip_texture_types.h} | 12 +- include/hip/nvcc_detail/hip_texture_types.h | 6 + src/hip_memory.cpp | 173 +- src/hip_texture.cpp | 668 ++ .../hip/nvcc_detail => src}/hip_texture.h | 14 +- 14 files changed, 8194 insertions(+), 188 deletions(-) delete mode 100644 include/hip/hcc_detail/hip_texture.h create mode 100644 include/hip/hcc_detail/hip_texture_types.h create mode 100644 include/hip/hcc_detail/texture_functions.h rename include/hip/{hip_texture.h => hip_texture_types.h} (83%) create mode 100644 include/hip/nvcc_detail/hip_texture_types.h create mode 100644 src/hip_texture.cpp rename {include/hip/nvcc_detail => src}/hip_texture.h (78%) diff --git a/CMakeLists.txt b/CMakeLists.txt index 7d4039b4b4..442beb3428 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -181,6 +181,7 @@ if(HIP_PLATFORM STREQUAL "hcc") src/hip_module.cpp src/hip_db.cpp src/grid_launch.cpp + src/hip_texture.cpp src/env.cpp) set(SOURCE_FILES_DEVICE diff --git a/bin/hipify-perl b/bin/hipify-perl index 27acc5bccc..4b6e6e89b9 100755 --- a/bin/hipify-perl +++ b/bin/hipify-perl @@ -283,6 +283,7 @@ while (@ARGV) { $ft{'mem'} += s/\bcudaMemcpy2D\b/hipMemcpy2D/g; $ft{'mem'} += s/\bcudaMemcpy2DToArray\b/hipMemcpy2DToArray/g; + $ft{'mem'} += s/\bcudaMemcpyToArray\b/hipMemcpyToArray/g; #-------- # Memory management: @@ -302,6 +303,7 @@ while (@ARGV) { $ft{'mem'} += s/\bcudaHostGetDevicePointer\b/hipHostGetDevicePointer/g; $ft{'mem'} += s/\bcudaMallocArray\b/hipMallocArray/g; + $ft{'mem'} += s/\bcudaFreeArray\b/hipFreeArray/g; $ft{'mem'} += s/\bcudaMallocPitch\b/hipMallocPitch/g; @@ -511,11 +513,14 @@ while (@ARGV) { $ft{'tex'} += s/\bcudaFilterModePoint\b/hipFilterModePoint/g; $ft{'tex'} += s/\bcudaReadModeElementType\b/hipReadModeElementType/g; - $ft{'tex'} += s/\bcudaArray\b/hipArrary/g; + $ft{'tex'} += s/\bcudaArray\b/hipArray/g; $ft{'tex'} += s/\bcudaCreateChannelDesc\b/hipCreateChannelDesc/g; $ft{'tex'} += s/\bcudaBindTexture\b/hipBindTexture/g; $ft{'tex'} += s/\bcudaBindTextureToArray\b/hipBindTextureToArray/g; $ft{'tex'} += s/\bcudaUnbindTexture\b/hipUnbindTexture/g; + $ft{'tex'} += s/\bcudaChannelFormatKindFloat\b/hipChannelFormatKindFloat/g; + $ft{'tex'} += s/\bcudaAddressMode/hipAddressMode/g; + $ft{'tex'} += s/\bcudaFilterMode/hipFilterMode/g; } diff --git a/include/hip/hcc_detail/driver_types.h b/include/hip/hcc_detail/driver_types.h index 3578ddc609..ce5e9789be 100644 --- a/include/hip/hcc_detail/driver_types.h +++ b/include/hip/hcc_detail/driver_types.h @@ -25,20 +25,220 @@ THE SOFTWARE. enum hipChannelFormatKind { - hipChannelFormatKindSigned = 0, - hipChannelFormatKindUnsigned = 1, - hipChannelFormatKindFloat = 2, - hipChannelFormatKindNone = 3 + hipChannelFormatKindSigned = 0, + hipChannelFormatKindUnsigned = 1, + hipChannelFormatKindFloat = 2, + hipChannelFormatKindNone = 3 }; struct hipChannelFormatDesc { - int x; - int y; - int z; - int w; - enum hipChannelFormatKind f; + int x; + int y; + int z; + int w; + enum hipChannelFormatKind f; }; +struct hipArray { + void* data; //FIXME: generalize this + struct hipChannelFormatDesc desc; + unsigned int type; + unsigned int width; + unsigned int height; + unsigned int depth; +}; + +typedef struct hipArray* hipArray_t; + +typedef const struct hipArray* hipArray_const_t; + +// TODO: It needs to be modified since it was just copied from hipArray. +struct hipMipmappedArray { + void* data; //FIXME: generalize this + struct hipChannelFormatDesc desc; + unsigned int width; + unsigned int height; + unsigned int depth; +}; + +typedef struct hipMipmappedArray *hipMipmappedArray_t; + +typedef const struct hipMipmappedArray *hipMipmappedArray_const_t; + +/** + * hip resource types + */ +enum hipResourceType +{ + hipResourceTypeArray = 0x00, + hipResourceTypeMipmappedArray = 0x01, + hipResourceTypeLinear = 0x02, + hipResourceTypePitch2D = 0x03 +}; + +/** + * hip texture resource view formats + */ +enum hipResourceViewFormat +{ + hipResViewFormatNone = 0x00, + hipResViewFormatUnsignedChar1 = 0x01, + hipResViewFormatUnsignedChar2 = 0x02, + hipResViewFormatUnsignedChar4 = 0x03, + hipResViewFormatSignedChar1 = 0x04, + hipResViewFormatSignedChar2 = 0x05, + hipResViewFormatSignedChar4 = 0x06, + hipResViewFormatUnsignedShort1 = 0x07, + hipResViewFormatUnsignedShort2 = 0x08, + hipResViewFormatUnsignedShort4 = 0x09, + hipResViewFormatSignedShort1 = 0x0a, + hipResViewFormatSignedShort2 = 0x0b, + hipResViewFormatSignedShort4 = 0x0c, + hipResViewFormatUnsignedInt1 = 0x0d, + hipResViewFormatUnsignedInt2 = 0x0e, + hipResViewFormatUnsignedInt4 = 0x0f, + hipResViewFormatSignedInt1 = 0x10, + hipResViewFormatSignedInt2 = 0x11, + hipResViewFormatSignedInt4 = 0x12, + hipResViewFormatHalf1 = 0x13, + hipResViewFormatHalf2 = 0x14, + hipResViewFormatHalf4 = 0x15, + hipResViewFormatFloat1 = 0x16, + hipResViewFormatFloat2 = 0x17, + hipResViewFormatFloat4 = 0x18, + hipResViewFormatUnsignedBlockCompressed1 = 0x19, + hipResViewFormatUnsignedBlockCompressed2 = 0x1a, + hipResViewFormatUnsignedBlockCompressed3 = 0x1b, + hipResViewFormatUnsignedBlockCompressed4 = 0x1c, + hipResViewFormatSignedBlockCompressed4 = 0x1d, + hipResViewFormatUnsignedBlockCompressed5 = 0x1e, + hipResViewFormatSignedBlockCompressed5 = 0x1f, + hipResViewFormatUnsignedBlockCompressed6H = 0x20, + hipResViewFormatSignedBlockCompressed6H = 0x21, + hipResViewFormatUnsignedBlockCompressed7 = 0x22 +}; + +/** + * HIP resource descriptor + */ +struct hipResourceDesc { + enum hipResourceType resType; + + union { + struct { + hipArray_t array; + } array; + struct { + hipMipmappedArray_t mipmap; + } mipmap; + struct { + void *devPtr; + struct hipChannelFormatDesc desc; + size_t sizeInBytes; + } linear; + struct { + void *devPtr; + struct hipChannelFormatDesc desc; + size_t width; + size_t height; + size_t pitchInBytes; + } pitch2D; + } res; +}; + +/** + * hip resource view descriptor + */ +struct hipResourceViewDesc +{ + enum hipResourceViewFormat format; + size_t width; + size_t height; + size_t depth; + unsigned int firstMipmapLevel; + unsigned int lastMipmapLevel; + unsigned int firstLayer; + unsigned int lastLayer; +}; + +/** + * Memory copy types + * + */ +typedef enum hipMemcpyKind { + hipMemcpyHostToHost = 0, ///< Host-to-Host Copy + hipMemcpyHostToDevice = 1, ///< Host-to-Device Copy + hipMemcpyDeviceToHost = 2, ///< Device-to-Host Copy + hipMemcpyDeviceToDevice =3, ///< Device-to-Device Copy + hipMemcpyDefault = 4 ///< Runtime will automatically determine copy-kind based on virtual addresses. +} hipMemcpyKind; + +struct hipPitchedPtr +{ + void *ptr; + size_t pitch; + size_t xsize; + size_t ysize; +}; + +struct hipExtent { + size_t width; // Width in elements when referring to array memory, in bytes when referring to linear memory + size_t height; + size_t depth; +}; + +struct hipPos { + size_t x; + size_t y; + size_t z; +}; + +struct hipMemcpy3DParms { + hipArray_t srcArray; + struct hipPos srcPos; + struct hipPitchedPtr srcPtr; + + hipArray_t dstArray; + struct hipPos dstPos; + struct hipPitchedPtr dstPtr; + + struct hipExtent extent; + enum hipMemcpyKind kind; +}; + +static __inline__ struct hipPitchedPtr make_hipPitchedPtr(void *d, size_t p, size_t xsz, size_t ysz) +{ + struct hipPitchedPtr s; + + s.ptr = d; + s.pitch = p; + s.xsize = xsz; + s.ysize = ysz; + + return s; +} + +static __inline__ struct hipPos make_hipPos(size_t x, size_t y, size_t z) +{ + struct hipPos p; + + p.x = x; + p.y = y; + p.z = z; + + return p; +} + +static __inline__ struct hipExtent make_hipExtent(size_t w, size_t h, size_t d) +{ + struct hipExtent e; + + e.width = w; + e.height = h; + e.depth = d; + + return e; +} #endif diff --git a/include/hip/hcc_detail/hip_runtime.h b/include/hip/hcc_detail/hip_runtime.h index 5536ded5e2..fecf79c496 100644 --- a/include/hip/hcc_detail/hip_runtime.h +++ b/include/hip/hcc_detail/hip_runtime.h @@ -81,12 +81,12 @@ namespace hip_impl extern int HIP_TRACE_API; #ifdef __cplusplus -//#include #include #endif #include #include #include +#include // TODO-HCC remove old definitions ; ~1602 hcc supports __HCC_ACCELERATOR__ define. #if defined (__KALMAR_ACCELERATOR__) && !defined (__HCC_ACCELERATOR__) diff --git a/include/hip/hcc_detail/hip_runtime_api.h b/include/hip/hcc_detail/hip_runtime_api.h index 2036aa5008..e28dbab82c 100644 --- a/include/hip/hcc_detail/hip_runtime_api.h +++ b/include/hip/hcc_detail/hip_runtime_api.h @@ -37,7 +37,8 @@ THE SOFTWARE. #include #include -#include +#include +#include #if defined (__HCC__) && (__hcc_workweek__ < 16155) #error("This version of HIP requires a newer version of HCC."); @@ -136,6 +137,11 @@ enum hipLimit_t #define hipDeviceMapHost 0x8 #define hipDeviceLmemResizeToMax 0x16 +#define hipArrayDefault 0x00 ///< Default HIP array allocation flag +#define hipArrayLayered 0x01 +#define hipArraySurfaceLoadStore 0x02 +#define hipArrayCubemap 0x04 +#define hipArrayTextureGather 0x08 /* * @brief hipJitOption @@ -200,27 +206,6 @@ typedef struct dim3 { } dim3; -/** - * Memory copy types - * - */ -typedef enum hipMemcpyKind { - hipMemcpyHostToHost = 0 ///< Host-to-Host Copy - ,hipMemcpyHostToDevice = 1 ///< Host-to-Device Copy - ,hipMemcpyDeviceToHost = 2 ///< Device-to-Host Copy - ,hipMemcpyDeviceToDevice =3 ///< Device-to-Device Copy - ,hipMemcpyDefault = 4, ///< Runtime will automatically determine copy-kind based on virtual addresses. -} hipMemcpyKind; - -typedef struct { - unsigned int width; - unsigned int height; - enum hipChannelFormatKind f; - void* data; //FIXME: generalize this -} hipArray; - - - // Doxygen end group GlobalDefs /** @} */ @@ -1287,6 +1272,19 @@ hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t s hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t stream); #endif +/** + * @brief Fills the memory area pointed to by dst with the constant value. + * + * @param[out] dst Pointer to device memory + * @param[in] pitch - data size in bytes + * @param[in] value - constant value to be set + * @param[in] width + * @param[in] height + * @return #hipSuccess, #hipErrorInvalidValue, #hipErrorMemoryFree + */ + +hipError_t hipMemset2D(void* dst, size_t pitch, int value, size_t width, size_t height); + /** * @brief Query memory info. * Return snapshot of free memory, and total allocatable memory on the device. @@ -1315,7 +1313,7 @@ hipError_t hipMemPtrGetInfo(void *ptr, size_t *size); */ #if __cplusplus hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, - size_t width, size_t height = 0, unsigned int flags = 0); + size_t width, size_t height = 0, unsigned int flags = hipArrayDefault); #else hipError_t hipMallocArray(hipArray** array, const struct hipChannelFormatDesc* desc, size_t width, size_t height, unsigned int flags); @@ -1330,6 +1328,22 @@ hipError_t hipMallocArray(hipArray** array, const struct hipChannelFormatDesc* d */ hipError_t hipFreeArray(hipArray* array); +/** + * @brief Allocate an array on the device. + * + * @param[out] array Pointer to allocated array in device memory + * @param[in] desc Requested channel format + * @param[in] extent Requested array allocation width, height and depth + * @param[in] flags Requested properties of allocated array + * @return #hipSuccess, #hipErrorMemoryAllocation + * + * @see hipMalloc, hipMallocPitch, hipFree, hipFreeArray, hipHostMalloc, hipHostFree + */ + +hipError_t hipMalloc3DArray(hipArray_t *array, + const struct hipChannelFormatDesc* desc, + struct hipExtent extent, + unsigned int flags); /** * @brief Copies data between host and device. * @@ -1402,6 +1416,7 @@ hipError_t hipMemcpyToArray(hipArray* dst, size_t wOffset, size_t hOffset, const void* src, size_t count, hipMemcpyKind kind); +hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p); // doxygen end Memory /** @@ -2150,6 +2165,24 @@ hipError_t hipIpcCloseMemHandle(void *devPtr); #endif #ifdef __cplusplus + +hipError_t hipBindTexture(size_t* offset, + textureReference* tex, + const void* devPtr, + const hipChannelFormatDesc* desc, + size_t size = UINT_MAX); + +hipError_t ihipBindTextureImpl(int dim, + enum hipTextureReadMode readMode, + size_t *offset, + const void *devPtr, + const struct hipChannelFormatDesc& desc, + size_t size, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords, + hipTextureObject_t& textureObject); + /* * @brief hipBindTexture Binds size bytes of the memory area pointed to by @p devPtr to the texture reference tex. * @@ -2164,15 +2197,15 @@ hipError_t hipIpcCloseMemHandle(void *devPtr); * @return #hipSuccess, #hipErrorInvalidValue, #hipErrorMemoryFree, #hipErrorUnknown **/ template -hipError_t hipBindTexture(size_t *offset, - struct texture &tex, - const void *devPtr, - const struct hipChannelFormatDesc *desc, - size_t size=UINT_MAX) +hipError_t hipBindTexture(size_t *offset, + struct texture& tex, + const void *devPtr, + const struct hipChannelFormatDesc& desc, + size_t size = UINT_MAX) { - tex._dataPtr = static_cast(devPtr); - - return hipSuccess; + return ihipBindTextureImpl(dim, readMode, offset, devPtr, desc, size, + tex.addressMode[0], tex.filterMode, tex.normalized, + tex.textureObject); } /* @@ -2188,20 +2221,115 @@ hipError_t hipBindTexture(size_t *offset, * @return #hipSuccess, #hipErrorInvalidValue, #hipErrorMemoryFree, #hipErrorUnknown **/ template -hipError_t hipBindTexture(size_t *offset, - struct texture &tex, - const void *devPtr, - size_t size=UINT_MAX) +hipError_t hipBindTexture(size_t *offset, + struct texture& tex, + const void *devPtr, + size_t size = UINT_MAX) { - return hipBindTexture(offset, tex, devPtr, &tex.channelDesc, size); + return ihipBindTextureImpl(dim, readMode, offset, devPtr, tex.channelDesc, size, + tex.addressMode[0], tex.filterMode, tex.normalized, + tex.textureObject); +} + +// C API +hipError_t hipBindTexture2D(size_t* offset, + textureReference* tex, + const void* devPtr, + const hipChannelFormatDesc* desc, + size_t width, + size_t height, + size_t pitch); + +hipError_t ihipBindTexture2DImpl(int dim, + enum hipTextureReadMode readMode, + size_t *offset, + const void *devPtr, + const struct hipChannelFormatDesc& desc, + size_t width, + size_t height, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords, + hipTextureObject_t& textureObject); + +template +hipError_t hipBindTexture2D(size_t *offset, + struct texture& tex, + const void *devPtr, + size_t width, + size_t height, + size_t pitch) +{ + return ihipBindTexture2DImpl(dim, readMode, offset, devPtr, tex.channelDesc, width, height, + tex.addressMode[0], tex.filterMode, tex.normalized, + tex.textureObject); } template -hipError_t hipBindTextureToArray(struct texture &tex, hipArray* array) { - tex.width = array->width; - tex.height = array->height; - tex._dataPtr = static_cast(array->data); - return hipSuccess; +hipError_t hipBindTexture2D(size_t *offset, + struct texture& tex, + const void *devPtr, + const struct hipChannelFormatDesc &desc, + size_t width, + size_t height, + size_t pitch) +{ + return ihipBindTexture2DImpl(dim, readMode, offset, devPtr, desc, width, height, + tex.addressMode[0], tex.filterMode, tex.normalized, + tex.textureObject); +} + +//C API +hipError_t hipBindTextureToArray(textureReference* tex, + hipArray_const_t array, + const hipChannelFormatDesc* desc); + +hipError_t ihipBindTextureToArrayImpl(int dim, + enum hipTextureReadMode readMode, + hipArray_const_t array, + const struct hipChannelFormatDesc& desc, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords, + hipTextureObject_t& textureObject); + +template +hipError_t hipBindTextureToArray(struct texture& tex, + hipArray_const_t array) +{ + return ihipBindTextureToArrayImpl(dim, readMode, array, tex.channelDesc, + tex.addressMode[0], tex.filterMode, tex.normalized, + tex.textureObject); +} + +template +hipError_t hipBindTextureToArray(struct texture& tex, + hipArray_const_t array, + const struct hipChannelFormatDesc& desc) +{ + return ihipBindTextureToArrayImpl(dim, readMode, array, desc, + tex.addressMode[0], tex.filterMode, tex.normalized, + tex.textureObject); +} + +//C API +hipError_t hipBindTextureToMipmappedArray(const textureReference* tex, + hipMipmappedArray_const_t mipmappedArray, + const hipChannelFormatDesc* desc); + +template +hipError_t hipBindTextureToMipmappedArray(const texture& tex, + hipMipmappedArray_const_t mipmappedArray) +{ + return hipSuccess; +} + +template +hipError_t hipBindTextureToMipmappedArray(const texture& tex, + hipMipmappedArray_const_t mipmappedArray, + const hipChannelFormatDesc& desc) +{ + return hipSuccess; } /* @@ -2211,15 +2339,30 @@ hipError_t hipBindTextureToArray(struct texture &tex, hipArray * * @return #hipSuccess **/ -template -hipError_t hipUnbindTexture(struct texture &tex) -{ - tex._dataPtr = NULL; +hipError_t hipUnbindTexture(const textureReference* tex); - return hipSuccess; +extern hipError_t ihipUnbindTextureImpl(const hipTextureObject_t& textureObject); + +template +hipError_t hipUnbindTexture(struct texture &tex) +{ + return ihipUnbindTextureImpl(tex.textureObject); } +hipError_t hipGetChannelDesc(hipChannelFormatDesc* desc, hipArray_const_t array); +hipError_t hipGetTextureAlignmentOffset (size_t* offset, const textureReference* texref); +hipError_t hipGetTextureReference(const textureReference** texref, const void* symbol); +hipError_t hipCreateTextureObject(hipTextureObject_t* pTexObject, + const hipResourceDesc* pResDesc, + const hipTextureDesc* pTexDesc, + const hipResourceViewDesc* pResViewDesc); + +hipError_t hipDestroyTextureObject(hipTextureObject_t textureObject); + +hipError_t hipGetTextureObjectResourceDesc(hipResourceDesc* pResDesc, hipTextureObject_t textureObject); +hipError_t hipGetTextureObjectResourceViewDesc(hipResourceViewDesc* pResViewDesc, hipTextureObject_t textureObject); +hipError_t hipGetTextureObjectTextureDesc(hipTextureDesc* pTexDesc, hipTextureObject_t textureObject); // doxygen end Texture /** diff --git a/include/hip/hcc_detail/hip_texture.h b/include/hip/hcc_detail/hip_texture.h deleted file mode 100644 index c6f5a1cfb2..0000000000 --- a/include/hip/hcc_detail/hip_texture.h +++ /dev/null @@ -1,107 +0,0 @@ -/* -Copyright (c) 2015 - present Advanced Micro Devices, Inc. All rights reserved. - -Permission is hereby granted, free of charge, to any person obtaining a copy -of this software and associated documentation files (the "Software"), to deal -in the Software without restriction, including without limitation the rights -to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -copies of the Software, and to permit persons to whom the Software is -furnished to do so, subject to the following conditions: - -The above copyright notice and this permission notice shall be included in -all copies or substantial portions of the Software. - -THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -THE SOFTWARE. -*/ - -//#pragma once - -#ifndef HIP_INCLUDE_HIP_HCC_DETAIL_HIP_TEXTURE_H -#define HIP_INCLUDE_HIP_HCC_DETAIL_HIP_TEXTURE_H - -/** - * @file hcc_detail/hip_texture.h - * @brief HIP C++ Texture API for hcc compiler - */ - -#include -#include -#include -#include -//#include - -//---- -//Texture - TODO - likely need to move this to a separate file only included with kernel compilation. -#define hipTextureType1D 1 - -#if __cplusplus -template -struct texture : public textureReference { - - const T * _dataPtr; // pointer to underlying data. - - //texture() : filterMode(hipFilterModePoint), normalized(false), _dataPtr(NULL) {}; - unsigned int width; - unsigned int height; - -}; -#endif - - -#define tex1Dfetch(_tex, _addr) (_tex._dataPtr[_addr]) - -#define tex2D(_tex, _dx, _dy) \ - _tex._dataPtr[(unsigned int)_dx + (unsigned int)_dy*(_tex.width)] - -/** - * @addtogroup API HIP API - * @{ - * - * Defines the HIP API. See the individual sections for more information. - */ - -// These are C++ APIs - maybe belong in separate file. -/** - *------------------------------------------------------------------------------------------------- - *------------------------------------------------------------------------------------------------- - * @defgroup Texture Texture Reference Management - * @{ - * - * - * @warning The HIP texture API implements a small subset of full texture API. Known limitations include: - * - Only point sampling is supported. - * - Only C++ APIs are provided. - * - Many APIs and modes are not implemented. - * - * The HIP texture support is intended to allow use of texture cache on hardware where this is beneficial. - * - * The following CUDA APIs are not currently supported: - * - cudaBindTexture2D - * - cudaBindTextureToArray - * - cudaBindTextureToMipmappedArray - * - cudaGetChannelDesc - * - cudaGetTextureReference - * - */ - -// C API: -#if 0 -hipChannelFormatDesc hipBindTexture(size_t *offset, struct textureReference *tex, const void *devPtr, const struct hipChannelFormatDesc *desc, size_t size=UINT_MAX) -{ - tex->_dataPtr = devPtr; -} -#endif - - -// End doxygen API: -/** - * @} - */ - -#endif diff --git a/include/hip/hcc_detail/hip_texture_types.h b/include/hip/hcc_detail/hip_texture_types.h new file mode 100644 index 0000000000..27f8747653 --- /dev/null +++ b/include/hip/hcc_detail/hip_texture_types.h @@ -0,0 +1,83 @@ +/* +Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN +THE SOFTWARE. +*/ + +/** + * @file hcc_detail/hip_texture_types.h + * @brief Defines the different newt vector types for HIP runtime. + */ + +#ifndef HIP_INCLUDE_HIP_HCC_DETAIL_HIP_TEXTURE_TYPES_H +#define HIP_INCLUDE_HIP_HCC_DETAIL_HIP_TEXTURE_TYPES_H + +/******************************************************************************* +* * +* * +* * +*******************************************************************************/ +#include +//#include +#include +#include + +#if __cplusplus + +/******************************************************************************* +* * +* * +* * +*******************************************************************************/ + +template +struct texture : public textureReference +{ + texture(int norm = 0, + enum hipTextureFilterMode fMode = hipFilterModePoint, + enum hipTextureAddressMode aMode = hipAddressModeClamp) + { + normalized = norm; + filterMode = fMode; + addressMode[0] = aMode; + addressMode[1] = aMode; + addressMode[2] = aMode; + channelDesc = hipCreateChannelDesc(); + sRGB = 0; + } + + texture(int norm, + enum hipTextureFilterMode fMode, + enum hipTextureAddressMode aMode, + struct hipChannelFormatDesc desc) + { + normalized = norm; + filterMode = fMode; + addressMode[0] = aMode; + addressMode[1] = aMode; + addressMode[2] = aMode; + channelDesc = desc; + sRGB = 0; + } +}; + +#endif /* __cplusplus */ + +#endif /* !HIP_INCLUDE_HIP_HCC_DETAIL_HIP_TEXTURE_TYPES_H */ + diff --git a/include/hip/hcc_detail/texture_functions.h b/include/hip/hcc_detail/texture_functions.h new file mode 100644 index 0000000000..d08b429fca --- /dev/null +++ b/include/hip/hcc_detail/texture_functions.h @@ -0,0 +1,6771 @@ +/* +Copyright (c) 2015 - present Advanced Micro Devices, Inc. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN +THE SOFTWARE. +*/ + +#ifndef HIP_INCLUDE_HIP_HCC_DETAIL_TEXTURE_FUNCTIONS_H +#define HIP_INCLUDE_HIP_HCC_DETAIL_TEXTURE_FUNCTIONS_H + +#include +#include +#include + +union TData { + hc::short_vector::float4::vector_value_type f; + hc::short_vector::int4::vector_value_type i; + hc::short_vector::uint4::vector_value_type u; +}; + +#define __TEXTURE_FUNCTIONS_DECL__ static __inline__ __device__ +#define ADDRESS_SPACE_2 __attribute__((address_space(2))) +#define TEXTURE_PARAMETERS_INIT \ + unsigned int ADDRESS_SPACE_2 *i = (unsigned int ADDRESS_SPACE_2*)textureObject; \ + unsigned int ADDRESS_SPACE_2 *s = i + HIP_SAMPLER_OBJECT_OFFSET_DWORD; \ + TData texel; + +#define TEXTURE_SET_FLOAT \ + *retVal = texel.f.x; + +#define TEXTURE_SET_SIGNED \ + *retVal = texel.i.x; + +#define TEXTURE_SET_UNSIGNED \ + *retVal = texel.u.x; + +#define TEXTURE_SET_FLOAT_X \ + retVal->x = texel.f.x; + +#define TEXTURE_SET_SIGNED_X \ + retVal->x = texel.i.x; + +#define TEXTURE_SET_UNSIGNED_X \ + retVal->x = texel.u.x; + +#define TEXTURE_SET_FLOAT_XY \ + retVal->x = texel.f.x; \ + retVal->y = texel.f.y; + +#define TEXTURE_SET_SIGNED_XY \ + retVal->x = texel.i.x; \ + retVal->y = texel.i.y; + +#define TEXTURE_SET_UNSIGNED_XY \ + retVal->x = texel.u.x; \ + retVal->y = texel.u.y; + +#define TEXTURE_SET_FLOAT_XYZW \ + retVal->x = texel.f.x; \ + retVal->y = texel.f.y; \ + retVal->z = texel.f.z; \ + retVal->w = texel.f.w; + +#define TEXTURE_SET_SIGNED_XYZW \ + retVal->x = texel.i.x; \ + retVal->y = texel.i.y; \ + retVal->z = texel.i.z; \ + retVal->w = texel.i.w; + +#define TEXTURE_SET_UNSIGNED_XYZW \ + retVal->x = texel.u.x; \ + retVal->y = texel.u.y; \ + retVal->z = texel.u.z; \ + retVal->w = texel.u.w; + +#define TEXTURE_RETURN_CHAR \ + return texel.i.x; + +#define TEXTURE_RETURN_UCHAR \ + return texel.u.x; + +#define TEXTURE_RETURN_SHORT \ + return texel.i.x; + +#define TEXTURE_RETURN_USHORT \ + return texel.u.x; + +#define TEXTURE_RETURN_INT \ + return texel.i.x; + +#define TEXTURE_RETURN_UINT \ + return texel.u.x; + +#define TEXTURE_RETURN_FLOAT \ + return texel.f.x; + +#define TEXTURE_RETURN_SIGNED \ + return texel.i.x; + +#define TEXTURE_RETURN_UNSIGNED \ + return texel.u.x; + +#define TEXTURE_RETURN_CHAR_X \ + return char1(texel.i.x); + +#define TEXTURE_RETURN_UCHAR_X \ + return uchar1(texel.u.x); + +#define TEXTURE_RETURN_SHORT_X \ + return short1(texel.i.x); + +#define TEXTURE_RETURN_USHORT_X \ + return ushort1(texel.u.x); + +#define TEXTURE_RETURN_INT_X \ + return int1(texel.i.x); + +#define TEXTURE_RETURN_UINT_X \ + return uint1(texel.u.x); + +#define TEXTURE_RETURN_FLOAT_X \ + return float1(texel.f.x); + +#define TEXTURE_RETURN_CHAR_XY \ + return char2(texel.i.x, texel.i.y); + +#define TEXTURE_RETURN_UCHAR_XY \ + return uchar2(texel.u.x, texel.u.y); + +#define TEXTURE_RETURN_SHORT_XY \ + return short2(texel.i.x, texel.i.y); + +#define TEXTURE_RETURN_USHORT_XY \ + return ushort2(texel.u.x, texel.u.y); + +#define TEXTURE_RETURN_INT_XY \ + return int2(texel.i.x, texel.i.y); + +#define TEXTURE_RETURN_UINT_XY \ + return uint2(texel.u.x, texel.u.y); + +#define TEXTURE_RETURN_FLOAT_XY \ + return float2(texel.f.x, texel.f.y); + +#define TEXTURE_RETURN_CHAR_XYZW \ + return char4(texel.i.x, texel.i.y, texel.i.z, texel.i.w); + +#define TEXTURE_RETURN_UCHAR_XYZW \ + return uchar4(texel.u.x, texel.u.y, texel.u.z, texel.u.w); + +#define TEXTURE_RETURN_SHORT_XYZW \ + return short4(texel.i.x, texel.i.y, texel.i.z, texel.i.w); + +#define TEXTURE_RETURN_USHORT_XYZW \ + return ushort4(texel.u.x, texel.u.y, texel.u.z, texel.u.w); + +#define TEXTURE_RETURN_INT_XYZW \ + return int4(texel.i.x, texel.i.y, texel.i.z, texel.i.w); + +#define TEXTURE_RETURN_UINT_XYZW \ + return uint4(texel.u.x, texel.u.y, texel.u.z, texel.u.w); + +#define TEXTURE_RETURN_FLOAT_XYZW \ + return float4(texel.f.x, texel.f.y, texel.f.z, texel.f.w); + +extern "C" { +hc::short_vector::float4::vector_value_type +__ockl_image_sample_1D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + float c) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_1Da(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_2D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c) [[hc]]; + + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_2Da(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c) [[hc]]; + +float +__ockl_image_sample_2Dad(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c) [[hc]]; + +float +__ockl_image_sample_2Dd(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_3D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_grad_1D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + float c, + float dx, + float dy) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_grad_1Da(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c, + float dx, + float dy) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_grad_2D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c, + hc::short_vector::float2::vector_value_type dx, + hc::short_vector::float2::vector_value_type dy) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_grad_2Da(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c, + hc::short_vector::float2::vector_value_type dx, + hc::short_vector::float2::vector_value_type dy) [[hc]]; + +float +__ockl_image_sample_grad_2Dad(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c, + hc::short_vector::float2::vector_value_type dx, + hc::short_vector::float2::vector_value_type dy) [[hc]]; + +float +__ockl_image_sample_grad_2Dd(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c, + hc::short_vector::float2::vector_value_type dx, + hc::short_vector::float2::vector_value_type dy) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_grad_3D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c, + hc::short_vector::float4::vector_value_type dx, + hc::short_vector::float4::vector_value_type dy) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_lod_1D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + float c, + float l) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_lod_1Da(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c, + float l) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_lod_2D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c, + float l) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_lod_2Da(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c, + float l) [[hc]]; + +float +__ockl_image_sample_lod_2Dad(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c, + float l) [[hc]]; + +float +__ockl_image_sample_lod_2Dd(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float2::vector_value_type c, + float l) [[hc]]; + +hc::short_vector::float4::vector_value_type +__ockl_image_sample_lod_3D(unsigned int ADDRESS_SPACE_2 * i, + unsigned int ADDRESS_SPACE_2 * s, + hc::short_vector::float4::vector_value_type c, + float l) [[hc]]; +} + +//////////////////////////////////////////////////////////// +// Texture object APIs +//////////////////////////////////////////////////////////// + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(char *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(char1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(char2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(char4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(unsigned char *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(uchar1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(uchar2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(uchar4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(short *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(short1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(short2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(short4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(unsigned short *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(ushort1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(ushort2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(ushort4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(int *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(int1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(int2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(int4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(unsigned int *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(uint1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(uint2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(uint4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(float *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(float1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(float2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1Dfetch(float4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex1Dfetch(hipTextureObject_t textureObject, int x) +{ + T ret; + tex1Dfetch(&ret, textureObject, x); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex1D(char *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(char1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(char2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(char4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(unsigned char *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(uchar1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(uchar2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(uchar4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(short *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(short1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(short2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(short4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(unsigned short *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(ushort1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(ushort2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(ushort4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(int *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(int1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(int2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(int4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(unsigned int *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(uint1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(uint2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(uint4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(float *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(float1 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(float2 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1D(float4 *retVal, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_SET_FLOAT_XYZW; +} +template +__TEXTURE_FUNCTIONS_DECL__ T tex1D(hipTextureObject_t textureObject, int x) +{ + T ret; + tex1D(&ret, textureObject, x); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(char *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(char1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(char2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(char4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(unsigned char *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(uchar1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(uchar2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(uchar4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(short *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(short1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(short2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(short4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(unsigned short *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(ushort1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(ushort2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(ushort4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(int *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(int1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(int2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(int4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(unsigned int *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(uint1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(uint2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(uint4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(float *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(float1 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(float2 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLod(float4 *retVal, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex1DLod(hipTextureObject_t textureObject, float x, float level) +{ + T ret; + tex1DLod(&ret, textureObject, x, level); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(char *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(char1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(char2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(char4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(unsigned char *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(uchar1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(uchar2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(uchar4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(short *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(short1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(short2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(short4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(unsigned short *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(ushort1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(ushort2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(ushort4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(int *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(int1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(int2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(int4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(unsigned int *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(uint1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(uint2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(uint4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(float *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(float1 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(float2 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DGrad(float4 *retVal, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex1DGrad(hipTextureObject_t textureObject, float x, float dx, float dy) +{ + T ret; + tex1DLod(&ret, textureObject, x, dx, dy); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex2D(char *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(char1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(char2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(char4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(unsigned char *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(uchar1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(uchar2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(uchar4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(short *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(short1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(short2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(short4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(unsigned short *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(ushort1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(ushort2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(ushort4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(int *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(int1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(int2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(int4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(unsigned int *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(uint1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(uint2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(uint4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(float *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(float1 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(float2 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2D(float4 *retVal, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex2D(hipTextureObject_t textureObject, float x, float y) +{ + T ret; + tex2D(&ret, textureObject, x, y); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(char *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(char1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(char2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(char4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(unsigned char *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(uchar1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(uchar2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(uchar4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(short *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(short1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(short2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(short4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(unsigned short *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(ushort1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(ushort2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(ushort4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(int *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(int1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(int2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(int4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(unsigned int *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(uint1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(uint2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(uint4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(float *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(float1 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(float2 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLod(float4 *retVal, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex2DLod(hipTextureObject_t textureObject, float x, float y, float level) +{ + T ret; + tex2DLod(&ret, textureObject, x, y, level); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex3D(char *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(char1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(char2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(char4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(unsigned char *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(uchar1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(uchar2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(uchar4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(short *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(short1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(short2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(short4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(unsigned short *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(ushort1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(ushort2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(ushort4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(int *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(int1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(int2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(int4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(unsigned int *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(uint1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(uint2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(uint4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(float *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(float1 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(float2 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3D(float4 *retVal, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex3D(hipTextureObject_t textureObject, float x, float y, float z) +{ + T ret; + tex3D(&ret, textureObject, x, y, z); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(char *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(char1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(char2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(char4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(unsigned char *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(uchar1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(uchar2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(uchar4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(short *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(short1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(short2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(short4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(unsigned short *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(ushort1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(ushort2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(ushort4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(int *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(int1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(int2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(int4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(unsigned int *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(uint1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(uint2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(uint4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(float *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(float1 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(float2 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex3DLod(float4 *retVal, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex3DLod(hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + T ret; + tex3DLod(&ret, textureObject, x, y, z, level); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(char *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(char1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(char2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(char4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(unsigned char *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(uchar1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(uchar2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(uchar4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(short *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(short1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(short2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(short4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(unsigned short *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(ushort1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(ushort2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(ushort4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(int *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(int1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(int2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(int4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(unsigned int *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(uint1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(uint2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(uint4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(float *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(float1 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(float2 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_FLOAT_XY; +} +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayered(float4 *retVal, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex1DLayered(hipTextureObject_t textureObject, float x, int layer) +{ + T ret; + tex1DLayered(&ret, textureObject, x, layer); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(char *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(char1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(char2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(char4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(unsigned char *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(uchar1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(uchar2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(uchar4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(short *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(short1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(short2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(short4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(unsigned short *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(ushort1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(ushort2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(ushort4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(int *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(int1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(int2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(int4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(unsigned int *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(uint1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(uint2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(uint4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(float *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(float1 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(float2 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredLod(float4 *retVal, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex1DLayeredLod(hipTextureObject_t textureObject, float x, int layer, float level) +{ + T ret; + tex1DLayeredLod(&ret, textureObject, x, layer, level); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(char *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(char1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(char2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(char4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(unsigned char *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(uchar1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(uchar2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(uchar4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(short *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(short1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(short2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(short4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(unsigned short *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(ushort1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(ushort2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(ushort4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(int *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(int1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(int2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(int4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(unsigned int *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(uint1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(uint2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(uint4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(float *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(float1 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(float2 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex1DLayeredGrad(float4 *retVal, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex1DLayeredGrad(hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + T ret; + tex1DLayeredGrad(&ret, textureObject, x, layer, dx, dy); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(char *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(char1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(char2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(char4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(unsigned char *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(uchar1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(uchar2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(uchar4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(short *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(short1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(short2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(short4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(unsigned short *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(ushort1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(ushort2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(ushort4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(int *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(int1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(int2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(int4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(unsigned int *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(uint1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(uint2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(uint4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(float *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(float1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(float2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayered(float4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex2DLayered(hipTextureObject_t textureObject, float x, float y, int layer) +{ + T ret; + tex2DLayered(&ret, textureObject, x, y, layer); + return ret; +} + +//////////////////////////////////////////////////////////// +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(char *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(char1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(char2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(char4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(unsigned char *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(uchar1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(uchar2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(uchar4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(short *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(short1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(short2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(short4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(unsigned short *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(ushort1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(ushort2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(ushort4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(int *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(int1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(int2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(int4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_SIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(unsigned int *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(uint1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(uint2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(uint4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_UNSIGNED_XYZW; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(float *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(float1 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT_X; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(float2 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT_XY; +} + +__TEXTURE_FUNCTIONS_DECL__ void tex2DLayeredLod(float4 *retVal, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_SET_FLOAT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ T tex2DLayeredLod(hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + T ret; + tex2DLayeredLod(&ret, textureObject, x, y, layer, level); + return ret; +} + +//////////////////////////////////////////////////////////// +// Texture Reference APIs +//////////////////////////////////////////////////////////// +template +__TEXTURE_FUNCTIONS_DECL__ char tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1Dfetch(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// +template +__TEXTURE_FUNCTIONS_DECL__ char tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1D(texture texRef, hipTextureObject_t textureObject, int x) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1D(i, s, x); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_CHAR_X; +} +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1DLod(texture texRef, hipTextureObject_t textureObject, float x, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1D(i, s, x, level); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1DGrad(texture texRef, hipTextureObject_t textureObject, float x, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1D(i, s, x, dx, dy); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex2D(texture texRef, hipTextureObject_t textureObject, float x, float y) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2D(i, s, hc::short_vector::float2(x, y).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex2DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2D(i, s, hc::short_vector::float2(x, y).get_vector(), level); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex2DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2D(i, s, + hc::short_vector::float2(x, y).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex3D(texture texRef, hipTextureObject_t textureObject, float x, float y, float z) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex3DLod(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_3D(i, s, hc::short_vector::float4(x, y, z, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// +template +__TEXTURE_FUNCTIONS_DECL__ char tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex3DGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, float z, float4 dx, float4 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f =__ockl_image_sample_grad_3D(i, s, + hc::short_vector::float4(x, y, z, 0.0f).get_vector(), + hc::short_vector::float4(dx.x, dx.y, dx.z, dx.w).get_vector(), + hc::short_vector::float4(dy.x, dy.y, dy.z, dy.w).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1DLayered(texture texRef, hipTextureObject_t textureObject, float x, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_1Da(i, s, hc::short_vector::float2(x, layer).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), level); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex1DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, int layer, float dx, float dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_1Da(i, s, hc::short_vector::float2(x, layer).get_vector(), dx, dy); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex2DLayered(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex2DLayeredLod(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float level) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_lod_2Da(i, s, hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), level); + TEXTURE_RETURN_FLOAT_XYZW; +} + +//////////////////////////////////////////////////////////// + +template +__TEXTURE_FUNCTIONS_DECL__ char tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ char4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_CHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned char tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uchar4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UCHAR_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ short4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_SHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned short tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ ushort4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_USHORT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ int4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_INT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ unsigned int tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ uint4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_UINT_XYZW; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float1 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT_X; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float2 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT_XY; +} + +template +__TEXTURE_FUNCTIONS_DECL__ float4 tex2DLayeredGrad(texture texRef, hipTextureObject_t textureObject, float x, float y, int layer, float2 dx, float2 dy) +{ + TEXTURE_PARAMETERS_INIT; + texel.f = __ockl_image_sample_grad_2Da(i, s, + hc::short_vector::float4(x, y, layer, 0.0f).get_vector(), + hc::short_vector::float2(dx.x, dx.y).get_vector(), + hc::short_vector::float2(dy.x, dy.y).get_vector()); + TEXTURE_RETURN_FLOAT_XYZW; +} +#endif diff --git a/include/hip/hcc_detail/texture_types.h b/include/hip/hcc_detail/texture_types.h index 74680bbc76..731ed12308 100644 --- a/include/hip/hcc_detail/texture_types.h +++ b/include/hip/hcc_detail/texture_types.h @@ -26,20 +26,91 @@ THE SOFTWARE. #include -enum hipTextureReadMode +#define hipTextureType1D 0x01 +#define hipTextureType2D 0x02 +#define hipTextureType3D 0x03 +#define hipTextureTypeCubemap 0x0C +#define hipTextureType1DLayered 0xF1 +#define hipTextureType2DLayered 0xF2 +#define hipTextureTypeCubemapLayered 0xFC + +/** + * Should be same as HSA_IMAGE_OBJECT_SIZE_DWORD/HSA_SAMPLER_OBJECT_SIZE_DWORD + */ +#define HIP_IMAGE_OBJECT_SIZE_DWORD 12 +#define HIP_SAMPLER_OBJECT_SIZE_DWORD 8 +#define HIP_SAMPLER_OBJECT_OFFSET_DWORD HIP_IMAGE_OBJECT_SIZE_DWORD +#define HIP_TEXTURE_OBJECT_SIZE_DWORD (HIP_IMAGE_OBJECT_SIZE_DWORD + HIP_SAMPLER_OBJECT_SIZE_DWORD) + +/** + * An opaque value that represents a hip texture object + */ +typedef unsigned long long hipTextureObject_t; + +/** + * hip texture address modes + */ +enum hipTextureAddressMode { - hipReadModeElementType = 0 + hipAddressModeWrap = 0, + hipAddressModeClamp = 1, + hipAddressModeMirror = 2, + hipAddressModeBorder = 3 }; +/** + * hip texture filter modes + */ enum hipTextureFilterMode { - hipFilterModePoint = 0 + hipFilterModePoint = 0, + hipFilterModeLinear = 1 }; -struct textureReference { - enum hipTextureFilterMode filterMode; - unsigned normalized; - struct hipChannelFormatDesc channelDesc; +/** + * hip texture read modes + */ +enum hipTextureReadMode +{ + hipReadModeElementType = 0, + hipReadModeNormalizedFloat = 1 +}; + +/** + * hip texture reference + */ +struct textureReference +{ + int normalized; + enum hipTextureFilterMode filterMode; + enum hipTextureAddressMode addressMode[3]; //Texture address mode for up to 3 dimensions + struct hipChannelFormatDesc channelDesc; + int sRGB; // Perform sRGB->linear conversion during texture read + unsigned int maxAnisotropy; // Limit to the anisotropy ratio + enum hipTextureFilterMode mipmapFilterMode; + float mipmapLevelBias; + float minMipmapLevelClamp; + float maxMipmapLevelClamp; + + hipTextureObject_t textureObject; +}; + +/** + * hip texture descriptor + */ +struct hipTextureDesc +{ + enum hipTextureAddressMode addressMode[3]; //Texture address mode for up to 3 dimensions + enum hipTextureFilterMode filterMode; + enum hipTextureReadMode readMode; + int sRGB; // Perform sRGB->linear conversion during texture read + float borderColor[4]; + int normalizedCoords; + unsigned int maxAnisotropy; + enum hipTextureFilterMode mipmapFilterMode; + float mipmapLevelBias; + float minMipmapLevelClamp; + float maxMipmapLevelClamp; }; #endif diff --git a/include/hip/hip_texture.h b/include/hip/hip_texture_types.h similarity index 83% rename from include/hip/hip_texture.h rename to include/hip/hip_texture_types.h index a15c5a1016..44dc7a9e0f 100644 --- a/include/hip/hip_texture.h +++ b/include/hip/hip_texture_types.h @@ -1,5 +1,5 @@ /* -Copyright (c) 2015 - present Advanced Micro Devices, Inc. All rights reserved. +Copyright (c) 2015-2017 Advanced Micro Devices, Inc. All rights reserved. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal @@ -20,13 +20,15 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ -#ifndef HIP_INCLUDE_HIP_HIP_TEXTURE_H -#define HIP_INCLUDE_HIP_HIP_TEXTURE_H + + +#ifndef HIP_INCLUDE_HIP_HIP_TEXTURE_TYPES_H +#define HIP_INCLUDE_HIP_HIP_TEXTURE_TYPES_H #if defined(__HIP_PLATFORM_HCC__) && !defined (__HIP_PLATFORM_NVCC__) -#include +#include #elif defined(__HIP_PLATFORM_NVCC__) && !defined (__HIP_PLATFORM_HCC__) -#include +#include #else #error("Must define exactly one of __HIP_PLATFORM_HCC__ or __HIP_PLATFORM_NVCC__"); #endif diff --git a/include/hip/nvcc_detail/hip_texture_types.h b/include/hip/nvcc_detail/hip_texture_types.h new file mode 100644 index 0000000000..751dd8e4d4 --- /dev/null +++ b/include/hip/nvcc_detail/hip_texture_types.h @@ -0,0 +1,6 @@ +#ifndef HIP_INCLUDE_HIP_NVCC_DETAIL_HIP_TEXTURE_TYPES_H +#define HIP_INCLUDE_HIP_NVCC_DETAIL_HIP_TEXTURE_TYPES_H + +#include + +#endif diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 6728ae42d3..0e557a5621 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -27,10 +27,6 @@ THE SOFTWARE. #include "hip/hip_runtime.h" #include "hip_hcc_internal.h" #include "trace_helper.h" -#include "hip/hcc_detail/hip_texture.h" -#include - - // Internal HIP APIS: namespace hip_internal { @@ -377,10 +373,11 @@ hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, auto ctx = ihipGetTlsDefaultCtx(); *array = (hipArray*)malloc(sizeof(hipArray)); + array[0]->type = flags; array[0]->width = width; array[0]->height = height; - - array[0]->f = desc->f; + array[0]->depth = 1; + array[0]->desc = *desc; void ** ptr = &array[0]->data; @@ -418,6 +415,62 @@ hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, return ihipLogStatus(hip_status); } +hipError_t hipMalloc3DArray(hipArray_t *array, + const struct hipChannelFormatDesc* desc, + struct hipExtent extent, + unsigned int flags) +{ + HIP_INIT(); + HIP_SET_DEVICE(); + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + + *array = (hipArray*)malloc(sizeof(hipArray)); + array[0]->type = flags; + array[0]->width = extent.width; + array[0]->height = extent.height; + array[0]->depth = extent.depth; + array[0]->desc = *desc; + + void ** ptr = &array[0]->data; + + if (ctx) { + const unsigned am_flags = 0; + const size_t size = extent.width*extent.height*extent.depth; + + size_t allocSize = 0; + switch(desc->f) { + case hipChannelFormatKindSigned: + allocSize = size * sizeof(int); + break; + case hipChannelFormatKindUnsigned: + allocSize = size * sizeof(unsigned int); + break; + case hipChannelFormatKindFloat: + allocSize = size * sizeof(float); + break; + case hipChannelFormatKindNone: + allocSize = size * sizeof(size_t); + break; + default: + hip_status = hipErrorUnknown; + break; + } + *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false, am_flags, 0); + + if (size && (*ptr == NULL)) { + hip_status = hipErrorMemoryAllocation; + } + + } else { + hip_status = hipErrorMemoryAllocation; + } + + //return ihipLogStatus(hip_status); + return hip_status; +} + hipError_t hipHostGetFlags(unsigned int* flagsPtr, void* hostPtr) { HIP_INIT_API(flagsPtr, hostPtr); @@ -857,7 +910,7 @@ hipError_t hipMemcpy2DToArray(hipArray* dst, size_t wOffset, size_t hOffset, con size_t byteSize; if(dst) { - switch(dst[0].f) { + switch(dst[0].desc.f) { case hipChannelFormatKindSigned: byteSize = sizeof(int); break; @@ -918,6 +971,56 @@ hipError_t hipMemcpyToArray(hipArray* dst, size_t wOffset, size_t hOffset, return ihipLogStatus(e); } +hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p) +{ + HIP_INIT_SPECIAL_API((TRACE_MCMD), p); + + hipStream_t stream = ihipSyncAndResolveStream(hipStreamNull); + + hc::completion_future marker; + + hipError_t e = hipSuccess; + + size_t byteSize; + if(p) { + switch(p->dstArray->desc.f) { + case hipChannelFormatKindSigned: + byteSize = sizeof(int); + break; + case hipChannelFormatKindUnsigned: + byteSize = sizeof(unsigned int); + break; + case hipChannelFormatKindFloat: + byteSize = sizeof(float); + break; + case hipChannelFormatKindNone: + byteSize = sizeof(size_t); + break; + default: + byteSize = 0; + break; + } + } else { + return ihipLogStatus(hipErrorUnknown); + } + + try { + for (int i = 0; i < p->extent.depth; i++) { + for(int j = 0; j < p->extent.height; j++) { + // TODO: p->srcPos or p->dstPos are not 0. + unsigned char* src = (unsigned char*)p->srcPtr.ptr + i*p->srcPtr.ysize*p->srcPtr.pitch + j*p->srcPtr.pitch; + unsigned char* dst = (unsigned char*)p->dstArray->data + i*p->dstArray->height*p->dstArray->width*byteSize + j*p->dstArray->width*byteSize; + stream->locked_copySync(dst, src, p->extent.width*byteSize, p->kind); + } + } + } + catch (ihipException ex) { + e = ex._code; + } + + return ihipLogStatus(e); +} + // TODO - make member function of stream? template void @@ -1006,7 +1109,7 @@ hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t s return ihipLogStatus(e); }; -hipError_t hipMemset(void* dst, int value, size_t sizeBytes ) +hipError_t hipMemset(void* dst, int value, size_t sizeBytes) { HIP_INIT_SPECIAL_API((TRACE_MCMD), dst, value, sizeBytes); @@ -1058,6 +1161,58 @@ hipError_t hipMemset(void* dst, int value, size_t sizeBytes ) return ihipLogStatus(e); } +hipError_t hipMemset2D(void* dst, size_t pitch, int value, size_t width, size_t height) +{ + HIP_INIT_SPECIAL_API((TRACE_MCMD), dst, pitch, value, width, height); + + hipError_t e = hipSuccess; + + hipStream_t stream = hipStreamNull; + // TODO - call an ihip memset so HIP_TRACE is correct. + stream = ihipSyncAndResolveStream(stream); + + if (stream) { + auto crit = stream->lockopen_preKernelCommand(); + + hc::completion_future cf ; + + size_t sizeBytes = pitch * height; + if ((sizeBytes & 0x3) == 0) { + // use a faster dword-per-workitem copy: + try { + value = value & 0xff; + uint32_t value32 = (value << 24) | (value << 16) | (value << 8) | (value) ; + ihipMemsetKernel (stream, crit, static_cast (dst), value32, sizeBytes/sizeof(uint32_t), &cf); + } + catch (std::exception &ex) { + e = hipErrorInvalidValue; + } + } else { + // use a slow byte-per-workitem copy: + try { + ihipMemsetKernel (stream, crit, static_cast (dst), value, sizeBytes, &cf); + } + catch (std::exception &ex) { + e = hipErrorInvalidValue; + } + } + // TODO - is hipMemset supposed to be async? + cf.wait(); + + stream->lockclose_postKernelCommand("hipMemset", &crit->_av); + + if (HIP_LAUNCH_BLOCKING) { + tprintf (DB_SYNC, "'%s' LAUNCH_BLOCKING wait for memset in %s.\n", __func__, ToString(stream).c_str()); + cf.wait(); + tprintf (DB_SYNC, "'%s' LAUNCH_BLOCKING memset completed in %s.\n", __func__, ToString(stream).c_str()); + } + } else { + e = hipErrorInvalidValue; + } + + return ihipLogStatus(e); +} + hipError_t hipMemsetD8(hipDeviceptr_t dst, unsigned char value, size_t sizeBytes ) { HIP_INIT_SPECIAL_API((TRACE_MCMD), dst, value, sizeBytes); @@ -1108,7 +1263,7 @@ hipError_t hipMemsetD8(hipDeviceptr_t dst, unsigned char value, size_t sizeByte return ihipLogStatus(e); } -hipError_t hipMemGetInfo (size_t *free, size_t *total) +hipError_t hipMemGetInfo(size_t *free, size_t *total) { HIP_INIT_API(free, total); diff --git a/src/hip_texture.cpp b/src/hip_texture.cpp new file mode 100644 index 0000000000..656c160f00 --- /dev/null +++ b/src/hip_texture.cpp @@ -0,0 +1,668 @@ + +#include + +#include + +#include "hsa/hsa.h" +#include "hsa/hsa_ext_amd.h" + +#include "hip/hip_runtime.h" +#include "hip_hcc_internal.h" +#include "trace_helper.h" + +#include "hip_texture.h" + +static std::map textureHash; + +void saveTextureInfo(const hipTexture* pTexture, + const hipResourceDesc* pResDesc, + const hipTextureDesc* pTexDesc, + const hipResourceViewDesc* pResViewDesc) +{ + if (pResDesc != nullptr) { + memcpy((void*)&(pTexture->resDesc), (void*)pResDesc, sizeof(hipResourceDesc)); + } + + if (pTexDesc != nullptr) { + memcpy((void*)&(pTexture->texDesc), (void*)pTexDesc, sizeof(hipTextureDesc)); + } + + if (pResViewDesc != nullptr) { + memcpy((void*)&(pTexture->resViewDesc), (void*)pResViewDesc, sizeof(hipResourceViewDesc)); + } +} + +void getChannelOrderAndType(const hipChannelFormatDesc& desc, + enum hipTextureReadMode readMode, + hsa_ext_image_channel_order_t& channelOrder, + hsa_ext_image_channel_type_t& channelType) +{ + if (desc.x != 0 && desc.y != 0 && desc.z != 0 && desc.w != 0) { + channelOrder = HSA_EXT_IMAGE_CHANNEL_ORDER_RGBA; + } else if (desc.x != 0 && desc.y != 0 && desc.z != 0 && desc.w == 0) { + channelOrder = HSA_EXT_IMAGE_CHANNEL_ORDER_RGB; + } else if (desc.x != 0 && desc.y != 0 && desc.z == 0 && desc.w == 0) { + channelOrder = HSA_EXT_IMAGE_CHANNEL_ORDER_RG; + } else if (desc.x != 0 && desc.y == 0 && desc.z == 0 && desc.w == 0) { + channelOrder = HSA_EXT_IMAGE_CHANNEL_ORDER_R; + } else { + } + + switch (desc.f) { + case hipChannelFormatKindUnsigned: + switch(desc.x) { + case 32: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_UNSIGNED_INT32; + break; + case 16: + channelType = readMode == hipReadModeNormalizedFloat ? HSA_EXT_IMAGE_CHANNEL_TYPE_UNORM_INT16 : + HSA_EXT_IMAGE_CHANNEL_TYPE_UNSIGNED_INT16; + break; + case 8: + channelType = readMode == hipReadModeNormalizedFloat ? HSA_EXT_IMAGE_CHANNEL_TYPE_UNORM_INT8 : + HSA_EXT_IMAGE_CHANNEL_TYPE_UNSIGNED_INT8; + break; + default: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_UNSIGNED_INT32; + } + break; + case hipChannelFormatKindSigned: + switch(desc.x) { + case 32: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_SIGNED_INT32; + break; + case 16: + channelType = readMode == hipReadModeNormalizedFloat ? HSA_EXT_IMAGE_CHANNEL_TYPE_SNORM_INT16 : + HSA_EXT_IMAGE_CHANNEL_TYPE_SIGNED_INT16; + break; + case 8: + channelType = readMode == hipReadModeNormalizedFloat ? HSA_EXT_IMAGE_CHANNEL_TYPE_SNORM_INT8 : + HSA_EXT_IMAGE_CHANNEL_TYPE_SIGNED_INT8; + break; + default: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_SIGNED_INT32; + } + break; + case hipChannelFormatKindFloat: + switch(desc.x) { + case 32: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_FLOAT; + break; + case 16: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_HALF_FLOAT; + break; + case 8: + break; + default: + channelType = HSA_EXT_IMAGE_CHANNEL_TYPE_FLOAT; + } + break; + case hipChannelFormatKindNone: + default: + break; + } +} + +void fillSamplerDescriptor(hsa_ext_sampler_descriptor_t& samplerDescriptor, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords) +{ + if (normalizedCoords) { + samplerDescriptor.coordinate_mode = HSA_EXT_SAMPLER_COORDINATE_MODE_NORMALIZED; + } else { + samplerDescriptor.coordinate_mode = HSA_EXT_SAMPLER_COORDINATE_MODE_UNNORMALIZED; + } + + switch (filterMode) { + case hipFilterModePoint: + samplerDescriptor.filter_mode = HSA_EXT_SAMPLER_FILTER_MODE_NEAREST; + break; + case hipFilterModeLinear: + samplerDescriptor.filter_mode = HSA_EXT_SAMPLER_FILTER_MODE_LINEAR; + break; + } + + switch (addressMode) { + case hipAddressModeWrap: + samplerDescriptor.address_mode = HSA_EXT_SAMPLER_ADDRESSING_MODE_REPEAT; + break; + case hipAddressModeClamp: + samplerDescriptor.address_mode = HSA_EXT_SAMPLER_ADDRESSING_MODE_CLAMP_TO_EDGE; + break; + case hipAddressModeMirror: + samplerDescriptor.address_mode = HSA_EXT_SAMPLER_ADDRESSING_MODE_MIRRORED_REPEAT; + break; + case hipAddressModeBorder: + samplerDescriptor.address_mode = HSA_EXT_SAMPLER_ADDRESSING_MODE_CLAMP_TO_BORDER; + break; + } +} + +bool getHipTextureObject(hipTextureObject_t* pTexObject, + hsa_ext_image_t& image, + hsa_ext_sampler_t sampler) +{ + unsigned int* texSRD; + hipMalloc((void **) &texSRD, HIP_TEXTURE_OBJECT_SIZE_DWORD * 4); + hipMemcpy(texSRD, (void *)image.handle, HIP_IMAGE_OBJECT_SIZE_DWORD * 4, hipMemcpyDeviceToDevice); + hipMemcpy(texSRD + HIP_SAMPLER_OBJECT_OFFSET_DWORD, (void *)sampler.handle, HIP_SAMPLER_OBJECT_SIZE_DWORD * 4, hipMemcpyDeviceToDevice); + *pTexObject = (hipTextureObject_t) texSRD; + +#ifdef DEBUG + unsigned int* srd = (unsigned int*) malloc(HIP_TEXTURE_OBJECT_SIZE_DWORD * 4); + hipMemcpy(srd, texSRD, HIP_TEXTURE_OBJECT_SIZE_DWORD * 4, hipMemcpyDeviceToHost); + printf("New SRD: \n"); + for (int i = 0; i < HIP_TEXTURE_OBJECT_SIZE_DWORD; i++) { + printf("SRD[%d]: %x\n", i, srd[i]); + } + printf("\n"); +#endif + return true; +} + +// Texture Object APIs +hipError_t hipCreateTextureObject(hipTextureObject_t* pTexObject, + const hipResourceDesc* pResDesc, + const hipTextureDesc* pTexDesc, + const hipResourceViewDesc* pResViewDesc) +{ + HIP_INIT_API(pTexObject, pResDesc, pTexDesc, pResViewDesc); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hc::accelerator acc = ctx->getDevice()->_acc; + auto device = ctx->getWriteableDevice(); + + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + hipTexture* pTexture = (hipTexture*) malloc(sizeof(hipTexture)); + if (pTexture != nullptr) { + memset(pTexture, 0, sizeof(hipTexture)); + saveTextureInfo(pTexture, pResDesc, pTexDesc, pResViewDesc); + } + + hsa_ext_image_descriptor_t imageDescriptor; + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + void* devPtr = nullptr; + + switch (pResDesc->resType) { + case hipResourceTypeArray: + devPtr = pResDesc->res.array.array->data; + imageDescriptor.width = pResDesc->res.array.array->width; + imageDescriptor.height = pResDesc->res.array.array->height; + switch (pResDesc->res.array.array->type) { + case hipArrayLayered: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2DA; + imageDescriptor.depth = 0; + imageDescriptor.array_size = pResDesc->res.array.array->depth; + break; + case hipArrayCubemap: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_3D; + imageDescriptor.depth = pResDesc->res.array.array->depth; + imageDescriptor.array_size = 0; + break; + case hipArraySurfaceLoadStore: + case hipArrayTextureGather: + case hipArrayDefault: + default: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + break; + } + getChannelOrderAndType(pResDesc->res.array.array->desc, pTexDesc->readMode, channelOrder, channelType); + break; + case hipResourceTypeMipmappedArray: + devPtr = pResDesc->res.mipmap.mipmap->data; + imageDescriptor.width = pResDesc->res.mipmap.mipmap->width; + imageDescriptor.height = pResDesc->res.mipmap.mipmap->height; + imageDescriptor.depth = pResDesc->res.mipmap.mipmap->depth; + imageDescriptor.array_size = 0; + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + getChannelOrderAndType(pResDesc->res.mipmap.mipmap->desc, pTexDesc->readMode, channelOrder, channelType); + break; + case hipResourceTypeLinear: + devPtr = pResDesc->res.linear.devPtr; + imageDescriptor.width = pResDesc->res.linear.sizeInBytes; + imageDescriptor.height = 1; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_1D; // ? HSA_EXT_IMAGE_DATA_LAYOUT_LINEAR + getChannelOrderAndType(pResDesc->res.linear.desc, pTexDesc->readMode, channelOrder, channelType); + break; + case hipResourceTypePitch2D: + devPtr = pResDesc->res.pitch2D.devPtr; + imageDescriptor.width = pResDesc->res.pitch2D.width; + imageDescriptor.height = pResDesc->res.pitch2D.height; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + getChannelOrderAndType(pResDesc->res.pitch2D.desc, pTexDesc->readMode, channelOrder, channelType); + break; + default: + break; + } + + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; + + hsa_ext_sampler_descriptor_t samplerDescriptor; + fillSamplerDescriptor(samplerDescriptor, pTexDesc->addressMode[0], pTexDesc->filterMode, pTexDesc->normalizedCoords); + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + if (HSA_STATUS_SUCCESS != hsa_ext_image_create(*agent, &imageDescriptor, devPtr, permission, &(pTexture->image)) || + HSA_STATUS_SUCCESS != hsa_ext_sampler_create(*agent, &samplerDescriptor, &(pTexture->sampler))) { + return ihipLogStatus(hipErrorRuntimeOther); + } + + getHipTextureObject(pTexObject, pTexture->image, pTexture->sampler); + + textureHash[*pTexObject] = pTexture; + } + + return ihipLogStatus(hip_status); +} + +hipError_t hipDestroyTextureObject(hipTextureObject_t textureObject) +{ + HIP_INIT_API(textureObject); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hc::accelerator acc = ctx->getDevice()->_acc; + auto device = ctx->getWriteableDevice(); + + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + hipTexture* pTexture = textureHash[textureObject]; + if (pTexture != nullptr) { + hsa_ext_image_destroy(*agent, pTexture->image); + hsa_ext_sampler_destroy(*agent, pTexture->sampler); + free(pTexture); + textureHash.erase(textureObject); + } + } + return ihipLogStatus(hip_status); +} + +hipError_t hipGetTextureObjectResourceDesc(hipResourceDesc* pResDesc, hipTextureObject_t textureObject) +{ + HIP_INIT_API(pResDesc, textureObject); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hipTexture* pTexture = textureHash[textureObject]; + if (pTexture != nullptr && pResDesc != nullptr) { + memcpy((void*)pResDesc, (void*)&(pTexture->resDesc), sizeof(hipResourceDesc)); + } + } + return ihipLogStatus(hip_status); +} + +hipError_t hipGetTextureObjectResourceViewDesc(hipResourceViewDesc* pResViewDesc, hipTextureObject_t textureObject) +{ + HIP_INIT_API(pResViewDesc, textureObject); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hipTexture* pTexture = textureHash[textureObject]; + if (pTexture != nullptr && pResViewDesc != nullptr) { + memcpy((void*)pResViewDesc, (void*)&(pTexture->resViewDesc), sizeof(hipResourceViewDesc)); + } + } + return ihipLogStatus(hip_status); +} + +hipError_t hipGetTextureObjectTextureDesc(hipTextureDesc* pTexDesc, hipTextureObject_t textureObject) +{ + HIP_INIT_API(pTexDesc, textureObject); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hipTexture* pTexture = textureHash[textureObject]; + if (pTexture != nullptr && pTexDesc != nullptr) { + memcpy((void*)pTexDesc, (void*)&(pTexture->texDesc), sizeof(hipTextureDesc)); + } + } + return ihipLogStatus(hip_status); +} + +// Texture Reference APIs +hipError_t ihipBindTextureImpl(int dim, + enum hipTextureReadMode readMode, + size_t *offset, + const void *devPtr, + const struct hipChannelFormatDesc& desc, + size_t size, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords, + hipTextureObject_t& textureObject) +{ + HIP_INIT_API(); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hc::accelerator acc = ctx->getDevice()->_acc; + auto device = ctx->getWriteableDevice(); + + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + hipTexture* pTexture = (hipTexture*) malloc(sizeof(hipTexture)); + if (pTexture != nullptr) { + memset(pTexture, 0, sizeof(hipTexture)); + } + + hsa_ext_image_descriptor_t imageDescriptor; + + assert(dim == hipTextureType1D); + + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_1D; + imageDescriptor.width = size; + imageDescriptor.height = 1; + imageDescriptor.depth = 1; + imageDescriptor.array_size = 0; + + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + getChannelOrderAndType(desc, readMode, channelOrder, channelType); + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; + + hsa_ext_sampler_descriptor_t samplerDescriptor; + fillSamplerDescriptor(samplerDescriptor, addressMode, filterMode, normalizedCoords); + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + + if (HSA_STATUS_SUCCESS != hsa_ext_image_create(*agent, &imageDescriptor, devPtr, permission, &(pTexture->image)) || + HSA_STATUS_SUCCESS != hsa_ext_sampler_create(*agent, &samplerDescriptor, &(pTexture->sampler))) { + return ihipLogStatus(hipErrorRuntimeOther); + } + getHipTextureObject(&textureObject, pTexture->image, pTexture->sampler); + textureHash[textureObject] = pTexture; + } + + return ihipLogStatus(hip_status); +} + +hipError_t hipBindTexture(size_t* offset, + textureReference* tex, + const void* devPtr, + const hipChannelFormatDesc* desc, + size_t size) +{ + // TODO: hipReadModeElementType is default. + return ihipBindTextureImpl(hipTextureType1D, hipReadModeElementType, + offset, devPtr, *desc, size, + tex->addressMode[0], tex->filterMode, tex->normalized, + tex->textureObject); +} + +hipError_t ihipBindTexture2DImpl(int dim, + enum hipTextureReadMode readMode, + size_t *offset, + const void *devPtr, + const struct hipChannelFormatDesc& desc, + size_t width, + size_t height, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords, + hipTextureObject_t& textureObject) +{ + HIP_INIT_API(); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hc::accelerator acc = ctx->getDevice()->_acc; + auto device = ctx->getWriteableDevice(); + + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + hipTexture* pTexture = (hipTexture*) malloc(sizeof(hipTexture)); + if (pTexture != nullptr) { + memset(pTexture, 0, sizeof(hipTexture)); + } + + hsa_ext_image_descriptor_t imageDescriptor; + + assert(dim == hipTextureType2D); + + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + imageDescriptor.width = width; + imageDescriptor.height = height; + imageDescriptor.depth = 1; + imageDescriptor.array_size = 0; + + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + getChannelOrderAndType(desc, readMode, channelOrder, channelType); + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; + + hsa_ext_sampler_descriptor_t samplerDescriptor; + fillSamplerDescriptor(samplerDescriptor, addressMode, filterMode, normalizedCoords); + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + + if (HSA_STATUS_SUCCESS != hsa_ext_image_create(*agent, &imageDescriptor, devPtr, permission, &(pTexture->image)) || + HSA_STATUS_SUCCESS != hsa_ext_sampler_create(*agent, &samplerDescriptor, &(pTexture->sampler))) { + return ihipLogStatus(hipErrorRuntimeOther); + } + getHipTextureObject(&textureObject, pTexture->image, pTexture->sampler); + textureHash[textureObject] = pTexture; + } + + return ihipLogStatus(hip_status); +} + +hipError_t hipBindTexture2D(size_t* offset, + textureReference* tex, + const void* devPtr, + const hipChannelFormatDesc* desc, + size_t width, + size_t height, + size_t pitch) +{ + // TODO: hipReadModeElementType is default. + return ihipBindTexture2DImpl(hipTextureType2D, hipReadModeElementType, + offset, devPtr, *desc, width, height, + tex->addressMode[0], tex->filterMode, tex->normalized, + tex->textureObject); +} + +hipError_t ihipBindTextureToArrayImpl(int dim, + enum hipTextureReadMode readMode, + hipArray_const_t array, + const struct hipChannelFormatDesc& desc, + enum hipTextureAddressMode addressMode, + enum hipTextureFilterMode filterMode, + int normalizedCoords, + hipTextureObject_t& textureObject) +{ + HIP_INIT_API(); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hc::accelerator acc = ctx->getDevice()->_acc; + auto device = ctx->getWriteableDevice(); + + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + hipTexture* pTexture = (hipTexture*) malloc(sizeof(hipTexture)); + if (pTexture != nullptr) { + memset(pTexture, 0, sizeof(hipTexture)); + } + + hsa_ext_image_descriptor_t imageDescriptor; + + imageDescriptor.width = array->width; + imageDescriptor.height = array->height; + imageDescriptor.depth = array->depth; + imageDescriptor.array_size = 0; + + switch (dim) { + case hipTextureType1D: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_1D; + imageDescriptor.height = 1; + imageDescriptor.depth = 1; + break; + case hipTextureType2D: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + imageDescriptor.depth = 1; + break; + case hipTextureType3D: + case hipTextureTypeCubemap: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_3D; + break; + case hipTextureType1DLayered: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_1DA; + imageDescriptor.height = 1; + imageDescriptor.array_size = array->height; + break; + case hipTextureType2DLayered: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2DA; + imageDescriptor.depth = 1; + imageDescriptor.array_size = array->depth; + break; + case hipTextureTypeCubemapLayered: + default: + break; + } + + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + getChannelOrderAndType(desc, readMode, channelOrder, channelType); + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; + + hsa_ext_sampler_descriptor_t samplerDescriptor; + fillSamplerDescriptor(samplerDescriptor, addressMode, filterMode, normalizedCoords); + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + + if (HSA_STATUS_SUCCESS != hsa_ext_image_create(*agent, &imageDescriptor, array->data, permission, &(pTexture->image)) || + HSA_STATUS_SUCCESS != hsa_ext_sampler_create(*agent, &samplerDescriptor, &(pTexture->sampler))) { + return ihipLogStatus(hipErrorRuntimeOther); + } + getHipTextureObject(&textureObject, pTexture->image, pTexture->sampler); + textureHash[textureObject] = pTexture; + } + + return ihipLogStatus(hip_status); +} + +hipError_t hipBindTextureToArray(textureReference* tex, + hipArray_const_t array, + const hipChannelFormatDesc* desc) +{ + // TODO: hipReadModeElementType is default. + return ihipBindTextureToArrayImpl(hipTextureType2D, hipReadModeElementType, + array, *desc, + tex->addressMode[0], tex->filterMode, tex->normalized, + tex->textureObject); +} + +hipError_t hipBindTextureToMipmappedArray(textureReference* tex, + hipMipmappedArray_const_t mipmappedArray, + const hipChannelFormatDesc* desc) +{ + return hipSuccess; +} + +hipError_t ihipUnbindTextureImpl(const hipTextureObject_t& textureObject) +{ + HIP_INIT_API(); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + hc::accelerator acc = ctx->getDevice()->_acc; + auto device = ctx->getWriteableDevice(); + + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + hipTexture* pTexture = textureHash[textureObject]; + if (pTexture != nullptr) { + hsa_ext_image_destroy(*agent, pTexture->image); + hsa_ext_sampler_destroy(*agent, pTexture->sampler); + free(pTexture); + textureHash.erase(textureObject); + } + } + + return ihipLogStatus(hip_status); +} + +hipError_t hipUnbindTexture(const textureReference* tex) +{ + return ihipUnbindTextureImpl(tex->textureObject); +} + +hipError_t hipGetChannelDesc(hipChannelFormatDesc* desc, hipArray_const_t array) +{ + HIP_INIT_API(desc, array); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + *desc = array->desc; + } + return ihipLogStatus(hip_status); +} + +hipError_t hipGetTextureAlignmentOffset(size_t* offset, const textureReference* tex) +{ + HIP_INIT_API(offset, tex); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + } + return ihipLogStatus(hip_status); +} + +hipError_t hipGetTextureReference(const textureReference** tex, const void* symbol) +{ + HIP_INIT_API(tex, symbol); + HIP_SET_DEVICE(); + + hipError_t hip_status = hipSuccess; + + auto ctx = ihipGetTlsDefaultCtx(); + if (ctx) { + } + return ihipLogStatus(hip_status); +} diff --git a/include/hip/nvcc_detail/hip_texture.h b/src/hip_texture.h similarity index 78% rename from include/hip/nvcc_detail/hip_texture.h rename to src/hip_texture.h index c669d62192..365a02e103 100644 --- a/include/hip/nvcc_detail/hip_texture.h +++ b/src/hip_texture.h @@ -20,9 +20,17 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ -#ifndef HIP_INCLUDE_HIP_NVCC_DETAIL_HIP_TEXTURE_H -#define HIP_INCLUDE_HIP_NVCC_DETAIL_HIP_TEXTURE_H +#ifndef HIP_INCLUDE_HCC_DETAIL_HIP_TEXTURE_H +#define HIP_INCLUDE_HCC_DETAIL_HIP_TEXTURE_H -#include +#include + +struct hipTexture { + hipResourceDesc resDesc; + hipTextureDesc texDesc; + hipResourceViewDesc resViewDesc; + hsa_ext_image_t image; + hsa_ext_sampler_t sampler; +}; #endif From 7ddb6b6b76fa5cb10047f02ad8537478c0eaf5d3 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Tue, 18 Jul 2017 11:01:02 +0530 Subject: [PATCH 030/108] Don't crash for hipEventDestroy(NULL) Change-Id: I3f4de1ad2712dbe8af81e10c5d3bef16d468c582 --- src/hip_event.cpp | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/src/hip_event.cpp b/src/hip_event.cpp index ab1c43a00b..71da9fdc30 100644 --- a/src/hip_event.cpp +++ b/src/hip_event.cpp @@ -164,13 +164,16 @@ hipError_t hipEventDestroy(hipEvent_t event) { HIP_INIT_API(event); - event->_state = hipEventStatusUnitialized; + if (event) { + event->_state = hipEventStatusUnitialized; - delete event; - event = NULL; + delete event; + event = NULL; - // TODO - examine return additional error codes - return ihipLogStatus(hipSuccess); + return ihipLogStatus(hipSuccess); + } else { + return ihipLogStatus(hipErrorInvalidResourceHandle); + } } hipError_t hipEventSynchronize(hipEvent_t event) From b3a9e202b1c9c1631fd5d8608b07c499639505f9 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Tue, 18 Jul 2017 11:17:02 +0530 Subject: [PATCH 031/108] [dt] temporarily disable broken tests Change-Id: I3ce23e410dc36e06d460263d768412f8368a6699 --- tests/src/runtimeApi/event/record_event.cpp | 2 +- tests/src/runtimeApi/stream/hipStreamSync2.cpp | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/src/runtimeApi/event/record_event.cpp b/tests/src/runtimeApi/event/record_event.cpp index bd8a3ada8e..c117f28f5c 100644 --- a/tests/src/runtimeApi/event/record_event.cpp +++ b/tests/src/runtimeApi/event/record_event.cpp @@ -19,7 +19,7 @@ THE SOFTWARE. /* HIT_START * BUILD: %t %s ../../test_common.cpp - * RUN: %t + * RUN: %t EXCLUDE_HIP_PLATFORM hcc * HIT_END */ diff --git a/tests/src/runtimeApi/stream/hipStreamSync2.cpp b/tests/src/runtimeApi/stream/hipStreamSync2.cpp index c6a58ce7d4..70b4e559fb 100644 --- a/tests/src/runtimeApi/stream/hipStreamSync2.cpp +++ b/tests/src/runtimeApi/stream/hipStreamSync2.cpp @@ -19,7 +19,7 @@ THE SOFTWARE. /* HIT_START * BUILD: %t %s ../../test_common.cpp - * RUN: %t + * RUN: %t EXCLUDE_HIP_PLATFORM hcc * HIT_END */ From 039f62213b1aa9cb29d1b8fb1d64fd89edb9eb08 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Tue, 18 Jul 2017 12:37:37 +0530 Subject: [PATCH 032/108] Enable memcpytosymbol family functions test --- tests/src/deviceLib/hipTestDeviceSymbol.cpp | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/tests/src/deviceLib/hipTestDeviceSymbol.cpp b/tests/src/deviceLib/hipTestDeviceSymbol.cpp index 2ee1e02dc0..140f3d97dc 100644 --- a/tests/src/deviceLib/hipTestDeviceSymbol.cpp +++ b/tests/src/deviceLib/hipTestDeviceSymbol.cpp @@ -18,7 +18,7 @@ THE SOFTWARE. */ /* HIT_START - * BUILD: %t %s EXCLUDE_HIP_PLATFORM all + * BUILD: %t %s * RUN: %t * HIT_END */ @@ -47,7 +47,7 @@ int main() A = new int[NUM]; B = new int[NUM]; C = new int[NUM]; - for(unsigned i=0;i Date: Wed, 19 Jul 2017 12:46:18 +0530 Subject: [PATCH 033/108] [dt] hip_texture.h -> hip_texture_types.h Change-Id: If106efa0a2e52105d4025706fdbc291c27352a74 --- tests/src/test_common.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/src/test_common.h b/tests/src/test_common.h index f585fb8bca..e3cc664d56 100644 --- a/tests/src/test_common.h +++ b/tests/src/test_common.h @@ -23,7 +23,7 @@ THE SOFTWARE. #include #include "hip/hip_runtime.h" -#include "hip/hip_texture.h" +#include "hip/hip_texture_types.h" #include "hip/hip_runtime_api.h" #define HC __attribute__((hc)) From 7a446dcbc54ef227ffda84f6667c3205837ceea7 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Thu, 20 Jul 2017 14:41:30 +0530 Subject: [PATCH 034/108] Renable frexp(f) device math function Change-Id: I53c022b8ddf38cd17ddb42eba457b9020db66395 --- include/hip/hcc_detail/math_functions.h | 4 ++-- tests/src/deviceLib/hipDoublePrecisionMathDevice.cpp | 2 +- tests/src/deviceLib/hipSinglePrecisionMathDevice.cpp | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/include/hip/hcc_detail/math_functions.h b/include/hip/hcc_detail/math_functions.h index 9faff2743a..79bacf274b 100644 --- a/include/hip/hcc_detail/math_functions.h +++ b/include/hip/hcc_detail/math_functions.h @@ -60,7 +60,7 @@ __device__ float fmaf(float x, float y, float z); __device__ float fmaxf(float x, float y); __device__ float fminf(float x, float y); __device__ float fmodf(float x, float y); -//__device__ float frexpf(float x, int* nptr); +__device__ float frexpf(float x, int* nptr); __device__ float hypotf(float x, float y); __device__ float ilogbf(float x); __device__ int isfinite(float a); @@ -146,7 +146,7 @@ __device__ double fma(double x, double y, double z); __device__ double fmax(double x, double y); __device__ double fmin(double x, double y); __device__ double fmod(double x, double y); -//__device__ double frexp(double x, int *nptr); +__device__ double frexp(double x, int *nptr); __device__ double hypot(double x, double y); __device__ double ilogb(double x); __device__ int isfinite(double x); diff --git a/tests/src/deviceLib/hipDoublePrecisionMathDevice.cpp b/tests/src/deviceLib/hipDoublePrecisionMathDevice.cpp index f4f7ab0479..0a81d111c2 100644 --- a/tests/src/deviceLib/hipDoublePrecisionMathDevice.cpp +++ b/tests/src/deviceLib/hipDoublePrecisionMathDevice.cpp @@ -69,7 +69,7 @@ __device__ void double_precision_math_functions() fmax(0.0, 0.0); fmin(0.0, 0.0); fmod(0.0, 1.0); -// frexp(0.0, &iX); + frexp(0.0, &iX); hypot(1.0, 0.0); ilogb(1.0); isfinite(0.0); diff --git a/tests/src/deviceLib/hipSinglePrecisionMathDevice.cpp b/tests/src/deviceLib/hipSinglePrecisionMathDevice.cpp index de3dec35ef..a52b1a22fe 100644 --- a/tests/src/deviceLib/hipSinglePrecisionMathDevice.cpp +++ b/tests/src/deviceLib/hipSinglePrecisionMathDevice.cpp @@ -70,7 +70,7 @@ __device__ void single_precision_math_functions() fmaxf(0.0f, 0.0f); fminf(0.0f, 0.0f); fmodf(0.0f, 1.0f); - //frexpf(0.0f, &iX); + frexpf(0.0f, &iX); hypotf(1.0f, 0.0f); ilogbf(1.0f); isfinite(0.0f); From 961440b59358d84f85b93e6f7d4e5ba1491a9025 Mon Sep 17 00:00:00 2001 From: Aditya Atluri Date: Fri, 21 Jul 2017 15:50:12 -0500 Subject: [PATCH 035/108] fixed device selection during compilation to use rocm_agent_enumerator 1. Changed hipcc to use rocm_agent_enumerator 2. Changed square sample test to use device variable --- bin/hipcc | 22 ++++++++++++++++++++++ samples/0_Intro/square/square.hipref.cpp | 5 +++-- 2 files changed, 25 insertions(+), 2 deletions(-) diff --git a/bin/hipcc b/bin/hipcc index 4a12b9e541..8a1874a36a 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -103,6 +103,28 @@ if ($HIP_PLATFORM eq "hcc") { $HIPLDFLAGS = `${HCC_HOME}/bin/hcc-config --ldflags`; + $ROCM_AGENT_ENUM = "${ROCM_PATH}/bin/rocm_agent_enumerator"; + + my $myAgents = `${ROCM_AGENT_ENUM} -t GPU`; + my @agentsLine = split('\n', $myAgents); + + foreach my $val (@agentsLine) { + if($val eq "gfx701") { + $target_gfx701 = 1; + } + if($val eq "gfx801") { + $target_gfx801 = 1; + } + if($val eq "gfx802") { + $target_gfx802 = 1; + } + if($val eq "gfx803") { + $target_gfx803 = 1; + } + if($val eq "gfx900") { + $target_gfx900 = 1; + } + } #### GCC system includes workaround #### $HCC_WA_FLAGS = " "; diff --git a/samples/0_Intro/square/square.hipref.cpp b/samples/0_Intro/square/square.hipref.cpp index e694bfb8a4..167cb135f4 100644 --- a/samples/0_Intro/square/square.hipref.cpp +++ b/samples/0_Intro/square/square.hipref.cpp @@ -54,9 +54,10 @@ int main(int argc, char *argv[]) float *A_h, *C_h; size_t N = 1000000; size_t Nbytes = N * sizeof(float); - + static int device = 0; + CHECK(hipSetDevice(device)); hipDeviceProp_t props; - CHECK(hipGetDeviceProperties(&props, 0/*deviceID*/)); + CHECK(hipGetDeviceProperties(&props, device/*deviceID*/)); printf ("info: running on device %s\n", props.name); #ifdef __HIP_PLATFORM_HCC__ printf ("info: architecture on AMD GPU device is: %d\n",props.gcnArch); From 4b8931f7173113cc99d15372fcd188e879d55fc2 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Mon, 24 Jul 2017 15:16:41 +0530 Subject: [PATCH 036/108] [ci] Exclude sporadically failing test in CI only Change-Id: If76bc89d39aa35f5c246cbcd41a997b580832450 --- Jenkinsfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Jenkinsfile b/Jenkinsfile index 250a8949a3..42f8b67c3c 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -119,7 +119,7 @@ def docker_build_inside_image( def build_image, String inside_args, String platf cd ${build_dir_rel} make install -j\$(nproc) make build_tests -j\$(nproc) - make test + ctest -E hipEnvVarDriver """ // If unit tests output a junit or xunit file in the future, jenkins can parse that file // to display test results on the dashboard From 81bc82d16604b29eafc6b22f7fd2585383cca053 Mon Sep 17 00:00:00 2001 From: sunway513 Date: Mon, 24 Jul 2017 17:16:13 +0000 Subject: [PATCH 037/108] Fix indentation, binary path and add wait on popen calls to hipEnvVar test --- tests/src/hipEnvVarDriver.cpp | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) diff --git a/tests/src/hipEnvVarDriver.cpp b/tests/src/hipEnvVarDriver.cpp index 1edb158aa3..dc1d451e0d 100644 --- a/tests/src/hipEnvVarDriver.cpp +++ b/tests/src/hipEnvVarDriver.cpp @@ -28,16 +28,21 @@ THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ #include #include #include "hip/hip_runtime.h" +#include +#include using namespace std; int getDeviceNumber(){ - FILE *in; - char buff[512]; - string str; - if(!(in = popen("./hipEnvVar -c", "r"))){ - return 1; + FILE *in; + char buff[512]; + string str; + std::this_thread::sleep_for(std::chrono::milliseconds(10)); + if(!(in = popen("./directed_tests/hipEnvVar -c", "r"))){ + return 1; + } + while(fgets(buff, sizeof(buff), in)!=NULL){ + cout << buff; } - fgets(buff, sizeof(buff), in); pclose(in); return atoi(buff); } @@ -46,12 +51,15 @@ int getDeviceNumber(){ int getDevicePCIBusNumRemote(int deviceID){ FILE *in; char buff[512]; - string str = "./hipEnvVar -d "; + string str = "./directed_tests/hipEnvVar -d "; str += std::to_string(deviceID); + std::this_thread::sleep_for(std::chrono::milliseconds(10)); if(!(in = popen(str.c_str(), "r"))){ return 1; } - fgets(buff, sizeof(buff), in); + while(fgets(buff, sizeof(buff), in)!=NULL){ + cout << buff; + } pclose(in); return atoi(buff); } From 1275ceed675dd4db0175e765a7224098f1c4543d Mon Sep 17 00:00:00 2001 From: sunway513 Date: Mon, 24 Jul 2017 18:53:46 +0000 Subject: [PATCH 038/108] Re-enable hipEnvVarDriver in Jenkinsfile --- Jenkinsfile | 1 - 1 file changed, 1 deletion(-) diff --git a/Jenkinsfile b/Jenkinsfile index 42f8b67c3c..08fea0b0bb 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -119,7 +119,6 @@ def docker_build_inside_image( def build_image, String inside_args, String platf cd ${build_dir_rel} make install -j\$(nproc) make build_tests -j\$(nproc) - ctest -E hipEnvVarDriver """ // If unit tests output a junit or xunit file in the future, jenkins can parse that file // to display test results on the dashboard From 6f359dfbe66aecc1be2b8665ae316831ee1a43a7 Mon Sep 17 00:00:00 2001 From: sunway513 Date: Mon, 24 Jul 2017 19:02:00 +0000 Subject: [PATCH 039/108] Update Jenkinsfile to run all the built tests --- Jenkinsfile | 1 + 1 file changed, 1 insertion(+) diff --git a/Jenkinsfile b/Jenkinsfile index 08fea0b0bb..526a825c65 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -119,6 +119,7 @@ def docker_build_inside_image( def build_image, String inside_args, String platf cd ${build_dir_rel} make install -j\$(nproc) make build_tests -j\$(nproc) + make test """ // If unit tests output a junit or xunit file in the future, jenkins can parse that file // to display test results on the dashboard From 67f5d2de3d5d1d57b61e8f929d397ac201263435 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Mon, 24 Jul 2017 18:57:19 -0500 Subject: [PATCH 040/108] Enable HCC_OPT_FLUSH=1 (if HCC compiler new enough) --- src/hip_hcc.cpp | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 8aaa7af02b..9d431df2dd 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -95,8 +95,12 @@ int HIP_SYNC_HOST_ALLOC = 1; int HIP_SYNC_NULL_STREAM = 1; // HIP needs to change some behavior based on HCC_OPT_FLUSH : -// TODO - set this to 1 +#if (__hcc_workweek__ >= 17296) +int HCC_OPT_FLUSH = 1; +#else +#warning "HIP disabled HCC_OPT_FLUSH since HCC version does not yet support" int HCC_OPT_FLUSH = 0; +#endif From 85e6593a1bfc80b582292874dbf5b1411548e8ea Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Wed, 26 Jul 2017 07:08:03 +0300 Subject: [PATCH 041/108] =?UTF-8?q?[HIPIFY]=20Finally=20finished=20syncing?= =?UTF-8?q?=20with=20CUDA=208.0.61=20Driver=20and=20Runtime=E2=80=A6=20(#1?= =?UTF-8?q?05)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit [hipify-clang] Finally finished syncing with CUDA 8.0.61 Driver and Runtime API (including missing data types, D3D, OpenGL, VDPAU and EGL interop). + All the Modules are supported now: 1) 4.1 – 4.31 from CUDA 8.0.61 Driver API 2) CUDA_Driver_API_functions_supported_by_HIP.md updated accordingly 3) 4.1 – 4.31 from CUDA 8.0.61 Runtime API 4) CUDA_Runtime_API_functions_supported_by_HIP.md updated accordingly + Typos fixing + Annotating --- ...A_Driver_API_functions_supported_by_HIP.md | 1359 ++++++++------- ..._Runtime_API_functions_supported_by_HIP.md | 1331 ++++++++------- hipify-clang/src/Cuda2Hip.cpp | 1459 ++++++++++------- 3 files changed, 2410 insertions(+), 1739 deletions(-) diff --git a/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md index d797b31832..bd6634dbe5 100644 --- a/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md @@ -2,690 +2,859 @@ ## **1. Data types used by CUDA driver** -| **type** | **CUDA** | **HIP** | **CUDA description** | -|-------------:|---------------------------------------------------------------|------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| struct | `CUDA_ARRAY3D_DESCRIPTOR` | | | -| struct | `CUDA_ARRAY_DESCRIPTOR` | | | -| struct | `CUDA_MEMCPY2D` | | | -| struct | `CUDA_MEMCPY3D` | | | -| struct | `CUDA_MEMCPY3D_PEER` | | | -| struct | `CUDA_POINTER_ATTRIBUTE_P2P_TOKENS` | | | -| struct | `CUDA_RESOURCE_DESC` | | | -| struct | `CUDA_RESOURCE_VIEW_DESC` | | | -| struct | `CUdevprop` | `hipDeviceProp_t` | | -| struct | `CUipcEventHandle` | | | -| struct | `CUipcMemHandle` | | | -| enum |***`CUaddress_mode`*** | | Texture reference addressing modes | -| 0 |*`CU_TR_ADDRESS_MODE_WRAP`* | | Wrapping address mode | -| 1 |*`CU_TR_ADDRESS_MODE_CLAMP`* | | Clamp to edge address mode | -| 2 |*`CU_TR_ADDRESS_MODE_MIRROR`* | | Mirror address mode | -| 3 |*`CU_TR_ADDRESS_MODE_BORDER`* | | Border address mode | -| enum |***`CUarray_cubemap_face`*** | | Array indices for cube faces | -| 0x00 |*`CU_CUBEMAP_FACE_POSITIVE_X`* | | Positive X face of cubemap | -| 0x01 |*`CU_CUBEMAP_FACE_NEGATIVE_X`* | | Negative X face of cubemap | -| 0x02 |*`CU_CUBEMAP_FACE_POSITIVE_Y`* | | Positive Y face of cubemap | -| 0x03 |*`CU_CUBEMAP_FACE_NEGATIVE_Y`* | | Negative Y face of cubemap | -| 0x04 |*`CU_CUBEMAP_FACE_POSITIVE_Z`* | | Positive Z face of cubemap | -| 0x05 |*`CU_CUBEMAP_FACE_NEGATIVE_Z`* | | Negative Z face of cubemap | -| enum |***`CUarray_format`*** | | Array formats | -| 0x01 |*`CU_AD_FORMAT_UNSIGNED_INT8`* | | Unsigned 8-bit integers | -| 0x02 |*`CU_AD_FORMAT_UNSIGNED_INT16`* | | Unsigned 16-bit integers | -| 0x03 |*`CU_AD_FORMAT_UNSIGNED_INT32`* | | Unsigned 32-bit integers | -| 0x08 |*`CU_AD_FORMAT_SIGNED_INT8`* | | Signed 8-bit integers | -| 0x09 |*`CU_AD_FORMAT_SIGNED_INT16`* | | Signed 16-bit integers | -| 0x0a |*`CU_AD_FORMAT_SIGNED_INT32`* | | Signed 32-bit integers | -| 0x10 |*`CU_AD_FORMAT_HALF`* | | 16-bit floating point | -| 0x20 |*`CU_AD_FORMAT_FLOAT`* | | 32-bit floating point | -| enum |***`CUctx_flags`*** | | Context creation flags | -| 0x00 |*`CU_CTX_SCHED_AUTO`* | | Automatic scheduling | -| 0x01 |*`CU_CTX_SCHED_SPIN`* | | Set spin as default scheduling | -| 0x02 |*`CU_CTX_SCHED_YIELD`* | | Set yield as default scheduling | -| 0x04 |*`CU_CTX_SCHED_BLOCKING_SYNC`* | | Set blocking synchronization as default scheduling | -| 0x04 |*`CU_CTX_BLOCKING_SYNC`* | | Set blocking synchronization as default scheduling Deprecated. This flag was deprecated as of CUDA 4.0 and was replaced with CU_CTX_SCHED_BLOCKING_SYNC.| -| 0x07 |*`CU_CTX_SCHED_MASK`* | | | -| 0x08 |*`CU_CTX_MAP_HOST`* | | Support mapped pinned allocations | -| 0x10 |*`CU_CTX_LMEM_RESIZE_TO_MAX`* | | Keep local memory allocation after launch | -| 0x1f |*`CU_CTX_FLAGS_MASK`* | | | -| enum |***`CUdevice_attribute`*** | | Device properties | -| 1 |*`CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK`* |*`hipDeviceAttributeMaxThreadsPerBlock`* | Maximum number of threads per block | -| 2 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X`* |*`hipDeviceAttributeMaxBlockDimX`* | Maximum block dimension X | -| 3 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y`* |*`hipDeviceAttributeMaxBlockDimY`* | Maximum block dimension Y | -| 4 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z`* |*`hipDeviceAttributeMaxBlockDimZ`* | Maximum block dimension Z | -| 5 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X`* |*`hipDeviceAttributeMaxGridDimX`* | Maximum grid dimension X | -| 6 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y`* |*`hipDeviceAttributeMaxGridDimY`* | Maximum grid dimension Y | -| 7 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z`* |*`hipDeviceAttributeMaxGridDimZ`* | Maximum grid dimension Y | -| 8 |*`CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | Maximum shared memory available per block in bytes | -| 8 |*`CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | Deprecated, use CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK | -| 9 |*`CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY`* |*`hipDeviceAttributeTotalConstantMemory`* | Memory available on device for __constant__ variables in a CUDA C kernel in bytes | -| 10 |*`CU_DEVICE_ATTRIBUTE_WARP_SIZE`* |*`hipDeviceAttributeWarpSize`* | Warp size in threads | -| 11 |*`CU_DEVICE_ATTRIBUTE_MAX_PITCH`* | | Maximum pitch in bytes allowed by memory copies | -| 12 |*`CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | Maximum number of 32-bit registers available per block | -| 12 |*`CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | Deprecated, use CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK | -| 13 |*`CU_DEVICE_ATTRIBUTE_CLOCK_RATE`* |*`hipDeviceAttributeClockRate`* | Typical clock frequency in kilohertz | -| 14 |*`CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT`* | | Alignment requirement for textures | -| 15 |*`CU_DEVICE_ATTRIBUTE_GPU_OVERLAP`* | | Device can possibly copy memory and execute a kernel concurrently. Deprecated. Use instead CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT| -| 16 |*`CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT`* |*`hipDeviceAttributeMultiprocessorCount`* | Number of multiprocessors on device | -| 17 |*`CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT`* | | Specifies whether there is a run time limit on kernels | -| 18 |*`CU_DEVICE_ATTRIBUTE_INTEGRATED`* | | Device is integrated with host memory | -| 19 |*`CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY`* | | Device can map host memory into CUDA address space | -| 20 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_MODE`* |*`hipDeviceAttributeComputeMode`* | Compute mode (See CUcomputemode for details) | -| 21 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH`* | | Maximum 1D texture width | -| 22 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH`* | | Maximum 2D texture width | -| 23 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT`* | | Maximum 2D texture height | -| 24 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH`* | | Maximum 3D texture width | -| 25 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT`* | | Maximum 3D texture height | -| 26 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH`* | | Maximum 3D texture depth | -| 27 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH`* | | Maximum 2D layered texture width | -| 28 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT`* | | Maximum 2D layered texture height | -| 29 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS`* | | Maximum layers in a 2D layered texture | -| 27 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH`* | | Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH | -| 28 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT`* | | Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT | -| 29 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES`* | | Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS | -| 30 |*`CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT`* | | Alignment requirement for surfaces | -| 31 |*`CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS`* |*`hipDeviceAttributeConcurrentKernels`* | Device can possibly execute multiple kernels concurrently | -| 32 |*`CU_DEVICE_ATTRIBUTE_ECC_ENABLED`* | | Device has ECC support enabled | -| 33 |*`CU_DEVICE_ATTRIBUTE_PCI_BUS_ID`* |*`hipDeviceAttributePciBusId`* | PCI bus ID of the device | -| 34 |*`CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID`* |*`hipDeviceAttributePciDeviceId`* | PCI device ID of the device | -| 35 |*`CU_DEVICE_ATTRIBUTE_TCC_DRIVER`* | | Device is using TCC driver model | -| 36 |*`CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE`* |*`hipDeviceAttributeMemoryClockRate`* | Peak memory clock frequency in kilohertz | -| 37 |*`CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH`* |*`hipDeviceAttributeMemoryBusWidth`* | Global memory bus width in bits | -| 38 |*`CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE`* |*`hipDeviceAttributeL2CacheSize`* | Size of L2 cache in bytes | -| 39 |*`CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR`* |*`hipDeviceAttributeMaxThreadsPerMultiProcessor`* | Maximum resident threads per multiprocessor | -| 40 |*`CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT`* | | Number of asynchronous engines | -| 41 |*`CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING`* | | Device shares a unified address space with the host | -| 42 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH`* | | Maximum 1D layered texture width | -| 43 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS`* | | Maximum layers in a 1D layered texture | -| 44 |*`CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER`* | | Deprecated, do not use | -| 45 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH`* | | Maximum 2D texture width if CUDA_ARRAY3D_TEXTURE_GATHER is set | -| 46 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT`* | | Maximum 2D texture height if CUDA_ARRAY3D_TEXTURE_GATHER is set | -| 47 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE`* | | Alternate maximum 3D texture width | -| 48 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE`* | | Alternate maximum 3D texture height | -| 49 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE`* | | Alternate maximum 3D texture depth | -| 50 |*`CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID`* | | PCI domain ID of the device | -| 51 |*`CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT`* | | Pitch alignment requirement for textures | -| 52 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH`* | | Maximum cubemap texture width/height | -| 53 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH`* | | Maximum cubemap layered texture width/height | -| 54 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS`* | | Maximum layers in a cubemap layered texture | -| 55 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH`* | | Maximum 1D surface width | -| 56 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH`* | | Maximum 2D surface width | -| 57 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT`* | | Maximum 2D surface height | -| 58 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH`* | | Maximum 3D surface width | -| 59 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT`* | | Maximum 3D surface height | -| 60 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH`* | | Maximum 3D surface depth | -| 61 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH`* | | Maximum 1D layered surface width | -| 62 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS`* | | Maximum layers in a 1D layered surface | -| 63 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH`* | | Maximum 2D layered surface width | -| 64 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT`* | | Maximum 2D layered surface height | -| 65 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS`* | | Maximum layers in a 2D layered surface | -| 66 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH`* | | Maximum cubemap surface width | -| 67 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH`* | | Maximum cubemap layered surface width | -| 68 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS`* | | Maximum layers in a cubemap layered surface | -| 69 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH`* | | Maximum 1D linear texture width | -| 70 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH`* | | Maximum 2D linear texture width | -| 71 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT`* | | Maximum 2D linear texture height | -| 72 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH`* | | Maximum 2D linear texture pitch in bytes | -| 73 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH`* | | Maximum mipmapped 2D texture width | -| 74 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT`* | | Maximum mipmapped 2D texture height | -| 75 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR`* |*`hipDeviceAttributeComputeCapabilityMajor`* | Major compute capability version number | -| 76 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR`* |*`hipDeviceAttributeComputeCapabilityMinor`* | Minor compute capability version number | -| 77 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH`* | | Maximum mipmapped 1D texture width | -| 78 |*`CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED`* | | Device supports stream priorities | -| 79 |*`CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED`* | | Device supports caching globals in L1 | -| 80 |*`CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED`* | | Device supports caching locals in L1 | -| 81 |*`CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR`* |*`hipDeviceAttributeMaxSharedMemoryPerMultiprocessor`* | Maximum shared memory available per multiprocessor in bytes | -| 82 |*`CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR`* | | Maximum number of 32-bit registers available per multiprocessor | -| 83 |*`CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY`* |*`hipDeviceAttributeManagedMemory`* | Device can allocate managed memory on this system | -| 84 |*`CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD`* | | Device is on a multi-GPU board | -| 85 |*`CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID`* | | Unique id for a group of devices on the same multi-GPU board | -| 86 |*`CU_DEVICE_ATTRIBUTE_MAX`* | | | -| enum |***`CUevent_flags`*** | | Event creation flags | -| 0x00 |*`CU_EVENT_DEFAULT`* |*`hipEventDefault`* | Default event flag | -| 0x01 |*`CU_EVENT_BLOCKING_SYNC`* |*`hipEventBlockingSync`* | Event uses blocking synchronization | -| 0x02 |*`CU_EVENT_DISABLE_TIMING`* |*`hipEventDisableTiming`* | Event will not record timing data | -| 0x04 |*`CU_EVENT_INTERPROCESS`* |*`hipEventInterprocess`* | Event is suitable for interprocess use. CU_EVENT_DISABLE_TIMING must be set | -| enum |***`CUfilter_mode`*** |***`hipTextureFilterMode`*** | Texture reference filtering modes | -| 0 |*`CU_TR_FILTER_MODE_POINT`* |*`hipFilterModePoint`* | Point filter mode | -| 1 |*`CU_TR_FILTER_MODE_LINEAR`* |*`hipFilterModeLinear`* | Linear filter mode | -| enum |***`CUfunc_cache`*** |***`hipFuncCache`*** | Function cache configurations | -| 0x00 |*`CU_FUNC_CACHE_PREFER_NONE`* |*`hipFuncCachePreferNone`* | no preference for shared memory or L1 (default) | -| 0x01 |*`CU_FUNC_CACHE_PREFER_SHARED`* |*`hipFuncCachePreferShared`* | prefer larger shared memory and smaller L1 cache | -| 0x02 |*`CU_FUNC_CACHE_PREFER_L1`* |*`hipFuncCachePreferL1`* | prefer larger L1 cache and smaller shared memory | -| 0x03 |*`CU_FUNC_CACHE_PREFER_EQUAL`* |*`hipFuncCachePreferEqual`* | prefer equal sized L1 cache and shared memory | -| enum |***`CUfunction_attribute`*** | | Function properties | -| 0 |*`CU_FUNC_ATTRIBUTE_MAX_THREADS_PER_BLOCK`* | | The maximum number of threads per block, beyond which a launch of the function would fail. This number depends on both the function and the device on which the function is currently loaded. | -| 1 |*`CU_FUNC_ATTRIBUTE_SHARED_SIZE_BYTES`* | | The size in bytes of statically-allocated shared memory required by this function. This does not include dynamically-allocated shared memory requested by the user at runtime. | -| 2 |*`CU_FUNC_ATTRIBUTE_CONST_SIZE_BYTES`* | | The size in bytes of user-allocated constant memory required by this function. | -| 3 |*`CU_FUNC_ATTRIBUTE_LOCAL_SIZE_BYTES`* | | The size in bytes of local memory used by each thread of this function. | -| 4 |*`CU_FUNC_ATTRIBUTE_NUM_REGS`* | | The number of registers used by each thread of this function. | -| 5 |*`CU_FUNC_ATTRIBUTE_PTX_VERSION`* | | The PTX virtual architecture version for which the function was compiled. This value is the major PTX version * 10 + the minor PTX version, so a PTX version 1.3 function would return the value 13. Note that this may return the undefined value of 0 for cubins compiled prior to CUDA 3.0. | -| 6 |*`CU_FUNC_ATTRIBUTE_BINARY_VERSION`* | | The binary architecture version for which the function was compiled. This value is the major binary version * 10 + the minor binary version, so a binary version 1.3 function would return the value 13. Note that this will return a value of 10 for legacy cubins that do not have a properly-encoded binary architecture version. | -| 7 |*`CU_FUNC_ATTRIBUTE_CACHE_MODE_CA`* | | The attribute to indicate whether the function has been compiled with user specified option "-Xptxas --dlcm=ca" set. | -| 8 |*`CU_FUNC_ATTRIBUTE_MAX`* | | | -| enum |***`CUgraphicsMapResourceFlags`*** | | Flags for mapping and unmapping interop resources | -| 0x00 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE`* | | | -| 0x01 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY`* | | | -| 0x02 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_WRITE_DISCARD`* | | | -| enum |***`CUgraphicsRegisterFlags`*** | | Flags to register a graphics resource | -| 0x00 |*`CU_GRAPHICS_REGISTER_FLAGS_NONE`* | | | -| 0x01 |*`CU_GRAPHICS_REGISTER_FLAGS_READ_ONLY`* | | | -| 0x02 |*`CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD`* | | | -| 0x04 |*`CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST`* | | | -| 0x08 |*`CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER`* | | | -| enum |***`CUipcMem_flags`*** | | CUDA Ipc Mem Flags | -| 0x1 |*`CU_IPC_MEM_LAZY_ENABLE_PEER_ACCESS`* |*`hipIpcMemLazyEnablePeerAccess`* | Automatically enable peer access between remote devices as needed | -| enum |***`CUjit_cacheMode`*** | | Caching modes for dlcm | -| 0 |*`CU_JIT_CACHE_OPTION_NONE`* | | Compile with no -dlcm flag specified | -| |*`CU_JIT_CACHE_OPTION_CG`* | | Compile with L1 cache disabled | -| |*`CU_JIT_CACHE_OPTION_CA`* | | Compile with L1 cache enabled | -| enum |***`CUjit_fallback`*** | | Cubin matching fallback strategies | -| 0 |*`CU_PREFER_PTX`* | | Prefer to compile ptx if exact binary match not found | -| |*`CU_PREFER_BINARY`* | | Prefer to fall back to compatible binary code if exact match not found | -| enum |***`CUjit_option`*** | | Online compiler and linker options | -| 0 |*`CU_JIT_MAX_REGISTERS`* | | Max number of registers that a thread may use. Option type: unsigned int Applies to: compiler only. | -| |*`CU_JIT_THREADS_PER_BLOCK`* | | IN: Specifies minimum number of threads per block to target compilation for OUT: Returns the number of threads the compiler actually targeted. This restricts the resource utilization fo the compiler (e.g. max registers) such that a block with the given number of threads should be able to launch based on register limitations. Note, this option does not currently take into account any other resource limitations, such as shared memory utilization. Cannot be combined with CU_JIT_TARGET. Option type: unsigned int Applies to: compiler only. | -| |*`CU_JIT_WALL_TIME`* | | Overwrites the option value with the total wall clock time, in milliseconds, spent in the compiler and linker Option type: float Applies to: compiler and linker. | -| |*`CU_JIT_INFO_LOG_BUFFER`* | | Pointer to a buffer in which to print any log messages that are informational in nature (the buffer size is specified via option CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES) Option type: char * Applies to: compiler and linker. | -| |*`CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES`* | | IN: Log buffer size in bytes. Log messages will be capped at this size (including null terminator) OUT: Amount of log buffer filled with messages Option type: unsigned int Applies to: compiler and linker. | -| |*`CU_JIT_OPTIMIZATION_LEVEL`* | | Level of optimizations to apply to generated code (0 - 4), with 4 being the default and highest level of optimizations. Option type: unsigned int Applies to: compiler only. | -| |*`CU_JIT_TARGET_FROM_CUCONTEXT`* | | No option value required. Determines the target based on the current attached context (default) Option type: No option value needed Applies to: compiler and linker. | -| |*`CU_JIT_TARGET`* | | Target is chosen based on supplied CUjit_target. Cannot be combined with CU_JIT_THREADS_PER_BLOCK. Option type: unsigned int for enumerated type CUjit_target Applies to: compiler and linker. | -| |*`CU_JIT_FALLBACK_STRATEGY`* | | Specifies choice of fallback strategy if matching cubin is not found. Choice is based on supplied CUjit_fallback. This option cannot be used with cuLink* APIs as the linker requires exact matches. Option type: unsigned int for enumerated type CUjit_fallback Applies to: compiler only. | -| |*`CU_JIT_GENERATE_DEBUG_INFO`* | | Specifies whether to create debug information in output (-g) (0: false, default) Option type: int Applies to: compiler and linker. | -| |*`CU_JIT_LOG_VERBOSE`* | | Generate verbose log messages (0: false, default) Option type: int Applies to: compiler and linker. | -| |*`CU_JIT_GENERATE_LINE_INFO`* | | Generate line number information (-lineinfo) (0: false, default) Option type: int Applies to: compiler only. | -| |*`CU_JIT_CACHE_MODE`* | | Specifies whether to enable caching explicitly (-dlcm) Choice is based on supplied CUjit_cacheMode_enum. Option type: unsigned int for enumerated type CUjit_cacheMode_enum Applies to: compiler only. | -| |*`CU_JIT_NUM_OPTIONS`* | | | -| enum |***`CUjit_target`*** | | Online compilation targets | -| 10 |*`CU_TARGET_COMPUTE_10`* | | Compute device class 1.0. | -| 11 |*`CU_TARGET_COMPUTE_11`* | | Compute device class 1.1. | -| 12 |*`CU_TARGET_COMPUTE_12`* | | Compute device class 1.2. | -| 13 |*`CU_TARGET_COMPUTE_13`* | | Compute device class 1.3. | -| 20 |*`CU_TARGET_COMPUTE_20`* | | Compute device class 2.0. | -| 21 |*`CU_TARGET_COMPUTE_21`* | | Compute device class 2.1. | -| 30 |*`CU_TARGET_COMPUTE_30`* | | Compute device class 3.0. | -| 32 |*`CU_TARGET_COMPUTE_32`* | | Compute device class 3.2. | -| 35 |*`CU_TARGET_COMPUTE_35`* | | Compute device class 3.5. | -| 37 |*`CU_TARGET_COMPUTE_37`* | | Compute device class 3.7. | -| 50 |*`CU_TARGET_COMPUTE_50`* | | Compute device class 5.0. | -| 52 |*`CU_TARGET_COMPUTE_52`* | | Compute device class 5.2. | -| enum |***`CUjitInputType`*** | | Device code formats | -| 0 |*`CU_JIT_INPUT_CUBIN`* | | Compiled device-class-specific device code Applicable options: none. | -| |*`CU_JIT_INPUT_PTX`* | | PTX source code Applicable options: PTX compiler options. | -| |*`CU_JIT_INPUT_FATBINARY`* | | Bundle of multiple cubins and/or PTX of some device code Applicable options: PTX compiler options, CU_JIT_FALLBACK_STRATEGY. | -| |*`CU_JIT_INPUT_OBJECT`* | | Host object with embedded device code Applicable options: PTX compiler options, CU_JIT_FALLBACK_STRATEGY. | -| |*`CU_JIT_INPUT_LIBRARY`* | | Archive of host objects with embedded device code Applicable options: PTX compiler options, CU_JIT_FALLBACK_STRATEGY. | -| |*`CU_JIT_NUM_INPUT_TYPES`* | | | -| enum |***`CUlimit`*** |***`hipLimit_t`*** | Limits | -| 0x00 |*`CU_LIMIT_STACK_SIZE`* | | GPU thread stack size. | -| 0x01 |*`CU_LIMIT_PRINTF_FIFO_SIZE`* | | GPU printf FIFO size. | -| 0x02 |*`CU_LIMIT_MALLOC_HEAP_SIZE`* |*`hipLimitMallocHeapSize`* | GPU malloc heap size. | -| 0x03 |*`CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH`* | | GPU device runtime launch synchronize depth. | -| 0x04 |*`CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT`* | | GPU device runtime pending launch count. | -| |*`CU_LIMIT_MAX`* | | | -| enum |***`CUmemAttach_flags`*** | | CUDA Mem Attach Flags | -| 0x1 |*`CU_MEM_ATTACH_GLOBAL`* | | Memory can be accessed by any stream on any device. | -| 0x2 |*`CU_MEM_ATTACH_HOST`* | | Memory cannot be accessed by any stream on any device. | -| 0x4 |*`CU_MEM_ATTACH_SINGLE`* | | Memory can only be accessed by a single stream on the associated device. | -| enum |***`CUmemorytype`*** | | Memory types | -| 0x01 |*`CU_MEMORYTYPE_HOST`* | | Host memory | -| 0x02 |*`CU_MEMORYTYPE_DEVICE`* | | Device memory | -| 0x03 |*`CU_MEMORYTYPE_ARRAY`* | | Array memory | -| 0x04 |*`CU_MEMORYTYPE_UNIFIED`* | | Unified device or host memory | -| enum |***`CUoccupancy_flags`*** | | Occupancy calculator flag | -| 0x00 |*`CU_OCCUPANCY_DEFAULT`* | | Default behavior | -| 0x01 |*`CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE`* | | Assume global caching is enabled and cannot be automatically turned off | -| enum |***`CUpointer_attribute`*** | | Pointer information | -| 1 |*`CU_POINTER_ATTRIBUTE_CONTEXT`* | | The CUcontext on which a pointer was allocated or registered | -| 2 |*`CU_POINTER_ATTRIBUTE_MEMORY_TYPE`* | | The CUmemorytype describing the physical location of a pointer | -| 3 |*`CU_POINTER_ATTRIBUTE_DEVICE_POINTER`* | | The address at which a pointer's memory may be accessed on the device | -| 4 |*`CU_POINTER_ATTRIBUTE_HOST_POINTER`* | | The address at which a pointer's memory may be accessed on the host | -| 5 |*`CU_POINTER_ATTRIBUTE_P2P_TOKENS`* | | A pair of tokens for use with the nv-p2p.h Linux kernel interface | -| 6 |*`CU_POINTER_ATTRIBUTE_SYNC_MEMOPS`* | | Synchronize every synchronous memory operation initiated on this region | -| 7 |*`CU_POINTER_ATTRIBUTE_BUFFER_ID`* | | A process-wide unique ID for an allocated memory region | -| 8 |*`CU_POINTER_ATTRIBUTE_IS_MANAGED`* | | Indicates if the pointer points to managed memory | -| enum |***`CUmemorytype`*** | | Resource types | -| 0x00 |*`CU_RESOURCE_TYPE_ARRAY`* | | Array resoure | -| 0x01 |*`CU_RESOURCE_TYPE_MIPMAPPED_ARRAY`* | | Mipmapped array resource | -| 0x02 |*`CU_RESOURCE_TYPE_LINEAR`* | | Linear resource | -| 0x03 |*`CU_RESOURCE_TYPE_PITCH2D`* | | Pitch 2D resource | -| enum |***`CUresourceViewFormat`*** | | Resource view format | -| 0x00 |*`CU_RES_VIEW_FORMAT_NONE`* | | No resource view format (use underlying resource format) | -| 0x01 |*`CU_RES_VIEW_FORMAT_UINT_1X8`* | | 1 channel unsigned 8-bit integers | -| 0x02 |*`CU_RES_VIEW_FORMAT_UINT_2X8`* | | 2 channel unsigned 8-bit integers | -| 0x03 |*`CU_RES_VIEW_FORMAT_UINT_4X8`* | | 4 channel unsigned 8-bit integers | -| 0x04 |*`CU_RES_VIEW_FORMAT_SINT_1X8`* | | 1 channel signed 8-bit integers | -| 0x05 |*`CU_RES_VIEW_FORMAT_SINT_2X8`* | | 2 channel signed 8-bit integers | -| 0x06 |*`CU_RES_VIEW_FORMAT_SINT_4X8`* | | 4 channel signed 8-bit integers | -| 0x07 |*`CU_RES_VIEW_FORMAT_UINT_1X16`* | | 1 channel unsigned 16-bit integers | -| 0x08 |*`CU_RES_VIEW_FORMAT_UINT_2X16`* | | 2 channel unsigned 16-bit integers | -| 0x09 |*`CU_RES_VIEW_FORMAT_UINT_4X16`* | | 4 channel unsigned 16-bit integers | -| 0x0a |*`CU_RES_VIEW_FORMAT_SINT_1X16`* | | 1 channel signed 16-bit integers | -| 0x0b |*`CU_RES_VIEW_FORMAT_SINT_2X16`* | | 2 channel signed 16-bit integers | -| 0x0c |*`CU_RES_VIEW_FORMAT_SINT_4X16`* | | 4 channel signed 16-bit integers | -| 0x0d |*`CU_RES_VIEW_FORMAT_UINT_1X32`* | | 1 channel unsigned 32-bit integers | -| 0x0e |*`CU_RES_VIEW_FORMAT_UINT_2X32`* | | 2 channel unsigned 32-bit integers | -| 0x0f |*`CU_RES_VIEW_FORMAT_UINT_4X32`* | | 4 channel unsigned 32-bit integers | -| 0x10 |*`CU_RES_VIEW_FORMAT_SINT_1X32`* | | 1 channel signed 32-bit integers | -| 0x11 |*`CU_RES_VIEW_FORMAT_SINT_2X32`* | | 2 channel signed 32-bit integers | -| 0x12 |*`CU_RES_VIEW_FORMAT_SINT_4X32`* | | 4 channel signed 32-bit integers | -| 0x13 |*`CU_RES_VIEW_FORMAT_FLOAT_1X16`* | | 1 channel 16-bit floating point | -| 0x14 |*`CU_RES_VIEW_FORMAT_FLOAT_2X16`* | | 2 channel 16-bit floating point | -| 0x15 |*`CU_RES_VIEW_FORMAT_FLOAT_4X16`* | | 4 channel 16-bit floating point | -| 0x16 |*`CU_RES_VIEW_FORMAT_FLOAT_1X32`* | | 1 channel 32-bit floating point | -| 0x17 |*`CU_RES_VIEW_FORMAT_FLOAT_2X32`* | | 2 channel 32-bit floating point | -| 0x18 |*`CU_RES_VIEW_FORMAT_FLOAT_4X32`* | | 4 channel 32-bit floating point | -| 0x19 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC1`* | | Block compressed 1 | -| 0x1a |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC3`* | | Block compressed 2 | -| 0x1b |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC3`* | | Block compressed 3 | -| 0x1c |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC4`* | | Block compressed 4 unsigned | -| 0x1d |*`CU_RES_VIEW_FORMAT_SIGNED_BC4`* | | Block compressed 4 signed | -| 0x1e |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC5`* | | Block compressed 5 unsigned | -| 0x1f |*`CU_RES_VIEW_FORMAT_SIGNED_BC5`* | | Block compressed 5 signed | -| 0x20 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC6H`* | | Block compressed 6 unsigned half-float | -| 0x21 |*`CU_RES_VIEW_FORMAT_SIGNED_BC6H`* | | Block compressed 6 signed half-float | -| 0x22 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC7`* | | Block compressed 7 | -| enum |***`CUresult`*** |***`hipError_t`*** | Error codes | -| 0 |*`CUDA_SUCCESS`* |*`hipSuccess`* | The API call returned with no errors. In the case of query calls, this can also mean that the operation being queried is complete (see cuEventQuery() and cuStreamQuery()). | -| 1 |*`CUDA_ERROR_INVALID_VALUE`* |*`hipErrorInvalidValue`* | This indicates that one or more of the parameters passed to the API call is not within an acceptable range of values. | -| 2 |*`CUDA_ERROR_OUT_OF_MEMORY`* |*`hipErrorMemoryAllocation`* | The API call failed because it was unable to allocate enough memory to perform the requested operation. | -| 3 |*`CUDA_ERROR_NOT_INITIALIZED`* |*`hipErrorNotInitialized`* | This indicates that the CUDA driver has not been initialized with cuInit() or that initialization has failed. | -| 4 |*`CUDA_ERROR_DEINITIALIZED`* |*`hipErrorDeinitialized`* | This indicates that the CUDA driver is in the process of shutting down. | -| 5 |*`CUDA_ERROR_PROFILER_DISABLED`* |*`hipErrorProfilerDisabled`* | This indicates profiler is not initialized for this run. This can happen when the application is running with external profiling tools like visual profiler. | -| 6 |*`CUDA_ERROR_PROFILER_NOT_INITIALIZED`* |*`hipErrorProfilerNotInitialized`* | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to attempt to enable/disable the profiling via cuProfilerStart or cuProfilerStop without initialization. | -| 7 |*`CUDA_ERROR_PROFILER_ALREADY_STARTED`* |*`hipErrorProfilerAlreadyStarted`* | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to call cuProfilerStart() when profiling is already enabled. | -| 8 |*`CUDA_ERROR_PROFILER_ALREADY_STOPPED`* |*`hipErrorProfilerAlreadyStopped`* | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to call cuProfilerStop() when profiling is already disabled. | -| 100 |*`CUDA_ERROR_NO_DEVICE`* |*`hipErrorNoDevice`* | This indicates that no CUDA-capable devices were detected by the installed CUDA driver. | -| 101 |*`CUDA_ERROR_INVALID_DEVICE`* |*`hipErrorInvalidDevice`* | This indicates that the device ordinal supplied by the user does not correspond to a valid CUDA device. | -| 200 |*`CUDA_ERROR_INVALID_IMAGE`* |*`hipErrorInvalidImage`* | This indicates that the device kernel image is invalid. This can also indicate an invalid CUDA module. | -| 201 |*`CUDA_ERROR_INVALID_CONTEXT`* |*`hipErrorInvalidContext`* | This most frequently indicates that there is no context bound to the current thread. This can also be returned if the context passed to an API call is not a valid handle (such as a context that has had cuCtxDestroy() invoked on it). This can also be returned if a user mixes different API versions (i.e. 3010 context with 3020 API calls). See cuCtxGetApiVersion() for more details. | -| 202 |*`CUDA_ERROR_CONTEXT_ALREADY_CURRENT`* |*`hipErrorContextAlreadyCurrent`* | This indicated that the context being supplied as a parameter to the API call was already the active context. Deprecated This error return is deprecated as of CUDA 3.2. It is no longer an error to attempt to push the active context via cuCtxPushCurrent(). | -| 205 |*`CUDA_ERROR_MAP_FAILED`* |*`hipErrorMapFailed`* | This indicates that a map or register operation has failed. | -| 206 |*`CUDA_ERROR_UNMAP_FAILED`* |*`hipErrorUnmapFailed`* | This indicates that an unmap or unregister operation has failed. | -| 207 |*`CUDA_ERROR_ARRAY_IS_MAPPED`* |*`hipErrorArrayIsMapped`* | This indicates that the specified array is currently mapped and thus cannot be destroyed. | -| 208 |*`CUDA_ERROR_ALREADY_MAPPED`* |*`hipErrorAlreadyMapped`* | This indicates that the resource is already mapped. | -| 209 |*`CUDA_ERROR_NO_BINARY_FOR_GPU`* |*`hipErrorNoBinaryForGpu* | This indicates that there is no kernel image available that is suitable for the device. This can occur when a user specifies code generation options for a particular CUDA source file that do not include the corresponding device configuration. | -| 210 |*`CUDA_ERROR_ALREADY_ACQUIRED`* |*`hipErrorAlreadyAcquired* | This indicates that a resource has already been acquired. | -| 211 |*`CUDA_ERROR_NOT_MAPPED`* |*`hipErrorNotMapped`* | This indicates that a resource is not mapped. | -| 212 |*`CUDA_ERROR_NOT_MAPPED_AS_ARRAY`* |*`hipErrorNotMappedAsArray`* | This indicates that a mapped resource is not available for access as an array. | -| 213 |*`CUDA_ERROR_NOT_MAPPED_AS_POINTER`* |*`hipErrorNotMappedAsPointer`* | This indicates that a mapped resource is not available for access as a pointer. | -| 214 |*`CUDA_ERROR_ECC_UNCORRECTABLE`* |*`hipErrorECCNotCorrectable`* | This indicates that an uncorrectable ECC error was detected during execution. | -| 215 |*`CUDA_ERROR_UNSUPPORTED_LIMIT`* |*`hipErrorUnsupportedLimit`* | This indicates that the CUlimit passed to the API call is not supported by the active device. | -| 216 |*`CUDA_ERROR_CONTEXT_ALREADY_IN_USE`* |*`hipErrorContextAlreadyInUse`* | This indicates that the CUcontext passed to the API call can only be bound to a single CPU thread at a time but is already bound to a CPU thread. | -| 217 |*`CUDA_ERROR_PEER_ACCESS_UNSUPPORTED`* |*`hipErrorPeerAccessUnsupported`* | This indicates that peer access is not supported across the given devices. | -| 218 |*`CUDA_ERROR_INVALID_PTX`* |*`hipErrorInvalidKernelFile`* | This indicates that a PTX JIT compilation failed. | -| 219 |*`CUDA_ERROR_INVALID_GRAPHICS_CONTEXT`* |*`hipErrorInvalidGraphicsContext`* | This indicates an error with OpenGL or DirectX context. | -| 300 |*`CUDA_ERROR_INVALID_SOURCE`* |*`hipErrorInvalidSource`* | This indicates that the device kernel source is invalid. | -| 301 |*`CUDA_ERROR_FILE_NOT_FOUND`* |*`hipErrorFileNotFound`* | This indicates that the file specified was not found. | -| 302 |*`CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND`* |*`hipErrorSharedObjectSymbolNotFound`* | This indicates that a link to a shared object failed to resolve. | -| 303 |*`CUDA_ERROR_SHARED_OBJECT_INIT_FAILED`* |*`hipErrorSharedObjectInitFailed`* | This indicates that initialization of a shared object failed. | -| 304 |*`CUDA_ERROR_OPERATING_SYSTEM`* |*`hipErrorOperatingSystem`* | This indicates that an OS call failed. | -| 400 |*`CUDA_ERROR_INVALID_HANDLE`* |*`hipErrorInvalidResourceHandle`* | This indicates that a resource handle passed to the API call was not valid. Resource handles are opaque types like CUstream and CUevent. | -| 500 |*`CUDA_ERROR_NOT_FOUND`* |*`hipErrorNotFound`* | This indicates that a named symbol was not found. Examples of symbols are global/constant variable names, texture names, and surface names. | -| 600 |*`CUDA_ERROR_NOT_READY`* |*`hipErrorNotReady`* | This indicates that asynchronous operations issued previously have not completed yet. This result is not actually an error, but must be indicated differently than CUDA_SUCCESS (which indicates completion). Calls that may return this value include cuEventQuery() and cuStreamQuery(). | -| 700 |*`CUDA_ERROR_ILLEGAL_ADDRESS`* |*`hipErrorIllegalAddress`* | While executing a kernel, the device encountered a load or store instruction on an invalid memory address. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 701 |*`CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES`* |*`hipErrorLaunchOutOfResources`* | This indicates that a launch did not occur because it did not have appropriate resources. This error usually indicates that the user has attempted to pass too many arguments to the device kernel, or the kernel launch specifies too many threads for the kernel's register count. Passing arguments of the wrong size (i.e. a 64-bit pointer when a 32-bit int is expected) is equivalent to passing too many arguments and can also result in this error. | -| 702 |*`CUDA_ERROR_LAUNCH_TIMEOUT`* |*`hipErrorLaunchTimeOut`* | This indicates that the device kernel took too long to execute. This can only occur if timeouts are enabled - see the device attribute CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT for more information. The context cannot be used (and must be destroyed similar to CUDA_ERROR_LAUNCH_FAILED). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 703 |*`CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING`* | | This error indicates a kernel launch that uses an incompatible texturing mode. | -| 704 |*`CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED`* |*`hipErrorPeerAccessAlreadyEnabled`* | This error indicates that a call to cuCtxEnablePeerAccess() is trying to re-enable peer access to a context which has already had peer access to it enabled. | -| 705 |*`CUDA_ERROR_PEER_ACCESS_NOT_ENABLED`* |*`hipErrorPeerAccessNotEnabled`* | This error indicates that cuCtxDisablePeerAccess() is trying to disable peer access which has not been enabled yet via cuCtxEnablePeerAccess(). | -| 708 |*`CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE`* | | This error indicates that cuCtxDisablePeerAccess() is trying to disable peer access which has not been enabled yet via cuCtxEnablePeerAccess(). | -| 709 |*`CUDA_ERROR_CONTEXT_IS_DESTROYED`* | | This error indicates that the context current to the calling thread has been destroyed using cuCtxDestroy, or is a primary context which has not yet been initialized. | -| 710 |*`CUDA_ERROR_ASSERT`* | | A device-side assert triggered during kernel execution. The context cannot be used anymore, and must be destroyed. All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 711 |*`CUDA_ERROR_TOO_MANY_PEERS`* | | This error indicates that the hardware resources required to enable peer access have been exhausted for one or more of the devices passed to cuCtxEnablePeerAccess(). | -| 712 |*`CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED`* |*`hipErrorHostMemoryAlreadyRegistered`* | This error indicates that the memory range passed to cuMemHostRegister() has already been registered. | -| 713 |*`CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED`* |*`hipErrorHostMemoryNotRegistered`* | This error indicates that the pointer passed to cuMemHostUnregister() does not correspond to any currently registered memory region. | -| 714 |*`CUDA_ERROR_HARDWARE_STACK_ERROR`* | | While executing a kernel, the device encountered a stack error. This can be due to stack corruption or exceeding the stack size limit. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 715 |*`CUDA_ERROR_ILLEGAL_INSTRUCTION`* | | While executing a kernel, the device encountered an illegal instruction. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 716 |*`CUDA_ERROR_MISALIGNED_ADDRESS`* | | While executing a kernel, the device encountered a load or store instruction on a memory address which is not aligned. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 717 |*`CUDA_ERROR_INVALID_ADDRESS_SPACE`* | | While executing a kernel, the device encountered an instruction which can only operate on memory locations in certain address spaces (global, shared, or local), but was supplied a memory address not belonging to an allowed address space. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 718 |*`CUDA_ERROR_INVALID_PC`* | | While executing a kernel, the device program counter wrapped its address space. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 719 |*`CUDA_ERROR_LAUNCH_FAILED`* | | An exception occurred on the device while executing a kernel. Common causes include dereferencing an invalid device pointer and accessing out of bounds shared memory. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA. | -| 800 |*`CUDA_ERROR_NOT_PERMITTED`* | | This error indicates that the attempted operation is not permitted. | -| 801 |*`CUDA_ERROR_NOT_SUPPORTED`* | | This error indicates that the attempted operation is not supported on the current system or device. | -| 999 |*`CUDA_ERROR_UNKNOWN`* | | This indicates that an unknown internal error has occurred. | -| enum |***`CUstream_flags`*** |***`hipStreamFlags`*** | Stream creation flags | -| 0x0 |*`CU_STREAM_DEFAULT`* |*`hipStreamDefault`* | Default stream flag | -| 0x1 |*`CU_STREAM_NON_BLOCKING`* |*`hipStreamNonBlocking`* | Stream does not synchronize with stream 0 (the NULL stream) | -| typedef | `CUarray` | `hipArray *` | CUDA array | -| struct | `CUarray_st` | `hipArray` | CUDA array | -| typedef | `CUcontext` | `hipCtx_t` | CUDA context | -| typedef | `CUdevice` | `hipDevice_t` | CUDA device | -| typedef | `CUdeviceptr` | `hipDeviceptr_t` | CUDA device pointer CUdeviceptr is defined as an unsigned integer type whose size matches the size of a pointer on the target platform. | -| typedef | `CUevent` | `hipEvent_t` | CUDA event | -| typedef | `CUfunction` | `hipFunction_t` | CUDA function | -| typedef | `CUgraphicsResource` | | CUDA graphics interop resource | -| typedef | `CUmipmappedArray` | | CUDA mipmapped array | -| typedef | `CUmodule` | `hipModule_t` | CUDA module | -| typedef | `CUstream` | `hipStream_t` | CUDA module | -| typedef | `CUstreamCallback` | `hipStreamCallback_t` | CUDA stream callback | -| typedef | `CUsurfObject` | | An opaque value that represents a CUDA surface object | -| typedef | `CUsurfref` | | CUDA surface reference | -| typedef | `CUtexObject` | | An opaque value that represents a CUDA texture object | -| typedef | `CUtexref` | | CUDA texture reference | -| define |`CU_IPC_HANDLE_SIZE` | | CUDA IPC handle size. | -| define |`CU_LAUNCH_PARAM_BUFFER_POINTER` | `HIP_LAUNCH_PARAM_BUFFER_POINTER` | Indicator that the next value in the extra parameter to cuLaunchKernel will be a pointer to a buffer containing all kernel parameters used for launching kernel f. This buffer needs to honor all alignment/padding requirements of the individual parameters. If CU_LAUNCH_PARAM_BUFFER_SIZE is not also specified in the extra array, then CU_LAUNCH_PARAM_BUFFER_POINTER will have no effect. | -| define |`CU_LAUNCH_PARAM_BUFFER_SIZE` | `HIP_LAUNCH_PARAM_BUFFER_SIZE` | Indicator that the next value in the extra parameter to cuLaunchKernel will be a pointer to a size_t which contains the size of the buffer specified with CU_LAUNCH_PARAM_BUFFER_POINTER. It is required that CU_LAUNCH_PARAM_BUFFER_POINTER also be specified in the extra array if the value associated with CU_LAUNCH_PARAM_BUFFER_SIZE is not zero. | -| define |`CU_LAUNCH_PARAM_END` | `HIP_LAUNCH_PARAM_END` | End of array terminator for the extra parameter to cuLaunchKernel. | -| define |`CU_MEMHOSTALLOC_DEVICEMAP` | | If set, host memory is mapped into CUDA address space and cuMemHostGetDevicePointer() may be called on the host pointer. Flag for cuMemHostAlloc(). | -| define |`CU_MEMHOSTALLOC_PORTABLE` | | If set, host memory is portable between CUDA contexts. Flag for cuMemHostAlloc(). | -| define |`CU_MEMHOSTALLOC_WRITECOMBINED` | | If set, host memory is allocated as write-combined - fast to write, faster to DMA, slow to read except via SSE4 streaming load instruction (MOVNTDQA). Flag for cuMemHostAlloc(). | -| define |`CU_MEMHOSTREGISTER_DEVICEMAP` | | If set, host memory is mapped into CUDA address space and cuMemHostGetDevicePointer() may be called on the host pointer. Flag for cuMemHostRegister(). | -| define |`CU_MEMHOSTREGISTER_IOMEMORY` | | If set, the passed memory pointer is treated as pointing to some memory-mapped I/O space, e.g. belonging to a third-party PCIe device. On Windows the flag is a no-op. On Linux that memory is marked as non cache-coherent for the GPU and is expected to be physically contiguous. It may return CUDA_ERROR_NOT_PERMITTED if run as an unprivileged user, CUDA_ERROR_NOT_SUPPORTED on older Linux kernel versions. On all other platforms, it is not supported and CUDA_ERROR_NOT_SUPPORTED is returned. Flag for cuMemHostRegister(). | -| define |`CU_MEMHOSTREGISTER_PORTABLE` | | If set, host memory is portable between CUDA contexts. Flag for cuMemHostRegister(). | -| define |`CU_PARAM_TR_DEFAULT` | | For texture references loaded into the module, use default texunit from texture reference. | -| define |`CU_STREAM_LEGACY` | | Legacy stream handle. Stream handle that can be passed as a CUstream to use an implicit stream with legacy synchronization behavior. See details of the synchronization behavior. | -| define |`CU_STREAM_PER_THREAD` | | Per-thread stream handle. Stream handle that can be passed as a CUstream to use an implicit stream with perthread synchronization behavior. See details of the synchronization behavior. | -| define |`CU_TRSA_OVERRIDE_FORMAT` | | Override the texref format with a format inferred from the array. Flag for cuTexRefSetArray(). | -| define |`CU_TRSF_NORMALIZED_COORDINATES` | | Use normalized texture coordinates in the range [0,1) instead of [0,dim). Flag for cuTexRefSetFlags(). | -| define |`CU_TRSF_SRGB` | | Perform sRGB->linear conversion during texture read. Flag for cuTexRefSetFlags(). | -| define |`CUDA_ARRAY3D_2DARRAY` | | Deprecated, use CUDA_ARRAY3D_LAYERED. | -| define |`CUDA_ARRAY3D_CUBEMAP` | | If set, the CUDA array is a collection of six 2D arrays, representing faces of a cube. The width of such a CUDA array must be equal to its height, and Depth must be six. If CUDA_ARRAY3D_LAYERED flag is also set, then the CUDA array is a collection of cubemaps and Depth must be a multiple of six. | -| define |`CUDA_ARRAY3D_DEPTH_TEXTURE` | | This flag if set indicates that the CUDA array is a DEPTH_TEXTURE. | -| define |`CUDA_ARRAY3D_LAYERED` | | If set, the CUDA array is a collection of layers, where each layer is either a 1D or a 2D array and the Depth member of CUDA_ARRAY3D_DESCRIPTOR specifies the number of layers, not the depth of a 3D array. | -| define |`CUDA_ARRAY3D_SURFACE_LDST` | | This flag must be set in order to bind a surface reference to the CUDA array. | -| define |`CUDA_ARRAY3D_TEXTURE_GATHER` | | This flag must be set in order to perform texture gather operations on a CUDA array. | -| define |`CUDA_VERSION` | | CUDA API version number. | +| **type** | **CUDA** | **HIP** | +|-------------:|---------------------------------------------------------------|------------------------------------------------------------| +| struct | `CUDA_ARRAY3D_DESCRIPTOR` | | +| struct | `CUDA_ARRAY_DESCRIPTOR` | | +| struct | `CUDA_MEMCPY2D` | | +| struct | `CUDA_MEMCPY3D` | | +| struct | `CUDA_MEMCPY3D_PEER` | | +| struct | `CUDA_POINTER_ATTRIBUTE_P2P_TOKENS` | | +| struct | `CUDA_RESOURCE_DESC` | | +| struct | `CUDA_RESOURCE_VIEW_DESC` | | +| struct | `CUdevprop` | `hipDeviceProp_t` | +| struct | `CUipcEventHandle` | | +| struct | `CUipcMemHandle` | | +| enum |***`CUaddress_mode`*** | | +| 0 |*`CU_TR_ADDRESS_MODE_WRAP`* | | +| 1 |*`CU_TR_ADDRESS_MODE_CLAMP`* | | +| 2 |*`CU_TR_ADDRESS_MODE_MIRROR`* | | +| 3 |*`CU_TR_ADDRESS_MODE_BORDER`* | | +| enum |***`CUarray_cubemap_face`*** | | +| 0x00 |*`CU_CUBEMAP_FACE_POSITIVE_X`* | | +| 0x01 |*`CU_CUBEMAP_FACE_NEGATIVE_X`* | | +| 0x02 |*`CU_CUBEMAP_FACE_POSITIVE_Y`* | | +| 0x03 |*`CU_CUBEMAP_FACE_NEGATIVE_Y`* | | +| 0x04 |*`CU_CUBEMAP_FACE_POSITIVE_Z`* | | +| 0x05 |*`CU_CUBEMAP_FACE_NEGATIVE_Z`* | | +| enum |***`CUarray_format`*** | | +| 0x01 |*`CU_AD_FORMAT_UNSIGNED_INT8`* | | +| 0x02 |*`CU_AD_FORMAT_UNSIGNED_INT16`* | | +| 0x03 |*`CU_AD_FORMAT_UNSIGNED_INT32`* | | +| 0x08 |*`CU_AD_FORMAT_SIGNED_INT8`* | | +| 0x09 |*`CU_AD_FORMAT_SIGNED_INT16`* | | +| 0x0a |*`CU_AD_FORMAT_SIGNED_INT32`* | | +| 0x10 |*`CU_AD_FORMAT_HALF`* | | +| 0x20 |*`CU_AD_FORMAT_FLOAT`* | | +| enum |***`CUctx_flags`*** | | +| 0x00 |*`CU_CTX_SCHED_AUTO`* | | +| 0x01 |*`CU_CTX_SCHED_SPIN`* | | +| 0x02 |*`CU_CTX_SCHED_YIELD`* | | +| 0x04 |*`CU_CTX_SCHED_BLOCKING_SYNC`* | | +| 0x04 |*`CU_CTX_BLOCKING_SYNC`* | | +| 0x07 |*`CU_CTX_SCHED_MASK`* | | +| 0x08 |*`CU_CTX_MAP_HOST`* | | +| 0x10 |*`CU_CTX_LMEM_RESIZE_TO_MAX`* | | +| 0x1f |*`CU_CTX_FLAGS_MASK`* | | +| enum |***`CUdevice_attribute`*** | | +| 1 |*`CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK`* |*`hipDeviceAttributeMaxThreadsPerBlock`* | +| 2 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X`* |*`hipDeviceAttributeMaxBlockDimX`* | +| 3 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y`* |*`hipDeviceAttributeMaxBlockDimY`* | +| 4 |*`CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z`* |*`hipDeviceAttributeMaxBlockDimZ`* | +| 5 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X`* |*`hipDeviceAttributeMaxGridDimX`* | +| 6 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y`* |*`hipDeviceAttributeMaxGridDimY`* | +| 7 |*`CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z`* |*`hipDeviceAttributeMaxGridDimZ`* | +| 8 |*`CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | +| 8 |*`CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | +| 9 |*`CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY`* |*`hipDeviceAttributeTotalConstantMemory`* | +| 10 |*`CU_DEVICE_ATTRIBUTE_WARP_SIZE`* |*`hipDeviceAttributeWarpSize`* | +| 11 |*`CU_DEVICE_ATTRIBUTE_MAX_PITCH`* | | +| 12 |*`CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | +| 12 |*`CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | +| 13 |*`CU_DEVICE_ATTRIBUTE_CLOCK_RATE`* |*`hipDeviceAttributeClockRate`* | +| 14 |*`CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT`* | | +| 15 |*`CU_DEVICE_ATTRIBUTE_GPU_OVERLAP`* | | +| 16 |*`CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT`* |*`hipDeviceAttributeMultiprocessorCount`* | +| 17 |*`CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT`* | | +| 18 |*`CU_DEVICE_ATTRIBUTE_INTEGRATED`* | | +| 19 |*`CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY`* | | +| 20 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_MODE`* |*`hipDeviceAttributeComputeMode`* | +| 21 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH`* | | +| 22 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH`* | | +| 23 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT`* | | +| 24 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH`* | | +| 25 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT`* | | +| 26 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH`* | | +| 27 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH`* | | +| 28 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT`* | | +| 29 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS`* | | +| 27 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH`* | | +| 28 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT`* | | +| 29 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES`* | | +| 30 |*`CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT`* | | +| 31 |*`CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS`* |*`hipDeviceAttributeConcurrentKernels`* | +| 32 |*`CU_DEVICE_ATTRIBUTE_ECC_ENABLED`* | | +| 33 |*`CU_DEVICE_ATTRIBUTE_PCI_BUS_ID`* |*`hipDeviceAttributePciBusId`* | +| 34 |*`CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID`* |*`hipDeviceAttributePciDeviceId`* | +| 35 |*`CU_DEVICE_ATTRIBUTE_TCC_DRIVER`* | | +| 36 |*`CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE`* |*`hipDeviceAttributeMemoryClockRate`* | +| 37 |*`CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH`* |*`hipDeviceAttributeMemoryBusWidth`* | +| 38 |*`CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE`* |*`hipDeviceAttributeL2CacheSize`* | +| 39 |*`CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR`* |*`hipDeviceAttributeMaxThreadsPerMultiProcessor`* | +| 40 |*`CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT`* | | +| 41 |*`CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING`* | | +| 42 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH`* | | +| 43 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS`* | | +| 44 |*`CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER`* | | +| 45 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH`* | | +| 46 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT`* | | +| 47 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE`* | | +| 48 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE`* | | +| 49 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE`* | | +| 50 |*`CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID`* | | +| 51 |*`CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT`* | | +| 52 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH`* | | +| 53 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH`* | | +| 54 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS`* | | +| 55 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH`* | | +| 56 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH`* | | +| 57 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT`* | | +| 58 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH`* | | +| 59 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT`* | | +| 60 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH`* | | +| 61 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH`* | | +| 62 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS`* | | +| 63 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH`* | | +| 64 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT`* | | +| 65 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS`* | | +| 66 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH`* | | +| 67 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH`* | | +| 68 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS`* | | +| 69 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH`* | | +| 70 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH`* | | +| 71 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT`* | | +| 72 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH`* | | +| 73 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH`* | | +| 74 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT`* | | +| 75 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR`* |*`hipDeviceAttributeComputeCapabilityMajor`* | +| 76 |*`CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR`* |*`hipDeviceAttributeComputeCapabilityMinor`* | +| 77 |*`CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH`* | | +| 78 |*`CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED`* | | +| 79 |*`CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED`* | | +| 80 |*`CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED`* | | +| 81 |*`CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR`* |*`hipDeviceAttributeMaxSharedMemoryPerMultiprocessor`* | +| 82 |*`CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR`* | | +| 83 |*`CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY`* |*`hipDeviceAttributeManagedMemory`* | +| 84 |*`CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD`* | | +| 85 |*`CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID`* | | +| 86 |*`CU_DEVICE_ATTRIBUTE_MAX`* | | +| enum |***`CUevent_flags`*** | | +| 0x00 |*`CU_EVENT_DEFAULT`* |*`hipEventDefault`* | +| 0x01 |*`CU_EVENT_BLOCKING_SYNC`* |*`hipEventBlockingSync`* | +| 0x02 |*`CU_EVENT_DISABLE_TIMING`* |*`hipEventDisableTiming`* | +| 0x04 |*`CU_EVENT_INTERPROCESS`* |*`hipEventInterprocess`* | +| enum |***`CUfilter_mode`*** |***`hipTextureFilterMode`*** | +| 0 |*`CU_TR_FILTER_MODE_POINT`* |*`hipFilterModePoint`* | +| 1 |*`CU_TR_FILTER_MODE_LINEAR`* |*`hipFilterModeLinear`* | +| enum |***`CUfunc_cache`*** |***`hipFuncCache`*** | +| 0x00 |*`CU_FUNC_CACHE_PREFER_NONE`* |*`hipFuncCachePreferNone`* | +| 0x01 |*`CU_FUNC_CACHE_PREFER_SHARED`* |*`hipFuncCachePreferShared`* | +| 0x02 |*`CU_FUNC_CACHE_PREFER_L1`* |*`hipFuncCachePreferL1`* | +| 0x03 |*`CU_FUNC_CACHE_PREFER_EQUAL`* |*`hipFuncCachePreferEqual`* | +| enum |***`CUfunction_attribute`*** | | +| 0 |*`CU_FUNC_ATTRIBUTE_MAX_THREADS_PER_BLOCK`* | | +| 1 |*`CU_FUNC_ATTRIBUTE_SHARED_SIZE_BYTES`* | | +| 2 |*`CU_FUNC_ATTRIBUTE_CONST_SIZE_BYTES`* | | +| 3 |*`CU_FUNC_ATTRIBUTE_LOCAL_SIZE_BYTES`* | | +| 4 |*`CU_FUNC_ATTRIBUTE_NUM_REGS`* | | +| 5 |*`CU_FUNC_ATTRIBUTE_PTX_VERSION`* | | +| 6 |*`CU_FUNC_ATTRIBUTE_BINARY_VERSION`* | | +| 7 |*`CU_FUNC_ATTRIBUTE_CACHE_MODE_CA`* | | +| 8 |*`CU_FUNC_ATTRIBUTE_MAX`* | | +| enum |***`CUgraphicsMapResourceFlags`*** | | +| 0x00 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE`* | | +| 0x01 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY`* | | +| 0x02 |*`CU_GRAPHICS_MAP_RESOURCE_FLAGS_WRITE_DISCARD`* | | +| enum |***`CUgraphicsRegisterFlags`*** | | +| 0x00 |*`CU_GRAPHICS_REGISTER_FLAGS_NONE`* | | +| 0x01 |*`CU_GRAPHICS_REGISTER_FLAGS_READ_ONLY`* | | +| 0x02 |*`CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD`* | | +| 0x04 |*`CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST`* | | +| 0x08 |*`CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER`* | | +| enum |***`CUipcMem_flags`*** | | +| 0x1 |*`CU_IPC_MEM_LAZY_ENABLE_PEER_ACCESS`* |*`hipIpcMemLazyEnablePeerAccess`* | +| enum |***`CUjit_cacheMode`*** | | +| 0 |*`CU_JIT_CACHE_OPTION_NONE`* | | +| |*`CU_JIT_CACHE_OPTION_CG`* | | +| |*`CU_JIT_CACHE_OPTION_CA`* | | +| enum |***`CUjit_fallback`*** | | +| 0 |*`CU_PREFER_PTX`* | | +| |*`CU_PREFER_BINARY`* | | +| enum |***`CUjit_option`*** | | +| 0 |*`CU_JIT_MAX_REGISTERS`* | | +| |*`CU_JIT_THREADS_PER_BLOCK`* | | +| |*`CU_JIT_WALL_TIME`* | | +| |*`CU_JIT_INFO_LOG_BUFFER`* | | +| |*`CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES`* | | +| |*`CU_JIT_OPTIMIZATION_LEVEL`* | | +| |*`CU_JIT_TARGET_FROM_CUCONTEXT`* | | +| |*`CU_JIT_TARGET`* | | +| |*`CU_JIT_FALLBACK_STRATEGY`* | | +| |*`CU_JIT_GENERATE_DEBUG_INFO`* | | +| |*`CU_JIT_LOG_VERBOSE`* | | +| |*`CU_JIT_GENERATE_LINE_INFO`* | | +| |*`CU_JIT_CACHE_MODE`* | | +| |*`CU_JIT_NUM_OPTIONS`* | | +| enum |***`CUjit_target`*** | | +| 10 |*`CU_TARGET_COMPUTE_10`* | | +| 11 |*`CU_TARGET_COMPUTE_11`* | | +| 12 |*`CU_TARGET_COMPUTE_12`* | | +| 13 |*`CU_TARGET_COMPUTE_13`* | | +| 20 |*`CU_TARGET_COMPUTE_20`* | | +| 21 |*`CU_TARGET_COMPUTE_21`* | | +| 30 |*`CU_TARGET_COMPUTE_30`* | | +| 32 |*`CU_TARGET_COMPUTE_32`* | | +| 35 |*`CU_TARGET_COMPUTE_35`* | | +| 37 |*`CU_TARGET_COMPUTE_37`* | | +| 50 |*`CU_TARGET_COMPUTE_50`* | | +| 52 |*`CU_TARGET_COMPUTE_52`* | | +| enum |***`CUjitInputType`*** | | +| 0 |*`CU_JIT_INPUT_CUBIN`* | | +| |*`CU_JIT_INPUT_PTX`* | | +| |*`CU_JIT_INPUT_FATBINARY`* | | +| |*`CU_JIT_INPUT_OBJECT`* | | +| |*`CU_JIT_INPUT_LIBRARY`* | | +| |*`CU_JIT_NUM_INPUT_TYPES`* | | +| enum |***`CUlimit`*** |***`hipLimit_t`*** | +| 0x00 |*`CU_LIMIT_STACK_SIZE`* | | +| 0x01 |*`CU_LIMIT_PRINTF_FIFO_SIZE`* | | +| 0x02 |*`CU_LIMIT_MALLOC_HEAP_SIZE`* |*`hipLimitMallocHeapSize`* | +| 0x03 |*`CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH`* | | +| 0x04 |*`CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT`* | | +| |*`CU_LIMIT_MAX`* | | +| enum |***`CUmemAttach_flags`*** | | +| 0x1 |*`CU_MEM_ATTACH_GLOBAL`* | | +| 0x2 |*`CU_MEM_ATTACH_HOST`* | | +| 0x4 |*`CU_MEM_ATTACH_SINGLE`* | | +| enum |***`CUmemorytype`*** | | +| 0x01 |*`CU_MEMORYTYPE_HOST`* | | +| 0x02 |*`CU_MEMORYTYPE_DEVICE`* | | +| 0x03 |*`CU_MEMORYTYPE_ARRAY`* | | +| 0x04 |*`CU_MEMORYTYPE_UNIFIED`* | | +| enum |***`CUoccupancy_flags`*** | | +| 0x00 |*`CU_OCCUPANCY_DEFAULT`* | | +| 0x01 |*`CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE`* | | +| enum |***`CUpointer_attribute`*** | | +| 1 |*`CU_POINTER_ATTRIBUTE_CONTEXT`* | | +| 2 |*`CU_POINTER_ATTRIBUTE_MEMORY_TYPE`* | | +| 3 |*`CU_POINTER_ATTRIBUTE_DEVICE_POINTER`* | | +| 4 |*`CU_POINTER_ATTRIBUTE_HOST_POINTER`* | | +| 5 |*`CU_POINTER_ATTRIBUTE_P2P_TOKENS`* | | +| 6 |*`CU_POINTER_ATTRIBUTE_SYNC_MEMOPS`* | | +| 7 |*`CU_POINTER_ATTRIBUTE_BUFFER_ID`* | | +| 8 |*`CU_POINTER_ATTRIBUTE_IS_MANAGED`* | | +| enum |***`CUmemorytype`*** | | +| 0x00 |*`CU_RESOURCE_TYPE_ARRAY`* | | +| 0x01 |*`CU_RESOURCE_TYPE_MIPMAPPED_ARRAY`* | | +| 0x02 |*`CU_RESOURCE_TYPE_LINEAR`* | | +| 0x03 |*`CU_RESOURCE_TYPE_PITCH2D`* | | +| enum |***`CUresourceViewFormat`*** | | +| 0x00 |*`CU_RES_VIEW_FORMAT_NONE`* | | +| 0x01 |*`CU_RES_VIEW_FORMAT_UINT_1X8`* | | +| 0x02 |*`CU_RES_VIEW_FORMAT_UINT_2X8`* | | +| 0x03 |*`CU_RES_VIEW_FORMAT_UINT_4X8`* | | +| 0x04 |*`CU_RES_VIEW_FORMAT_SINT_1X8`* | | +| 0x05 |*`CU_RES_VIEW_FORMAT_SINT_2X8`* | | +| 0x06 |*`CU_RES_VIEW_FORMAT_SINT_4X8`* | | +| 0x07 |*`CU_RES_VIEW_FORMAT_UINT_1X16`* | | +| 0x08 |*`CU_RES_VIEW_FORMAT_UINT_2X16`* | | +| 0x09 |*`CU_RES_VIEW_FORMAT_UINT_4X16`* | | +| 0x0a |*`CU_RES_VIEW_FORMAT_SINT_1X16`* | | +| 0x0b |*`CU_RES_VIEW_FORMAT_SINT_2X16`* | | +| 0x0c |*`CU_RES_VIEW_FORMAT_SINT_4X16`* | | +| 0x0d |*`CU_RES_VIEW_FORMAT_UINT_1X32`* | | +| 0x0e |*`CU_RES_VIEW_FORMAT_UINT_2X32`* | | +| 0x0f |*`CU_RES_VIEW_FORMAT_UINT_4X32`* | | +| 0x10 |*`CU_RES_VIEW_FORMAT_SINT_1X32`* | | +| 0x11 |*`CU_RES_VIEW_FORMAT_SINT_2X32`* | | +| 0x12 |*`CU_RES_VIEW_FORMAT_SINT_4X32`* | | +| 0x13 |*`CU_RES_VIEW_FORMAT_FLOAT_1X16`* | | +| 0x14 |*`CU_RES_VIEW_FORMAT_FLOAT_2X16`* | | +| 0x15 |*`CU_RES_VIEW_FORMAT_FLOAT_4X16`* | | +| 0x16 |*`CU_RES_VIEW_FORMAT_FLOAT_1X32`* | | +| 0x17 |*`CU_RES_VIEW_FORMAT_FLOAT_2X32`* | | +| 0x18 |*`CU_RES_VIEW_FORMAT_FLOAT_4X32`* | | +| 0x19 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC1`* | | +| 0x1a |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC3`* | | +| 0x1b |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC3`* | | +| 0x1c |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC4`* | | +| 0x1d |*`CU_RES_VIEW_FORMAT_SIGNED_BC4`* | | +| 0x1e |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC5`* | | +| 0x1f |*`CU_RES_VIEW_FORMAT_SIGNED_BC5`* | | +| 0x20 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC6H`* | | +| 0x21 |*`CU_RES_VIEW_FORMAT_SIGNED_BC6H`* | | +| 0x22 |*`CU_RES_VIEW_FORMAT_UNSIGNED_BC7`* | | +| enum |***`CUresult`*** |***`hipError_t`*** | +| 0 |*`CUDA_SUCCESS`* |*`hipSuccess`* | +| 1 |*`CUDA_ERROR_INVALID_VALUE`* |*`hipErrorInvalidValue`* | +| 2 |*`CUDA_ERROR_OUT_OF_MEMORY`* |*`hipErrorMemoryAllocation`* | +| 3 |*`CUDA_ERROR_NOT_INITIALIZED`* |*`hipErrorNotInitialized`* | +| 4 |*`CUDA_ERROR_DEINITIALIZED`* |*`hipErrorDeinitialized`* | +| 5 |*`CUDA_ERROR_PROFILER_DISABLED`* |*`hipErrorProfilerDisabled`* | +| 6 |*`CUDA_ERROR_PROFILER_NOT_INITIALIZED`* |*`hipErrorProfilerNotInitialized`* | +| 7 |*`CUDA_ERROR_PROFILER_ALREADY_STARTED`* |*`hipErrorProfilerAlreadyStarted`* | +| 8 |*`CUDA_ERROR_PROFILER_ALREADY_STOPPED`* |*`hipErrorProfilerAlreadyStopped`* | +| 100 |*`CUDA_ERROR_NO_DEVICE`* |*`hipErrorNoDevice`* | +| 101 |*`CUDA_ERROR_INVALID_DEVICE`* |*`hipErrorInvalidDevice`* | +| 200 |*`CUDA_ERROR_INVALID_IMAGE`* |*`hipErrorInvalidImage`* | +| 201 |*`CUDA_ERROR_INVALID_CONTEXT`* |*`hipErrorInvalidContext`* | +| 202 |*`CUDA_ERROR_CONTEXT_ALREADY_CURRENT`* |*`hipErrorContextAlreadyCurrent`* | +| 205 |*`CUDA_ERROR_MAP_FAILED`* |*`hipErrorMapFailed`* | +| 206 |*`CUDA_ERROR_UNMAP_FAILED`* |*`hipErrorUnmapFailed`* | +| 207 |*`CUDA_ERROR_ARRAY_IS_MAPPED`* |*`hipErrorArrayIsMapped`* | +| 208 |*`CUDA_ERROR_ALREADY_MAPPED`* |*`hipErrorAlreadyMapped`* | +| 209 |*`CUDA_ERROR_NO_BINARY_FOR_GPU`* |*`hipErrorNoBinaryForGpu* | +| 210 |*`CUDA_ERROR_ALREADY_ACQUIRED`* |*`hipErrorAlreadyAcquired* | +| 211 |*`CUDA_ERROR_NOT_MAPPED`* |*`hipErrorNotMapped`* | +| 212 |*`CUDA_ERROR_NOT_MAPPED_AS_ARRAY`* |*`hipErrorNotMappedAsArray`* | +| 213 |*`CUDA_ERROR_NOT_MAPPED_AS_POINTER`* |*`hipErrorNotMappedAsPointer`* | +| 214 |*`CUDA_ERROR_ECC_UNCORRECTABLE`* |*`hipErrorECCNotCorrectable`* | +| 215 |*`CUDA_ERROR_UNSUPPORTED_LIMIT`* |*`hipErrorUnsupportedLimit`* | +| 216 |*`CUDA_ERROR_CONTEXT_ALREADY_IN_USE`* |*`hipErrorContextAlreadyInUse`* | +| 217 |*`CUDA_ERROR_PEER_ACCESS_UNSUPPORTED`* |*`hipErrorPeerAccessUnsupported`* | +| 218 |*`CUDA_ERROR_INVALID_PTX`* |*`hipErrorInvalidKernelFile`* | +| 219 |*`CUDA_ERROR_INVALID_GRAPHICS_CONTEXT`* |*`hipErrorInvalidGraphicsContext`* | +| 300 |*`CUDA_ERROR_INVALID_SOURCE`* |*`hipErrorInvalidSource`* | +| 301 |*`CUDA_ERROR_FILE_NOT_FOUND`* |*`hipErrorFileNotFound`* | +| 302 |*`CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND`* |*`hipErrorSharedObjectSymbolNotFound`* | +| 303 |*`CUDA_ERROR_SHARED_OBJECT_INIT_FAILED`* |*`hipErrorSharedObjectInitFailed`* | +| 304 |*`CUDA_ERROR_OPERATING_SYSTEM`* |*`hipErrorOperatingSystem`* | +| 400 |*`CUDA_ERROR_INVALID_HANDLE`* |*`hipErrorInvalidResourceHandle`* | +| 500 |*`CUDA_ERROR_NOT_FOUND`* |*`hipErrorNotFound`* | +| 600 |*`CUDA_ERROR_NOT_READY`* |*`hipErrorNotReady`* | +| 700 |*`CUDA_ERROR_ILLEGAL_ADDRESS`* |*`hipErrorIllegalAddress`* | +| 701 |*`CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES`* |*`hipErrorLaunchOutOfResources`* | +| 702 |*`CUDA_ERROR_LAUNCH_TIMEOUT`* |*`hipErrorLaunchTimeOut`* | +| 703 |*`CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING`* | | +| 704 |*`CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED`* |*`hipErrorPeerAccessAlreadyEnabled`* | +| 705 |*`CUDA_ERROR_PEER_ACCESS_NOT_ENABLED`* |*`hipErrorPeerAccessNotEnabled`* | +| 708 |*`CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE`* | | +| 709 |*`CUDA_ERROR_CONTEXT_IS_DESTROYED`* | | +| 710 |*`CUDA_ERROR_ASSERT`* | | +| 711 |*`CUDA_ERROR_TOO_MANY_PEERS`* | | +| 712 |*`CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED`* |*`hipErrorHostMemoryAlreadyRegistered`* | +| 713 |*`CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED`* |*`hipErrorHostMemoryNotRegistered`* | +| 714 |*`CUDA_ERROR_HARDWARE_STACK_ERROR`* | | +| 715 |*`CUDA_ERROR_ILLEGAL_INSTRUCTION`* | | +| 716 |*`CUDA_ERROR_MISALIGNED_ADDRESS`* | | +| 717 |*`CUDA_ERROR_INVALID_ADDRESS_SPACE`* | | +| 718 |*`CUDA_ERROR_INVALID_PC`* | | +| 719 |*`CUDA_ERROR_LAUNCH_FAILED`* | | +| 800 |*`CUDA_ERROR_NOT_PERMITTED`* | | +| 801 |*`CUDA_ERROR_NOT_SUPPORTED`* | | +| 999 |*`CUDA_ERROR_UNKNOWN`* | | +| enum |***`CUstream_flags`*** |***`hipStreamFlags`*** | +| 0x0 |*`CU_STREAM_DEFAULT`* |*`hipStreamDefault`* | +| 0x1 |*`CU_STREAM_NON_BLOCKING`* |*`hipStreamNonBlocking`* | +| enum |***`CUGLDeviceList`*** | | +| 0x01 |*`CU_GL_DEVICE_LIST_ALL`* | | +| 0x02 |*`CU_GL_DEVICE_LIST_CURRENT_FRAME`* | | +| 0x03 |*`CU_GL_DEVICE_LIST_NEXT_FRAME`* | | +| enum |***`CUGLmap_flags`*** | | +| 0x00 |*`CU_GL_MAP_RESOURCE_FLAGS_NONE`* | | +| 0x01 |*`CU_GL_MAP_RESOURCE_FLAGS_READ_ONLY`* | | +| 0x02 |*`CU_GL_MAP_RESOURCE_FLAGS_WRITE_DISCARD`* | | +| enum |***`CUd3d9DeviceList`*** | | +| 0x01 |*`CU_D3D9_DEVICE_LIST_ALL`* | | +| 0x02 |*`CU_D3D9_DEVICE_LIST_CURRENT_FRAME`* | | +| 0x03 |*`CU_D3D9_DEVICE_LIST_NEXT_FRAME`* | | +| enum |***`CUd3d9map_flags`*** | | +| 0x00 |*`CU_D3D9_MAPRESOURCE_FLAGS_NONE`* | | +| 0x01 |*`CU_D3D9_MAPRESOURCE_FLAGS_READONLY`* | | +| 0x02 |*`CU_D3D9_MAPRESOURCE_FLAGS_WRITEDISCARD`* | | +| enum |***`CUd3d9register_flags`*** | | +| 0x00 |*`CU_D3D9_REGISTER_FLAGS_NONE`* | | +| 0x01 |*`CU_D3D9_REGISTER_FLAGS_ARRAY`* | | +| enum |***`CUd3d10DeviceList`*** | | +| 0x01 |*`CU_D3D10_DEVICE_LIST_ALL`* | | +| 0x02 |*`CU_D3D10_DEVICE_LIST_CURRENT_FRAME`* | | +| 0x03 |*`CU_D3D10_DEVICE_LIST_NEXT_FRAME`* | | +| enum |***`CUd3d10map_flags`*** | | +| 0x00 |*`CU_D3D10_MAPRESOURCE_FLAGS_NONE`* | | +| 0x01 |*`CU_D3D10_MAPRESOURCE_FLAGS_READONLY`* | | +| 0x02 |*`CU_D3D10_MAPRESOURCE_FLAGS_WRITEDISCARD`* | | +| enum |***`CUd3d10register_flags`*** | | +| 0x00 |*`CU_D3D10_REGISTER_FLAGS_NONE`* | | +| 0x01 |*`CU_D3D10_REGISTER_FLAGS_ARRAY`* | | +| enum |***`CUd3d11DeviceList`*** | | +| 0x01 |*`CU_D3D11_DEVICE_LIST_ALL`* | | +| 0x02 |*`CU_D3D11_DEVICE_LIST_CURRENT_FRAME`* | | +| 0x03 |*`CU_D3D11_DEVICE_LIST_NEXT_FRAME`* | | +| typedef | `CUarray` | `hipArray *` | +| struct | `CUarray_st` | `hipArray` | +| typedef | `CUcontext` | `hipCtx_t` | +| typedef | `CUdevice` | `hipDevice_t` | +| typedef | `CUdeviceptr` | `hipDeviceptr_t` | +| typedef | `CUevent` | `hipEvent_t` | +| typedef | `CUfunction` | `hipFunction_t` | +| typedef | `CUgraphicsResource` | | +| typedef | `CUmipmappedArray` | | +| typedef | `CUmodule` | `hipModule_t` | +| typedef | `CUstream` | `hipStream_t` | +| typedef | `CUstreamCallback` | `hipStreamCallback_t` | +| typedef | `CUsurfObject` | | +| typedef | `CUsurfref` | | +| typedef | `CUtexObject` | | +| typedef | `CUtexref` | | +| define |`CU_IPC_HANDLE_SIZE` | | +| define |`CU_LAUNCH_PARAM_BUFFER_POINTER` | `HIP_LAUNCH_PARAM_BUFFER_POINTER` | +| define |`CU_LAUNCH_PARAM_BUFFER_SIZE` | `HIP_LAUNCH_PARAM_BUFFER_SIZE` | +| define |`CU_LAUNCH_PARAM_END` | `HIP_LAUNCH_PARAM_END` | +| define |`CU_MEMHOSTALLOC_DEVICEMAP` | | +| define |`CU_MEMHOSTALLOC_PORTABLE` | | +| define |`CU_MEMHOSTALLOC_WRITECOMBINED` | | +| define |`CU_MEMHOSTREGISTER_DEVICEMAP` | | +| define |`CU_MEMHOSTREGISTER_IOMEMORY` | | +| define |`CU_MEMHOSTREGISTER_PORTABLE` | | +| define |`CU_PARAM_TR_DEFAULT` | | +| define |`CU_STREAM_LEGACY` | | +| define |`CU_STREAM_PER_THREAD` | | +| define |`CU_TRSA_OVERRIDE_FORMAT` | | +| define |`CU_TRSF_NORMALIZED_COORDINATES` | | +| define |`CU_TRSF_SRGB` | | +| define |`CUDA_ARRAY3D_2DARRAY` | | +| define |`CUDA_ARRAY3D_CUBEMAP` | | +| define |`CUDA_ARRAY3D_DEPTH_TEXTURE` | | +| define |`CUDA_ARRAY3D_LAYERED` | | +| define |`CUDA_ARRAY3D_SURFACE_LDST` | | +| define |`CUDA_ARRAY3D_TEXTURE_GATHER` | | +| define |`CUDA_VERSION` | | ## **2. Error Handling** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuGetErrorName` | | Gets the string representation of an error code enum name. | -| `cuGetErrorString` | | Gets the string description of an error code. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuGetErrorName` | | +| `cuGetErrorString` | | ## **3. Initialization** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuInit` | `hipInit` | Initialize the CUDA driver API. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuInit` | `hipInit` | ## **4. Version Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuDriverGetVersion` | `hipDriverGetVersion` | Returns the CUDA driver version. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuDriverGetVersion` | `hipDriverGetVersion` | ## **5. Device Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuDriverGetVersion` | `hipGetDevice` | Returns a handle to a compute device. | -| `cuDeviceGetAttribute` | `hipDeviceGetAttribute` | Returns information about the device. | -| `cuDeviceGetCount` | `hipGetDeviceCount` | Returns the number of compute-capable devices. | -| `cuDeviceGetName` | `hipDeviceGetName` | Returns an identifer string for the device. | -| `cuDeviceTotalMem` | `hipDeviceTotalMem` | Returns the total amount of memory on the device. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuDriverGetVersion` | `hipGetDevice` | +| `cuDeviceGetAttribute` | `hipDeviceGetAttribute` | +| `cuDeviceGetCount` | `hipGetDeviceCount` | +| `cuDeviceGetName` | `hipDeviceGetName` | +| `cuDeviceTotalMem` | `hipDeviceTotalMem` | ## **6. Device Management [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuDeviceComputeCapability` | `hipDeviceComputeCapability` | Returns the compute capability of the device. | -| `cuDeviceGetProperties` | `hipGetDeviceProperties` | Returns properties for a selected device. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuDeviceComputeCapability` | `hipDeviceComputeCapability` | +| `cuDeviceGetProperties` | `hipGetDeviceProperties` | ## **7. Primary Context Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuDevicePrimaryCtxGetState` | `hipDevicePrimaryCtxGetState` | Get the state of the primary context. | -| `cuDevicePrimaryCtxRelease` | `hipDevicePrimaryCtxRelease` | Release the primary context on the GPU. | -| `cuDevicePrimaryCtxReset` | `hipDevicePrimaryCtxReset` | Destroy all allocations and reset all state on the primary context. | -| `cuDevicePrimaryCtxRetain` | `hipDevicePrimaryCtxRetain` | Retain the primary context on the GPU. | -| `cuDevicePrimaryCtxSetFlags` | `hipDevicePrimaryCtxSetFlags` | Set flags for the primary context. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuDevicePrimaryCtxGetState` | `hipDevicePrimaryCtxGetState` | +| `cuDevicePrimaryCtxRelease` | `hipDevicePrimaryCtxRelease` | +| `cuDevicePrimaryCtxReset` | `hipDevicePrimaryCtxReset` | +| `cuDevicePrimaryCtxRetain` | `hipDevicePrimaryCtxRetain` | +| `cuDevicePrimaryCtxSetFlags` | `hipDevicePrimaryCtxSetFlags` | ## **8. Context Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuCtxCreate` | `hipCtxCreate` | Create a CUDA context. | -| `cuCtxDestroy` | `hipCtxDestroy` | Destroy a CUDA context. | -| `cuCtxGetApiVersion` | `hipCtxGetApiVersion` | Gets the context's API version. | -| `cuCtxGetCacheConfig` | `hipCtxGetCacheConfig` | Returns the preferred cache configuration for the current context. | -| `cuCtxGetCurrent` | `hipCtxGetCurrent` | Returns the CUDA context bound to the calling CPU thread. | -| `cuCtxGetDevice` | `hipCtxGetDevice` | Returns the device ID for the current context. | -| `cuCtxGetFlags` | `hipCtxGetFlags` | Returns the flags for the current context. | -| `cuCtxGetLimit` | | Returns resource limits. | -| `cuCtxGetSharedMemConfig` | `hipCtxGetSharedMemConfig` | Returns the current shared memory configuration for the current context. | -| `cuCtxGetStreamPriorityRange` | | Returns numerical values that correspond to the least and greatest stream priorities. | -| `cuCtxPopCurrent` | `hipCtxPopCurrent` | Pops the current CUDA context from the current CPU thread. | -| `cuCtxPushCurrent` | `hipCtxPushCurrent` | Pushes a context on the current CPU thread. | -| `cuCtxSetCacheConfig` | `hipCtxSetCacheConfig` | Sets the preferred cache configuration for the current context. | -| `cuCtxSetCurrent` | `hipCtxSetCurrent` | Binds the specified CUDA context to the calling CPU thread. | -| `cuCtxSetLimit` | | Set resource limits. | -| `cuCtxSetSharedMemConfig` | `hipCtxSetSharedMemConfig` | Sets the shared memory configuration for the current context. | -| `cuCtxSynchronize` | `hipCtxSynchronize` | Block for a context's tasks to complete. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuCtxCreate` | `hipCtxCreate` | +| `cuCtxDestroy` | `hipCtxDestroy` | +| `cuCtxGetApiVersion` | `hipCtxGetApiVersion` | +| `cuCtxGetCacheConfig` | `hipCtxGetCacheConfig` | +| `cuCtxGetCurrent` | `hipCtxGetCurrent` | +| `cuCtxGetDevice` | `hipCtxGetDevice` | +| `cuCtxGetFlags` | `hipCtxGetFlags` | +| `cuCtxGetLimit` | | +| `cuCtxGetSharedMemConfig` | `hipCtxGetSharedMemConfig` | +| `cuCtxGetStreamPriorityRange` | | +| `cuCtxPopCurrent` | `hipCtxPopCurrent` | +| `cuCtxPushCurrent` | `hipCtxPushCurrent` | +| `cuCtxSetCacheConfig` | `hipCtxSetCacheConfig` | +| `cuCtxSetCurrent` | `hipCtxSetCurrent` | +| `cuCtxSetLimit` | | +| `cuCtxSetSharedMemConfig` | `hipCtxSetSharedMemConfig` | +| `cuCtxSynchronize` | `hipCtxSynchronize` | ## **9. Context Management [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuCtxAttach` | | Increment a context's usage-count. | -| `cuCtxDetach` | | Decrement a context's usage-count. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuCtxAttach` | | +| `cuCtxDetach` | | ## **10. Module Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuLinkAddData` | | Add an input to a pending linker invocation. | -| `cuLinkAddFile` | | Add a file input to a pending linker invocation. | -| `cuLinkComplete` | | Complete a pending linker invocation. | -| `cuLinkCreate` | | Creates a pending JIT linker invocation. | -| `cuLinkDestroy` | | Destroys state for a JIT linker invocation. | -| `cuModuleGetFunction` | `hipModuleGetFunction` | Returns a function handle. | -| `cuModuleGetGlobal` | `hipModuleGetGlobal` | Returns a global pointer from a module. | -| `cuModuleGetSurfRef` | | Returns a handle to a surface reference. | -| `cuModuleGetTexRef` | | Returns a handle to a texture reference. | -| `cuModuleLoad` | `hipModuleLoad` | Loads a compute module. | -| `cuModuleLoadData` | `hipModuleLoadData` | Load a module's data. | -| `cuModuleLoadDataEx` | `hipModuleLoadDataEx` | Load a module's data with options. | -| `cuModuleLoadFatBinary` | | Load a module's data. | -| `cuModuleUnload` | `hipModuleUnload` | Unloads a module. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuLinkAddData` | | +| `cuLinkAddFile` | | +| `cuLinkComplete` | | +| `cuLinkCreate` | | +| `cuLinkDestroy` | | +| `cuModuleGetFunction` | `hipModuleGetFunction` | +| `cuModuleGetGlobal` | `hipModuleGetGlobal` | +| `cuModuleGetSurfRef` | | +| `cuModuleGetTexRef` | | +| `cuModuleLoad` | `hipModuleLoad` | +| `cuModuleLoadData` | `hipModuleLoadData` | +| `cuModuleLoadDataEx` | `hipModuleLoadDataEx` | +| `cuModuleLoadFatBinary` | | +| `cuModuleUnload` | `hipModuleUnload` | ## **11. Memory Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuArray3DCreate` | | Creates a 3D CUDA array. | -| `cuArray3DGetDescriptor` | | Get a 3D CUDA array descriptor. | -| `cuArrayCreate` | | Creates a 1D or 2D CUDA array. | -| `cuArrayDestroy` | | Destroys a CUDA array. | -| `cuArrayGetDescriptor` | | Get a 1D or 2D CUDA array descriptor. | -| `cuDeviceGetByPCIBusId` | `hipDeviceGetByPCIBusId` | Returns a handle to a compute device. | -| `cuDeviceGetPCIBusId` | `hipDeviceGetPCIBusId` | Returns a PCI Bus Id string for the device. | -| `cuIpcCloseMemHandle` | | Close memory mapped with cuIpcOpenMemHandle. | -| `cuIpcGetEventHandle` | | Gets an interprocess handle for a previously allocated event. | -| `cuIpcGetMemHandle` | | Gets an interprocess memory handle for an existing device memory allocation. | -| `cuIpcOpenEventHandle` | | Opens an interprocess event handle for use in the current process. | -| `cuIpcOpenMemHandle` | | Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process. | -| `cuMemAlloc` | `hipMalloc` | Allocates device memory. | -| `cuMemAllocHost` | | Allocates page-locked host memory. | -| `cuMemAllocManaged` | | Allocates memory that will be automatically managed by the Unified Memory system. | -| `cuMemAllocPitch` | | Allocates pitched device memory. | -| `cuMemcpy` | | Copies memory. | -| `cuMemcpy2D` | | Copies memory for 2D arrays. | -| `cuMemcpy2DAsync` | | Copies memory for 2D arrays. | -| `cuMemcpy2DUnaligned` | | Copies memory for 2D arrays. | -| `cuMemcpy3D` | | Copies memory for 3D arrays. | -| `cuMemcpy3DAsync` | | Copies memory for 3D arrays. | -| `cuMemcpy3DPeer` | | Copies memory between contexts. | -| `cuMemcpy3DPeerAsync` | | Copies memory between contexts asynchronously. | -| `cuMemcpyAsync` | | Copies memory asynchronously. | -| `cuMemcpyAtoA` | | Copies memory from Array to Array. | -| `cuMemcpyAtoD` | | Copies memory from Array to Device. | -| `cuMemcpyAtoH` | | Copies memory from Array to Host. | -| `cuMemcpyAtoHAsync` | | Copies memory from Array to Host. | -| `cuMemcpyDtoA` | | Copies memory from Device to Array. | -| `cuMemcpyDtoD` | `hipMemcpyDtoD` | Copies memory from Device to Device. | -| `cuMemcpyDtoDAsync` | `hipMemcpyDtoDAsync` | Copies memory from Device to Device. | -| `cuMemcpyDtoH` | `hipMemcpyDtoH` | Copies memory from Device to Host. | -| `cuMemcpyDtoHAsync` | `hipMemcpyDtoHAsync` | Copies memory from Device to Host. | -| `cuMemcpyHtoA` | | Copies memory from Host to Array. | -| `cuMemcpyHtoAAsync` | | Copies memory from Host to Array. | -| `cuMemcpyHtoD` | `hipMemcpyHtoD` | Copies memory from Host to Device. | -| `cuMemcpyHtoDAsync` | `hipMemcpyHtoDAsync` | Copies memory from Host to Device. | -| `cuMemcpyPeer` | | Copies device memory between two contexts. | -| `cuMemcpyPeerAsync` | | Copies device memory between two contexts asynchronously. | -| `cuMemFree` | `hipFree` | Frees device memory. | -| `cuMemFreeHost` | `hipFreeHost` | Frees page-locked host memory. | -| `cuMemGetAddressRange` | | Get information on memory allocations. | -| `cuMemGetInfo` | `hipMemGetInfo` | Gets free and total memory. | -| `cuMemHostAlloc` | `hipHostMalloc` | Allocates page-locked host memory. | -| `cuMemHostGetDevicePointer` | | Passes back device pointer of mapped pinned memory. | -| `cuMemHostGetFlags` | | Passes back flags that were used for a pinned allocation. | -| `cuMemHostRegister` | `hipHostRegister` | Registers an existing host memory range for use by CUDA. | -| `cuMemHostUnregister` | `hipHostUnregister` | Unregisters a memory range that was registered with cuMemHostRegister. | -| `cuMemsetD16` | | Initializes device memory. | -| `cuMemsetD16Async` | | Sets device memory. | -| `cuMemsetD2D16` | | Initializes device memory. | -| `cuMemsetD2D16Async` | | Sets device memory. | -| `cuMemsetD2D32` | | Initializes device memory. | -| `cuMemsetD2D32Async` | | Sets device memory. | -| `cuMemsetD2D8` | | Initializes device memory. | -| `cuMemsetD2D8Async` | | Sets device memory. | -| `cuMemsetD32` | `hipMemset` | Initializes device memory. | -| `cuMemsetD32Async` | `hipMemsetAsync` | Sets device memory. | -| `cuMemsetD2D8` | | Initializes device memory. | -| `cuMemsetD2D8Async` | | Sets device memory. | -| `cuMipmappedArrayCreate` | | Creates a CUDA mipmapped array. | -| `cuMipmappedArrayDestroy` | | Destroys a CUDA mipmapped array. | -| `cuMipmappedArrayGetLevel` | | Gets a mipmap level of a CUDA mipmapped array. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuArray3DCreate` | | +| `cuArray3DGetDescriptor` | | +| `cuArrayCreate` | | +| `cuArrayDestroy` | | +| `cuArrayGetDescriptor` | | +| `cuDeviceGetByPCIBusId` | `hipDeviceGetByPCIBusId` | +| `cuDeviceGetPCIBusId` | `hipDeviceGetPCIBusId` | +| `cuIpcCloseMemHandle` | | +| `cuIpcGetEventHandle` | | +| `cuIpcGetMemHandle` | | +| `cuIpcOpenEventHandle` | | +| `cuIpcOpenMemHandle` | | +| `cuMemAlloc` | `hipMalloc` | +| `cuMemAllocHost` | | +| `cuMemAllocManaged` | | +| `cuMemAllocPitch` | | +| `cuMemcpy` | | +| `cuMemcpy2D` | | +| `cuMemcpy2DAsync` | | +| `cuMemcpy2DUnaligned` | | +| `cuMemcpy3D` | | +| `cuMemcpy3DAsync` | | +| `cuMemcpy3DPeer` | | +| `cuMemcpy3DPeerAsync` | | +| `cuMemcpyAsync` | | +| `cuMemcpyAtoA` | | +| `cuMemcpyAtoD` | | +| `cuMemcpyAtoH` | | +| `cuMemcpyAtoHAsync` | | +| `cuMemcpyDtoA` | | +| `cuMemcpyDtoD` | `hipMemcpyDtoD` | +| `cuMemcpyDtoDAsync` | `hipMemcpyDtoDAsync` | +| `cuMemcpyDtoH` | `hipMemcpyDtoH` | +| `cuMemcpyDtoHAsync` | `hipMemcpyDtoHAsync` | +| `cuMemcpyHtoA` | | +| `cuMemcpyHtoAAsync` | | +| `cuMemcpyHtoD` | `hipMemcpyHtoD` | +| `cuMemcpyHtoDAsync` | `hipMemcpyHtoDAsync` | +| `cuMemcpyPeer` | | +| `cuMemcpyPeerAsync` | | +| `cuMemFree` | `hipFree` | +| `cuMemFreeHost` | `hipFreeHost` | +| `cuMemGetAddressRange` | | +| `cuMemGetInfo` | `hipMemGetInfo` | +| `cuMemHostAlloc` | `hipHostMalloc` | +| `cuMemHostGetDevicePointer` | | +| `cuMemHostGetFlags` | | +| `cuMemHostRegister` | `hipHostRegister` | +| `cuMemHostUnregister` | `hipHostUnregister` | +| `cuMemsetD16` | | +| `cuMemsetD16Async` | | +| `cuMemsetD2D16` | | +| `cuMemsetD2D16Async` | | +| `cuMemsetD2D32` | | +| `cuMemsetD2D32Async` | | +| `cuMemsetD2D8` | | +| `cuMemsetD2D8Async` | | +| `cuMemsetD32` | `hipMemset` | +| `cuMemsetD32Async` | `hipMemsetAsync` | +| `cuMemsetD2D8` | | +| `cuMemsetD2D8Async` | | +| `cuMipmappedArrayCreate` | | +| `cuMipmappedArrayDestroy` | | +| `cuMipmappedArrayGetLevel` | | ## **12. Unified Addressing** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuMemAdvise` | | Advise about the usage of a given memory range. | -| `cuMemPrefetchAsync` | | Prefetches memory to the specified destination device. | -| `cuMemRangeGetAttribute` | | Query an attribute of a given memory range. | -| `cuMemRangeGetAttributes` | | Query attributes of a given memory range. | -| `cuPointerGetAttribute` | | Returns information about a pointer. | -| `cuPointerGetAttributes` | | Returns information about a pointer. | -| `cuPointerSetAttribute` | | Set attributes on a previously allocated memory region. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuMemAdvise` | | +| `cuMemPrefetchAsync` | | +| `cuMemRangeGetAttribute` | | +| `cuMemRangeGetAttributes` | | +| `cuPointerGetAttribute` | | +| `cuPointerGetAttributes` | | +| `cuPointerSetAttribute` | | ## **13. Stream Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuStreamAddCallback` | | Add a callback to a compute stream. | -| `cuStreamAttachMemAsync` | | Attach memory to a stream asynchronously. | -| `cuStreamCreate` | | Create a stream. | -| `cuStreamCreateWithPriority` | | Create a stream with the given priority. | -| `cuStreamDestroy` | `hipStreamDestroy` | Destroys a stream. | -| `cuStreamGetFlags` | `hipStreamGetFlags` | Query the flags of a given stream. | -| `cuStreamGetPriority` | `hipStreamGetPriority` | Query the priority of a given stream. | -| `cuStreamQuery` | `hipStreamQuery` | Determine status of a compute stream. | -| `cuStreamSynchronize` | `hipStreamSynchronize` | Wait until a stream's tasks are completed. | -| `cuStreamWaitEvent` | `hipStreamWaitEvent` | Make a compute stream wait on an event. | -| `cuStreamBatchMemOp` | | Batch operations to synchronize the stream via memory operations. | -| `cuStreamWaitValue32` | | Wait on a memory location. | -| `cuStreamWriteValue32` | | Write a value to memory. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuStreamAddCallback` | | +| `cuStreamAttachMemAsync` | | +| `cuStreamCreate` | | +| `cuStreamCreateWithPriority` | | +| `cuStreamDestroy` | `hipStreamDestroy` | +| `cuStreamGetFlags` | `hipStreamGetFlags` | +| `cuStreamGetPriority` | `hipStreamGetPriority` | +| `cuStreamQuery` | `hipStreamQuery` | +| `cuStreamSynchronize` | `hipStreamSynchronize` | +| `cuStreamWaitEvent` | `hipStreamWaitEvent` | +| `cuStreamBatchMemOp` | | +| `cuStreamWaitValue32` | | +| `cuStreamWriteValue32` | | ## **14. Event Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuEventCreate` | `hipEventCreate` | Creates an event. | -| `cuEventDestroy` | `hipEventDestroy` | Destroys an event. | -| `cuEventElapsedTime` | `hipEventElapsedTime` | Computes the elapsed time between two events. | -| `cuEventQuery` | `hipEventQuery` | Queries an event's status. | -| `cuEventRecord` | `hipEventRecord` | Records an event. | -| `cuEventSynchronize` | `hipEventSynchronize` | Waits for an event to complete. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuEventCreate` | `hipEventCreate` | +| `cuEventDestroy` | `hipEventDestroy` | +| `cuEventElapsedTime` | `hipEventElapsedTime` | +| `cuEventQuery` | `hipEventQuery` | +| `cuEventRecord` | `hipEventRecord` | +| `cuEventSynchronize` | `hipEventSynchronize` | ## **15. Execution Control** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cuFuncGetAttribute` | | Returns information about a function. | -| `cuFuncSetCacheConfig` | `hipFuncSetCacheConfig` | Sets the preferred cache configuration for a device function. | -| `cuFuncSetSharedMemConfig` | | Sets the shared memory configuration for a device function. | -| `cuLaunchKernel` | `hipModuleLaunchKernel` | Launches a CUDA function. | - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuFuncGetAttribute` | | +| `cuFuncSetCacheConfig` | `hipFuncSetCacheConfig` | +| `cuFuncSetSharedMemConfig` | | +| `cuLaunchKernel` | `hipModuleLaunchKernel` | ## **16. Execution Control [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuFuncSetBlockShape` | | +| `cuFuncSetSharedSize` | | +| `cuLaunch` | | +| `cuLaunchGrid` | | +| `cuLaunchGridAsync` | | +| `cuParamSetf` | | +| `cuParamSeti` | | +| `cuParamSetTexRef` | | +| `cuParamSetv` | | ## **17. Occupancy** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|---------------------------------------------------------| +| `cuOccupancyMaxActiveBlocksPerMultiprocessor` | `hipOccupancyMaxActiveBlocksPerMultiprocessor` | +| `cuOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` | | +| `cuOccupancyMaxPotentialBlockSize` | `hipOccupancyMaxPotentialBlockSize` | +| `cuOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` | | ## **18. Texture Reference Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuTexRefGetAddress` | | +| `cuTexRefGetAddressMode` | | +| `cuTexRefGetArray` | | +| `cuTexRefGetBorderColor` | | +| `cuTexRefGetFilterMode` | | +| `cuTexRefGetFlags` | | +| `cuTexRefGetFormat` | | +| `cuTexRefGetMaxAnisotropy` | | +| `cuTexRefGetMipmapFilterMode` | | +| `cuTexRefGetMipmapLevelBias` | | +| `cuTexRefGetMipmapLevelClamp` | | +| `cuTexRefGetMipmappedArray` | | +| `cuTexRefSetAddress` | | +| `cuTexRefSetAddress2D` | | +| `cuTexRefSetAddressMode` | | +| `cuTexRefSetArray` | | +| `cuTexRefSetBorderColor` | | +| `cuTexRefSetFilterMode` | | +| `cuTexRefSetFlags` | | +| `cuTexRefSetFormat` | | +| `cuTexRefSetMaxAnisotropy` | | +| `cuTexRefSetMipmapFilterMode` | | +| `cuTexRefSetMipmapLevelBias` | | +| `cuTexRefSetMipmapLevelClamp` | | +| `cuTexRefSetMipmappedArray` | | ## **19. Texture Reference Management [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuTexRefCreate` | | +| `cuTexRefDestroy` | | ## **20. Surface Reference Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuSurfRefGetArray` | | +| `cuSurfRefSetArray` | | ## **21. Texture Object Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuTexObjectCreate` | | +| `cuTexObjectDestroy` | | +| `cuTexObjectGetResourceDesc` | | +| `cuTexObjectGetResourceViewDesc` | | +| `cuTexObjectGetTextureDesc` | | ## **22. Surface Object Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuSurfObjectCreate` | | +| `cuSurfObjectDestroy` | | +| `cuSurfObjectGetResourceDesc` | | ## **23. Peer Context Memory Access** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuCtxEnablePeerAccess` | `hipCtxEnablePeerAccess` | +| `cuCtxDisablePeerAccess` | `hipCtxDisablePeerAccess` | +| `cuDeviceCanAccessPeer` | `hipDeviceCanAccessPeer` | +| `cuDeviceGetP2PAttribute` | | ## **24. Graphics Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuGraphicsMapResources` | | +| `cuGraphicsResourceGetMappedMipmappedArray` | | +| `cuGraphicsResourceGetMappedPointer` | | +| `cuGraphicsResourceSetMapFlags` | | +| `cuGraphicsSubResourceGetMappedArray` | | +| `cuGraphicsUnmapResources` | | +| `cuGraphicsUnregisterResource` | | ## **25. Profiler Control** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| - +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuProfilerInitialize` | | +| `cuProfilerStart` | `hipProfilerStart` | +| `cuProfilerStop` | `hipProfilerStop` | ## **26. OpenGL Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuGLGetDevices` | | +| `cuGraphicsGLRegisterBuffer` | | +| `cuGraphicsGLRegisterImage` | | +| `cuWGLGetDevice` | | +## **26.1. OpenGL Interoperability [DEPRECATED]** +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuGLCtxCreate` | | +| `cuGLInit` | | +| `cuGLMapBufferObject` | | +| `cuGLMapBufferObjectAsync` | | +| `cuGLRegisterBufferObject` | | +| `cuGLSetBufferObjectMapFlags` | | +| `cuGLUnmapBufferObject` | | +| `cuGLUnmapBufferObjectAsync` | | +| `cuGLUnregisterBufferObject` | | ## **27. Direct3D 9 Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuD3D9CtxCreate` | | +| `cuD3D9CtxCreateOnDevice` | | +| `cuD3D9GetDevice` | | +| `cuD3D9GetDevices` | | +| `cuD3D9GetDirect3DDevice` | | +| `cuGraphicsD3D9RegisterResource` | | +## **27.1. Direct3D 9 Interoperability [DEPRECATED]** +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuD3D9MapResources` | | +| `cuD3D9RegisterResource` | | +| `cuD3D9ResourceGetMappedArray` | | +| `cuD3D9ResourceGetMappedPitch` | | +| `cuD3D9ResourceGetMappedPointer` | | +| `cuD3D9ResourceGetMappedSize` | | +| `cuD3D9ResourceGetSurfaceDimensions` | | +| `cuD3D9ResourceSetMapFlags` | | +| `cuD3D9UnmapResources` | | +| `cuD3D9UnregisterResource` | | ## **28. Direct3D 10 Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuD3D10GetDevice` | | +| `cuD3D10GetDevices` | | +| `cuGraphicsD3D10RegisterResource` | | +## **28.1. Direct3D 10 Interoperability [DEPRECATED]** +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuD3D10CtxCreate` | | +| `cuD3D10CtxCreateOnDevice` | | +| `cuD3D10GetDirect3DDevice` | | +| `cuD3D10MapResources` | | +| `cuD3D10RegisterResource` | | +| `cuD3D10ResourceGetMappedArray` | | +| `cuD3D10ResourceGetMappedPitch` | | +| `cuD3D10ResourceGetMappedPointer` | | +| `cuD3D10ResourceGetMappedSize` | | +| `cuD3D10ResourceGetSurfaceDimensions` | | +| `cuD3D10ResourceSetMapFlags` | | +| `cuD3D10UnmapResources` | | +| `cuD3D10UnregisterResource` | | ## **29. Direct3D 11 Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuD3D11GetDevice` | | +| `cuD3D11GetDevices` | | +| `cuGraphicsD3D11RegisterResource` | | +## **29.1. Direct3D 11 Interoperability [DEPRECATED]** +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuD3D11CtxCreate` | | +| `cuD3D11CtxCreateOnDevice` | | +| `cuD3D11GetDirect3DDevice` | | ## **30. VDPAU Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuGraphicsVDPAURegisterOutputSurface` | | +| `cuGraphicsVDPAURegisterVideoSurface` | | +| `cuVDPAUCtxCreate` | | +| `cuVDPAUGetDevice` | | +## **31. EGL Interoperability** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cuEGLStreamConsumerAcquireFrame` | | +| `cuEGLStreamConsumerConnect` | | +| `cuEGLStreamConsumerConnectWithFlags` | | +| `cuEGLStreamConsumerDisconnect` | | +| `cuEGLStreamConsumerReleaseFrame` | | +| `cuEGLStreamProducerConnect` | | +| `cuEGLStreamProducerDisconnect` | | +| `cuEGLStreamProducerPresentFrame` | | +| `cuEGLStreamProducerReturnFrame` | | +| `cuGraphicsEGLRegisterImage` | | +| `cuGraphicsResourceGetMappedEglFrame` | | diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index c5df7f6bcd..b4c004e2dc 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -2,636 +2,783 @@ ## **1. Device Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaChooseDevice` | `hipChooseDevice` | Select compute-device which best matches criteria. | -| `cudaDeviceGetAttribute` | `hipDeviceGetAttribute` | Returns information about the device. | -| `cudaDeviceGetByPCIBusId` | `hipDeviceGetByPCIBusId` | Returns a handle to a compute device. | -| `cudaDeviceGetCacheConfig` | `hipDeviceGetCacheConfig` | Returns the preferred cache configuration for the current device. | -| `cudaDeviceGetLimit` | `hipDeviceGetLimit` | Returns resource limits. | -| `cudaDeviceGetPCIBusId` | `hipDeviceGetPCIBusId` | Returns a PCI Bus Id string for the device. | -| `cudaDeviceGetSharedMemConfig` | `hipDeviceGetSharedMemConfig` | Returns the shared memory configuration for the current device. | -| `cudaDeviceGetStreamPriorityRange` | | Returns numerical values that correspond to the least and greatest stream priorities. | -| `cudaDeviceReset` | `hipDeviceReset` | Destroy all allocations and reset all state on the current device in the current process. | -| `cudaDeviceSetCacheConfig` | `hipDeviceSetCacheConfig` | Sets the preferred cache configuration for the current device. | -| `cudaDeviceSetLimit` | `hipDeviceSetLimit` | Set resource limits. | -| `cudaDeviceSetSharedMemConfig` | `hipDeviceSetSharedMemConfig` | Sets the shared memory configuration for the current device. | -| `cudaDeviceSynchronize` | `hipDeviceSynchronize` | Wait for compute device to finish. | -| `cudaGetDevice` | `hipGetDevice` | Returns which device is currently being used. | -| `cudaGetDeviceCount` | `hipGetDeviceCount` | Returns the number of compute-capable devices. | -| `cudaGetDeviceFlags` | | Gets the flags for the current device. | -| `cudaGetDeviceProperties` | `hipGetDeviceProperties` | Returns information about the compute-device. | -| `cudaIpcCloseMemHandle` | `hipIpcCloseMemHandle` | Close memory mapped with cudaIpcOpenMemHandle. | -| `cudaIpcGetEventHandle` | `hipIpcGetEventHandle` | Gets an interprocess handle for a previously allocated event. | -| `cudaIpcGetMemHandle` | `hipIpcGetMemHandle` | Gets an interprocess memory handle for an existing device memory allocation. | -| `cudaIpcOpenEventHandle` | `hipIpcOpenEventHandle` | Opens an interprocess event handle for use in the current process. | -| `cudaIpcOpenMemHandle` | `hipIpcOpenMemHandle` | Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process. | -| `cudaSetDevice` | `hipSetDevice` | Set device to be used for GPU executions. | -| `cudaSetDeviceFlags` | `hipSetDeviceFlags` | Sets flags to be used for device executions. | -| `cudaSetValidDevices` | | Set a list of devices that can be used for CUDA. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaChooseDevice` | `hipChooseDevice` | +| `cudaDeviceGetAttribute` | `hipDeviceGetAttribute` | +| `cudaDeviceGetByPCIBusId` | `hipDeviceGetByPCIBusId` | +| `cudaDeviceGetCacheConfig` | `hipDeviceGetCacheConfig` | +| `cudaDeviceGetLimit` | `hipDeviceGetLimit` | +| `cudaDeviceGetPCIBusId` | `hipDeviceGetPCIBusId` | +| `cudaDeviceGetSharedMemConfig` | `hipDeviceGetSharedMemConfig` | +| `cudaDeviceGetStreamPriorityRange` | | +| `cudaDeviceReset` | `hipDeviceReset` | +| `cudaDeviceSetCacheConfig` | `hipDeviceSetCacheConfig` | +| `cudaDeviceSetLimit` | `hipDeviceSetLimit` | +| `cudaDeviceSetSharedMemConfig` | `hipDeviceSetSharedMemConfig` | +| `cudaDeviceSynchronize` | `hipDeviceSynchronize` | +| `cudaGetDevice` | `hipGetDevice` | +| `cudaGetDeviceCount` | `hipGetDeviceCount` | +| `cudaGetDeviceFlags` | | +| `cudaGetDeviceProperties` | `hipGetDeviceProperties` | +| `cudaIpcCloseMemHandle` | `hipIpcCloseMemHandle` | +| `cudaIpcGetEventHandle` | `hipIpcGetEventHandle` | +| `cudaIpcGetMemHandle` | `hipIpcGetMemHandle` | +| `cudaIpcOpenEventHandle` | `hipIpcOpenEventHandle` | +| `cudaIpcOpenMemHandle` | `hipIpcOpenMemHandle` | +| `cudaSetDevice` | `hipSetDevice` | +| `cudaSetDeviceFlags` | `hipSetDeviceFlags` | +| `cudaSetValidDevices` | | -## **2. Error Handling** +## **2. Thread Management [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaGetErrorName` | `hipGetErrorName` | Returns the string representation of an error code enum name. | -| `cudaGetErrorString` | `hipGetErrorString` | Returns the description string for an error code. | -| `cudaGetLastError` | `hipGetLastError` | Returns the last error from a runtime call. | -| `cudaPeekAtLastError` | `hipPeekAtLastError` | Returns the last error from a runtime call. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaThreadExit` | `hipDeviceReset` | +| `cudaThreadGetCacheConfig` | `hipDeviceGetCacheConfig` | +| `cudaThreadGetLimit` | | +| `cudaThreadSetCacheConfig` | `hipDeviceSetCacheConfig` | +| `cudaThreadSetLimit` | | +| `cudaThreadSynchronize` | `hipDeviceSynchronize` | -## **3. Stream Management** +## **3. Error Handling** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaStreamAddCallback` | `hipStreamAddCallback` | Add a callback to a compute stream. | -| `cudaStreamAttachMemAsync` | | Attach managed memory to a stream asynchronously. | -| `cudaStreamCreate` | `hipStreamCreate` | Create an asynchronous stream. | -| `cudaStreamCreateWithFlags` | `hipStreamCreateWithFlags` | Create an asynchronous stream. | -| `cudaStreamCreateWithPriority` | | Create an asynchronous stream with the specified priority. | -| `cudaStreamDestroy` | `hipStreamDestroy` | Destroys and cleans up an asynchronous stream. | -| `cudaStreamGetFlags` | `hipStreamGetFlags` | Query the flags of a stream. | -| `cudaStreamGetPriority` | | Query the priority of a stream. | -| `cudaStreamQuery` | `hipStreamQuery` | Queries an asynchronous stream for completion status. | -| `cudaStreamSynchronize` | `hipStreamSynchronize` | Waits for stream tasks to complete. | -| `cudaStreamWaitEvent` | `hipStreamWaitEvent` | Make a compute stream wait on an event. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaGetErrorName` | `hipGetErrorName` | +| `cudaGetErrorString` | `hipGetErrorString` | +| `cudaGetLastError` | `hipGetLastError` | +| `cudaPeekAtLastError` | `hipPeekAtLastError` | -## **4. Event Management** +## **4. Stream Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaEventCreate` | `hipEventCreate` | Creates an event object. | -| `cudaEventCreateWithFlags` | `hipEventCreateWithFlags` | Creates an event object with the specified flags. | -| `cudaEventDestroy` | `hipEventDestroy` | Destroys an event object. | -| `cudaEventElapsedTime` | `hipEventElapsedTime` | Computes the elapsed time between events. | -| `cudaEventQuery` | `hipEventQuery` | Queries an event's status. | -| `cudaEventRecord` | `hipEventRecord` | Records an event. | -| `cudaEventSynchronize` | `hipEventSynchronize` | Waits for an event to complete. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaStreamAddCallback` | `hipStreamAddCallback` | +| `cudaStreamAttachMemAsync` | | +| `cudaStreamCreate` | `hipStreamCreate` | +| `cudaStreamCreateWithFlags` | `hipStreamCreateWithFlags` | +| `cudaStreamCreateWithPriority` | | +| `cudaStreamDestroy` | `hipStreamDestroy` | +| `cudaStreamGetFlags` | `hipStreamGetFlags` | +| `cudaStreamGetPriority` | | +| `cudaStreamQuery` | `hipStreamQuery` | +| `cudaStreamSynchronize` | `hipStreamSynchronize` | +| `cudaStreamWaitEvent` | `hipStreamWaitEvent` | -## **5. Execution Control** +## **5. Event Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaFuncGetAttributes` | | Find out attributes for a given function. | -| `cudaFuncSetCacheConfig` | `hipFuncSetCacheConfig` | Sets the preferred cache configuration for a device function. | -| `cudaFuncSetSharedMemConfig` | | Sets the shared memory configuration for a device function. | -| `cudaGetParameterBuffer` | | Obtains a parameter buffer. | -| `cudaGetParameterBufferV2` | | Launches a specified kernel. | -| `cudaLaunchKernel` | `hipLaunchKernel` | Launches a device function. | -| `cudaSetDoubleForDevice` | | Converts a double argument to be executed on a device. | -| `cudaSetDoubleForHost` | | Converts a double argument after execution on a device. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaEventCreate` | `hipEventCreate` | +| `cudaEventCreateWithFlags` | `hipEventCreateWithFlags` | +| `cudaEventDestroy` | `hipEventDestroy` | +| `cudaEventElapsedTime` | `hipEventElapsedTime` | +| `cudaEventQuery` | `hipEventQuery` | +| `cudaEventRecord` | `hipEventRecord` | +| `cudaEventSynchronize` | `hipEventSynchronize` | -## **6. Occupancy** +## **6. Execution Control** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaOccupancyMaxActiveBlocksPerMultiprocessor` | `hipOccupancyMaxActiveBlocksPerMultiprocessor`| Returns occupancy for a device function. | -| `cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` | | Returns occupancy for a device function with the specified flags. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaFuncGetAttributes` | | +| `cudaFuncSetCacheConfig` | `hipFuncSetCacheConfig` | +| `cudaFuncSetSharedMemConfig` | | +| `cudaGetParameterBuffer` | | +| `cudaGetParameterBufferV2` | | +| `cudaLaunchKernel` | `hipLaunchKernel` | +| `cudaSetDoubleForDevice` | | +| `cudaSetDoubleForHost` | | -## **7. Execution Control [deprecated since 7.0]** +## **7. Occupancy** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaConfigureCall` | | Configure a device-launch. | -| `cudaLaunch` | | Launches a device function. | -| `cudaSetupArgument` | | Configure a device launch. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaOccupancyMaxActiveBlocksPerMultiprocessor` | `hipOccupancyMaxActiveBlocksPerMultiprocessor`| +| `cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` | | -## **8. Memory Management** +## **8. Execution Control [deprecated since 7.0]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaArrayGetInfo` | | Gets info about the specified cudaArray. | -| `cudaFree` | `hipFree` | Frees memory on the device. | -| `cudaFreeArray` | `hipFreeArray` | Frees an array on the device. | -| `cudaFreeHost` | `hipHostFree` | Frees page-locked memory. | -| `cudaFreeMipmappedArray` | | Frees a mipmapped array on the device. | -| `cudaGetMipmappedArrayLevel` | | Gets a mipmap level of a CUDA mipmapped array. | -| `cudaGetSymbolAddress` | | Finds the address associated with a CUDA symbol. | -| `cudaGetSymbolSize` | | Finds the size of the object associated with a CUDA symbol. | -| `cudaHostAlloc` | `hipHostMalloc` | Allocates page-locked memory on the host. | -| `cudaHostGetDevicePointer` | `hipHostGetDevicePointer` | Passes back device pointer of mapped host memory allocated by cudaHostAlloc or registered by cudaHostRegister. | -| `cudaHostGetFlags` | `hipHostGetFlags` | Passes back flags used to allocate pinned host memory allocated by cudaHostAlloc. | -| `cudaHostRegister` | `hipHostRegister` | Registers an existing host memory range for use by CUDA. | -| `cudaHostUnregister` | `hipHostUnregister` | Unregisters a memory range that was registered with cudaHostRegister. | -| `cudaMalloc` | `hipMalloc` | Allocate memory on the device. | -| `cudaMalloc3D` | | Allocates logical 1D, 2D, or 3D memory objects on the device. | -| `cudaMalloc3DArray` | | Allocate an array on the device. | -| `cudaMallocArray` | `hipMallocArray` | Allocate an array on the device. | -| `cudaMallocHost` | `hipHostMalloc` | Allocates page-locked memory on the host. | -| `cudaMallocManaged` | | Allocates memory that will be automatically managed by the Unified Memory system. | -| `cudaMallocMipmappedArray` | | Allocate a mipmapped array on the device. | -| `cudaMallocPitch` | | Allocates pitched memory on the device. | -| `cudaMemGetInfo` | `hipMemGetInfo` | Gets free and total device memory. | -| `cudaMemcpy` | `hipMemcpy` | Copies data between host and device. | -| `cudaMemcpy2D` | `hipMemcpy2D` | Copies data between host and device. | -| `cudaMemcpy2DArrayToArray` | | Copies data between host and device. | -| `cudaMemcpy2DAsync` | | Copies data between host and device. | -| `cudaMemcpy2DFromArray` | | Copies data between host and device. | -| `cudaMemcpy2DFromArrayAsync` | | Copies data between host and device. | -| `cudaMemcpy2DToArray` | `hipMemcpy2DToArray` | Copies data between host and device. | -| `cudaMemcpy2DToArrayAsync` | | Copies data between host and device. | -| `cudaMemcpy3D` | | Copies data between 3D objects. | -| `cudaMemcpy3DAsync` | | Copies data between 3D objects. | -| `cudaMemcpy3DPeer` | | Copies memory between devices. | -| `cudaMemcpy3DPeerAsync` | | Copies memory between devices asynchronously. | -| `cudaMemcpyArrayToArray` | | Copies data between host and device. | -| `cudaMemcpyAsync` | `hipMemcpyAsync` | Copies data between host and device. | -| `cudaMemcpyFromArray` | `MemcpyFromArray` | Copies data between host and device. | -| `cudaMemcpyFromArrayAsync` | | Copies data between host and device. | -| `cudaMemcpyFromSymbol` | `hipMemcpyFromSymbol` | Copies data from the given symbol on the device. | -| `cudaMemcpyFromSymbolAsync` | | Copies data from the given symbol on the device. | -| `cudaMemcpyPeer` | `hipMemcpyPeer` | Copies memory between two devices. | -| `cudaMemcpyPeerAsync` | `hipMemcpyPeerAsync` | Copies memory between two devices asynchronously. | -| `cudaMemcpyToArray` | `hipMemcpyToArray` | Copies data between host and device. | -| `cudaMemcpyToArrayAsync` | | Copies data between host and device. | -| `cudaMemcpyToSymbol` | `hipMemcpyToSymbol` | Copies data to the given symbol on the device. | -| `cudaMemcpyToSymbolAsync` | `hipMemcpyToSymbolAsync` | Copies data to the given symbol on the device. | -| `cudaMemset` | `hipMemset` | Initializes or sets device memory to a value. | -| `cudaMemset2D` | | Initializes or sets device memory to a value. | -| `cudaMemset2DAsync` | | Initializes or sets device memory to a value. | -| `cudaMemset3D` | | Initializes or sets device memory to a value. | -| `cudaMemset3DAsync` | | Initializes or sets device memory to a value. | -| `cudaMemsetAsync` | `hipMemsetAsync` | Initializes or sets device memory to a value. | -| `make_cudaExtent` | | Returns a cudaExtent based on input parameters. | -| `make_cudaPitchedPtr` | | Returns a cudaPitchedPtr based on input parameters. | -| `make_cudaPos` | | Returns a cudaPos based on input parameters. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaConfigureCall` | | +| `cudaLaunch` | | +| `cudaSetupArgument` | | -## **9. Unified Addressing** +## **9. Memory Management** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaPointerGetAttributes` | `hipPointerGetAttributes` | Returns attributes about a specified pointer. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaArrayGetInfo` | | +| `cudaFree` | `hipFree` | +| `cudaFreeArray` | `hipFreeArray` | +| `cudaFreeHost` | `hipHostFree` | +| `cudaFreeMipmappedArray` | | +| `cudaGetMipmappedArrayLevel` | | +| `cudaGetSymbolAddress` | | +| `cudaGetSymbolSize` | | +| `cudaHostAlloc` | `hipHostMalloc` | +| `cudaHostGetDevicePointer` | `hipHostGetDevicePointer` | +| `cudaHostGetFlags` | `hipHostGetFlags` | +| `cudaHostRegister` | `hipHostRegister` | +| `cudaHostUnregister` | `hipHostUnregister` | +| `cudaMalloc` | `hipMalloc` | +| `cudaMalloc3D` | | +| `cudaMalloc3DArray` | | +| `cudaMallocArray` | `hipMallocArray` | +| `cudaMallocHost` | `hipHostMalloc` | +| `cudaMallocManaged` | | +| `cudaMallocMipmappedArray` | | +| `cudaMallocPitch` | | +| `cudaMemGetInfo` | `hipMemGetInfo` | +| `cudaMemcpy` | `hipMemcpy` | +| `cudaMemcpy2D` | `hipMemcpy2D` | +| `cudaMemcpy2DArrayToArray` | | +| `cudaMemcpy2DAsync` | | +| `cudaMemcpy2DFromArray` | | +| `cudaMemcpy2DFromArrayAsync` | | +| `cudaMemcpy2DToArray` | `hipMemcpy2DToArray` | +| `cudaMemcpy2DToArrayAsync` | | +| `cudaMemcpy3D` | | +| `cudaMemcpy3DAsync` | | +| `cudaMemcpy3DPeer` | | +| `cudaMemcpy3DPeerAsync` | | +| `cudaMemcpyArrayToArray` | | +| `cudaMemcpyAsync` | `hipMemcpyAsync` | +| `cudaMemcpyFromArray` | `MemcpyFromArray` | +| `cudaMemcpyFromArrayAsync` | | +| `cudaMemcpyFromSymbol` | `hipMemcpyFromSymbol` | +| `cudaMemcpyFromSymbolAsync` | | +| `cudaMemcpyPeer` | `hipMemcpyPeer` | +| `cudaMemcpyPeerAsync` | `hipMemcpyPeerAsync` | +| `cudaMemcpyToArray` | `hipMemcpyToArray` | +| `cudaMemcpyToArrayAsync` | | +| `cudaMemcpyToSymbol` | `hipMemcpyToSymbol` | +| `cudaMemcpyToSymbolAsync` | `hipMemcpyToSymbolAsync` | +| `cudaMemset` | `hipMemset` | +| `cudaMemset2D` | | +| `cudaMemset2DAsync` | | +| `cudaMemset3D` | | +| `cudaMemset3DAsync` | | +| `cudaMemsetAsync` | `hipMemsetAsync` | +| `make_cudaExtent` | | +| `make_cudaPitchedPtr` | | +| `make_cudaPos` | | -## **10. Peer Device Memory Access** +## **10. Unified Addressing** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaDeviceCanAccessPeer` | `hipDeviceCanAccessPeer` | Queries if a device may directly access a peer device's memory. | -| `cudaDeviceDisablePeerAccess` | `hipDeviceDisablePeerAccess` | Disables direct access to memory allocations on a peer device. | -| `cudaDeviceEnablePeerAccess` | `hipDeviceEnablePeerAccess` | Enables direct access to memory allocations on a peer device. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaPointerGetAttributes` | `hipPointerGetAttributes` | -## **11. OpenGL Interoperability** +## **11. Peer Device Memory Access** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaGLGetDevices` | | Gets the CUDA devices associated with the current OpenGL context. | -| `cudaGraphicsGLRegisterBuffer` | | Registers an OpenGL buffer object. | -| `cudaGraphicsGLRegisterImage` | | Register an OpenGL texture or renderbuffer object. | -| `cudaWGLGetDevice` | | Gets the CUDA device associated with hGpu. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaDeviceCanAccessPeer` | `hipDeviceCanAccessPeer` | +| `cudaDeviceDisablePeerAccess` | `hipDeviceDisablePeerAccess` | +| `cudaDeviceEnablePeerAccess` | `hipDeviceEnablePeerAccess` | -## **12. Graphics Interoperability** +## **12. OpenGL Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaGraphicsMapResources` | | Map graphics resources for access by CUDA. | -| `cudaGraphicsResourceGetMappedMipmappedArray` | | Get a mipmapped array through which to access a mapped graphics resource. | -| `cudaGraphicsResourceGetMappedPointer` | | Get a device pointer through which to access a mapped graphics resource. | -| `cudaGraphicsResourceSetMapFlags` | | Set usage flags for mapping a graphics resource. | -| `cudaGraphicsSubResourceGetMappedArray` | | Get an array through which to access a subresource of a mapped graphics resource. | -| `cudaGraphicsUnmapResources` | | Unmap graphics resources. | -| `cudaGraphicsUnregisterResource` | | Unregisters a graphics resource for access by CUDA. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaGLGetDevices` | | +| `cudaGraphicsGLRegisterBuffer` | | +| `cudaGraphicsGLRegisterImage` | | +| `cudaWGLGetDevice` | | -## **13. Texture Reference Management** +## **13. OpenGL Interoperability [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaBindTexture` | | Binds a memory area to a texture. | -| `cudaBindTexture2D` | | Binds a 2D memory area to a texture. | -| `cudaBindTextureToArray` | | Binds an array to a texture. | -| `cudaBindTextureToMipmappedArray` | | Binds a mipmapped array to a texture. | -| `cudaCreateChannelDesc` | | Returns a channel descriptor using the specified format. | -| `cudaGetChannelDesc` | | Get the channel descriptor of an array. | -| `cudaGetTextureAlignmentOffset` | | Get the alignment offset of a texture. | -| `cudaGetTextureReference` | | Get the texture reference associated with a symbol. | -| `cudaUnbindTexture` | | Unbinds a texture. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaGLMapBufferObject` | | +| `cudaGLMapBufferObjectAsync` | | +| `cudaGLRegisterBufferObject` | | +| `cudaGLSetBufferObjectMapFlags` | | +| `cudaGLSetGLDevice` | | +| `cudaGLUnmapBufferObject` | | +| `cudaGLUnmapBufferObjectAsync` | | +| `cudaGLUnregisterBufferObject` | | -## **14. Surface Reference Management** +## **14. Direct3D 9 Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaBindSurfaceToArray` | | Binds an array to a surface. | -| `cudaGetSurfaceReference` | | Get the surface reference associated with a symbol. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaD3D9GetDevice` | | +| `cudaD3D9GetDevices` | | +| `cudaD3D9GetDirect3DDevice` | | +| `cudaD3D9SetDirect3DDevice` | | +| `cudaGraphicsD3D9RegisterResource` | | -## **15. Texture Object Management** +## **15. Direct3D 9 Interoperability [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaCreateTextureObject` | | Creates a texture object. | -| `cudaDestroyTextureObject` | | Destroys a texture object. | -| `cudaGetTextureObjectResourceDesc` | | Returns a texture object's resource descriptor. | -| `cudaGetTextureObjectResourceViewDesc` | | Returns a texture object's resource view descriptor. | -| `cudaGetTextureObjectTextureDesc` | | Returns a texture object's texture descriptor. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaD3D9MapResources` | | +| `cudaD3D9RegisterResource` | | +| `cudaD3D9ResourceGetMappedArray` | | +| `cudaD3D9ResourceGetMappedPitch` | | +| `cudaD3D9ResourceGetMappedPointer` | | +| `cudaD3D9ResourceGetMappedSize` | | +| `cudaD3D9ResourceGetSurfaceDimensions` | | +| `cudaD3D9ResourceSetMapFlags` | | +| `cudaD3D9UnmapResources` | | +| `cudaD3D9UnregisterResource` | | -## **16. Surface Object Management** +## **16. Direct3D 10 Interoperability** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaCreateSurfaceObject` | | Creates a surface object. | -| `cudaDestroySurfaceObject` | | Destroys a surface object. | -| `cudaGetSurfaceObjectResourceDesc` | | Returns a surface object's resource descriptor Returns the resource descriptor for the surface object specified by surfObject. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaD3D10GetDevice` | | +| `cudaD3D10GetDevices` | | +| `cudaGraphicsD3D10RegisterResource` | | -## **17. Version Management** +## **17. Direct3D 10 Interoperability [DEPRECATED]** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaDriverGetVersion` | `hipDriverGetVersion` | Returns the CUDA driver version. | -| `cudaRuntimeGetVersion` | `hipRuntimeGetVersion` | Returns the CUDA Runtime version. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaD3D10GetDirect3DDevice` | | +| `cudaD3D10MapResources` | | +| `cudaD3D10RegisterResource` | | +| `cudaD3D10ResourceGetMappedArray` | | +| `cudaD3D10ResourceGetMappedPitch` | | +| `cudaD3D10ResourceGetMappedPointer` | | +| `cudaD3D10ResourceGetMappedSize` | | +| `cudaD3D10ResourceGetSurfaceDimensions` | | +| `cudaD3D10ResourceSetMapFlags` | | +| `cudaD3D10SetDirect3DDevice` | | +| `cudaD3D10UnmapResources` | | +| `cudaD3D10UnregisterResource` | | -## **18. C++ API Routines** +## **18. Direct3D 11 Interoperability** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaD3D11GetDevice` | | +| `cudaD3D11GetDevices` | | +| `cudaGraphicsD3D11RegisterResource` | | + +## **19. Direct3D 11 Interoperability [DEPRECATED]** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaD3D11GetDirect3DDevice` | | +| `cudaD3D11SetDirect3DDevice` | | + +## **20. VDPAU Interoperability** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaGraphicsVDPAURegisterOutputSurface` | | +| `cudaGraphicsVDPAURegisterVideoSurface` | | +| `cudaVDPAUGetDevice` | | +| `cudaVDPAUSetVDPAUDevice` | | + +## **21. EGL Interoperability** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaEGLStreamConsumerAcquireFrame` | | +| `cudaEGLStreamConsumerConnect` | | +| `cudaEGLStreamConsumerConnectWithFlags` | | +| `cudaEGLStreamConsumerReleaseFrame` | | +| `cudaEGLStreamProducerConnect` | | +| `cudaEGLStreamProducerDisconnect` | | +| `cudaEGLStreamProducerPresentFrame` | | +| `cudaEGLStreamProducerReturnFrame` | | +| `cudaGraphicsEGLRegisterImage` | | +| `cudaGraphicsResourceGetMappedEglFrame` | | + +## **22. Graphics Interoperability** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaGraphicsMapResources` | | +| `cudaGraphicsResourceGetMappedMipmappedArray` | | +| `cudaGraphicsResourceGetMappedPointer` | | +| `cudaGraphicsResourceSetMapFlags` | | +| `cudaGraphicsSubResourceGetMappedArray` | | +| `cudaGraphicsUnmapResources` | | +| `cudaGraphicsUnregisterResource` | | + +## **23. Texture Reference Management** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaBindTexture` | | +| `cudaBindTexture2D` | | +| `cudaBindTextureToArray` | | +| `cudaBindTextureToMipmappedArray` | | +| `cudaCreateChannelDesc` | | +| `cudaGetChannelDesc` | | +| `cudaGetTextureAlignmentOffset` | | +| `cudaGetTextureReference` | | +| `cudaUnbindTexture` | | + +## **24. Surface Reference Management** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaBindSurfaceToArray` | | +| `cudaGetSurfaceReference` | | + +## **25. Texture Object Management** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaCreateTextureObject` | | +| `cudaDestroyTextureObject` | | +| `cudaGetTextureObjectResourceDesc` | | +| `cudaGetTextureObjectResourceViewDesc` | | +| `cudaGetTextureObjectTextureDesc` | | + +## **26. Surface Object Management** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaCreateSurfaceObject` | | +| `cudaDestroySurfaceObject` | | +| `cudaGetSurfaceObjectResourceDesc` | | + +## **27. Version Management** + +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaDriverGetVersion` | `hipDriverGetVersion` | +| `cudaRuntimeGetVersion` | `hipRuntimeGetVersion` | + +## **28. C++ API Routines** *(7.0 contains, 7.5 doesn’t)* -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaBindSurfaceToArray` | | Binds an array to a surface. | -| `cudaBindTexture` | `hipBindTexture` | Binds a memory area to a texture. | -| `cudaBindTexture2D` | | Binds a 2D memory area to a texture. | -| `cudaBindTextureToArray` | | Binds an array to a texture. | -| `cudaBindTextureToMipmappedArray` | | Binds a mipmapped array to a texture. | -| `cudaCreateChannelDesc` | `hipCreateChannelDesc` | Returns a channel descriptor using the specified format. | -| `cudaFuncGetAttributes` | | Find out attributes for a given function. | -| `cudaFuncSetCacheConfig` | | Sets the preferred cache configuration for a device function. | -| `cudaGetSymbolAddress` | | Finds the address associated with a CUDA symbol | -| `cudaGetSymbolSize` | | Finds the size of the object associated with a CUDA symbol. | -| `cudaGetTextureAlignmentOffset` | | Get the alignment offset of a texture. | -| `cudaLaunch` | | Launches a device function. | -| `cudaLaunchKernel` | | Launches a device function. | -| `cudaMallocHost` | | Allocates page-locked memory on the host | -| `cudaMallocManaged` | | Allocates memory that will be automatically managed by the Unified Memory system. | -| `cudaMemcpyFromSymbol` | | Copies data from the given symbol on the device. | -| `cudaMemcpyFromSymbolAsync` | | Copies data from the given symbol on the device. | -| `cudaMemcpyToSymbol` | | Copies data to the given symbol on the device. | -| `cudaMemcpyToSymbolAsync` | | Async copies data to the given symbol on the device. | -| `cudaOccupancyMaxActiveBlocksPerMultiprocessor` | `hipOccupancyMaxActiveBlocksPerMultiprocessor` | Returns occupancy for a device function. | -| `cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` | | Returns occupancy for a device function with the specified flags. | -| `cudaOccupancyMaxPotentialBlockSize` | `hipOccupancyMaxPotentialBlockSize` | Returns grid and block size that achieves maximum potential occupancy for a device function. | -| `cudaOccupancyMaxPotentialBlockSizeVariableSMem` | | Returns grid and block size that achieves maximum potential occupancy for a device function. | -| `cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags` | | Returns grid and block size that achieves maximum potential occupancy for a device function. | -| `cudaOccupancyMaxPotentialBlockSizeWithFlags` | | Returns grid and block size that achived maximum potential occupancy for a device function with the specified flags. | -| `cudaSetupArgument` | | Configure a device launch. | -| `cudaStreamAttachMemAsync` | | Attach memory to a stream asynchronously. | -| `cudaUnbindTexture` | `hipUnbindTexture` | Unbinds a texture. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaBindSurfaceToArray` | | +| `cudaBindTexture` | `hipBindTexture` | +| `cudaBindTexture2D` | | +| `cudaBindTextureToArray` | | +| `cudaBindTextureToMipmappedArray` | | +| `cudaCreateChannelDesc` | `hipCreateChannelDesc` | +| `cudaFuncGetAttributes` | | +| `cudaFuncSetCacheConfig` | | +| `cudaGetSymbolAddress` | | +| `cudaGetSymbolSize` | | +| `cudaGetTextureAlignmentOffset` | | +| `cudaLaunch` | | +| `cudaLaunchKernel` | | +| `cudaMallocHost` | | +| `cudaMallocManaged` | | +| `cudaMemcpyFromSymbol` | | +| `cudaMemcpyFromSymbolAsync` | | +| `cudaMemcpyToSymbol` | | +| `cudaMemcpyToSymbolAsync` | | +| `cudaOccupancyMaxActiveBlocksPerMultiprocessor` | `hipOccupancyMaxActiveBlocksPerMultiprocessor` | +| `cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` | | +| `cudaOccupancyMaxPotentialBlockSize` | `hipOccupancyMaxPotentialBlockSize` | +| `cudaOccupancyMaxPotentialBlockSizeVariableSMem` | | +| `cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags` | | +| `cudaOccupancyMaxPotentialBlockSizeWithFlags` | | +| `cudaSetupArgument` | | +| `cudaStreamAttachMemAsync` | | +| `cudaUnbindTexture` | `hipUnbindTexture` | -## **19. Profiler Control** +## **30. Profiler Control** -| **CUDA** | **HIP** | **CUDA description** | -|-----------------------------------------------------------|-------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| `cudaProfilerInitialize` | | Initialize the CUDA profiler. | -| `cudaProfilerStart` | `hipProfilerStart` | Enable profiling. | -| `cudaProfilerStop` | `hipProfilerStop` | Disable profiling. | +| **CUDA** | **HIP** | +|-----------------------------------------------------------|-------------------------------| +| `cudaProfilerInitialize` | | +| `cudaProfilerStart` | `hipProfilerStart` | +| `cudaProfilerStop` | `hipProfilerStop` | # Data types used by CUDA Runtime API and supported by HIP -## **20. Data types** +## **31. Data types** -| **type** | **CUDA** | **HIP** | **CUDA description** | -|-------------:|-----------------------------------------------|------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------| -| struct | `cudaChannelFormatDesc` | `hipChannelFormatDesc` | CUDA Channel format descriptor. | -| struct | `cudaDeviceProp` | `hipDeviceProp_t` | CUDA device properties. | -| struct | `cudaExtent` | | CUDA extent (width, height, depth). | -| struct | `cudaFuncAttributes` | | CUDA function attributes. | -| struct | `cudaIpcEventHandle_t` | `hipIpcEventHandle_t` | CUDA IPC event handle. | -| struct | `cudaIpcMemHandle_t` | `hipIpcMemHandle_t` | CUDA IPC memory handle. | -| struct | `cudaMemcpy3DParms` | | CUDA 3D memory copying parameters. | -| struct | `cudaMemcpy3DPeerParms` | | CUDA 3D cross-device memory copying parameters. | -| struct | `cudaPitchedPtr` | | CUDA Pitched memory pointer. | -| struct | `cudaPointerAttributes` | `hipPointerAttribute_t` | CUDA pointer attributes. | -| struct | `cudaPos` | | CUDA 3D position. | -| struct | `cudaResourceDesc` | | CUDA resource descriptor. | -| struct | `cudaResourceViewDesc` | | CUDA resource view descriptor. | -| struct | `cudaTextureDesc` | | CUDA texture descriptor. | -| struct | `surfaceReference` | | CUDA Surface reference. | -| struct | `textureReference` | `textureReference` | CUDA texture reference. | -| enum |***`cudaChannelFormatKind`*** |***`hipChannelFormatKind`*** | Channel format kind. | -| 0 |*`cudaChannelFormatKindSigned`* |*`hipChannelFormatKindSigned`* | Signed channel format. | -| 1 |*`cudaChannelFormatKindUnsigned`* |*`hipChannelFormatKindUnsigned`* | Unsigned channel format. | -| 2 |*`cudaChannelFormatKindFloat`* |*`hipChannelFormatKindFloat`* | Float channel format. | -| 3 |*`cudaChannelFormatKindNone`* |*`hipChannelFormatKindNone`* | No channel format. | -| enum |***`cudaComputeMode`*** | | CUDA device compute modes. | -| 0 |*`cudaComputeModeDefault`* | | Default compute mode (Multiple threads can use ::cudaSetDevice() with this device). | -| 1 |*`cudaComputeModeExclusive`* | | Compute-exclusive-thread mode (Only one thread in one process will be able to use ::cudaSetDevice() with this device). | -| 2 |*`cudaComputeModeProhibited`* | | Compute-prohibited mode (No threads can use ::cudaSetDevice() with this device). | -| 3 |*`cudaComputeModeExclusiveProcess`* | | Compute-exclusive-process mode (Many threads in one process will be able to use ::cudaSetDevice() with this device). | -| enum |***`cudaDeviceAttr`*** |***`hipDeviceAttribute_t`*** | CUDA device attributes. | -| 1 |*`cudaDevAttrMaxThreadsPerBlock`* |*`hipDeviceAttributeMaxThreadsPerBlock`* | Maximum number of threads per block. | -| 2 |*`cudaDevAttrMaxBlockDimX`* |*`hipDeviceAttributeMaxBlockDimX`* | Maximum block dimension X. | -| 3 |*`cudaDevAttrMaxBlockDimY`* |*`hipDeviceAttributeMaxBlockDimY`* | Maximum block dimension Y. | -| 4 |*`cudaDevAttrMaxBlockDimZ`* |*`hipDeviceAttributeMaxBlockDimZ`* | Maximum block dimension Y. | -| 5 |*`cudaDevAttrMaxGridDimX`* |*`hipDeviceAttributeMaxGridDimX`* | Maximum grid dimension X. | -| 6 |*`cudaDevAttrMaxGridDimY`* |*`hipDeviceAttributeMaxGridDimY`* | Maximum grid dimension Y. | -| 7 |*`cudaDevAttrMaxGridDimZ`* |*`hipDeviceAttributeMaxGridDimZ`* | Maximum grid dimension Y. | -| 8 |*`cudaDevAttrMaxSharedMemoryPerBlock`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | Maximum shared memory available per block in bytes. | -| 9 |*`cudaDevAttrTotalConstantMemory`* |*`hipDeviceAttributeTotalConstantMemory`* | Memory available on device for \__constant__ variables in a CUDA C kernel in bytes. | -| 10 |*`cudaDevAttrWarpSize`* |*`hipDeviceAttributeWarpSize`* | Warp size in threads. | -| 11 |*`cudaDevAttrMaxPitch`* | | Maximum pitch in bytes allowed by memory copies. | -| 12 |*`cudaDevAttrMaxRegistersPerBlock`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | Maximum number of 32-bit registers available per block. | -| 13 |*`cudaDevAttrClockRate`* |*`hipDeviceAttributeClockRate`* | Peak clock frequency in kilohertz. | -| 14 |*`cudaDevAttrTextureAlignment`* | | Alignment requirement for textures. | -| 15 |*`cudaDevAttrGpuOverlap`* | | Device can possibly copy memory and execute a kernel concurrently. | -| 16 |*`cudaDevAttrMultiProcessorCount`* |*`hipDeviceAttributeMultiprocessorCount`* | Number of multiprocessors on device. | -| 17 |*`cudaDevAttrKernelExecTimeout`* | | Specifies whether there is a run time limit on kernels. | -| 18 |*`cudaDevAttrIntegrated`* | | Device is integrated with host memory. | -| 19 |*`cudaDevAttrCanMapHostMemory`* | | Device can map host memory into CUDA address space. | -| 20 |*`cudaDevAttrComputeMode`* |*`hipDeviceAttributeComputeMode`* | Compute mode (See cudaComputeMode for details). | -| 21 |*`cudaDevAttrMaxTexture1DWidth`* | | Maximum 1D texture width. | -| 22 |*`cudaDevAttrMaxTexture2DWidth`* | | Maximum 2D texture width. | -| 23 |*`cudaDevAttrMaxTexture2DHeight`* | | Maximum 2D texture height. | -| 24 |*`cudaDevAttrMaxTexture3DWidth`* | | Maximum 3D texture width. | -| 25 |*`cudaDevAttrMaxTexture3DHeight`* | | Maximum 3D texture height. | -| 26 |*`cudaDevAttrMaxTexture3DDepth`* | | Maximum 3D texture depth. | -| 27 |*`cudaDevAttrMaxTexture2DLayeredWidth`* | | Maximum 2D layered texture width. | -| 28 |*`cudaDevAttrMaxTexture2DLayeredHeight`* | | Maximum 2D layered texture height. | -| 29 |*`cudaDevAttrMaxTexture2DLayeredLayers`* | | Maximum layers in a 2D layered texture. | -| 30 |*`cudaDevAttrSurfaceAlignment`* | | Alignment requirement for surfaces. | -| 31 |*`cudaDevAttrConcurrentKernels`* |*`hipDeviceAttributeConcurrentKernels`* | Device can possibly execute multiple kernels concurrently. | -| 32 |*`cudaDevAttrEccEnabled`* | | Device has ECC support enabled. | -| 33 |*`cudaDevAttrPciBusId`* |*`hipDeviceAttributePciBusId`* | PCI bus ID of the device. | -| 34 |*`cudaDevAttrPciDeviceId`* |*`hipDeviceAttributePciDeviceId`* | PCI device ID of the device. | -| 35 |*`cudaDevAttrTccDriver`* | | Device is using TCC driver model. | -| 36 |*`cudaDevAttrMemoryClockRate`* |*`hipDeviceAttributeMemoryClockRate`* | Peak memory clock frequency in kilohertz. | -| 37 |*`cudaDevAttrGlobalMemoryBusWidth`* |*`hipDeviceAttributeMemoryBusWidth`* | Global memory bus width in bits. | -| 38 |*`cudaDevAttrL2CacheSize`* |*`hipDeviceAttributeL2CacheSize`* | Size of L2 cache in bytes. | -| 39 |*`cudaDevAttrMaxThreadsPerMultiProcessor`* |*`hipDeviceAttributeMaxThreadsPerMultiProcessor`* | Maximum resident threads per multiprocessor. | -| 40 |*`cudaDevAttrAsyncEngineCount`* | | Number of asynchronous engines. | -| 41 |*`cudaDevAttrUnifiedAddressing`* | | Device shares a unified address space with the host. | -| 42 |*`cudaDevAttrMaxTexture1DLayeredWidth`* | | Maximum 1D layered texture width. | -| 43 |*`cudaDevAttrMaxTexture1DLayeredLayers`* | | Maximum layers in a 1D layered texture. | -| 44 | | | *reserved* | -| 45 |*`cudaDevAttrMaxTexture2DGatherWidth`* | | Maximum 2D texture width if cudaArrayTextureGather is set. | -| 46 |*`cudaDevAttrMaxTexture2DGatherHeight`* | | Maximum 2D texture height if cudaArrayTextureGather is set. | -| 47 |*`cudaDevAttrMaxTexture3DWidthAlt`* | | Alternate maximum 3D texture width. | -| 48 |*`cudaDevAttrMaxTexture3DHeightAlt`* | | Alternate maximum 3D texture height. | -| 49 |*`cudaDevAttrMaxTexture3DDepthAlt`* | | Alternate maximum 3D texture depth. | -| 50 |*`cudaDevAttrPciDomainId`* | | PCI domain ID of the device. | -| 51 |*`cudaDevAttrTexturePitchAlignment`* | | Pitch alignment requirement for textures. | -| 52 |*`cudaDevAttrMaxTextureCubemapWidth`* | | Maximum cubemap texture width/height. | -| 53 |*`cudaDevAttrMaxTextureCubemapLayeredWidth`* | | Maximum cubemap layered texture width/height. | -| 54 |*`cudaDevAttrMaxTextureCubemapLayeredLayers`* | | Maximum layers in a cubemap layered texture. | -| 55 |*`cudaDevAttrMaxSurface1DWidth`* | | Maximum 1D surface width. | -| 56 |*`cudaDevAttrMaxSurface2DWidth`* | | Maximum 2D surface width. | -| 57 |*`cudaDevAttrMaxSurface2DHeight`* | | Maximum 2D surface height. | -| 58 |*`cudaDevAttrMaxSurface3DWidth`* | | Maximum 3D surface width. | -| 59 |*`cudaDevAttrMaxSurface3DHeight`* | | Maximum 3D surface height. | -| 60 |*`cudaDevAttrMaxSurface3DDepth`* | | Maximum 3D surface depth. | -| 61 |*`cudaDevAttrMaxSurface1DLayeredWidth`* | | Maximum 1D layered surface width. | -| 62 |*`cudaDevAttrMaxSurface1DLayeredLayers`* | | Maximum layers in a 1D layered surface. | -| 63 |*`cudaDevAttrMaxSurface2DLayeredWidth`* | | Maximum 2D layered surface width. | -| 64 |*`cudaDevAttrMaxSurface2DLayeredHeight`* | | Maximum 2D layered surface height. | -| 65 |*`cudaDevAttrMaxSurface2DLayeredLayers`* | | Maximum layers in a 2D layered surface. | -| 66 |*`cudaDevAttrMaxSurfaceCubemapWidth`* | | Maximum cubemap surface width. | -| 67 |*`cudaDevAttrMaxSurfaceCubemapLayeredWidth`* | | Maximum cubemap layered surface width. | -| 68 |*`cudaDevAttrMaxSurfaceCubemapLayeredLayers`* | | Maximum layers in a cubemap layered surface. | -| 69 |*`cudaDevAttrMaxTexture1DLinearWidth`* | | Maximum 1D linear texture width. | -| 70 |*`cudaDevAttrMaxTexture2DLinearWidth`* | | Maximum 2D linear texture width. | -| 71 |*`cudaDevAttrMaxTexture2DLinearHeight`* | | Maximum 2D linear texture height. | -| 72 |*`cudaDevAttrMaxTexture2DLinearPitch`* | | Maximum 2D linear texture pitch in bytes. | -| 73 |*`cudaDevAttrMaxTexture2DMipmappedWidth`* | | Maximum mipmapped 2D texture width. | -| 74 |*`cudaDevAttrMaxTexture2DMipmappedHeight`* | | Maximum mipmapped 2D texture height. | -| 75 |*`cudaDevAttrComputeCapabilityMajor`* |*`hipDeviceAttributeComputeCapabilityMajor`* | Major compute capability version number. | -| 76 |*`cudaDevAttrComputeCapabilityMinor`* |*`hipDeviceAttributeComputeCapabilityMinor`* | Minor compute capability version number. | -| 77 |*`cudaDevAttrMaxTexture1DMipmappedWidth`* | | Maximum mipmapped 1D texture width. | -| 78 |*`cudaDevAttrStreamPrioritiesSupported`* | | Device supports stream priorities. | -| 79 |*`cudaDevAttrGlobalL1CacheSupported`* | | Device supports caching globals in L1. | -| 80 |*`cudaDevAttrLocalL1CacheSupported`* | | Device supports caching locals in L1. | -| 81 |*`cudaDevAttrMaxSharedMemoryPerMultiprocessor`*|*`hipDeviceAttributeMaxSharedMemoryPerMultiprocessor`*| Maximum shared memory available per multiprocessor in bytes. | -| 82 |*`cudaDevAttrMaxRegistersPerMultiprocessor`* | | Maximum number of 32-bit registers available per multiprocessor. | -| 83 |*`cudaDevAttrManagedMemory`* | | Device can allocate managed memory on this system. | -| 84 |*`cudaDevAttrIsMultiGpuBoard`* |*`hipDeviceAttributeIsMultiGpuBoard`* | Device is on a multi-GPU board. | -| 85 |*`cudaDevAttrMultiGpuBoardGroupID`* | | Unique identifier for a group of devices on the same multi-GPU board. | -| enum |***`cudaError`*** |***`hipError_t`*** | CUDA Error types. | -| enum |***`cudaError_t`*** |***`hipError_t`*** | CUDA Error types. | -| 0 |*`cudaSuccess`* |*`hipSuccess`* | The API call returned with no errors. In the case of query calls, this can also mean that the operation being queried is complete.| -| 1 |*`cudaErrorMissingConfiguration`* | | The device function being invoked (usually via cudaLaunchKernel()) was not previously configured via the cudaConfigureCall() function.| -| 2 |*`cudaErrorMemoryAllocation`* |*`hipErrorMemoryAllocation`* | The API call failed because it was unable to allocate enough memory to perform the requested operation. | -| 3 |*`cudaErrorInitializationError`* |*`hipErrorInitializationError`* | The API call failed because the CUDA driver and runtime could not be initialized. | -| 4 |*`cudaErrorLaunchFailure`* | | An exception occurred on the device while executing a kernel. Common causes include dereferencing an invalid device pointer and accessing out of bounds shared memory. The device cannot be used until cudaThreadExit() is called. All existing device memory allocations are invalid and must be reconstructed if the program is to continue using CUDA. | -| 5 |*`cudaErrorPriorLaunchFailure`* | | This indicated that a previous kernel launch failed. This was previously used for device emulation of kernel launches. Deprecated This error return is deprecated as of CUDA 3.1. Device emulation mode was removed with the CUDA 3.1 release.| -| 6 |*`cudaErrorLaunchTimeout`* | | This indicates that the device kernel took too long to execute. This can only occur if timeouts are enabled - see the device property kernelExecTimeoutEnabled for more information. The device cannot be used until cudaThreadExit() is called. All existing device memory allocations are invalid and must be reconstructed if the program is to continue using CUDA.| -| 7 |*`cudaErrorLaunchOutOfResources`* |*`hipErrorLaunchOutOfResources`* | This indicates that a launch did not occur because it did not have appropriate resources. Although this error is similar to cudaErrorInvalidConfiguration, this error usually indicates that the user has attempted to pass too many arguments to the device kernel, or the kernel launch specifies too many threads for the kernel's register count.| -| 8 |*`cudaErrorInvalidDeviceFunction`* | | The requested device function does not exist or is not compiled for the proper device architecture. | -| 9 |*`cudaErrorInvalidConfiguration`* | | This indicates that a kernel launch is requesting resources that can never be satisfied by the current device. Requesting more shared memory per block than the device supports will trigger this error, as will requesting too many threads or blocks. See cudaDeviceProp for more device limitations.| -| 10 |*`cudaErrorInvalidDevice`* |*`hipErrorInvalidDevice`* | This indicates that the device ordinal supplied by the user does not correspond to a valid CUDA device. | -| 11 |*`cudaErrorInvalidValue`* |*`hipErrorInvalidValue`* | This indicates that one or more of the parameters passed to the API call is not within an acceptable range of values. | -| 12 |*`cudaErrorInvalidPitchValue`* | | This indicates that one or more of the pitch-related parameters passed to the API call is not within the acceptable range for pitch.| -| 13 |*`cudaErrorInvalidSymbol`* | | This indicates that the symbol name/identifier passed to the API call is not a valid name or identifier. | -| 14 |*`cudaErrorMapBufferObjectFailed`* | | This indicates that the buffer object could not be mapped. | -| 15 |*`cudaErrorUnmapBufferObjectFailed`* | | This indicates that the buffer object could not be unmapped. | -| 16 |*`cudaErrorInvalidHostPointer`* | | This indicates that at least one host pointer passed to the API call is not a valid host pointer. | -| 17 |*`cudaErrorInvalidDevicePointer`* |*`hipErrorInvalidDevicePointer`* | This indicates that at least one host pointer passed to the API call is not a valid host pointer. | -| 18 |*`cudaErrorInvalidTexture`* | | This indicates that the texture passed to the API call is not a valid texture. | -| 19 |*`cudaErrorInvalidTextureBinding`* | | This indicates that the texture binding is not valid. This occurs if you call cudaGetTextureAlignmentOffset() with an unbound texture.| -| 20 |*`cudaErrorInvalidChannelDescriptor`* | | This indicates that the channel descriptor passed to the API call is not valid. This occurs if the format is not one of the formats specified by cudaChannelFormatKind, or if one of the dimensions is invalid.| -| 21 |*`cudaErrorInvalidMemcpyDirection`* | | This indicates that the direction of the memcpy passed to the API call is not one of the types specified by cudaMemcpyKind. | -| 22 |*`cudaErrorAddressOfConstant`* | | This indicated that the user has taken the address of a constant variable, which was forbidden up until the CUDA 3.1 release. Deprecated This error return is deprecated as of CUDA 3.1. Variables in constant memory may now have their address taken by the runtime via cudaGetSymbolAddress().| -| 23 |*`cudaErrorTextureFetchFailed`* | | This indicated that a texture fetch was not able to be performed. This was previously used for device emulation of texture operations. Deprecated This error return is deprecated as of CUDA 3.1. Device emulation mode was removed with the CUDA 3.1 release.| -| 24 |*`cudaErrorTextureNotBound`* | | This indicated that a texture was not bound for access. This was previously used for device emulation of texture operations. Deprecated This error return is deprecated as of CUDA 3.1. Device emulation mode was removed with the CUDA 3.1 release.| -| 25 |*`cudaErrorSynchronizationError`* | | This indicated that a synchronization operation had failed. This was previously used for some device emulation functions. Deprecated This error return is deprecated as of CUDA 3.1. Device emulation mode was removed with the CUDA 3.1 release.| -| 26 |*`cudaErrorInvalidFilterSetting`* | | This indicates that a non-float texture was being accessed with linear filtering. This is not supported by CUDA. | -| 27 |*`cudaErrorInvalidNormSetting`* | | This indicates that an attempt was made to read a non-float texture as a normalized float. This is not supported by CUDA. | -| 28 |*`cudaErrorMixedDeviceExecution`* | | Mixing of device and device emulation code was not allowed. Deprecated This error return is deprecated as of CUDA 3.1. Device emulation mode was removed with the CUDA 3.1 release.| -| 29 |*`cudaErrorCudartUnloading`* | | This indicates that a CUDA Runtime API call cannot be executed because it is being called during process shut down, at a point in time after CUDA driver has been unloaded.| -| 30 |*`cudaErrorUnknown`* |*`hipErrorUnknown`* | This indicates that an unknown internal error has occurred. | -| 31 |*`cudaErrorNotYetImplemented`* | | This indicates that the API call is not yet implemented. Production releases of CUDA will never return this error. Deprecated This error return is deprecated as of CUDA 4.1.| -| 32 |*`cudaErrorMemoryValueTooLarge`* | | This indicated that an emulated device pointer exceeded the 32-bit address range. Deprecated This error return is deprecated as of CUDA 3.1. Device emulation mode was removed with the CUDA 3.1 release.| -| 33 |*`cudaErrorInvalidResourceHandle`* |*`hipErrorInvalidResourceHandle`* | This indicates that a resource handle passed to the API call was not valid. Resource handles are opaque types like cudaStream_t and cudaEvent_t.| -| 34 |*`cudaErrorNotReady`* |*`hipErrorNotReady`* | This indicates that asynchronous operations issued previously have not completed yet. This result is not actually an error, but must be indicated differently than cudaSuccess (which indicates completion). Calls that may return this value include cudaEventQuery() and cudaStreamQuery().| -| 35 |*`cudaErrorInsufficientDriver`* | | This indicates that the installed NVIDIA CUDA driver is older than the CUDA runtime library. This is not a supported configuration. Users should install an updated NVIDIA display driver to allow the application to run.| -| 36 |*`cudaErrorSetOnActiveProcess`* | | This indicates that the user has called cudaSetValidDevices(), cudaSetDeviceFlags(), cudaD3D9SetDirect3DDevice(), cudaD3D10SetDirect3DDevice, cudaD3D11SetDirect3DDevice(), or cudaVDPAUSetVDPAUDevice() after initializing the CUDA runtime by calling non-device management operations (allocating memory and launching kernels are examples of non-device management operations). This error can also be returned if using runtime/driver interoperability and there is an existing CUcontext active on the host thread.| -| 37 |*`cudaErrorInvalidSurface`* | | This indicates that the surface passed to the API call is not a valid surface. | -| 38 |*`cudaErrorNoDevice`* |*`hipErrorNoDevice`* | This indicates that no CUDA-capable devices were detected by the installed CUDA driver. | -| 39 |*`cudaErrorECCUncorrectable`* | | This indicates that an uncorrectable ECC error was detected during execution. | -| 40 |*`cudaErrorSharedObjectSymbolNotFound`* | | This indicates that a link to a shared object failed to resolve. | -| 41 |*`cudaErrorSharedObjectInitFailed`* | | This indicates that initialization of a shared object failed. | -| 42 |*`cudaErrorUnsupportedLimit`* |*`hipErrorUnsupportedLimit`* | This indicates that the cudaLimit passed to the API call is not supported by the active device. | -| 43 |*`cudaErrorDuplicateVariableName`* | | This indicates that multiple global or constant variables (across separate CUDA source files in the application) share the same string name.| -| 44 |*`cudaErrorDuplicateTextureName`* | | This indicates that multiple textures (across separate CUDA source files in the application) share the same string name. | -| 45 |*`cudaErrorDuplicateSurfaceName`* | | This indicates that multiple surfaces (across separate CUDA source files in the application) share the same string name. | -| 46 |*`cudaErrorDevicesUnavailable`* | | This indicates that all CUDA devices are busy or unavailable at the current time. Devices are often busy/unavailable due to use of cudaComputeModeExclusive, cudaComputeModeProhibited or when long running CUDA kernels have filled up the GPU and are blocking new work from starting. They can also be unavailable due to memory constraints on a device that already has active CUDA work being performed.| -| 47 |*`cudaErrorInvalidKernelImage`* | | This indicates that the device kernel image is invalid. | -| 48 |*`cudaErrorNoKernelImageForDevice`* | | This indicates that there is no kernel image available that is suitable for the device. This can occur when a user specifies code generation options for a particular CUDA source file that do not include the corresponding device configuration.| -| 49 |*`cudaErrorIncompatibleDriverContext`* | | This indicates that the current context is not compatible with this the CUDA Runtime. This can only occur if you are using CUDA Runtime/Driver interoperability and have created an existing Driver context using the driver API. The Driver context may be incompatible either because the Driver context was created using an older version of the API, because the Runtime API call expects a primary driver context and the Driver context is not primary, or because the Driver context has been destroyed. Please see Interactions with the CUDA Driver API" for more information.| -| 50 |*`cudaErrorPeerAccessAlreadyEnabled`* |*`hipErrorPeerAccessAlreadyEnabled`* | This error indicates that a call to cudaDeviceEnablePeerAccess() is trying to re-enable peer addressing on from a context which has already had peer addressing enabled.| -| 51 |*`cudaErrorPeerAccessNotEnabled`* |*`hipErrorPeerAccessNotEnabled`* | This error indicates that a call to cudaDeviceEnablePeerAccess() is trying to re-enable peer addressing on from a context which has already had peer addressing enabled.| -| 52 | | | *reserved* | -| 53 | | | *reserved* | -| 54 |*`cudaErrorDeviceAlreadyInUse`* | | This indicates that a call tried to access an exclusive-thread device that is already in use by a different thread. | -| 55 |*`cudaErrorProfilerDisabled`* | | This indicates profiler is not initialized for this run. This can happen when the application is running with external profiling tools like visual profiler.| -| 56 |*`cudaErrorProfilerNotInitialized`* | | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to attempt to enable/disable the profiling via cudaProfilerStart or cudaProfilerStop without initialization.| -| 57 |*`cudaErrorProfilerAlreadyStarted`* | | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to call cudaProfilerStart() when profiling is already enabled.| -| 58 |*`cudaErrorProfilerAlreadyStopped`* | | Deprecated This error return is deprecated as of CUDA 5.0. It is no longer an error to call cudaProfilerStop() when profiling is already disabled.| -| 59 |*`cudaErrorAssert`* | | An assert triggered in device code during kernel execution. The device cannot be used again until cudaThreadExit() is called. All existing allocations are invalid and must be reconstructed if the program is to continue using CUDA.| -| 60 |*`cudaErrorTooManyPeers`* | | This error indicates that the hardware resources required to enable peer access have been exhausted for one or more of the devices passed to cudaEnablePeerAccess().| -| 61 |*`cudaErrorHostMemoryAlreadyRegistered`* | *`hipErrorHostMemoryAlreadyRegistered`* | This error indicates that the memory range passed to cudaHostRegister() has already been registered. | -| 62 |*`cudaErrorHostMemoryNotRegistered`* | *`hipErrorHostMemoryNotRegistered`* | This error indicates that the pointer passed to cudaHostUnregister() does not correspond to any currently registered memory region.| -| 63 |*`cudaErrorOperatingSystem`* | | This error indicates that an OS call failed. | -| 64 |*`cudaErrorPeerAccessUnsupported`* | | This error indicates that P2P access is not supported across the given devices. | -| 65 |*`cudaErrorLaunchMaxDepthExceeded`* | | This error indicates that a device runtime grid launch did not occur because the depth of the child grid would exceed the maximum supported number of nested grid launches.| -| 66 |*`cudaErrorLaunchFileScopedTex`* | | This error indicates that a grid launch did not occur because the kernel uses filescoped textures which are unsupported by the device runtime. Kernels launched via the device runtime only support textures created with the Texture Object API's.| -| 67 |*`cudaErrorLaunchFileScopedSurf`* | | This error indicates that a grid launch did not occur because the kernel uses filescoped surfaces which are unsupported by the device runtime. Kernels launched via the device runtime only support surfaces created with the Surface Object API's.| -| 68 |*`cudaErrorSyncDepthExceeded`* | | This error indicates that a call to cudaDeviceSynchronize made from the device runtime failed because the call was made at grid depth greater than than either the default (2 levels of grids) or user specified device limit cudaLimitDevRuntimeSyncDepth. To be able to synchronize on launched grids at a greater depth successfully, the maximum nested depth at which cudaDeviceSynchronize will be called must be specified with the cudaLimitDevRuntimeSyncDepth limit to the cudaDeviceSetLimit api before the host-side launch of a kernel using the device runtime. Keep in mind that additional levels of sync depth require the runtime to reserve large amounts of device memory that cannot be used for user allocations.| -| 69 |*`cudaErrorLaunchPendingCountExceeded`* | | This error indicates that a device runtime grid launch failed because the launch would exceed the limit cudaLimitDevRuntimePendingLaunchCount. For this launch to proceed successfully, cudaDeviceSetLimit must be called to set the cudaLimitDevRuntimePendingLaunchCount to be higher than the upper bound of outstanding launches that can be issued to the device runtime. Keep in mind that raising the limit of pending device runtime launches will require the runtime to reserve device memory that cannot be used for user allocations.| -| 70 |*`cudaErrorNotPermitted`* | | This error indicates the attempted operation is not permitted. | -| 71 |*`cudaErrorNotSupported`* | | This error indicates the attempted operation is not supported on the current system or device. | -| 72 |*`cudaErrorHardwareStackError`* | | Device encountered an error in the call stack during kernel execution, possibly due to stack corruption or exceeding the stack size limit. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA.| -| 73 |*`cudaErrorIllegalInstruction`* | | The device encountered an illegal instruction during kernel execution The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA.| -| 74 |*`cudaErrorMisalignedAddress`* | | The device encountered a load or store instruction on a memory address which is not aligned. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA.| -| 75 |*`cudaErrorInvalidAddressSpace`* | | While executing a kernel, the device encountered an instruction which can only operate on memory locations in certain address spaces (global, shared, or local), but was supplied a memory address not belonging to an allowed address space. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA.| -| 76 |*`cudaErrorInvalidPc`* | | The device encountered an invalid program counter. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA.| -| 77 |*`cudaErrorIllegalAddress`* | | The device encountered a load or store instruction on an invalid memory address. The context cannot be used, so it must be destroyed (and a new one should be created). All existing device memory allocations from this context are invalid and must be reconstructed if the program is to continue using CUDA.| -| 78 |*`cudaErrorInvalidPtx`* | | A PTX compilation failed. The runtime may fall back to compiling PTX if an application does not contain a suitable binary for the current device.| -| 79 |*`cudaErrorInvalidGraphicsContext`* | | This indicates an error with the OpenGL or DirectX context. | -| 0x7f |*`cudaErrorStartupFailure`* | | This indicates an internal startup failure in the CUDA runtime. | -| 1000 |*`cudaErrorApiFailureBase`* | | Any unhandled CUDA driver error is added to this value and returned via the runtime. Production releases of CUDA should not return such errors. Deprecated This error return is deprecated as of CUDA 4.1.| -| enum |***`cudaFuncCache`*** |***`hipFuncCache_t`*** | CUDA function cache configurations. | -| 0 |*`cudaFuncCachePreferNone`* |*`hipFuncCachePreferNone`* | Default function cache configuration, no preference. | -| 1 |*`cudaFuncCachePreferShared`* |*`hipFuncCachePreferShared`* | Prefer larger shared memory and smaller L1 cache. | -| 2 |*`cudaFuncCachePreferL1`* |*`hipFuncCachePreferL1`* | Prefer larger L1 cache and smaller shared memory. | -| 3 |*`cudaFuncCachePreferEqual`* |*`hipFuncCachePreferEqual`* | Prefer equal size L1 cache and shared memory. | -| enum |***`cudaGraphicsCubeFace`*** | | CUDA graphics interop array indices for cube maps. | -| 0x00 |*`cudaGraphicsCubeFacePositiveX`* | | Positive X face of cubemap. | -| 0x01 |*`cudaGraphicsCubeFaceNegativeX`* | | Negative X face of cubemap. | -| 0x02 |*`cudaGraphicsCubeFacePositiveY`* | | Positive Y face of cubemap. | -| 0x03 |*`cudaGraphicsCubeFaceNegativeY`* | | Negative Y face of cubemap. | -| 0x04 |*`cudaGraphicsCubeFacePositiveZ`* | | Positive Z face of cubemap. | -| 0x05 |*`cudaGraphicsCubeFaceNegativeZ`* | | Negative Z face of cubemap. | -| enum |***`cudaGraphicsMapFlags`*** | | CUDA graphics interop map flags. | -| 0 |*`cudaGraphicsMapFlagsNone`* | | Default; Assume resource can be read/written. | -| 1 |*`cudaGraphicsMapFlagsReadOnly`* | | CUDA will not write to this resource. | -| 2 |*`cudaGraphicsMapFlagsWriteDiscard`* | | CUDA will only write to and will not read from this resource. | -| enum |***`cudaGraphicsRegisterFlags`*** | | CUDA graphics interop register flags. | -| 0 |*`cudaGraphicsRegisterFlagsNone`* | | Default. | -| 1 |*`cudaGraphicsRegisterFlagsReadOnly`* | | CUDA will not write to this resource. | -| 2 |*`cudaGraphicsRegisterFlagsWriteDiscard`* | | CUDA will only write to and will not read from this resource. | -| 4 |*`cudaGraphicsRegisterFlagsSurfaceLoadStore`* | | CUDA will bind this resource to a surface reference. | -| 8 |*`cudaGraphicsRegisterFlagsTextureGather`* | | CUDA will perform texture gather operations on this resource. | -| enum |***`cudaLimit`*** |***`hipLimit_t`*** | CUDA Limits. | -| 0x00 |*`cudaLimitStackSize`* | | GPU thread stack size. | -| 0x01 |*`cudaLimitPrintfFifoSize`* | | GPU printf/fprintf FIFO size. | -| 0x02 |*`cudaLimitMallocHeapSize`* |*`hipLimitMallocHeapSize`* | GPU malloc heap size. | -| 0x03 |*`cudaLimitDevRuntimeSyncDepth`* | | GPU device runtime synchronize depth. | -| 0x04 |*`cudaLimitDevRuntimePendingLaunchCount`* | | GPU device runtime pending launch count. | -| enum |***`cudaMemcpyKind`*** |***`hipMemcpyKind`*** | CUDA memory copy types. | -| 0 |*`cudaMemcpyHostToHost`* |*`hipMemcpyHostToHost`* | Host -> Host. | -| 1 |*`cudaMemcpyHostToDevice`* |*`hipMemcpyHostToDevice`* | Host -> Device. | -| 2 |*`cudaMemcpyDeviceToHost`* |*`hipMemcpyDeviceToHost`* | Device -> Host. | -| 3 |*`cudaMemcpyDeviceToDevice`* |*`hipMemcpyDeviceToDevice`* | Device -> Device. | -| 4 |*`cudaMemcpyDefault`* |*`hipMemcpyDefault`* | Default based unified virtual address space. | -| enum |***`cudaMemoryType`*** |***`hipMemoryType`*** | CUDA memory types. | -| 1 |*`cudaMemoryTypeHost`* |*`hipMemoryTypeHost`* | Host memory. | -| 2 |*`cudaMemoryTypeDevice`* |*`hipMemoryTypeDevice`* | Device memory. | -| enum |***`cudaResourceType`*** | | CUDA resource types. | -| 0 |*`cudaResourceTypeArray`* | | Array resource. | -| 1 |*`cudaResourceTypeMipmappedArray`* | | Mipmapped array resource. | -| 2 |*`cudaResourceTypeLinear`* | | Linear resource. | -| 3 |*`cudaResourceTypePitch2D`* | | Pitch 2D resource. | -| enum |***`cudaResourceViewFormat`*** | | CUDA texture resource view formats. | -| 0x00 |*`cudaResViewFormatNone`* | | No resource view format (use underlying resource format). | -| 0x01 |*`cudaResViewFormatUnsignedChar1`* | | 1 channel unsigned 8-bit integers. | -| 0x02 |*`cudaResViewFormatUnsignedChar2`* | | 2 channel unsigned 8-bit integers. | -| 0x03 |*`cudaResViewFormatUnsignedChar4`* | | 4 channel unsigned 8-bit integers. | -| 0x04 |*`cudaResViewFormatSignedChar1`* | | 1 channel signed 8-bit integers. | -| 0x05 |*`cudaResViewFormatSignedChar2`* | | 2 channel signed 8-bit integers. | -| 0x06 |*`cudaResViewFormatSignedChar4`* | | 4 channel signed 8-bit integers. | -| 0x07 |*`cudaResViewFormatUnsignedShort1`* | | 1 channel unsigned 16-bit integers. | -| 0x08 |*`cudaResViewFormatUnsignedShort2`* | | 2 channel unsigned 16-bit integers. | -| 0x09 |*`cudaResViewFormatUnsignedShort4`* | | 4 channel unsigned 16-bit integers. | -| 0x0a |*`cudaResViewFormatSignedShort1`* | | 1 channel signed 16-bit integers. | -| 0x0b |*`cudaResViewFormatSignedShort2`* | | 2 channel signed 16-bit integers. | -| 0x0c |*`cudaResViewFormatSignedShort4`* | | 4 channel signed 16-bit integers. | -| 0x0d |*`cudaResViewFormatUnsignedInt1`* | | 1 channel unsigned 32-bit integers. | -| 0x0e |*`cudaResViewFormatUnsignedInt2`* | | 2 channel unsigned 32-bit integers. | -| 0x0f |*`cudaResViewFormatUnsignedInt4`* | | 4 channel unsigned 32-bit integers. | -| 0x10 |*`cudaResViewFormatSignedInt1`* | | 1 channel signed 32-bit integers. | -| 0x11 |*`cudaResViewFormatSignedInt2`* | | 2 channel signed 32-bit integers. | -| 0x12 |*`cudaResViewFormatSignedInt4`* | | 4 channel signed 32-bit integers. | -| 0x13 |*`cudaResViewFormatHalf1`* | | 1 channel 16-bit floating point. | -| 0x14 |*`cudaResViewFormatHalf2`* | | 2 channel 16-bit floating point. | -| 0x15 |*`cudaResViewFormatHalf4`* | | 4 channel 16-bit floating point. | -| 0x16 |*`cudaResViewFormatFloat1`* | | 1 channel 32-bit floating point. | -| 0x17 |*`cudaResViewFormatFloat2`* | | 2 channel 32-bit floating point. | -| 0x18 |*`cudaResViewFormatFloat4`* | | 4 channel 32-bit floating point. | -| 0x19 |*`cudaResViewFormatUnsignedBlockCompressed1`* | | Block compressed 1. | -| 0x1a |*`cudaResViewFormatUnsignedBlockCompressed2`* | | Block compressed 2. | -| 0x1b |*`cudaResViewFormatUnsignedBlockCompressed3`* | | Block compressed 3. | -| 0x1c |*`cudaResViewFormatUnsignedBlockCompressed4`* | | Block compressed 4 unsigned. | -| 0x1d |*`cudaResViewFormatSignedBlockCompressed4`* | | Block compressed 4 signed. | -| 0x1e |*`cudaResViewFormatUnsignedBlockCompressed5`* | | Block compressed 5 unsigned. | -| 0x1f |*`cudaResViewFormatSignedBlockCompressed5`* | | Block compressed 5 signed. | -| 0x20 |*`cudaResViewFormatUnsignedBlockCompressed6H`* | | Block compressed 6 unsigned half-float. | -| 0x21 |*`cudaResViewFormatSignedBlockCompressed6H`* | | Block compressed 6 signed half-float. | -| 0x22 |*`cudaResViewFormatUnsignedBlockCompressed7`* | | Block compressed 7. | -| enum |***`cudaSharedMemConfig`*** |***`hipSharedMemConfig`*** | CUDA shared memory configuration. | -| 0 |*`cudaSharedMemBankSizeDefault`* |*`hipSharedMemBankSizeDefault`* | | -| 1 |*`cudaSharedMemBankSizeFourByte`* |*`hipSharedMemBankSizeFourByte`* | | -| 2 |*`cudaSharedMemBankSizeEightByte`* |*`hipSharedMemBankSizeEightByte`* | | -| enum |***`cudaSurfaceBoundaryMode`*** | | CUDA Surface boundary modes. | -| 0 |*`cudaBoundaryModeZero`* | | Zero boundary mode. | -| 1 |*`cudaBoundaryModeClamp`* | | Clamp boundary mode. | -| 2 |*`cudaBoundaryModeTrap`* | | Trap boundary mode. | -| enum |***`cudaSurfaceFormatMode`*** | | CUDA Surface format modes. | -| 0 |*`cudaFormatModeForced`* | | Forced format mode. | -| 1 |*`cudaFormatModeAuto`* | | Auto format mode. | -| enum |***`cudaTextureAddressMode`*** | | CUDA texture address modes. | -| 0 |*`cudaAddressModeWrap`* | | Wrapping address mode. | -| 1 |*`cudaAddressModeClamp`* | | Clamp to edge address mode. | -| 2 |*`cudaAddressModeMirror`* | | Mirror address mode. | -| 3 |*`cudaAddressModeBorder`* | | Border address mode. | -| enum |***`cudaTextureFilterMode`*** |***`hipTextureFilterMode`*** | Point filter mode. | -| 0 |*`cudaFilterModePoint`* |*`hipFilterModePoint`* | Linear filter mode. | -| 1 |*`cudaFilterModeLinear`* | | Clamp to edge address mode. | -| enum |***`cudaTextureReadMode`*** |***`hipTextureReadMode`*** | CUDA texture read modes. | -| 0 |*`cudaReadModeElementType`* |*`hipReadModeElementType`* | Read texture as specified element type. | -| 1 |*`cudaReadModeNormalizedFloat`* | | Read texture as normalized float. | -| struct | `cudaArray` | `hipArray` | CUDA array [opaque]. | -| typedef | `cudaArray_t` | `hipArray *` | CUDA array pointer. | -| typedef | `cudaArray_const_t` | `const hipArray *` | CUDA array (as source copy argument). | -| enum | `cudaError` | `hipError_t` | CUDA Error types. | -| typedef | `cudaError_t` | `hipError_t` | CUDA Error types. | -| typedef | `cudaEvent_t` | `hipEvent_t` | CUDA event types. | -| typedef | `cudaGraphicsResource_t` | | CUDA graphics resource types. | -| typedef | `cudaMipmappedArray_t` | | CUDA mipmapped array. | -| typedef | `cudaMipmappedArray_const_t` | | CUDA mipmapped array (as source argument). | -| enum |***`cudaOutputMode`*** | | CUDA Profiler Output modes. | -| 0x00 |*`cudaKeyValuePair`* | | Output mode Key-Value pair format. | -| 0x01 |*`cudaCSV`* | | Output mode Comma separated values format. | -| typedef | `cudaOutputMode_t` | | CUDA output file modes. | -| typedef | `cudaStream_t` | `hipStream_t` | CUDA stream. | -| typedef | `cudaSurfaceObject_t` | | An opaque value that represents a CUDA Surface object. | -| typedef | `cudaTextureObject_t` | | An opaque value that represents a CUDA texture object. | -| typedef | `CUuuid_stcudaUUID_t` | | CUDA UUID types. | -| define | `CUDA_IPC_HANDLE_SIZE` | | CUDA IPC Handle Size. | -| define | `cudaArrayCubemap` | | Must be set in cudaMalloc3DArray to create a cubemap CUDA array. | -| define | `cudaArrayDefault` | | Default CUDA array allocation flag. | -| define | `cudaArrayLayered` | | Must be set in cudaMalloc3DArray to create a layered CUDA array. | -| define | `cudaArraySurfaceLoadStore` | | Must be set in cudaMallocArray or cudaMalloc3DArray in order to bind surfaces to the CUDA array. | -| define | `cudaArrayTextureGather` | | Must be set in cudaMallocArray or cudaMalloc3DArray in order to perform texture gather operations on the CUDA array. | -| define | `cudaDeviceBlockingSync` | `hipDeviceScheduleBlockingSync` | Device flag - Use blocking synchronization. Deprecated as of CUDA 4.0 and replaced with cudaDeviceScheduleBlockingSync. | -| define | `cudaDeviceLmemResizeToMax` | | Device flag - Keep local memory allocation after launch. | -| define | `cudaDeviceMapHost` | | Device flag - Support mapped pinned allocations. | -| define | `cudaDeviceMask` | | Device flags mask. | -| define | `cudaDevicePropDontCare` | | Empty device properties. | -| define | `cudaDeviceScheduleAuto` | `hipDeviceScheduleAuto` | Device flag - Automatic scheduling. | -| define | `cudaDeviceScheduleBlockingSync` | `hipDeviceScheduleBlockingSync` | Device flag - Use blocking synchronization. | -| define | `cudaDeviceScheduleMask` | `hipDeviceScheduleMask` | Device schedule flags mask. | -| define | `cudaDeviceScheduleSpin` | `hipDeviceScheduleSpin` | Device flag - Spin default scheduling. | -| define | `cudaDeviceScheduleYield` | `hipDeviceScheduleYield` | Device flag - Yield default scheduling. | -| define | `cudaEventDefault` | `hipEventDefault` | Default event flag. | -| define | `cudaEventDisableTiming` | `hipEventDisableTiming` | Event will not record timing data. | -| define | `cudaEventInterprocess` | `hipEventInterprocess` | Event is suitable for interprocess use. cudaEventDisableTiming must be set. | -| define | `cudaHostAllocDefault` | `hipHostMallocDefault` | Default page-locked allocation flag. | -| define | `cudaHostAllocMapped` | `hipHostMallocMapped` | Map allocation into device space. | -| define | `cudaHostAllocPortable` | `hipHostMallocPortable` | Pinned memory accessible by all CUDA contexts. | -| define | `cudaHostAllocWriteCombined` | `hipHostMallocWriteCombined` | Write-combined memory. | -| define | `cudaHostRegisterDefault` | `hipHostRegisterDefault` | Default host memory registration flag. | -| define | `cudaHostRegisterIoMemory` | `hipHostRegisterIoMemory` | Memory-mapped I/O space. | -| define | `cudaHostRegisterMapped` | `hipHostRegisterMapped` | Map registered memory into device space. | -| define | `cudaHostRegisterPortable` | `hipHostRegisterPortable` | Pinned memory accessible by all CUDA contexts. | -| define | `cudaIpcMemLazyEnablePeerAccess` | `hipIpcMemLazyEnablePeerAccess` | Automatically enable peer access between remote devices as needed. | -| define | `cudaMemAttachGlobal` | | Memory can be accessed by any stream on any device. | -| define | `cudaMemAttachHost` | | Memory cannot be accessed by any stream on any device. | -| define | `cudaMemAttachSingle` | | Memory can only be accessed by a single stream on the associated device. | -| define | `cudaOccupancyDefault` | | Default behavior. | -| define | `cudaOccupancyDisableCachingOverride` | | Assume global caching is enabled and cannot be automatically turned off. | -| define | `cudaPeerAccessDefault` | | Default peer addressing enable flag. | -| define | `cudaStreamDefault` | `hipStreamDefault` | Default stream flag. | -| define | `cudaStreamLegacy` | | Default stream flag. | -| define | `cudaStreamNonBlocking` | `hipStreamNonBlocking` | Stream does not synchronize with stream 0 (the NULL stream). | -| define | `cudaStreamPerThread` | | Per-thread stream handle. | +| **type** | **CUDA** | **HIP** | +|-------------:|-----------------------------------------------|------------------------------------------------------| +| struct | `cudaChannelFormatDesc` | `hipChannelFormatDesc` | +| struct | `cudaDeviceProp` | `hipDeviceProp_t` | +| struct | `cudaExtent` | | +| struct | `cudaFuncAttributes` | | +| struct | `cudaIpcEventHandle_t` | `hipIpcEventHandle_t` | +| struct | `cudaIpcMemHandle_t` | `hipIpcMemHandle_t` | +| struct | `cudaMemcpy3DParms` | | +| struct | `cudaMemcpy3DPeerParms` | | +| struct | `cudaPitchedPtr` | | +| struct | `cudaPointerAttributes` | `hipPointerAttribute_t` | +| struct | `cudaPos` | | +| struct | `cudaResourceDesc` | | +| struct | `cudaResourceViewDesc` | | +| struct | `cudaTextureDesc` | | +| struct | `surfaceReference` | | +| struct | `textureReference` | `textureReference` | +| enum |***`cudaChannelFormatKind`*** |***`hipChannelFormatKind`*** | +| 0 |*`cudaChannelFormatKindSigned`* |*`hipChannelFormatKindSigned`* | +| 1 |*`cudaChannelFormatKindUnsigned`* |*`hipChannelFormatKindUnsigned`* | +| 2 |*`cudaChannelFormatKindFloat`* |*`hipChannelFormatKindFloat`* | +| 3 |*`cudaChannelFormatKindNone`* |*`hipChannelFormatKindNone`* | +| enum |***`cudaComputeMode`*** | | +| 0 |*`cudaComputeModeDefault`* | | +| 1 |*`cudaComputeModeExclusive`* | | +| 2 |*`cudaComputeModeProhibited`* | | +| 3 |*`cudaComputeModeExclusiveProcess`* | | +| enum |***`cudaDeviceAttr`*** |***`hipDeviceAttribute_t`*** | +| 1 |*`cudaDevAttrMaxThreadsPerBlock`* |*`hipDeviceAttributeMaxThreadsPerBlock`* | +| 2 |*`cudaDevAttrMaxBlockDimX`* |*`hipDeviceAttributeMaxBlockDimX`* | +| 3 |*`cudaDevAttrMaxBlockDimY`* |*`hipDeviceAttributeMaxBlockDimY`* | +| 4 |*`cudaDevAttrMaxBlockDimZ`* |*`hipDeviceAttributeMaxBlockDimZ`* | +| 5 |*`cudaDevAttrMaxGridDimX`* |*`hipDeviceAttributeMaxGridDimX`* | +| 6 |*`cudaDevAttrMaxGridDimY`* |*`hipDeviceAttributeMaxGridDimY`* | +| 7 |*`cudaDevAttrMaxGridDimZ`* |*`hipDeviceAttributeMaxGridDimZ`* | +| 8 |*`cudaDevAttrMaxSharedMemoryPerBlock`* |*`hipDeviceAttributeMaxSharedMemoryPerBlock`* | +| 9 |*`cudaDevAttrTotalConstantMemory`* |*`hipDeviceAttributeTotalConstantMemory`* | +| 10 |*`cudaDevAttrWarpSize`* |*`hipDeviceAttributeWarpSize`* | +| 11 |*`cudaDevAttrMaxPitch`* | | +| 12 |*`cudaDevAttrMaxRegistersPerBlock`* |*`hipDeviceAttributeMaxRegistersPerBlock`* | +| 13 |*`cudaDevAttrClockRate`* |*`hipDeviceAttributeClockRate`* | +| 14 |*`cudaDevAttrTextureAlignment`* | | +| 15 |*`cudaDevAttrGpuOverlap`* | | +| 16 |*`cudaDevAttrMultiProcessorCount`* |*`hipDeviceAttributeMultiprocessorCount`* | +| 17 |*`cudaDevAttrKernelExecTimeout`* | | +| 18 |*`cudaDevAttrIntegrated`* | | +| 19 |*`cudaDevAttrCanMapHostMemory`* | | +| 20 |*`cudaDevAttrComputeMode`* |*`hipDeviceAttributeComputeMode`* | +| 21 |*`cudaDevAttrMaxTexture1DWidth`* | | +| 22 |*`cudaDevAttrMaxTexture2DWidth`* | | +| 23 |*`cudaDevAttrMaxTexture2DHeight`* | | +| 24 |*`cudaDevAttrMaxTexture3DWidth`* | | +| 25 |*`cudaDevAttrMaxTexture3DHeight`* | | +| 26 |*`cudaDevAttrMaxTexture3DDepth`* | | +| 27 |*`cudaDevAttrMaxTexture2DLayeredWidth`* | | +| 28 |*`cudaDevAttrMaxTexture2DLayeredHeight`* | | +| 29 |*`cudaDevAttrMaxTexture2DLayeredLayers`* | | +| 30 |*`cudaDevAttrSurfaceAlignment`* | | +| 31 |*`cudaDevAttrConcurrentKernels`* |*`hipDeviceAttributeConcurrentKernels`* | +| 32 |*`cudaDevAttrEccEnabled`* | | +| 33 |*`cudaDevAttrPciBusId`* |*`hipDeviceAttributePciBusId`* | +| 34 |*`cudaDevAttrPciDeviceId`* |*`hipDeviceAttributePciDeviceId`* | +| 35 |*`cudaDevAttrTccDriver`* | | +| 36 |*`cudaDevAttrMemoryClockRate`* |*`hipDeviceAttributeMemoryClockRate`* | +| 37 |*`cudaDevAttrGlobalMemoryBusWidth`* |*`hipDeviceAttributeMemoryBusWidth`* | +| 38 |*`cudaDevAttrL2CacheSize`* |*`hipDeviceAttributeL2CacheSize`* | +| 39 |*`cudaDevAttrMaxThreadsPerMultiProcessor`* |*`hipDeviceAttributeMaxThreadsPerMultiProcessor`* | +| 40 |*`cudaDevAttrAsyncEngineCount`* | | +| 41 |*`cudaDevAttrUnifiedAddressing`* | | +| 42 |*`cudaDevAttrMaxTexture1DLayeredWidth`* | | +| 43 |*`cudaDevAttrMaxTexture1DLayeredLayers`* | | +| 44 | | | +| 45 |*`cudaDevAttrMaxTexture2DGatherWidth`* | | +| 46 |*`cudaDevAttrMaxTexture2DGatherHeight`* | | +| 47 |*`cudaDevAttrMaxTexture3DWidthAlt`* | | +| 48 |*`cudaDevAttrMaxTexture3DHeightAlt`* | | +| 49 |*`cudaDevAttrMaxTexture3DDepthAlt`* | | +| 50 |*`cudaDevAttrPciDomainId`* | | +| 51 |*`cudaDevAttrTexturePitchAlignment`* | | +| 52 |*`cudaDevAttrMaxTextureCubemapWidth`* | | +| 53 |*`cudaDevAttrMaxTextureCubemapLayeredWidth`* | | +| 54 |*`cudaDevAttrMaxTextureCubemapLayeredLayers`* | | +| 55 |*`cudaDevAttrMaxSurface1DWidth`* | | +| 56 |*`cudaDevAttrMaxSurface2DWidth`* | | +| 57 |*`cudaDevAttrMaxSurface2DHeight`* | | +| 58 |*`cudaDevAttrMaxSurface3DWidth`* | | +| 59 |*`cudaDevAttrMaxSurface3DHeight`* | | +| 60 |*`cudaDevAttrMaxSurface3DDepth`* | | +| 61 |*`cudaDevAttrMaxSurface1DLayeredWidth`* | | +| 62 |*`cudaDevAttrMaxSurface1DLayeredLayers`* | | +| 63 |*`cudaDevAttrMaxSurface2DLayeredWidth`* | | +| 64 |*`cudaDevAttrMaxSurface2DLayeredHeight`* | | +| 65 |*`cudaDevAttrMaxSurface2DLayeredLayers`* | | +| 66 |*`cudaDevAttrMaxSurfaceCubemapWidth`* | | +| 67 |*`cudaDevAttrMaxSurfaceCubemapLayeredWidth`* | | +| 68 |*`cudaDevAttrMaxSurfaceCubemapLayeredLayers`* | | +| 69 |*`cudaDevAttrMaxTexture1DLinearWidth`* | | +| 70 |*`cudaDevAttrMaxTexture2DLinearWidth`* | | +| 71 |*`cudaDevAttrMaxTexture2DLinearHeight`* | | +| 72 |*`cudaDevAttrMaxTexture2DLinearPitch`* | | +| 73 |*`cudaDevAttrMaxTexture2DMipmappedWidth`* | | +| 74 |*`cudaDevAttrMaxTexture2DMipmappedHeight`* | | +| 75 |*`cudaDevAttrComputeCapabilityMajor`* |*`hipDeviceAttributeComputeCapabilityMajor`* | +| 76 |*`cudaDevAttrComputeCapabilityMinor`* |*`hipDeviceAttributeComputeCapabilityMinor`* | +| 77 |*`cudaDevAttrMaxTexture1DMipmappedWidth`* | | +| 78 |*`cudaDevAttrStreamPrioritiesSupported`* | | +| 79 |*`cudaDevAttrGlobalL1CacheSupported`* | | +| 80 |*`cudaDevAttrLocalL1CacheSupported`* | | +| 81 |*`cudaDevAttrMaxSharedMemoryPerMultiprocessor`*|*`hipDeviceAttributeMaxSharedMemoryPerMultiprocessor`*| +| 82 |*`cudaDevAttrMaxRegistersPerMultiprocessor`* | | +| 83 |*`cudaDevAttrManagedMemory`* | | +| 84 |*`cudaDevAttrIsMultiGpuBoard`* |*`hipDeviceAttributeIsMultiGpuBoard`* | +| 85 |*`cudaDevAttrMultiGpuBoardGroupID`* | | +| enum |***`cudaError`*** |***`hipError_t`*** | +| enum |***`cudaError_t`*** |***`hipError_t`*** | +| 0 |*`cudaSuccess`* |*`hipSuccess`* | +| 1 |*`cudaErrorMissingConfiguration`* | | +| 2 |*`cudaErrorMemoryAllocation`* |*`hipErrorMemoryAllocation`* | +| 3 |*`cudaErrorInitializationError`* |*`hipErrorInitializationError`* | +| 4 |*`cudaErrorLaunchFailure`* | | +| 5 |*`cudaErrorPriorLaunchFailure`* | | +| 6 |*`cudaErrorLaunchTimeout`* | | +| 7 |*`cudaErrorLaunchOutOfResources`* |*`hipErrorLaunchOutOfResources`* | +| 8 |*`cudaErrorInvalidDeviceFunction`* | | +| 9 |*`cudaErrorInvalidConfiguration`* | | +| 10 |*`cudaErrorInvalidDevice`* |*`hipErrorInvalidDevice`* | +| 11 |*`cudaErrorInvalidValue`* |*`hipErrorInvalidValue`* | +| 12 |*`cudaErrorInvalidPitchValue`* | | +| 13 |*`cudaErrorInvalidSymbol`* | | +| 14 |*`cudaErrorMapBufferObjectFailed`* | | +| 15 |*`cudaErrorUnmapBufferObjectFailed`* | | +| 16 |*`cudaErrorInvalidHostPointer`* | | +| 17 |*`cudaErrorInvalidDevicePointer`* |*`hipErrorInvalidDevicePointer`* | +| 18 |*`cudaErrorInvalidTexture`* | | +| 19 |*`cudaErrorInvalidTextureBinding`* | | +| 20 |*`cudaErrorInvalidChannelDescriptor`* | | +| 21 |*`cudaErrorInvalidMemcpyDirection`* | | +| 22 |*`cudaErrorAddressOfConstant`* | | +| 23 |*`cudaErrorTextureFetchFailed`* | | +| 24 |*`cudaErrorTextureNotBound`* | | +| 25 |*`cudaErrorSynchronizationError`* | | +| 26 |*`cudaErrorInvalidFilterSetting`* | | +| 27 |*`cudaErrorInvalidNormSetting`* | | +| 28 |*`cudaErrorMixedDeviceExecution`* | | +| 29 |*`cudaErrorCudartUnloading`* | | +| 30 |*`cudaErrorUnknown`* |*`hipErrorUnknown`* | +| 31 |*`cudaErrorNotYetImplemented`* | | +| 32 |*`cudaErrorMemoryValueTooLarge`* | | +| 33 |*`cudaErrorInvalidResourceHandle`* |*`hipErrorInvalidResourceHandle`* | +| 34 |*`cudaErrorNotReady`* |*`hipErrorNotReady`* | +| 35 |*`cudaErrorInsufficientDriver`* | | +| 36 |*`cudaErrorSetOnActiveProcess`* | | +| 37 |*`cudaErrorInvalidSurface`* | | +| 38 |*`cudaErrorNoDevice`* |*`hipErrorNoDevice`* | +| 39 |*`cudaErrorECCUncorrectable`* | | +| 40 |*`cudaErrorSharedObjectSymbolNotFound`* | | +| 41 |*`cudaErrorSharedObjectInitFailed`* | | +| 42 |*`cudaErrorUnsupportedLimit`* |*`hipErrorUnsupportedLimit`* | +| 43 |*`cudaErrorDuplicateVariableName`* | | +| 44 |*`cudaErrorDuplicateTextureName`* | | +| 45 |*`cudaErrorDuplicateSurfaceName`* | | +| 46 |*`cudaErrorDevicesUnavailable`* | | +| 47 |*`cudaErrorInvalidKernelImage`* | | +| 48 |*`cudaErrorNoKernelImageForDevice`* | | +| 49 |*`cudaErrorIncompatibleDriverContext`* | | +| 50 |*`cudaErrorPeerAccessAlreadyEnabled`* |*`hipErrorPeerAccessAlreadyEnabled`* | +| 51 |*`cudaErrorPeerAccessNotEnabled`* |*`hipErrorPeerAccessNotEnabled`* | +| 52 | | | +| 53 | | | +| 54 |*`cudaErrorDeviceAlreadyInUse`* | | +| 55 |*`cudaErrorProfilerDisabled`* | | +| 56 |*`cudaErrorProfilerNotInitialized`* | | +| 57 |*`cudaErrorProfilerAlreadyStarted`* | | +| 58 |*`cudaErrorProfilerAlreadyStopped`* | | +| 59 |*`cudaErrorAssert`* | | +| 60 |*`cudaErrorTooManyPeers`* | | +| 61 |*`cudaErrorHostMemoryAlreadyRegistered`* | *`hipErrorHostMemoryAlreadyRegistered`* | +| 62 |*`cudaErrorHostMemoryNotRegistered`* | *`hipErrorHostMemoryNotRegistered`* | +| 63 |*`cudaErrorOperatingSystem`* | | +| 64 |*`cudaErrorPeerAccessUnsupported`* | | +| 65 |*`cudaErrorLaunchMaxDepthExceeded`* | | +| 66 |*`cudaErrorLaunchFileScopedTex`* | | +| 67 |*`cudaErrorLaunchFileScopedSurf`* | | +| 68 |*`cudaErrorSyncDepthExceeded`* | | +| 69 |*`cudaErrorLaunchPendingCountExceeded`* | | +| 70 |*`cudaErrorNotPermitted`* | | +| 71 |*`cudaErrorNotSupported`* | | +| 72 |*`cudaErrorHardwareStackError`* | | +| 73 |*`cudaErrorIllegalInstruction`* | | +| 74 |*`cudaErrorMisalignedAddress`* | | +| 75 |*`cudaErrorInvalidAddressSpace`* | | +| 76 |*`cudaErrorInvalidPc`* | | +| 77 |*`cudaErrorIllegalAddress`* | | +| 78 |*`cudaErrorInvalidPtx`* | | +| 79 |*`cudaErrorInvalidGraphicsContext`* | | +| 0x7f |*`cudaErrorStartupFailure`* | | +| 1000 |*`cudaErrorApiFailureBase`* | | +| enum |***`cudaFuncCache`*** |***`hipFuncCache_t`*** | +| 0 |*`cudaFuncCachePreferNone`* |*`hipFuncCachePreferNone`* | +| 1 |*`cudaFuncCachePreferShared`* |*`hipFuncCachePreferShared`* | +| 2 |*`cudaFuncCachePreferL1`* |*`hipFuncCachePreferL1`* | +| 3 |*`cudaFuncCachePreferEqual`* |*`hipFuncCachePreferEqual`* | +| enum |***`cudaGraphicsCubeFace`*** | | +| 0x00 |*`cudaGraphicsCubeFacePositiveX`* | | +| 0x01 |*`cudaGraphicsCubeFaceNegativeX`* | | +| 0x02 |*`cudaGraphicsCubeFacePositiveY`* | | +| 0x03 |*`cudaGraphicsCubeFaceNegativeY`* | | +| 0x04 |*`cudaGraphicsCubeFacePositiveZ`* | | +| 0x05 |*`cudaGraphicsCubeFaceNegativeZ`* | | +| enum |***`cudaGraphicsMapFlags`*** | | +| 0 |*`cudaGraphicsMapFlagsNone`* | | +| 1 |*`cudaGraphicsMapFlagsReadOnly`* | | +| 2 |*`cudaGraphicsMapFlagsWriteDiscard`* | | +| enum |***`cudaGraphicsRegisterFlags`*** | | +| 0 |*`cudaGraphicsRegisterFlagsNone`* | | +| 1 |*`cudaGraphicsRegisterFlagsReadOnly`* | | +| 2 |*`cudaGraphicsRegisterFlagsWriteDiscard`* | | +| 4 |*`cudaGraphicsRegisterFlagsSurfaceLoadStore`* | | +| 8 |*`cudaGraphicsRegisterFlagsTextureGather`* | | +| enum |***`cudaLimit`*** |***`hipLimit_t`*** | +| 0x00 |*`cudaLimitStackSize`* | | +| 0x01 |*`cudaLimitPrintfFifoSize`* | | +| 0x02 |*`cudaLimitMallocHeapSize`* |*`hipLimitMallocHeapSize`* | +| 0x03 |*`cudaLimitDevRuntimeSyncDepth`* | | +| 0x04 |*`cudaLimitDevRuntimePendingLaunchCount`* | | +| enum |***`cudaMemcpyKind`*** |***`hipMemcpyKind`*** | +| 0 |*`cudaMemcpyHostToHost`* |*`hipMemcpyHostToHost`* | +| 1 |*`cudaMemcpyHostToDevice`* |*`hipMemcpyHostToDevice`* | +| 2 |*`cudaMemcpyDeviceToHost`* |*`hipMemcpyDeviceToHost`* | +| 3 |*`cudaMemcpyDeviceToDevice`* |*`hipMemcpyDeviceToDevice`* | +| 4 |*`cudaMemcpyDefault`* |*`hipMemcpyDefault`* | +| enum |***`cudaMemoryType`*** |***`hipMemoryType`*** | +| 1 |*`cudaMemoryTypeHost`* |*`hipMemoryTypeHost`* | +| 2 |*`cudaMemoryTypeDevice`* |*`hipMemoryTypeDevice`* | +| enum |***`cudaResourceType`*** | | +| 0 |*`cudaResourceTypeArray`* | | +| 1 |*`cudaResourceTypeMipmappedArray`* | | +| 2 |*`cudaResourceTypeLinear`* | | +| 3 |*`cudaResourceTypePitch2D`* | | +| enum |***`cudaResourceViewFormat`*** | | +| 0x00 |*`cudaResViewFormatNone`* | | +| 0x01 |*`cudaResViewFormatUnsignedChar1`* | | +| 0x02 |*`cudaResViewFormatUnsignedChar2`* | | +| 0x03 |*`cudaResViewFormatUnsignedChar4`* | | +| 0x04 |*`cudaResViewFormatSignedChar1`* | | +| 0x05 |*`cudaResViewFormatSignedChar2`* | | +| 0x06 |*`cudaResViewFormatSignedChar4`* | | +| 0x07 |*`cudaResViewFormatUnsignedShort1`* | | +| 0x08 |*`cudaResViewFormatUnsignedShort2`* | | +| 0x09 |*`cudaResViewFormatUnsignedShort4`* | | +| 0x0a |*`cudaResViewFormatSignedShort1`* | | +| 0x0b |*`cudaResViewFormatSignedShort2`* | | +| 0x0c |*`cudaResViewFormatSignedShort4`* | | +| 0x0d |*`cudaResViewFormatUnsignedInt1`* | | +| 0x0e |*`cudaResViewFormatUnsignedInt2`* | | +| 0x0f |*`cudaResViewFormatUnsignedInt4`* | | +| 0x10 |*`cudaResViewFormatSignedInt1`* | | +| 0x11 |*`cudaResViewFormatSignedInt2`* | | +| 0x12 |*`cudaResViewFormatSignedInt4`* | | +| 0x13 |*`cudaResViewFormatHalf1`* | | +| 0x14 |*`cudaResViewFormatHalf2`* | | +| 0x15 |*`cudaResViewFormatHalf4`* | | +| 0x16 |*`cudaResViewFormatFloat1`* | | +| 0x17 |*`cudaResViewFormatFloat2`* | | +| 0x18 |*`cudaResViewFormatFloat4`* | | +| 0x19 |*`cudaResViewFormatUnsignedBlockCompressed1`* | | +| 0x1a |*`cudaResViewFormatUnsignedBlockCompressed2`* | | +| 0x1b |*`cudaResViewFormatUnsignedBlockCompressed3`* | | +| 0x1c |*`cudaResViewFormatUnsignedBlockCompressed4`* | | +| 0x1d |*`cudaResViewFormatSignedBlockCompressed4`* | | +| 0x1e |*`cudaResViewFormatUnsignedBlockCompressed5`* | | +| 0x1f |*`cudaResViewFormatSignedBlockCompressed5`* | | +| 0x20 |*`cudaResViewFormatUnsignedBlockCompressed6H`* | | +| 0x21 |*`cudaResViewFormatSignedBlockCompressed6H`* | | +| 0x22 |*`cudaResViewFormatUnsignedBlockCompressed7`* | | +| enum |***`cudaSharedMemConfig`*** |***`hipSharedMemConfig`*** | +| 0 |*`cudaSharedMemBankSizeDefault`* |*`hipSharedMemBankSizeDefault`* | +| 1 |*`cudaSharedMemBankSizeFourByte`* |*`hipSharedMemBankSizeFourByte`* | +| 2 |*`cudaSharedMemBankSizeEightByte`* |*`hipSharedMemBankSizeEightByte`* | +| enum |***`cudaSurfaceBoundaryMode`*** | | +| 0 |*`cudaBoundaryModeZero`* | | +| 1 |*`cudaBoundaryModeClamp`* | | +| 2 |*`cudaBoundaryModeTrap`* | | +| enum |***`cudaSurfaceFormatMode`*** | | +| 0 |*`cudaFormatModeForced`* | | +| 1 |*`cudaFormatModeAuto`* | | +| enum |***`cudaTextureAddressMode`*** | | +| 0 |*`cudaAddressModeWrap`* | | +| 1 |*`cudaAddressModeClamp`* | | +| 2 |*`cudaAddressModeMirror`* | | +| 3 |*`cudaAddressModeBorder`* | | +| enum |***`cudaTextureFilterMode`*** |***`hipTextureFilterMode`*** | +| 0 |*`cudaFilterModePoint`* |*`hipFilterModePoint`* | +| 1 |*`cudaFilterModeLinear`* | | +| enum |***`cudaTextureReadMode`*** |***`hipTextureReadMode`*** | +| 0 |*`cudaReadModeElementType`* |*`hipReadModeElementType`* | +| 1 |*`cudaReadModeNormalizedFloat`* | | +| enum |***`cudaGLDeviceList`*** | | +| 0x01 |*`cudaGLDeviceListAll`* | | +| 0x02 |*`cudaGLDeviceListCurrentFrame`* | | +| 0x03 |*`cudaGLDeviceListNextFrame`* | | +| enum |***`cudaGLMapFlags`*** | | +| 0x00 |*`cudaGLMapFlagsNone`* | | +| 0x01 |*`cudaGLMapFlagsReadOnly`* | | +| 0x02 |*`cudaGLMapFlagsWriteDiscard`* | | +| enum |***`cudaD3D9DeviceList`*** | | +| 1 |*`cudaD3D9DeviceListAll`* | | +| 2 |*`cudaD3D9DeviceListCurrentFrame`* | | +| 3 |*`cudaD3D9DeviceListNextFrame`* | | +| enum |***`cudaD3D9MapFlags`*** | | +| 0 |*`cudaD3D9MapFlagsNone`* | | +| 1 |*`cudaD3D9MapFlagsReadOnly`* | | +| 2 |*`cudaD3D9MapFlagsWriteDiscard`* | | +| enum |***`cudaD3D9RegisterFlags`*** | | +| 0 |*`cudaD3D9RegisterFlagsNone`* | | +| 1 |*`cudaD3D9RegisterFlagsArray`* | | +| enum |***`cudaD3D10DeviceList`*** | | +| 1 |*`cudaD3D10DeviceListAll`* | | +| 2 |*`cudaD3D10DeviceListCurrentFrame`* | | +| 3 |*`cudaD3D10DeviceListNextFrame`* | | +| enum |***`cudaD3D10MapFlags`*** | | +| 0 |*`cudaD3D10MapFlagsNone`* | | +| 1 |*`cudaD3D10MapFlagsReadOnly`* | | +| 2 |*`cudaD3D10MapFlagsWriteDiscard`* | | +| enum |***`cudaD3D10RegisterFlags`*** | | +| 0 |*`cudaD3D10RegisterFlagsNone`* | | +| 1 |*`cudaD3D10RegisterFlagsArray`* | | +| enum |***`cudaD3D11DeviceList`*** | | +| 1 |*`cudaD3D11DeviceListAll`* | | +| 2 |*`cudaD3D11DeviceListCurrentFrame`* | | +| 3 |*`cudaD3D11DeviceListNextFrame`* | | +| struct | `cudaArray` | `hipArray` | +| typedef | `cudaArray_t` | `hipArray *` | +| typedef | `cudaArray_const_t` | `const hipArray *` | +| enum | `cudaError` | `hipError_t` | +| typedef | `cudaError_t` | `hipError_t` | +| typedef | `cudaEvent_t` | `hipEvent_t` | +| typedef | `cudaGraphicsResource_t` | | +| typedef | `cudaMipmappedArray_t` | | +| typedef | `cudaMipmappedArray_const_t` | | +| enum |***`cudaOutputMode`*** | | +| 0x00 |*`cudaKeyValuePair`* | | +| 0x01 |*`cudaCSV`* | | +| typedef | `cudaOutputMode_t` | | +| typedef | `cudaStream_t` | `hipStream_t` | +| typedef | `cudaSurfaceObject_t` | | +| typedef | `cudaTextureObject_t` | | +| typedef | `CUuuid_stcudaUUID_t` | | +| define | `CUDA_IPC_HANDLE_SIZE` | | +| define | `cudaArrayCubemap` | | +| define | `cudaArrayDefault` | | +| define | `cudaArrayLayered` | | +| define | `cudaArraySurfaceLoadStore` | | +| define | `cudaArrayTextureGather` | | +| define | `cudaDeviceBlockingSync` | `hipDeviceScheduleBlockingSync` | +| define | `cudaDeviceLmemResizeToMax` | | +| define | `cudaDeviceMapHost` | | +| define | `cudaDeviceMask` | | +| define | `cudaDevicePropDontCare` | | +| define | `cudaDeviceScheduleAuto` | `hipDeviceScheduleAuto` | +| define | `cudaDeviceScheduleBlockingSync` | `hipDeviceScheduleBlockingSync` | +| define | `cudaDeviceScheduleMask` | `hipDeviceScheduleMask` | +| define | `cudaDeviceScheduleSpin` | `hipDeviceScheduleSpin` | +| define | `cudaDeviceScheduleYield` | `hipDeviceScheduleYield` | +| define | `cudaEventDefault` | `hipEventDefault` | +| define | `cudaEventDisableTiming` | `hipEventDisableTiming` | +| define | `cudaEventInterprocess` | `hipEventInterprocess` | +| define | `cudaHostAllocDefault` | `hipHostMallocDefault` | +| define | `cudaHostAllocMapped` | `hipHostMallocMapped` | +| define | `cudaHostAllocPortable` | `hipHostMallocPortable` | +| define | `cudaHostAllocWriteCombined` | `hipHostMallocWriteCombined` | +| define | `cudaHostRegisterDefault` | `hipHostRegisterDefault` | +| define | `cudaHostRegisterIoMemory` | `hipHostRegisterIoMemory` | +| define | `cudaHostRegisterMapped` | `hipHostRegisterMapped` | +| define | `cudaHostRegisterPortable` | `hipHostRegisterPortable` | +| define | `cudaIpcMemLazyEnablePeerAccess` | `hipIpcMemLazyEnablePeerAccess` | +| define | `cudaMemAttachGlobal` | | +| define | `cudaMemAttachHost` | | +| define | `cudaMemAttachSingle` | | +| define | `cudaOccupancyDefault` | | +| define | `cudaOccupancyDisableCachingOverride` | | +| define | `cudaPeerAccessDefault` | | +| define | `cudaStreamDefault` | `hipStreamDefault` | +| define | `cudaStreamLegacy` | | +| define | `cudaStreamNonBlocking` | `hipStreamNonBlocking` | +| define | `cudaStreamPerThread` | | diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index 9b58173899..138c3434a6 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -61,8 +61,9 @@ using namespace llvm; #define HIP_UNSUPPORTED -1 enum ConvTypes { - CONV_DRIVER = 0, - CONV_DEV, + CONV_VERSION = 0, + CONV_INIT, + CONV_DEVICE, CONV_MEM, CONV_KERN, CONV_COORD_FUNC, @@ -72,16 +73,23 @@ enum ConvTypes { CONV_EVENT, CONV_OCCUPANCY, CONV_CONTEXT, + CONV_PEER, CONV_MODULE, CONV_CACHE, CONV_EXEC, - CONV_ERR, + CONV_ERROR, CONV_DEF, CONV_TEX, CONV_GL, CONV_GRAPHICS, CONV_SURFACE, CONV_JIT, + CONV_D3D9, + CONV_D3D10, + CONV_D3D11, + CONV_VDPAU, + CONV_EGL, + CONV_THREAD, CONV_OTHER, CONV_INCLUDE, CONV_INCLUDE_CUDA_MAIN_H, @@ -92,11 +100,12 @@ enum ConvTypes { }; const char *counterNames[CONV_LAST] = { - "driver", "dev", "mem", "kern", "coord_func", "math_func", - "special_func", "stream", "event", "occupancy", "ctx", "module", - "cache", "exec", "err", "def", "tex", "gl", - "graphics", "surface", "jit", "other", "include", "include_cuda_main_header", - "type", "literal", "numeric_literal"}; + "version", "init", "device", "mem", "kern", "coord_func", "math_func", + "special_func", "stream", "event", "occupancy", "ctx", "peer", "module", + "cache", "exec", "err", "def", "tex", "gl", "graphics", + "surface", "jit", "d3d9", "d3d10", "d3d11", "vdpau", "egl", + "thread", "other", "include", "include_cuda_main_header", "type", "literal", + "numeric_literal"}; enum ApiTypes { API_DRIVER = 0, @@ -196,199 +205,199 @@ struct cuda2hipMap { cuda2hipRename["cudaError"] = {"hipError_t", CONV_TYPE, API_RUNTIME}; // CUDA Driver API error codes only - cuda2hipRename["CUDA_ERROR_INVALID_CONTEXT"] = {"hipErrorInvalidContext", CONV_ERR, API_DRIVER}; // 201 - cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_CURRENT"] = {"hipErrorContextAlreadyCurrent", CONV_ERR, API_DRIVER}; // 202 - cuda2hipRename["CUDA_ERROR_ARRAY_IS_MAPPED"] = {"hipErrorArrayIsMapped", CONV_ERR, API_DRIVER}; // 207 - cuda2hipRename["CUDA_ERROR_ALREADY_MAPPED"] = {"hipErrorAlreadyMapped", CONV_ERR, API_DRIVER}; // 208 - cuda2hipRename["CUDA_ERROR_ALREADY_ACQUIRED"] = {"hipErrorAlreadyAcquired", CONV_ERR, API_DRIVER}; // 210 - cuda2hipRename["CUDA_ERROR_NOT_MAPPED"] = {"hipErrorNotMapped", CONV_ERR, API_DRIVER}; // 211 - cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_ARRAY"] = {"hipErrorNotMappedAsArray", CONV_ERR, API_DRIVER}; // 212 - cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_POINTER"] = {"hipErrorNotMappedAsPointer", CONV_ERR, API_DRIVER}; // 213 - cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_IN_USE"] = {"hipErrorContextAlreadyInUse", CONV_ERR, API_DRIVER}; // 216 - cuda2hipRename["CUDA_ERROR_INVALID_SOURCE"] = {"hipErrorInvalidSource", CONV_ERR, API_DRIVER}; // 300 - cuda2hipRename["CUDA_ERROR_FILE_NOT_FOUND"] = {"hipErrorFileNotFound", CONV_ERR, API_DRIVER}; // 301 - cuda2hipRename["CUDA_ERROR_NOT_FOUND"] = {"hipErrorNotFound", CONV_ERR, API_DRIVER}; // 500 - cuda2hipRename["CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING"] = {"hipErrorLaunchIncompatibleTexturing", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 703 - cuda2hipRename["CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE"] = {"hipErrorPrimaryContextActive", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 708 - cuda2hipRename["CUDA_ERROR_CONTEXT_IS_DESTROYED"] = {"hipErrorContextIsDestroyed", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 709 - cuda2hipRename["CUDA_ERROR_NOT_PERMITTED"] = {"hipErrorNotPermitted", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 800 - cuda2hipRename["CUDA_ERROR_NOT_SUPPORTED"] = {"hipErrorNotSupported", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 801 + cuda2hipRename["CUDA_ERROR_INVALID_CONTEXT"] = {"hipErrorInvalidContext", CONV_TYPE, API_DRIVER}; // 201 + cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_CURRENT"] = {"hipErrorContextAlreadyCurrent", CONV_TYPE, API_DRIVER}; // 202 + cuda2hipRename["CUDA_ERROR_ARRAY_IS_MAPPED"] = {"hipErrorArrayIsMapped", CONV_TYPE, API_DRIVER}; // 207 + cuda2hipRename["CUDA_ERROR_ALREADY_MAPPED"] = {"hipErrorAlreadyMapped", CONV_TYPE, API_DRIVER}; // 208 + cuda2hipRename["CUDA_ERROR_ALREADY_ACQUIRED"] = {"hipErrorAlreadyAcquired", CONV_TYPE, API_DRIVER}; // 210 + cuda2hipRename["CUDA_ERROR_NOT_MAPPED"] = {"hipErrorNotMapped", CONV_TYPE, API_DRIVER}; // 211 + cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_ARRAY"] = {"hipErrorNotMappedAsArray", CONV_TYPE, API_DRIVER}; // 212 + cuda2hipRename["CUDA_ERROR_NOT_MAPPED_AS_POINTER"] = {"hipErrorNotMappedAsPointer", CONV_TYPE, API_DRIVER}; // 213 + cuda2hipRename["CUDA_ERROR_CONTEXT_ALREADY_IN_USE"] = {"hipErrorContextAlreadyInUse", CONV_TYPE, API_DRIVER}; // 216 + cuda2hipRename["CUDA_ERROR_INVALID_SOURCE"] = {"hipErrorInvalidSource", CONV_TYPE, API_DRIVER}; // 300 + cuda2hipRename["CUDA_ERROR_FILE_NOT_FOUND"] = {"hipErrorFileNotFound", CONV_TYPE, API_DRIVER}; // 301 + cuda2hipRename["CUDA_ERROR_NOT_FOUND"] = {"hipErrorNotFound", CONV_TYPE, API_DRIVER}; // 500 + cuda2hipRename["CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING"] = {"hipErrorLaunchIncompatibleTexturing", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 703 + cuda2hipRename["CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE"] = {"hipErrorPrimaryContextActive", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 708 + cuda2hipRename["CUDA_ERROR_CONTEXT_IS_DESTROYED"] = {"hipErrorContextIsDestroyed", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 709 + cuda2hipRename["CUDA_ERROR_NOT_PERMITTED"] = {"hipErrorNotPermitted", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 800 + cuda2hipRename["CUDA_ERROR_NOT_SUPPORTED"] = {"hipErrorNotSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 801 // CUDA RT API error code only - cuda2hipRename["cudaErrorMissingConfiguration"] = {"hipErrorMissingConfiguration", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 1 - cuda2hipRename["cudaErrorPriorLaunchFailure"] = {"hipErrorPriorLaunchFailure", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 5 - cuda2hipRename["cudaErrorInvalidDeviceFunction"] = {"hipErrorInvalidDeviceFunction", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 8 - cuda2hipRename["cudaErrorInvalidConfiguration"] = {"hipErrorInvalidConfiguration", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 9 - cuda2hipRename["cudaErrorInvalidPitchValue"] = {"hipErrorInvalidPitchValue", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 12 - cuda2hipRename["cudaErrorInvalidSymbol"] = {"hipErrorInvalidSymbol", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 13 - cuda2hipRename["cudaErrorInvalidHostPointer"] = {"hipErrorInvalidHostPointer", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 16 - cuda2hipRename["cudaErrorInvalidDevicePointer"] = {"hipErrorInvalidDevicePointer", CONV_ERR, API_RUNTIME}; // 17 - cuda2hipRename["cudaErrorInvalidTexture"] = {"hipErrorInvalidTexture", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 18 - cuda2hipRename["cudaErrorInvalidTextureBinding"] = {"hipErrorInvalidTextureBinding", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 19 - cuda2hipRename["cudaErrorInvalidChannelDescriptor"] = {"hipErrorInvalidChannelDescriptor", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 20 - cuda2hipRename["cudaErrorInvalidMemcpyDirection"] = {"hipErrorInvalidMemcpyDirection", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 21 - cuda2hipRename["cudaErrorAddressOfConstant"] = {"hipErrorAddressOfConstant", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 22 - cuda2hipRename["cudaErrorTextureFetchFailed"] = {"hipErrorTextureFetchFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 23 - cuda2hipRename["cudaErrorTextureNotBound"] = {"hipErrorTextureNotBound", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 24 - cuda2hipRename["cudaErrorSynchronizationError"] = {"hipErrorSynchronizationError", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 25 - cuda2hipRename["cudaErrorInvalidFilterSetting"] = {"hipErrorInvalidFilterSetting", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 26 - cuda2hipRename["cudaErrorInvalidNormSetting"] = {"hipErrorInvalidNormSetting", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 27 - cuda2hipRename["cudaErrorMixedDeviceExecution"] = {"hipErrorMixedDeviceExecution", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 28 + cuda2hipRename["cudaErrorMissingConfiguration"] = {"hipErrorMissingConfiguration", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 1 + cuda2hipRename["cudaErrorPriorLaunchFailure"] = {"hipErrorPriorLaunchFailure", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 5 + cuda2hipRename["cudaErrorInvalidDeviceFunction"] = {"hipErrorInvalidDeviceFunction", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 8 + cuda2hipRename["cudaErrorInvalidConfiguration"] = {"hipErrorInvalidConfiguration", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 9 + cuda2hipRename["cudaErrorInvalidPitchValue"] = {"hipErrorInvalidPitchValue", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 12 + cuda2hipRename["cudaErrorInvalidSymbol"] = {"hipErrorInvalidSymbol", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 13 + cuda2hipRename["cudaErrorInvalidHostPointer"] = {"hipErrorInvalidHostPointer", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 16 + cuda2hipRename["cudaErrorInvalidDevicePointer"] = {"hipErrorInvalidDevicePointer", CONV_TYPE, API_RUNTIME}; // 17 + cuda2hipRename["cudaErrorInvalidTexture"] = {"hipErrorInvalidTexture", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 18 + cuda2hipRename["cudaErrorInvalidTextureBinding"] = {"hipErrorInvalidTextureBinding", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 19 + cuda2hipRename["cudaErrorInvalidChannelDescriptor"] = {"hipErrorInvalidChannelDescriptor", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 20 + cuda2hipRename["cudaErrorInvalidMemcpyDirection"] = {"hipErrorInvalidMemcpyDirection", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 21 + cuda2hipRename["cudaErrorAddressOfConstant"] = {"hipErrorAddressOfConstant", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 22 + cuda2hipRename["cudaErrorTextureFetchFailed"] = {"hipErrorTextureFetchFailed", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 23 + cuda2hipRename["cudaErrorTextureNotBound"] = {"hipErrorTextureNotBound", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 24 + cuda2hipRename["cudaErrorSynchronizationError"] = {"hipErrorSynchronizationError", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 25 + cuda2hipRename["cudaErrorInvalidFilterSetting"] = {"hipErrorInvalidFilterSetting", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 26 + cuda2hipRename["cudaErrorInvalidNormSetting"] = {"hipErrorInvalidNormSetting", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 27 + cuda2hipRename["cudaErrorMixedDeviceExecution"] = {"hipErrorMixedDeviceExecution", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 28 // Deprecated as of CUDA 4.1 - cuda2hipRename["cudaErrorNotYetImplemented"] = {"hipErrorNotYetImplemented", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 31 + cuda2hipRename["cudaErrorNotYetImplemented"] = {"hipErrorNotYetImplemented", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 31 // Deprecated as of CUDA 3.1 - cuda2hipRename["cudaErrorMemoryValueTooLarge"] = {"hipErrorMemoryValueTooLarge", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 32 - cuda2hipRename["cudaErrorInsufficientDriver"] = {"hipErrorInsufficientDriver", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 35 - cuda2hipRename["cudaErrorSetOnActiveProcess"] = {"hipErrorSetOnActiveProcess", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 36 - cuda2hipRename["cudaErrorInvalidSurface"] = {"hipErrorInvalidSurface", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 37 - cuda2hipRename["cudaErrorDuplicateVariableName"] = {"hipErrorDuplicateVariableName", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 43 - cuda2hipRename["cudaErrorDuplicateTextureName"] = {"hipErrorDuplicateTextureName", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 44 - cuda2hipRename["cudaErrorDuplicateSurfaceName"] = {"hipErrorDuplicateSurfaceName", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 45 - cuda2hipRename["cudaErrorDevicesUnavailable"] = {"hipErrorDevicesUnavailable", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 46 - cuda2hipRename["cudaErrorIncompatibleDriverContext"] = {"hipErrorIncompatibleDriverContext", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 49 - cuda2hipRename["cudaErrorDeviceAlreadyInUse"] = {"hipErrorDeviceAlreadyInUse", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 54 - cuda2hipRename["cudaErrorLaunchMaxDepthExceeded"] = {"hipErrorLaunchMaxDepthExceeded", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 65 - cuda2hipRename["cudaErrorLaunchFileScopedTex"] = {"hipErrorLaunchFileScopedTex", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 66 - cuda2hipRename["cudaErrorLaunchFileScopedSurf"] = {"hipErrorLaunchFileScopedSurf", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 67 - cuda2hipRename["cudaErrorSyncDepthExceeded"] = {"hipErrorSyncDepthExceeded", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 68 - cuda2hipRename["cudaErrorLaunchPendingCountExceeded"] = {"hipErrorLaunchPendingCountExceeded", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 69 - cuda2hipRename["cudaErrorNotPermitted"] = {"hipErrorNotPermitted", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 70 - cuda2hipRename["cudaErrorNotSupported"] = {"hipErrorNotSupported", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 71 - cuda2hipRename["cudaErrorStartupFailure"] = {"hipErrorStartupFailure", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 0x7f + cuda2hipRename["cudaErrorMemoryValueTooLarge"] = {"hipErrorMemoryValueTooLarge", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 32 + cuda2hipRename["cudaErrorInsufficientDriver"] = {"hipErrorInsufficientDriver", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 35 + cuda2hipRename["cudaErrorSetOnActiveProcess"] = {"hipErrorSetOnActiveProcess", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 36 + cuda2hipRename["cudaErrorInvalidSurface"] = {"hipErrorInvalidSurface", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 37 + cuda2hipRename["cudaErrorDuplicateVariableName"] = {"hipErrorDuplicateVariableName", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 43 + cuda2hipRename["cudaErrorDuplicateTextureName"] = {"hipErrorDuplicateTextureName", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 44 + cuda2hipRename["cudaErrorDuplicateSurfaceName"] = {"hipErrorDuplicateSurfaceName", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 45 + cuda2hipRename["cudaErrorDevicesUnavailable"] = {"hipErrorDevicesUnavailable", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 46 + cuda2hipRename["cudaErrorIncompatibleDriverContext"] = {"hipErrorIncompatibleDriverContext", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 49 + cuda2hipRename["cudaErrorDeviceAlreadyInUse"] = {"hipErrorDeviceAlreadyInUse", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 54 + cuda2hipRename["cudaErrorLaunchMaxDepthExceeded"] = {"hipErrorLaunchMaxDepthExceeded", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 65 + cuda2hipRename["cudaErrorLaunchFileScopedTex"] = {"hipErrorLaunchFileScopedTex", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 66 + cuda2hipRename["cudaErrorLaunchFileScopedSurf"] = {"hipErrorLaunchFileScopedSurf", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 67 + cuda2hipRename["cudaErrorSyncDepthExceeded"] = {"hipErrorSyncDepthExceeded", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 68 + cuda2hipRename["cudaErrorLaunchPendingCountExceeded"] = {"hipErrorLaunchPendingCountExceeded", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 69 + cuda2hipRename["cudaErrorNotPermitted"] = {"hipErrorNotPermitted", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 70 + cuda2hipRename["cudaErrorNotSupported"] = {"hipErrorNotSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 71 + cuda2hipRename["cudaErrorStartupFailure"] = {"hipErrorStartupFailure", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x7f // Deprecated as of CUDA 4.1 - cuda2hipRename["cudaErrorApiFailureBase"] = {"hipErrorApiFailureBase", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 10000 + cuda2hipRename["cudaErrorApiFailureBase"] = {"hipErrorApiFailureBase", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 10000 - cuda2hipRename["CUDA_SUCCESS"] = {"hipSuccess", CONV_ERR, API_DRIVER}; // 0 - cuda2hipRename["cudaSuccess"] = {"hipSuccess", CONV_ERR, API_RUNTIME}; // 0 + cuda2hipRename["CUDA_SUCCESS"] = {"hipSuccess", CONV_TYPE, API_DRIVER}; // 0 + cuda2hipRename["cudaSuccess"] = {"hipSuccess", CONV_TYPE, API_RUNTIME}; // 0 - cuda2hipRename["CUDA_ERROR_INVALID_VALUE"] = {"hipErrorInvalidValue", CONV_ERR, API_DRIVER}; // 1 - cuda2hipRename["cudaErrorInvalidValue"] = {"hipErrorInvalidValue", CONV_ERR, API_RUNTIME}; // 11 + cuda2hipRename["CUDA_ERROR_INVALID_VALUE"] = {"hipErrorInvalidValue", CONV_TYPE, API_DRIVER}; // 1 + cuda2hipRename["cudaErrorInvalidValue"] = {"hipErrorInvalidValue", CONV_TYPE, API_RUNTIME}; // 11 - cuda2hipRename["CUDA_ERROR_OUT_OF_MEMORY"] = {"hipErrorMemoryAllocation", CONV_ERR, API_DRIVER}; // 2 - cuda2hipRename["cudaErrorMemoryAllocation"] = {"hipErrorMemoryAllocation", CONV_ERR, API_RUNTIME}; // 2 + cuda2hipRename["CUDA_ERROR_OUT_OF_MEMORY"] = {"hipErrorMemoryAllocation", CONV_TYPE, API_DRIVER}; // 2 + cuda2hipRename["cudaErrorMemoryAllocation"] = {"hipErrorMemoryAllocation", CONV_TYPE, API_RUNTIME}; // 2 - cuda2hipRename["CUDA_ERROR_NOT_INITIALIZED"] = {"hipErrorNotInitialized", CONV_ERR, API_DRIVER}; // 3 - cuda2hipRename["cudaErrorInitializationError"] = {"hipErrorInitializationError", CONV_ERR, API_RUNTIME}; // 3 + cuda2hipRename["CUDA_ERROR_NOT_INITIALIZED"] = {"hipErrorNotInitialized", CONV_TYPE, API_DRIVER}; // 3 + cuda2hipRename["cudaErrorInitializationError"] = {"hipErrorInitializationError", CONV_TYPE, API_RUNTIME}; // 3 - cuda2hipRename["CUDA_ERROR_DEINITIALIZED"] = {"hipErrorDeinitialized", CONV_ERR, API_DRIVER}; // 4 + cuda2hipRename["CUDA_ERROR_DEINITIALIZED"] = {"hipErrorDeinitialized", CONV_TYPE, API_DRIVER}; // 4 // TODO: double check, that these errors match - cuda2hipRename["cudaErrorCudartUnloading"] = {"hipErrorDeinitialized", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 29 + cuda2hipRename["cudaErrorCudartUnloading"] = {"hipErrorDeinitialized", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 29 - cuda2hipRename["CUDA_ERROR_PROFILER_DISABLED"] = {"hipErrorProfilerDisabled", CONV_ERR, API_DRIVER}; // 5 - cuda2hipRename["cudaErrorProfilerDisabled"] = {"hipErrorProfilerDisabled", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 55 + cuda2hipRename["CUDA_ERROR_PROFILER_DISABLED"] = {"hipErrorProfilerDisabled", CONV_TYPE, API_DRIVER}; // 5 + cuda2hipRename["cudaErrorProfilerDisabled"] = {"hipErrorProfilerDisabled", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 55 - cuda2hipRename["CUDA_ERROR_PROFILER_NOT_INITIALIZED"] = {"hipErrorProfilerNotInitialized", CONV_ERR, API_DRIVER}; // 6 + cuda2hipRename["CUDA_ERROR_PROFILER_NOT_INITIALIZED"] = {"hipErrorProfilerNotInitialized", CONV_TYPE, API_DRIVER}; // 6 // Deprecated as of CUDA 5.0 - cuda2hipRename["cudaErrorProfilerNotInitialized"] = {"hipErrorProfilerNotInitialized", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 56 + cuda2hipRename["cudaErrorProfilerNotInitialized"] = {"hipErrorProfilerNotInitialized", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 56 - cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STARTED"] = {"hipErrorProfilerAlreadyStarted", CONV_ERR, API_DRIVER}; // 7 + cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STARTED"] = {"hipErrorProfilerAlreadyStarted", CONV_TYPE, API_DRIVER}; // 7 // Deprecated as of CUDA 5.0 - cuda2hipRename["cudaErrorProfilerAlreadyStarted"] = {"hipErrorProfilerAlreadyStarted", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 57 + cuda2hipRename["cudaErrorProfilerAlreadyStarted"] = {"hipErrorProfilerAlreadyStarted", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 57 - cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STOPPED"] = {"hipErrorProfilerAlreadyStopped", CONV_ERR, API_DRIVER}; // 8 + cuda2hipRename["CUDA_ERROR_PROFILER_ALREADY_STOPPED"] = {"hipErrorProfilerAlreadyStopped", CONV_TYPE, API_DRIVER}; // 8 // Deprecated as of CUDA 5.0 - cuda2hipRename["cudaErrorProfilerAlreadyStopped"] = {"hipErrorProfilerAlreadyStopped", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 58 + cuda2hipRename["cudaErrorProfilerAlreadyStopped"] = {"hipErrorProfilerAlreadyStopped", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 58 - cuda2hipRename["CUDA_ERROR_NO_DEVICE"] = {"hipErrorNoDevice", CONV_ERR, API_DRIVER}; // 100 - cuda2hipRename["cudaErrorNoDevice"] = {"hipErrorNoDevice", CONV_ERR, API_RUNTIME}; // 38 + cuda2hipRename["CUDA_ERROR_NO_DEVICE"] = {"hipErrorNoDevice", CONV_TYPE, API_DRIVER}; // 100 + cuda2hipRename["cudaErrorNoDevice"] = {"hipErrorNoDevice", CONV_TYPE, API_RUNTIME}; // 38 - cuda2hipRename["CUDA_ERROR_INVALID_DEVICE"] = {"hipErrorInvalidDevice", CONV_ERR, API_DRIVER}; // 101 - cuda2hipRename["cudaErrorInvalidDevice"] = {"hipErrorInvalidDevice", CONV_ERR, API_RUNTIME}; // 10 + cuda2hipRename["CUDA_ERROR_INVALID_DEVICE"] = {"hipErrorInvalidDevice", CONV_TYPE, API_DRIVER}; // 101 + cuda2hipRename["cudaErrorInvalidDevice"] = {"hipErrorInvalidDevice", CONV_TYPE, API_RUNTIME}; // 10 - cuda2hipRename["CUDA_ERROR_INVALID_IMAGE"] = {"hipErrorInvalidImage", CONV_ERR, API_DRIVER}; // 200 - cuda2hipRename["cudaErrorInvalidKernelImage"] = {"hipErrorInvalidImage", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 47 + cuda2hipRename["CUDA_ERROR_INVALID_IMAGE"] = {"hipErrorInvalidImage", CONV_TYPE, API_DRIVER}; // 200 + cuda2hipRename["cudaErrorInvalidKernelImage"] = {"hipErrorInvalidImage", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 47 - cuda2hipRename["CUDA_ERROR_MAP_FAILED"] = {"hipErrorMapFailed", CONV_ERR, API_DRIVER}; // 205 + cuda2hipRename["CUDA_ERROR_MAP_FAILED"] = {"hipErrorMapFailed", CONV_TYPE, API_DRIVER}; // 205 // TODO: double check, that these errors match - cuda2hipRename["cudaErrorMapBufferObjectFailed"] = {"hipErrorMapFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 14 + cuda2hipRename["cudaErrorMapBufferObjectFailed"] = {"hipErrorMapFailed", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 14 - cuda2hipRename["CUDA_ERROR_UNMAP_FAILED"] = {"hipErrorUnmapFailed", CONV_ERR, API_DRIVER}; // 206 + cuda2hipRename["CUDA_ERROR_UNMAP_FAILED"] = {"hipErrorUnmapFailed", CONV_TYPE, API_DRIVER}; // 206 // TODO: double check, that these errors match - cuda2hipRename["cudaErrorUnmapBufferObjectFailed"] = {"hipErrorUnmapFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 15 + cuda2hipRename["cudaErrorUnmapBufferObjectFailed"] = {"hipErrorUnmapFailed", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 15 - cuda2hipRename["CUDA_ERROR_NO_BINARY_FOR_GPU"] = {"hipErrorNoBinaryForGpu", CONV_ERR, API_DRIVER}; // 209 - cuda2hipRename["cudaErrorNoKernelImageForDevice"] = {"hipErrorNoBinaryForGpu", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 48 + cuda2hipRename["CUDA_ERROR_NO_BINARY_FOR_GPU"] = {"hipErrorNoBinaryForGpu", CONV_TYPE, API_DRIVER}; // 209 + cuda2hipRename["cudaErrorNoKernelImageForDevice"] = {"hipErrorNoBinaryForGpu", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 48 - cuda2hipRename["CUDA_ERROR_ECC_UNCORRECTABLE"] = {"hipErrorECCNotCorrectable", CONV_ERR, API_DRIVER}; // 214 - cuda2hipRename["cudaErrorECCUncorrectable"] = {"hipErrorECCNotCorrectable", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 39 + cuda2hipRename["CUDA_ERROR_ECC_UNCORRECTABLE"] = {"hipErrorECCNotCorrectable", CONV_TYPE, API_DRIVER}; // 214 + cuda2hipRename["cudaErrorECCUncorrectable"] = {"hipErrorECCNotCorrectable", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 39 - cuda2hipRename["CUDA_ERROR_UNSUPPORTED_LIMIT"] = {"hipErrorUnsupportedLimit", CONV_ERR, API_DRIVER}; // 215 - cuda2hipRename["cudaErrorUnsupportedLimit"] = {"hipErrorUnsupportedLimit", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 42 + cuda2hipRename["CUDA_ERROR_UNSUPPORTED_LIMIT"] = {"hipErrorUnsupportedLimit", CONV_TYPE, API_DRIVER}; // 215 + cuda2hipRename["cudaErrorUnsupportedLimit"] = {"hipErrorUnsupportedLimit", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 42 - cuda2hipRename["CUDA_ERROR_PEER_ACCESS_UNSUPPORTED"] = {"hipErrorPeerAccessUnsupported", CONV_ERR, API_DRIVER}; // 217 - cuda2hipRename["cudaErrorPeerAccessUnsupported"] = {"hipErrorPeerAccessUnsupported", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 64 + cuda2hipRename["CUDA_ERROR_PEER_ACCESS_UNSUPPORTED"] = {"hipErrorPeerAccessUnsupported", CONV_TYPE, API_DRIVER}; // 217 + cuda2hipRename["cudaErrorPeerAccessUnsupported"] = {"hipErrorPeerAccessUnsupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 64 - cuda2hipRename["CUDA_ERROR_INVALID_PTX"] = {"hipErrorInvalidKernelFile", CONV_ERR, API_DRIVER}; // 218 - cuda2hipRename["cudaErrorInvalidPtx"] = {"hipErrorInvalidKernelFile", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 78 + cuda2hipRename["CUDA_ERROR_INVALID_PTX"] = {"hipErrorInvalidKernelFile", CONV_TYPE, API_DRIVER}; // 218 + cuda2hipRename["cudaErrorInvalidPtx"] = {"hipErrorInvalidKernelFile", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 78 - cuda2hipRename["CUDA_ERROR_INVALID_GRAPHICS_CONTEXT"] = {"hipErrorInvalidGraphicsContext", CONV_ERR, API_DRIVER}; // 219 - cuda2hipRename["cudaErrorInvalidGraphicsContext"] = {"hipErrorInvalidGraphicsContext", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 79 + cuda2hipRename["CUDA_ERROR_INVALID_GRAPHICS_CONTEXT"] = {"hipErrorInvalidGraphicsContext", CONV_TYPE, API_DRIVER}; // 219 + cuda2hipRename["cudaErrorInvalidGraphicsContext"] = {"hipErrorInvalidGraphicsContext", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 79 - cuda2hipRename["CUDA_ERROR_NVLINK_UNCORRECTABLE"] = {"hipErrorNvlinkUncorrectable", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 220 [CUDA 8.0.44] - cuda2hipRename["cudaErrorNvlinkUncorrectable"] = {"hipErrorNvlinkUncorrectable", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 80 [CUDA 8.0.44] + cuda2hipRename["CUDA_ERROR_NVLINK_UNCORRECTABLE"] = {"hipErrorNvlinkUncorrectable", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 220 [CUDA 8.0.44] + cuda2hipRename["cudaErrorNvlinkUncorrectable"] = {"hipErrorNvlinkUncorrectable", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 80 [CUDA 8.0.44] - cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND"] = {"hipErrorSharedObjectSymbolNotFound", CONV_ERR, API_DRIVER}; // 302 - cuda2hipRename["cudaErrorSharedObjectSymbolNotFound"] = {"hipErrorSharedObjectSymbolNotFound", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 40 + cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND"] = {"hipErrorSharedObjectSymbolNotFound", CONV_TYPE, API_DRIVER}; // 302 + cuda2hipRename["cudaErrorSharedObjectSymbolNotFound"] = {"hipErrorSharedObjectSymbolNotFound", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 40 - cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_INIT_FAILED"] = {"hipErrorSharedObjectInitFailed", CONV_ERR, API_DRIVER}; // 303 - cuda2hipRename["cudaErrorSharedObjectInitFailed"] = {"hipErrorSharedObjectInitFailed", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 41 + cuda2hipRename["CUDA_ERROR_SHARED_OBJECT_INIT_FAILED"] = {"hipErrorSharedObjectInitFailed", CONV_TYPE, API_DRIVER}; // 303 + cuda2hipRename["cudaErrorSharedObjectInitFailed"] = {"hipErrorSharedObjectInitFailed", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 41 - cuda2hipRename["CUDA_ERROR_OPERATING_SYSTEM"] = {"hipErrorOperatingSystem", CONV_ERR, API_DRIVER}; // 304 - cuda2hipRename["cudaErrorOperatingSystem"] = {"hipErrorOperatingSystem", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 63 + cuda2hipRename["CUDA_ERROR_OPERATING_SYSTEM"] = {"hipErrorOperatingSystem", CONV_TYPE, API_DRIVER}; // 304 + cuda2hipRename["cudaErrorOperatingSystem"] = {"hipErrorOperatingSystem", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 63 - cuda2hipRename["CUDA_ERROR_INVALID_HANDLE"] = {"hipErrorInvalidResourceHandle", CONV_ERR, API_DRIVER}; // 400 - cuda2hipRename["cudaErrorInvalidResourceHandle"] = {"hipErrorInvalidResourceHandle", CONV_ERR, API_RUNTIME}; // 33 + cuda2hipRename["CUDA_ERROR_INVALID_HANDLE"] = {"hipErrorInvalidResourceHandle", CONV_TYPE, API_DRIVER}; // 400 + cuda2hipRename["cudaErrorInvalidResourceHandle"] = {"hipErrorInvalidResourceHandle", CONV_TYPE, API_RUNTIME}; // 33 - cuda2hipRename["CUDA_ERROR_NOT_READY"] = {"hipErrorNotReady", CONV_ERR, API_DRIVER}; // 600 - cuda2hipRename["cudaErrorNotReady"] = {"hipErrorNotReady", CONV_ERR, API_RUNTIME}; // 34 + cuda2hipRename["CUDA_ERROR_NOT_READY"] = {"hipErrorNotReady", CONV_TYPE, API_DRIVER}; // 600 + cuda2hipRename["cudaErrorNotReady"] = {"hipErrorNotReady", CONV_TYPE, API_RUNTIME}; // 34 - cuda2hipRename["CUDA_ERROR_ILLEGAL_ADDRESS"] = {"hipErrorIllegalAddress", CONV_ERR, API_DRIVER}; // 700 - cuda2hipRename["cudaErrorIllegalAddress"] = {"hipErrorIllegalAddress", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 77 + cuda2hipRename["CUDA_ERROR_ILLEGAL_ADDRESS"] = {"hipErrorIllegalAddress", CONV_TYPE, API_DRIVER}; // 700 + cuda2hipRename["cudaErrorIllegalAddress"] = {"hipErrorIllegalAddress", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 77 - cuda2hipRename["CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES"] = {"hipErrorLaunchOutOfResources", CONV_ERR, API_DRIVER}; // 701 - cuda2hipRename["cudaErrorLaunchOutOfResources"] = {"hipErrorLaunchOutOfResources", CONV_ERR, API_RUNTIME}; // 7 + cuda2hipRename["CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES"] = {"hipErrorLaunchOutOfResources", CONV_TYPE, API_DRIVER}; // 701 + cuda2hipRename["cudaErrorLaunchOutOfResources"] = {"hipErrorLaunchOutOfResources", CONV_TYPE, API_RUNTIME}; // 7 - cuda2hipRename["CUDA_ERROR_LAUNCH_TIMEOUT"] = {"hipErrorLaunchTimeOut", CONV_ERR, API_DRIVER}; // 702 - cuda2hipRename["cudaErrorLaunchTimeout"] = {"hipErrorLaunchTimeOut", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 6 + cuda2hipRename["CUDA_ERROR_LAUNCH_TIMEOUT"] = {"hipErrorLaunchTimeOut", CONV_TYPE, API_DRIVER}; // 702 + cuda2hipRename["cudaErrorLaunchTimeout"] = {"hipErrorLaunchTimeOut", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 6 - cuda2hipRename["CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED"] = {"hipErrorPeerAccessAlreadyEnabled", CONV_ERR, API_DRIVER}; // 704 - cuda2hipRename["cudaErrorPeerAccessAlreadyEnabled"] = {"hipErrorPeerAccessAlreadyEnabled", CONV_ERR, API_RUNTIME}; // 50 + cuda2hipRename["CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED"] = {"hipErrorPeerAccessAlreadyEnabled", CONV_TYPE, API_DRIVER}; // 704 + cuda2hipRename["cudaErrorPeerAccessAlreadyEnabled"] = {"hipErrorPeerAccessAlreadyEnabled", CONV_TYPE, API_RUNTIME}; // 50 - cuda2hipRename["CUDA_ERROR_PEER_ACCESS_NOT_ENABLED"] = {"hipErrorPeerAccessNotEnabled", CONV_ERR, API_DRIVER}; // 705 - cuda2hipRename["cudaErrorPeerAccessNotEnabled"] = {"hipErrorPeerAccessNotEnabled", CONV_ERR, API_RUNTIME}; // 51 + cuda2hipRename["CUDA_ERROR_PEER_ACCESS_NOT_ENABLED"] = {"hipErrorPeerAccessNotEnabled", CONV_TYPE, API_DRIVER}; // 705 + cuda2hipRename["cudaErrorPeerAccessNotEnabled"] = {"hipErrorPeerAccessNotEnabled", CONV_TYPE, API_RUNTIME}; // 51 - cuda2hipRename["CUDA_ERROR_ASSERT"] = {"hipErrorAssert", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 710 - cuda2hipRename["cudaErrorAssert"] = {"hipErrorAssert", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 59 + cuda2hipRename["CUDA_ERROR_ASSERT"] = {"hipErrorAssert", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 710 + cuda2hipRename["cudaErrorAssert"] = {"hipErrorAssert", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 59 - cuda2hipRename["CUDA_ERROR_TOO_MANY_PEERS"] = {"hipErrorTooManyPeers", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 711 - cuda2hipRename["cudaErrorTooManyPeers"] = {"hipErrorTooManyPeers", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 60 + cuda2hipRename["CUDA_ERROR_TOO_MANY_PEERS"] = {"hipErrorTooManyPeers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 711 + cuda2hipRename["cudaErrorTooManyPeers"] = {"hipErrorTooManyPeers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 60 - cuda2hipRename["CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED"] = {"hipErrorHostMemoryAlreadyRegistered", CONV_ERR, API_DRIVER}; // 712 - cuda2hipRename["cudaErrorHostMemoryAlreadyRegistered"] = {"hipErrorHostMemoryAlreadyRegistered", CONV_ERR, API_RUNTIME}; // 61 + cuda2hipRename["CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED"] = {"hipErrorHostMemoryAlreadyRegistered", CONV_TYPE, API_DRIVER}; // 712 + cuda2hipRename["cudaErrorHostMemoryAlreadyRegistered"] = {"hipErrorHostMemoryAlreadyRegistered", CONV_TYPE, API_RUNTIME}; // 61 - cuda2hipRename["CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED"] = {"hipErrorHostMemoryNotRegistered", CONV_ERR, API_DRIVER}; // 713 - cuda2hipRename["cudaErrorHostMemoryNotRegistered"] = {"hipErrorHostMemoryNotRegistered", CONV_ERR, API_RUNTIME}; // 62 + cuda2hipRename["CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED"] = {"hipErrorHostMemoryNotRegistered", CONV_TYPE, API_DRIVER}; // 713 + cuda2hipRename["cudaErrorHostMemoryNotRegistered"] = {"hipErrorHostMemoryNotRegistered", CONV_TYPE, API_RUNTIME}; // 62 - cuda2hipRename["CUDA_ERROR_HARDWARE_STACK_ERROR"] = {"hipErrorHardwareStackError", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 714 - cuda2hipRename["cudaErrorHardwareStackError"] = {"hipErrorHardwareStackError", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 72 + cuda2hipRename["CUDA_ERROR_HARDWARE_STACK_ERROR"] = {"hipErrorHardwareStackError", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 714 + cuda2hipRename["cudaErrorHardwareStackError"] = {"hipErrorHardwareStackError", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 72 - cuda2hipRename["CUDA_ERROR_ILLEGAL_INSTRUCTION"] = {"hipErrorIllegalInstruction", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 715 - cuda2hipRename["cudaErrorIllegalInstruction"] = {"hipErrorIllegalInstruction", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 73 + cuda2hipRename["CUDA_ERROR_ILLEGAL_INSTRUCTION"] = {"hipErrorIllegalInstruction", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 715 + cuda2hipRename["cudaErrorIllegalInstruction"] = {"hipErrorIllegalInstruction", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 73 - cuda2hipRename["CUDA_ERROR_MISALIGNED_ADDRESS"] = {"hipErrorMisalignedAddress", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 716 - cuda2hipRename["cudaErrorMisalignedAddress"] = {"hipErrorMisalignedAddress", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 74 + cuda2hipRename["CUDA_ERROR_MISALIGNED_ADDRESS"] = {"hipErrorMisalignedAddress", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 716 + cuda2hipRename["cudaErrorMisalignedAddress"] = {"hipErrorMisalignedAddress", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 74 - cuda2hipRename["CUDA_ERROR_INVALID_ADDRESS_SPACE"] = {"hipErrorInvalidAddressSpace", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 717 - cuda2hipRename["cudaErrorInvalidAddressSpace"] = {"hipErrorInvalidAddressSpace", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 75 + cuda2hipRename["CUDA_ERROR_INVALID_ADDRESS_SPACE"] = {"hipErrorInvalidAddressSpace", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 717 + cuda2hipRename["cudaErrorInvalidAddressSpace"] = {"hipErrorInvalidAddressSpace", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 75 - cuda2hipRename["CUDA_ERROR_INVALID_PC"] = {"hipErrorInvalidPc", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 718 - cuda2hipRename["cudaErrorInvalidPc"] = {"hipErrorInvalidPc", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 76 + cuda2hipRename["CUDA_ERROR_INVALID_PC"] = {"hipErrorInvalidPc", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 718 + cuda2hipRename["cudaErrorInvalidPc"] = {"hipErrorInvalidPc", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 76 - cuda2hipRename["CUDA_ERROR_LAUNCH_FAILED"] = {"hipErrorLaunchFailure", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 719 - cuda2hipRename["cudaErrorLaunchFailure"] = {"hipErrorLaunchFailure", CONV_ERR, API_RUNTIME, HIP_UNSUPPORTED}; // 4 + cuda2hipRename["CUDA_ERROR_LAUNCH_FAILED"] = {"hipErrorLaunchFailure", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 719 + cuda2hipRename["cudaErrorLaunchFailure"] = {"hipErrorLaunchFailure", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 4 - cuda2hipRename["CUDA_ERROR_UNKNOWN"] = {"hipErrorUnknown", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // 999 - cuda2hipRename["cudaErrorUnknown"] = {"hipErrorUnknown", CONV_ERR, API_RUNTIME}; // 30 + cuda2hipRename["CUDA_ERROR_UNKNOWN"] = {"hipErrorUnknown", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 999 + cuda2hipRename["cudaErrorUnknown"] = {"hipErrorUnknown", CONV_TYPE, API_RUNTIME}; // 30 ///////////////////////////// CUDA DRIVER API ///////////////////////////// // structs @@ -428,11 +437,11 @@ struct cuda2hipMap { cuda2hipRename["CU_AD_FORMAT_HALF"] = {"HIP_AD_FORMAT_HALF", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 cuda2hipRename["CU_AD_FORMAT_FLOAT"] = {"HIP_AD_FORMAT_FLOAT", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x20 // Compute mode - cuda2hipRename["CUcomputemode"] = {"hipComputemode", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_RUNTIME ANALOGUE (cudaComputeMode) - cuda2hipRename["CU_COMPUTEMODE_DEFAULT"] = {"hipComputeModeDefault", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0 // API_RUNTIME ANALOGUE (cudaComputeModeDefault = 0) - cuda2hipRename["CU_COMPUTEMODE_EXCLUSIVE"] = {"hipComputeModeExclusive", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_RUNTIME ANALOGUE (cudaComputeModeExclusive = 1) - cuda2hipRename["CU_COMPUTEMODE_PROHIBITED"] = {"hipComputeModeProhibited", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_RUNTIME ANALOGUE (cudaComputeModeProhibited = 2) - cuda2hipRename["CU_COMPUTEMODE_EXCLUSIVE_PROCESS"] = {"hipComputeModeExclusiveProcess", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_RUNTIME ANALOGUE (cudaComputeModeExclusiveProcess = 3) + cuda2hipRename["CUcomputemode"] = {"hipComputemode", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_RUNTIME ANALOGUE (cudaComputeMode) + cuda2hipRename["CU_COMPUTEMODE_DEFAULT"] = {"hipComputeModeDefault", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0 // API_RUNTIME ANALOGUE (cudaComputeModeDefault = 0) + cuda2hipRename["CU_COMPUTEMODE_EXCLUSIVE"] = {"hipComputeModeExclusive", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_RUNTIME ANALOGUE (cudaComputeModeExclusive = 1) + cuda2hipRename["CU_COMPUTEMODE_PROHIBITED"] = {"hipComputeModeProhibited", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_RUNTIME ANALOGUE (cudaComputeModeProhibited = 2) + cuda2hipRename["CU_COMPUTEMODE_EXCLUSIVE_PROCESS"] = {"hipComputeModeExclusiveProcess", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_RUNTIME ANALOGUE (cudaComputeModeExclusiveProcess = 3) // unsupported yet by HIP [CUDA 8.0.44] // Memory advise values @@ -465,31 +474,31 @@ struct cuda2hipMap { cuda2hipRename["CU_CTX_FLAGS_MASK"] = {"HIP_CTX_FLAGS_MASK", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x1f // Defines - cuda2hipRename["CU_LAUNCH_PARAM_BUFFER_POINTER"] = {"HIP_LAUNCH_PARAM_BUFFER_POINTER", CONV_DEV, API_DRIVER}; // ((void*)0x01) - cuda2hipRename["CU_LAUNCH_PARAM_BUFFER_SIZE"] = {"HIP_LAUNCH_PARAM_BUFFER_SIZE", CONV_DEV, API_DRIVER}; // ((void*)0x02) - cuda2hipRename["CU_LAUNCH_PARAM_END"] = {"HIP_LAUNCH_PARAM_END", CONV_DEV, API_DRIVER}; // ((void*)0x00) - cuda2hipRename["CU_IPC_HANDLE_SIZE"] = {"HIP_LAUNCH_PARAM_END", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 64 - cuda2hipRename["CU_MEMHOSTALLOC_DEVICEMAP"] = {"HIP_MEMHOSTALLOC_DEVICEMAP", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 - cuda2hipRename["CU_MEMHOSTALLOC_PORTABLE"] = {"HIP_MEMHOSTALLOC_PORTABLE", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 - cuda2hipRename["CU_MEMHOSTALLOC_WRITECOMBINED"] = {"HIP_MEMHOSTALLOC_WRITECOMBINED", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 - cuda2hipRename["CU_MEMHOSTREGISTER_DEVICEMAP"] = {"HIP_MEMHOSTREGISTER_DEVICEMAP", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 - cuda2hipRename["CU_MEMHOSTREGISTER_IOMEMORY"] = {"HIP_MEMHOSTREGISTER_IOMEMORY", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 - cuda2hipRename["CU_MEMHOSTREGISTER_PORTABLE"] = {"HIP_MEMHOSTREGISTER_PORTABLE", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 - cuda2hipRename["CU_PARAM_TR_DEFAULT"] = {"HIP_PARAM_TR_DEFAULT", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // -1 - cuda2hipRename["CU_STREAM_LEGACY"] = {"HIP_STREAM_LEGACY", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // ((CUstream)0x1) - cuda2hipRename["CU_STREAM_PER_THREAD"] = {"HIP_STREAM_PER_THREAD", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // ((CUstream)0x2) - cuda2hipRename["CU_TRSA_OVERRIDE_FORMAT"] = {"HIP_TRSA_OVERRIDE_FORMAT", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 - cuda2hipRename["CU_TRSF_NORMALIZED_COORDINATES"] = {"HIP_TRSF_NORMALIZED_COORDINATES", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED};// 0x02 - cuda2hipRename["CU_TRSF_READ_AS_INTEGER"] = {"HIP_TRSF_READ_AS_INTEGER", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 - cuda2hipRename["CU_TRSF_SRGB"] = {"HIP_TRSF_SRGB", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 + cuda2hipRename["CU_LAUNCH_PARAM_BUFFER_POINTER"] = {"HIP_LAUNCH_PARAM_BUFFER_POINTER", CONV_TYPE, API_DRIVER}; // ((void*)0x01) + cuda2hipRename["CU_LAUNCH_PARAM_BUFFER_SIZE"] = {"HIP_LAUNCH_PARAM_BUFFER_SIZE", CONV_TYPE, API_DRIVER}; // ((void*)0x02) + cuda2hipRename["CU_LAUNCH_PARAM_END"] = {"HIP_LAUNCH_PARAM_END", CONV_TYPE, API_DRIVER}; // ((void*)0x00) + cuda2hipRename["CU_IPC_HANDLE_SIZE"] = {"HIP_LAUNCH_PARAM_END", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 64 + cuda2hipRename["CU_MEMHOSTALLOC_DEVICEMAP"] = {"HIP_MEMHOSTALLOC_DEVICEMAP", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 + cuda2hipRename["CU_MEMHOSTALLOC_PORTABLE"] = {"HIP_MEMHOSTALLOC_PORTABLE", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_MEMHOSTALLOC_WRITECOMBINED"] = {"HIP_MEMHOSTALLOC_WRITECOMBINED", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 + cuda2hipRename["CU_MEMHOSTREGISTER_DEVICEMAP"] = {"HIP_MEMHOSTREGISTER_DEVICEMAP", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 + cuda2hipRename["CU_MEMHOSTREGISTER_IOMEMORY"] = {"HIP_MEMHOSTREGISTER_IOMEMORY", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 + cuda2hipRename["CU_MEMHOSTREGISTER_PORTABLE"] = {"HIP_MEMHOSTREGISTER_PORTABLE", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_PARAM_TR_DEFAULT"] = {"HIP_PARAM_TR_DEFAULT", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // -1 + cuda2hipRename["CU_STREAM_LEGACY"] = {"HIP_STREAM_LEGACY", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // ((CUstream)0x1) + cuda2hipRename["CU_STREAM_PER_THREAD"] = {"HIP_STREAM_PER_THREAD", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // ((CUstream)0x2) + cuda2hipRename["CU_TRSA_OVERRIDE_FORMAT"] = {"HIP_TRSA_OVERRIDE_FORMAT", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_TRSF_NORMALIZED_COORDINATES"] = {"HIP_TRSF_NORMALIZED_COORDINATES", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED};// 0x02 + cuda2hipRename["CU_TRSF_READ_AS_INTEGER"] = {"HIP_TRSF_READ_AS_INTEGER", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CU_TRSF_SRGB"] = {"HIP_TRSF_SRGB", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 // Deprecated, use CUDA_ARRAY3D_LAYERED - cuda2hipRename["CUDA_ARRAY3D_2DARRAY"] = {"HIP_ARRAY3D_LAYERED", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 - cuda2hipRename["CUDA_ARRAY3D_CUBEMAP"] = {"HIP_ARRAY3D_CUBEMAP", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 - cuda2hipRename["CUDA_ARRAY3D_DEPTH_TEXTURE"] = {"HIP_ARRAY3D_DEPTH_TEXTURE", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 - cuda2hipRename["CUDA_ARRAY3D_LAYERED"] = {"HIP_ARRAY3D_LAYERED", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 - cuda2hipRename["CUDA_ARRAY3D_SURFACE_LDST"] = {"HIP_ARRAY3D_SURFACE_LDST", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 - cuda2hipRename["CUDA_ARRAY3D_TEXTURE_GATHER"] = {"HIP_ARRAY3D_TEXTURE_GATHER", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x08 - cuda2hipRename["CUDA_VERSION"] = {"HIP_VERSION", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 7050 + cuda2hipRename["CUDA_ARRAY3D_2DARRAY"] = {"HIP_ARRAY3D_LAYERED", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CUDA_ARRAY3D_CUBEMAP"] = {"HIP_ARRAY3D_CUBEMAP", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x04 + cuda2hipRename["CUDA_ARRAY3D_DEPTH_TEXTURE"] = {"HIP_ARRAY3D_DEPTH_TEXTURE", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x10 + cuda2hipRename["CUDA_ARRAY3D_LAYERED"] = {"HIP_ARRAY3D_LAYERED", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 + cuda2hipRename["CUDA_ARRAY3D_SURFACE_LDST"] = {"HIP_ARRAY3D_SURFACE_LDST", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 + cuda2hipRename["CUDA_ARRAY3D_TEXTURE_GATHER"] = {"HIP_ARRAY3D_TEXTURE_GATHER", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x08 + cuda2hipRename["CUDA_VERSION"] = {"HIP_VERSION", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 7050 // Types // NOTE: CUdevice might be changed to typedef int in the future. @@ -500,115 +509,115 @@ struct cuda2hipMap { // CUDA: "The types::CUarray and struct ::cudaArray * represent the same data type and may be used interchangeably by casting the two types between each other." // typedef struct cudaArray *cudaArray_t; // typedef struct CUarray_st *CUarray; - cuda2hipRename["CUarray_st"] = {"hipArray", CONV_MEM, API_RUNTIME}; // API_Runtime ANALOGUE (cudaArray) + cuda2hipRename["CUarray_st"] = {"hipArray", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaArray) cuda2hipRename["CUarray"] = {"hipArray *", CONV_TYPE, API_DRIVER}; // API_Runtime ANALOGUE (cudaArray_t) // unsupported yet by HIP - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (cudaDevAttrMaxThreadsPerBlock = 1) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X"] = {"hipDeviceAttributeMaxBlockDimX", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimX = 2) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y"] = {"hipDeviceAttributeMaxBlockDimY", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimY = 3) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 4 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimZ = 4) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X"] = {"hipDeviceAttributeMaxGridDimX", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 5 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimX =5) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y"] = {"hipDeviceAttributeMaxGridDimY", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 6 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimY = 6) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z"] = {"hipDeviceAttributeMaxGridDimZ", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 7 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimZ - 7) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 8 // API_Runtime ANALOGUE (cudaDevAttrMaxSharedMemoryPerBlock = 8) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (cudaDevAttrMaxThreadsPerBlock = 1) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X"] = {"hipDeviceAttributeMaxBlockDimX", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimX = 2) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y"] = {"hipDeviceAttributeMaxBlockDimY", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimY = 3) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 4 // API_Runtime ANALOGUE (cudaDevAttrMaxBlockDimZ = 4) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X"] = {"hipDeviceAttributeMaxGridDimX", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 5 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimX =5) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Y"] = {"hipDeviceAttributeMaxGridDimY", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 6 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimY = 6) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_Z"] = {"hipDeviceAttributeMaxGridDimZ", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 7 // API_Runtime ANALOGUE (cudaDevAttrMaxGridDimZ - 7) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 8 // API_Runtime ANALOGUE (cudaDevAttrMaxSharedMemoryPerBlock = 8) // Deprecated, use CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK - cuda2hipRename["CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 8 - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY"] = {"hipDeviceAttributeTotalConstantMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 9 // API_Runtime ANALOGUE (cudaDevAttrTotalConstantMemory = 9) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_WARP_SIZE"] = {"hipDeviceAttributeWarpSize", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 10 // API_Runtime ANALOGUE (cudaDevAttrWarpSize = 10) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_PITCH"] = {"hipDeviceAttributeMaxPitch", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 11 // API_Runtime ANALOGUE (cudaDevAttrMaxPitch = 11) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 12 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerBlock = 12) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 12 - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CLOCK_RATE"] = {"hipDeviceAttributeClockRate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 13 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerBlock = 13) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT"] = {"hipDeviceAttributeTextureAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 14 // API_Runtime ANALOGUE (cudaDevAttrTextureAlignment = 14) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_SHARED_MEMORY_PER_BLOCK"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 8 + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY"] = {"hipDeviceAttributeTotalConstantMemory", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 9 // API_Runtime ANALOGUE (cudaDevAttrTotalConstantMemory = 9) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_WARP_SIZE"] = {"hipDeviceAttributeWarpSize", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 10 // API_Runtime ANALOGUE (cudaDevAttrWarpSize = 10) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_PITCH"] = {"hipDeviceAttributeMaxPitch", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 11 // API_Runtime ANALOGUE (cudaDevAttrMaxPitch = 11) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 12 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerBlock = 12) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_REGISTERS_PER_BLOCK"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 12 + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CLOCK_RATE"] = {"hipDeviceAttributeClockRate", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 13 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerBlock = 13) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT"] = {"hipDeviceAttributeTextureAlignment", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 14 // API_Runtime ANALOGUE (cudaDevAttrTextureAlignment = 14) // Deprecated. Use instead CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT - cuda2hipRename["CU_DEVICE_ATTRIBUTE_GPU_OVERLAP"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 15 // API_Runtime ANALOGUE (cudaDevAttrGpuOverlap = 15) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT"] = {"hipDeviceAttributeMultiprocessorCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 16 // API_Runtime ANALOGUE (cudaDevAttrMultiProcessorCount = 16) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT"] = {"hipDeviceAttributeKernelExecTimeout", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 17 // API_Runtime ANALOGUE (cudaDevAttrKernelExecTimeout = 17) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_INTEGRATED"] = {"hipDeviceAttributeIntegrated", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 18 // API_Runtime ANALOGUE (cudaDevAttrIntegrated = 18) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY"] = {"hipDeviceAttributeCanMapHostMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 19 // API_Runtime ANALOGUE (cudaDevAttrCanMapHostMemory = 19) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_MODE"] = {"hipDeviceAttributeComputeMode", CONV_DEV, API_DRIVER}; // 20 // API_Runtime ANALOGUE (cudaDevAttrComputeMode = 20) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 21 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DWidth = 21) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 22 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DWidth = 22) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 23 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DHeight = 23) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 24 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DWidth = 24) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 25 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DHeight = 25) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 26 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DDepth = 26) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 27 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredWidth = 27) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 28 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredHeight = 28) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 29 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredLayers = 29) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_GPU_OVERLAP"] = {"hipDeviceAttributeAsyncEngineCount", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 15 // API_Runtime ANALOGUE (cudaDevAttrGpuOverlap = 15) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT"] = {"hipDeviceAttributeMultiprocessorCount", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 16 // API_Runtime ANALOGUE (cudaDevAttrMultiProcessorCount = 16) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT"] = {"hipDeviceAttributeKernelExecTimeout", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 17 // API_Runtime ANALOGUE (cudaDevAttrKernelExecTimeout = 17) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_INTEGRATED"] = {"hipDeviceAttributeIntegrated", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 18 // API_Runtime ANALOGUE (cudaDevAttrIntegrated = 18) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY"] = {"hipDeviceAttributeCanMapHostMemory", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 19 // API_Runtime ANALOGUE (cudaDevAttrCanMapHostMemory = 19) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_MODE"] = {"hipDeviceAttributeComputeMode", CONV_TYPE, API_DRIVER}; // 20 // API_Runtime ANALOGUE (cudaDevAttrComputeMode = 20) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 21 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DWidth = 21) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 22 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DWidth = 22) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 23 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DHeight = 23) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 24 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DWidth = 24) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 25 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DHeight = 25) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 26 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DDepth = 26) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 27 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredWidth = 27) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 28 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredHeight = 28) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 29 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredLayers = 29) // Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 27 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredWidth = 27) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 27 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredWidth = 27) // Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 28 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredHeight = 28) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 28 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredHeight = 28) // Deprecated, use CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 29 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredLayers = 29) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT"] = {"hipDeviceAttributeSurfaceAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 30 // API_Runtime ANALOGUE (cudaDevAttrSurfaceAlignment = 30) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS"] = {"hipDeviceAttributeConcurrentKernels", CONV_DEV, API_DRIVER}; // 31 // API_Runtime ANALOGUE (cudaDevAttrConcurrentKernels = 31) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_ECC_ENABLED"] = {"hipDeviceAttributeEccEnabled", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 32 // API_Runtime ANALOGUE (cudaDevAttrEccEnabled = 32) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_BUS_ID"] = {"hipDeviceAttributePciBusId", CONV_DEV, API_DRIVER}; // 33 // API_Runtime ANALOGUE (cudaDevAttrPciBusId = 33) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID"] = {"hipDeviceAttributePciDeviceId", CONV_DEV, API_DRIVER}; // 34 // API_Runtime ANALOGUE (cudaDevAttrPciDeviceId = 34) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TCC_DRIVER"] = {"hipDeviceAttributeTccDriver", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 35 // API_Runtime ANALOGUE (cudaDevAttrTccDriver = 35) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE"] = {"hipDeviceAttributeMemoryClockRate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 36 // API_Runtime ANALOGUE (cudaDevAttrMemoryClockRate = 36) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH"] = {"hipDeviceAttributeMemoryBusWidth", CONV_DEV, API_DRIVER}; // 37 // API_Runtime ANALOGUE (cudaDevAttrGlobalMemoryBusWidth = 37) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE"] = {"hipDeviceAttributeL2CacheSize", CONV_DEV, API_DRIVER}; // 38 // API_Runtime ANALOGUE (cudaDevAttrL2CacheSize = 38) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_DEV, API_DRIVER}; // 39 // API_Runtime ANALOGUE (cudaDevAttrMaxThreadsPerMultiProcessor = 39) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 40 // API_Runtime ANALOGUE (cudaDevAttrAsyncEngineCount = 40) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING"] = {"hipDeviceAttributeUnifiedAddressing", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 41 // API_Runtime ANALOGUE (cudaDevAttrUnifiedAddressing = 41) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 42 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLayeredWidth = 42) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 43 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLayeredLayers = 43) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_ARRAY_NUMSLICES"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 29 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLayeredLayers = 29) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT"] = {"hipDeviceAttributeSurfaceAlignment", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 30 // API_Runtime ANALOGUE (cudaDevAttrSurfaceAlignment = 30) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS"] = {"hipDeviceAttributeConcurrentKernels", CONV_TYPE, API_DRIVER}; // 31 // API_Runtime ANALOGUE (cudaDevAttrConcurrentKernels = 31) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_ECC_ENABLED"] = {"hipDeviceAttributeEccEnabled", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 32 // API_Runtime ANALOGUE (cudaDevAttrEccEnabled = 32) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_BUS_ID"] = {"hipDeviceAttributePciBusId", CONV_TYPE, API_DRIVER}; // 33 // API_Runtime ANALOGUE (cudaDevAttrPciBusId = 33) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID"] = {"hipDeviceAttributePciDeviceId", CONV_TYPE, API_DRIVER}; // 34 // API_Runtime ANALOGUE (cudaDevAttrPciDeviceId = 34) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TCC_DRIVER"] = {"hipDeviceAttributeTccDriver", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 35 // API_Runtime ANALOGUE (cudaDevAttrTccDriver = 35) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE"] = {"hipDeviceAttributeMemoryClockRate", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 36 // API_Runtime ANALOGUE (cudaDevAttrMemoryClockRate = 36) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH"] = {"hipDeviceAttributeMemoryBusWidth", CONV_TYPE, API_DRIVER}; // 37 // API_Runtime ANALOGUE (cudaDevAttrGlobalMemoryBusWidth = 37) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE"] = {"hipDeviceAttributeL2CacheSize", CONV_TYPE, API_DRIVER}; // 38 // API_Runtime ANALOGUE (cudaDevAttrL2CacheSize = 38) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_TYPE, API_DRIVER}; // 39 // API_Runtime ANALOGUE (cudaDevAttrMaxThreadsPerMultiProcessor = 39) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT"] = {"hipDeviceAttributeAsyncEngineCount", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 40 // API_Runtime ANALOGUE (cudaDevAttrAsyncEngineCount = 40) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING"] = {"hipDeviceAttributeUnifiedAddressing", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 41 // API_Runtime ANALOGUE (cudaDevAttrUnifiedAddressing = 41) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 42 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLayeredWidth = 42) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 43 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLayeredLayers = 43) // deprecated, do not use - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER"] = {"hipDeviceAttributeCanTex2DGather", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 44 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 45 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DGatherWidth = 45) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 46 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DGatherHeight = 46) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 47 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DWidthAlt = 47) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 48 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DHeightAlt = 48) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 49 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DDepthAlt = 49) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID"] = {"hipDeviceAttributePciDomainId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 50 // API_Runtime ANALOGUE (cudaDevAttrPciDomainId = 50) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 51 // API_Runtime ANALOGUE (cudaDevAttrTexturePitchAlignment = 51) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 52 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapWidth = 52) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 53 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapLayeredWidth = 53) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 54 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapLayeredLayers = 54) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 55 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DWidth = 55) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 56 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DWidth = 56) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 57 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DHeight = 57) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 58 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DWidth = 58) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 59 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DHeight = 59) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 60 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DDepth = 60) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 61 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DLayeredWidth = 61) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 62 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DLayeredLayers = 62) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 63 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredWidth = 63) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 64 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredHeight = 64) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 65 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredLayers = 65) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 66 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapWidth = 66) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 67 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapLayeredWidth = 67) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 68 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapLayeredLayers = 68) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 69 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLinearWidth = 69) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 70 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearWidth = 70) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 71 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearHeight = 71) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 72 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearPitch = 72) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 73 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DMipmappedWidth = 73) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 74 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DMipmappedHeight = 74) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_DEV, API_DRIVER}; // 75 // API_Runtime ANALOGUE (cudaDevAttrComputeCapabilityMajor = 75) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_DEV, API_DRIVER}; // 76 // API_Runtime ANALOGUE (cudaDevAttrComputeCapabilityMinor = 76) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 77 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DMipmappedWidth = 77) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 78 // API_Runtime ANALOGUE (cudaDevAttrStreamPrioritiesSupported = 78) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 79 // API_Runtime ANALOGUE (cudaDevAttrGlobalL1CacheSupported = 79) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 80 // API_Runtime ANALOGUE (cudaDevAttrLocalL1CacheSupported = 80) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_DEV, API_DRIVER}; // 81 // API_Runtime ANALOGUE (cudaDevAttrMaxSharedMemoryPerMultiprocessor = 81) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 82 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerMultiprocessor = 82) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY"] = {"hipDeviceAttributeManagedMemory", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 83 // API_Runtime ANALOGUE (cudaDevAttrManagedMemory = 83) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_DEV, API_DRIVER}; // 84 // API_Runtime ANALOGUE (cudaDevAttrIsMultiGpuBoard = 84) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID"] = {"hipDeviceAttributeMultiGpuBoardGroupId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 85 // API_Runtime ANALOGUE (cudaDevAttrMultiGpuBoardGroupID = 85) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_TEX2D_GATHER"] = {"hipDeviceAttributeCanTex2DGather", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 44 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 45 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DGatherWidth = 45) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 46 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DGatherHeight = 46) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 47 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DWidthAlt = 47) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 48 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DHeightAlt = 48) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 49 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture3DDepthAlt = 49) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID"] = {"hipDeviceAttributePciDomainId", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 50 // API_Runtime ANALOGUE (cudaDevAttrPciDomainId = 50) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 51 // API_Runtime ANALOGUE (cudaDevAttrTexturePitchAlignment = 51) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 52 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapWidth = 52) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 53 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapLayeredWidth = 53) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 54 // API_Runtime ANALOGUE (cudaDevAttrMaxTextureCubemapLayeredLayers = 54) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 55 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DWidth = 55) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 56 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DWidth = 56) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 57 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DHeight = 57) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 58 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DWidth = 58) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 59 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DHeight = 59) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 60 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface3DDepth = 60) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 61 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DLayeredWidth = 61) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 62 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface1DLayeredLayers = 62) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 63 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredWidth = 63) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 64 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredHeight = 64) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 65 // API_Runtime ANALOGUE (cudaDevAttrMaxSurface2DLayeredLayers = 65) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 66 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapWidth = 66) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 67 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapLayeredWidth = 67) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 68 // API_Runtime ANALOGUE (cudaDevAttrMaxSurfaceCubemapLayeredLayers = 68) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 69 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DLinearWidth = 69) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 70 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearWidth = 70) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 71 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearHeight = 71) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 72 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DLinearPitch = 72) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 73 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DMipmappedWidth = 73) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 74 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture2DMipmappedHeight = 74) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_TYPE, API_DRIVER}; // 75 // API_Runtime ANALOGUE (cudaDevAttrComputeCapabilityMajor = 75) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_TYPE, API_DRIVER}; // 76 // API_Runtime ANALOGUE (cudaDevAttrComputeCapabilityMinor = 76) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 77 // API_Runtime ANALOGUE (cudaDevAttrMaxTexture1DMipmappedWidth = 77) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 78 // API_Runtime ANALOGUE (cudaDevAttrStreamPrioritiesSupported = 78) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 79 // API_Runtime ANALOGUE (cudaDevAttrGlobalL1CacheSupported = 79) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 80 // API_Runtime ANALOGUE (cudaDevAttrLocalL1CacheSupported = 80) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_TYPE, API_DRIVER}; // 81 // API_Runtime ANALOGUE (cudaDevAttrMaxSharedMemoryPerMultiprocessor = 81) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 82 // API_Runtime ANALOGUE (cudaDevAttrMaxRegistersPerMultiprocessor = 82) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY"] = {"hipDeviceAttributeManagedMemory", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 83 // API_Runtime ANALOGUE (cudaDevAttrManagedMemory = 83) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_TYPE, API_DRIVER}; // 84 // API_Runtime ANALOGUE (cudaDevAttrIsMultiGpuBoard = 84) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID"] = {"hipDeviceAttributeMultiGpuBoardGroupId", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 85 // API_Runtime ANALOGUE (cudaDevAttrMultiGpuBoardGroupID = 85) // unsupported yet by HIP [CUDA 8.0.44] - cuda2hipRename["CU_DEVICE_ATTRIBUTE_HOST_NATIVE_ATOMIC_SUPPORTED"] = {"hipDeviceAttributeHostNativeAtomicSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 86 // API_Runtime ANALOGUE (cudaDevAttrHostNativeAtomicSupported = 86) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_SINGLE_TO_DOUBLE_PRECISION_PERF_RATIO"] = {"hipDeviceAttributeSingleToDoublePrecisionPerfRatio", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 87 // API_Runtime ANALOGUE (cudaDevAttrSingleToDoublePrecisionPerfRatio = 87) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_PAGEABLE_MEMORY_ACCESS"] = {"hipDeviceAttributePageableMemoryAccess", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 88 // API_Runtime ANALOGUE (cudaDevAttrPageableMemoryAccess = 88) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CONCURRENT_MANAGED_ACCESS"] = {"hipDeviceAttributeConcurrentManagedAccess", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 89 // API_Runtime ANALOGUE (cudaDevAttrConcurrentManagedAccess = 89) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_PREEMPTION_SUPPORTED"] = {"hipDeviceAttributeComputePreemptionSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 90 // API_Runtime ANALOGUE (cudaDevAttrComputePreemptionSupported = 90) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_USE_HOST_POINTER_FOR_REGISTERED_MEM"] = {"hipDeviceAttributeCanUseHostPointerForRegisteredMem", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 91 // API_Runtime ANALOGUE (cudaDevAttrCanUseHostPointerForRegisteredMem = 91) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_HOST_NATIVE_ATOMIC_SUPPORTED"] = {"hipDeviceAttributeHostNativeAtomicSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 86 // API_Runtime ANALOGUE (cudaDevAttrHostNativeAtomicSupported = 86) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_SINGLE_TO_DOUBLE_PRECISION_PERF_RATIO"] = {"hipDeviceAttributeSingleToDoublePrecisionPerfRatio", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 87 // API_Runtime ANALOGUE (cudaDevAttrSingleToDoublePrecisionPerfRatio = 87) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_PAGEABLE_MEMORY_ACCESS"] = {"hipDeviceAttributePageableMemoryAccess", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 88 // API_Runtime ANALOGUE (cudaDevAttrPageableMemoryAccess = 88) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CONCURRENT_MANAGED_ACCESS"] = {"hipDeviceAttributeConcurrentManagedAccess", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 89 // API_Runtime ANALOGUE (cudaDevAttrConcurrentManagedAccess = 89) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_COMPUTE_PREEMPTION_SUPPORTED"] = {"hipDeviceAttributeComputePreemptionSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 90 // API_Runtime ANALOGUE (cudaDevAttrComputePreemptionSupported = 90) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_CAN_USE_HOST_POINTER_FOR_REGISTERED_MEM"] = {"hipDeviceAttributeCanUseHostPointerForRegisteredMem", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 91 // API_Runtime ANALOGUE (cudaDevAttrCanUseHostPointerForRegisteredMem = 91) - cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX"] = {"hipDeviceAttributeMax", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 92 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_DEVICE_ATTRIBUTE_MAX"] = {"hipDeviceAttributeMax", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 92 // API_Runtime ANALOGUE (no) cuda2hipRename["CUdevprop_st"] = {"hipDeviceProp_t", CONV_TYPE, API_DRIVER}; cuda2hipRename["CUdevprop"] = {"hipDeviceProp_t", CONV_TYPE, API_DRIVER}; @@ -617,14 +626,14 @@ struct cuda2hipMap { // TODO: Do for Pointer Attributes the same as for Device Attributes. // cuda2hipRename["CUpointer_attribute_enum"] = {"hipPointerAttribute_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) // cuda2hipRename["CUpointer_attribute"] = {"hipPointerAttribute_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_CONTEXT"] = {"hipPointerAttributeContext", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_MEMORY_TYPE"] = {"hipPointerAttributeMemoryType", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_DEVICE_POINTER"] = {"hipPointerAttributeDevicePointer", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_HOST_POINTER"] = {"hipPointerAttributeHostPointer", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 4 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_P2P_TOKENS"] = {"hipPointerAttributeP2pTokens", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 5 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_SYNC_MEMOPS"] = {"hipPointerAttributeSyncMemops", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 6 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_BUFFER_ID"] = {"hipPointerAttributeBufferId", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 7 // API_Runtime ANALOGUE (no) - cuda2hipRename["CU_POINTER_ATTRIBUTE_IS_MANAGED"] = {"hipPointerAttributeIsManaged", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 8 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_CONTEXT"] = {"hipPointerAttributeContext", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_MEMORY_TYPE"] = {"hipPointerAttributeMemoryType", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 2 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_DEVICE_POINTER"] = {"hipPointerAttributeDevicePointer", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 3 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_HOST_POINTER"] = {"hipPointerAttributeHostPointer", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 4 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_P2P_TOKENS"] = {"hipPointerAttributeP2pTokens", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 5 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_SYNC_MEMOPS"] = {"hipPointerAttributeSyncMemops", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 6 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_BUFFER_ID"] = {"hipPointerAttributeBufferId", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 7 // API_Runtime ANALOGUE (no) + cuda2hipRename["CU_POINTER_ATTRIBUTE_IS_MANAGED"] = {"hipPointerAttributeIsManaged", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 8 // API_Runtime ANALOGUE (no) // pointer to CUfunc_st cuda2hipRename["CUfunction"] = {"hipFunction_t", CONV_TYPE, API_DRIVER}; @@ -826,9 +835,9 @@ struct cuda2hipMap { cuda2hipRename["CUsharedconfig"] = {"hipSharedMemConfig", CONV_TYPE, API_DRIVER}; cuda2hipRename["CUsharedconfig_enum"] = {"hipSharedMemConfig", CONV_TYPE, API_DRIVER}; - cuda2hipRename["CU_SHARED_MEM_CONFIG_DEFAULT_BANK_SIZE"] = {"hipSharedMemBankSizeDefault", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_SHARED_MEM_CONFIG_FOUR_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeFourByte", CONV_DEV, API_DRIVER}; - cuda2hipRename["CU_SHARED_MEM_CONFIG_EIGHT_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeEightByte", CONV_DEV, API_DRIVER}; + cuda2hipRename["CU_SHARED_MEM_CONFIG_DEFAULT_BANK_SIZE"] = {"hipSharedMemBankSizeDefault", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CU_SHARED_MEM_CONFIG_FOUR_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeFourByte", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CU_SHARED_MEM_CONFIG_EIGHT_BYTE_BANK_SIZE"] = {"hipSharedMemBankSizeEightByte", CONV_TYPE, API_DRIVER}; cuda2hipRename["CUcontext"] = {"hipCtx_t", CONV_TYPE, API_DRIVER}; // TODO: move "typedef struct ihipCtx_t *hipCtx_t;" from hcc_details to HIP @@ -857,40 +866,40 @@ struct cuda2hipMap { // cuda2hipRename["CUtexref_st"] = {"ihipTextureReference_t", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // Stream Flags enum - cuda2hipRename["CUstream_flags"] = {"hipStreamFlags", CONV_STREAM, API_DRIVER}; - // cuda2hipRename["CUstream_flags_enum"] = {"hipStreamFlags", CONV_STREAM, API_DRIVER}; - cuda2hipRename["CU_STREAM_DEFAULT"] = {"hipStreamDefault", CONV_STREAM, API_DRIVER}; - cuda2hipRename["CU_STREAM_NON_BLOCKING"] = {"hipStreamNonBlocking", CONV_STREAM, API_DRIVER}; + cuda2hipRename["CUstream_flags"] = {"hipStreamFlags", CONV_TYPE, API_DRIVER}; + // cuda2hipRename["CUstream_flags_enum"] = {"hipStreamFlags", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CU_STREAM_DEFAULT"] = {"hipStreamDefault", CONV_TYPE, API_DRIVER}; + cuda2hipRename["CU_STREAM_NON_BLOCKING"] = {"hipStreamNonBlocking", CONV_TYPE, API_DRIVER}; // unsupported yet by HIP [CUDA 8.0.44] // Flags for ::cuStreamWaitValue32 - cuda2hipRename["CUstreamWaitValue_flags"] = {"hipStreamWaitValueFlags", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - // cuda2hipRename["CUstreamWaitValue_flags_enum"] = {"hipStreamWaitValueFlags", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_STREAM_WAIT_VALUE_GEQ"] = {"hipStreamWaitValueGeq", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 0x0 - cuda2hipRename["CU_STREAM_WAIT_VALUE_EQ"] = {"hipStreamWaitValueEq", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 0x1 - cuda2hipRename["CU_STREAM_WAIT_VALUE_AND"] = {"hipStreamWaitValueAnd", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 0x2 - cuda2hipRename["CU_STREAM_WAIT_VALUE_FLUSH"] = {"hipStreamWaitValueFlush", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 1<<30 + cuda2hipRename["CUstreamWaitValue_flags"] = {"hipStreamWaitValueFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + // cuda2hipRename["CUstreamWaitValue_flags_enum"] = {"hipStreamWaitValueFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_STREAM_WAIT_VALUE_GEQ"] = {"hipStreamWaitValueGeq", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x0 + cuda2hipRename["CU_STREAM_WAIT_VALUE_EQ"] = {"hipStreamWaitValueEq", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x1 + cuda2hipRename["CU_STREAM_WAIT_VALUE_AND"] = {"hipStreamWaitValueAnd", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x2 + cuda2hipRename["CU_STREAM_WAIT_VALUE_FLUSH"] = {"hipStreamWaitValueFlush", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 1<<30 // Flags for ::cuStreamWriteValue32 - cuda2hipRename["CUstreamWriteValue_flags"] = {"hipStreamWriteValueFlags", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - // cuda2hipRename["CUstreamWriteValue_flags"] = {"hipStreamWriteValueFlags", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_STREAM_WRITE_VALUE_DEFAULT"] = {"hipStreamWriteValueDefault", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 0x0 - cuda2hipRename["CU_STREAM_WRITE_VALUE_NO_MEMORY_BARRIER"] = {"hipStreamWriteValueNoMemoryBarrier", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 0x1 + cuda2hipRename["CUstreamWriteValue_flags"] = {"hipStreamWriteValueFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + // cuda2hipRename["CUstreamWriteValue_flags"] = {"hipStreamWriteValueFlags", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_STREAM_WRITE_VALUE_DEFAULT"] = {"hipStreamWriteValueDefault", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x0 + cuda2hipRename["CU_STREAM_WRITE_VALUE_NO_MEMORY_BARRIER"] = {"hipStreamWriteValueNoMemoryBarrier", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x1 // Flags for ::cuStreamBatchMemOp - cuda2hipRename["CUstreamBatchMemOpType"] = {"hipStreamBatchMemOpType", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - // cuda2hipRename["CUstreamBatchMemOpType_enum"] = {"hipStreamBatchMemOpType", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_STREAM_MEM_OP_WAIT_VALUE_32"] = {"hipStreamBatchMemOpWaitValue32", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 1 - cuda2hipRename["CU_STREAM_MEM_OP_WRITE_VALUE_32"] = {"hipStreamBatchMemOpWriteValue32", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 2 - cuda2hipRename["CU_STREAM_MEM_OP_FLUSH_REMOTE_WRITES"] = {"hipStreamBatchMemOpFlushRemoteWrites", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // 3 + cuda2hipRename["CUstreamBatchMemOpType"] = {"hipStreamBatchMemOpType", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + // cuda2hipRename["CUstreamBatchMemOpType_enum"] = {"hipStreamBatchMemOpType", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_STREAM_MEM_OP_WAIT_VALUE_32"] = {"hipStreamBatchMemOpWaitValue32", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 1 + cuda2hipRename["CU_STREAM_MEM_OP_WRITE_VALUE_32"] = {"hipStreamBatchMemOpWriteValue32", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 2 + cuda2hipRename["CU_STREAM_MEM_OP_FLUSH_REMOTE_WRITES"] = {"hipStreamBatchMemOpFlushRemoteWrites", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 3 // Error Handling - cuda2hipRename["cuGetErrorName"] = {"hipGetErrorName___", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // cudaGetErrorName (hipGetErrorName) has different signature - cuda2hipRename["cuGetErrorString"] = {"hipGetErrorString___", CONV_ERR, API_DRIVER, HIP_UNSUPPORTED}; // cudaGetErrorString (hipGetErrorString) has different signature + cuda2hipRename["cuGetErrorName"] = {"hipGetErrorName___", CONV_ERROR, API_DRIVER, HIP_UNSUPPORTED}; // cudaGetErrorName (hipGetErrorName) has different signature + cuda2hipRename["cuGetErrorString"] = {"hipGetErrorString___", CONV_ERROR, API_DRIVER, HIP_UNSUPPORTED}; // cudaGetErrorString (hipGetErrorString) has different signature // Init - cuda2hipRename["cuInit"] = {"hipInit", CONV_DRIVER, API_DRIVER}; + cuda2hipRename["cuInit"] = {"hipInit", CONV_INIT, API_DRIVER}; // Driver - cuda2hipRename["cuDriverGetVersion"] = {"hipDriverGetVersion", CONV_DRIVER, API_DRIVER}; + cuda2hipRename["cuDriverGetVersion"] = {"hipDriverGetVersion", CONV_VERSION, API_DRIVER}; // Context Management cuda2hipRename["cuCtxCreate_v2"] = {"hipCtxCreate", CONV_CONTEXT, API_DRIVER}; @@ -915,10 +924,10 @@ struct cuda2hipMap { cuda2hipRename["cuCtxDetach"] = {"hipCtxDetach", CONV_CONTEXT, API_DRIVER, HIP_UNSUPPORTED}; // Peer Context Memory Access - cuda2hipRename["cuCtxEnablePeerAccess"] = {"hipCtxEnablePeerAccess", CONV_CONTEXT, API_DRIVER}; - cuda2hipRename["cuCtxDisablePeerAccess"] = {"hipCtxDisablePeerAccess", CONV_CONTEXT, API_DRIVER}; - cuda2hipRename["cuDeviceCanAccessPeer"] = {"hipDeviceCanAccessPeer", CONV_DEV, API_DRIVER}; - + cuda2hipRename["cuCtxEnablePeerAccess"] = {"hipCtxEnablePeerAccess", CONV_PEER, API_DRIVER}; + cuda2hipRename["cuCtxDisablePeerAccess"] = {"hipCtxDisablePeerAccess", CONV_PEER, API_DRIVER}; + cuda2hipRename["cuDeviceCanAccessPeer"] = {"hipDeviceCanAccessPeer", CONV_PEER, API_DRIVER}; + cuda2hipRename["cuDeviceGetP2PAttribute"] = {"hipDeviceGetP2PAttribute", CONV_PEER, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaDeviceGetP2PAttribute) // Primary Context Management cuda2hipRename["cuDevicePrimaryCtxGetState"] = {"hipDevicePrimaryCtxGetState", CONV_CONTEXT, API_DRIVER}; @@ -928,28 +937,28 @@ struct cuda2hipMap { cuda2hipRename["cuDevicePrimaryCtxSetFlags"] = {"hipDevicePrimaryCtxSetFlags", CONV_CONTEXT, API_DRIVER}; // Device Management - cuda2hipRename["cuDeviceGet"] = {"hipGetDevice", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceGetName"] = {"hipDeviceGetName", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceGetCount"] = {"hipGetDeviceCount", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceGetAttribute"] = {"hipDeviceGetAttribute", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceGetPCIBusId"] = {"hipDeviceGetPCIBusId", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceGetByPCIBusId"] = {"hipDeviceGetByPCIBusId", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceTotalMem_v2"] = {"hipDeviceTotalMem", CONV_DEV, API_DRIVER}; + cuda2hipRename["cuDeviceGet"] = {"hipGetDevice", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceGetName"] = {"hipDeviceGetName", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceGetCount"] = {"hipGetDeviceCount", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceGetAttribute"] = {"hipDeviceGetAttribute", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceGetPCIBusId"] = {"hipDeviceGetPCIBusId", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceGetByPCIBusId"] = {"hipDeviceGetByPCIBusId", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceTotalMem_v2"] = {"hipDeviceTotalMem", CONV_DEVICE, API_DRIVER}; // Device Management [DEPRECATED] - cuda2hipRename["cuDeviceComputeCapability"] = {"hipDeviceComputeCapability", CONV_DEV, API_DRIVER}; - cuda2hipRename["cuDeviceGetProperties"] = {"hipGetDeviceProperties", CONV_DEV, API_DRIVER}; + cuda2hipRename["cuDeviceComputeCapability"] = {"hipDeviceComputeCapability", CONV_DEVICE, API_DRIVER}; + cuda2hipRename["cuDeviceGetProperties"] = {"hipGetDeviceProperties", CONV_DEVICE, API_DRIVER}; // Module Management - cuda2hipRename["cuLinkAddData"] = {"hipLinkAddData", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuLinkAddFile"] = {"hipLinkAddFile", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuLinkComplete"] = {"hipLinkComplete", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuLinkCreate"] = {"hipLinkCreate", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuLinkDestroy"] = {"hipLinkDestroy", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLinkAddData"] = {"hipLinkAddData", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLinkAddFile"] = {"hipLinkAddFile", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLinkComplete"] = {"hipLinkComplete", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLinkCreate"] = {"hipLinkCreate", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLinkDestroy"] = {"hipLinkDestroy", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuModuleGetFunction"] = {"hipModuleGetFunction", CONV_MODULE, API_DRIVER}; cuda2hipRename["cuModuleGetGlobal_v2"] = {"hipModuleGetGlobal", CONV_MODULE, API_DRIVER}; - cuda2hipRename["cuModuleGetSurfRef"] = {"hipModuleGetSurfRef", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuModuleGetTexRef"] = {"hipModuleGetTexRef", CONV_EVENT, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuModuleGetSurfRef"] = {"hipModuleGetSurfRef", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuModuleGetTexRef"] = {"hipModuleGetTexRef", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuModuleLoad"] = {"hipModuleLoad", CONV_MODULE, API_DRIVER}; cuda2hipRename["cuModuleLoadData"] = {"hipModuleLoadData", CONV_MODULE, API_DRIVER}; cuda2hipRename["cuModuleLoadDataEx"] = {"hipModuleLoadDataEx", CONV_MODULE, API_DRIVER}; @@ -958,13 +967,11 @@ struct cuda2hipMap { // unsupported yet by HIP [CUDA 8.0.44] // P2P Attributes - cuda2hipRename["CUdevice_P2PAttribute"] = {"hipDeviceP2PAttribute", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaDeviceP2PAttr) - // cuda2hipRename["CUdevice_P2PAttribute_enum"] = {"hipDeviceP2PAttribute", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["CU_DEVICE_P2P_ATTRIBUTE_PERFORMANCE_RANK"] = {"hipDeviceP2PAttributePerformanceRank", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaDevP2PAttrPerformanceRank = 0x01) - cuda2hipRename["CU_DEVICE_P2P_ATTRIBUTE_ACCESS_SUPPORTED"] = {"hipDeviceP2PAttributeAccessSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaDevP2PAttrAccessSupported = 0x02) - cuda2hipRename["CU_DEVICE_P2P_ATTRIBUTE_NATIVE_ATOMIC_SUPPORTED"] = {"hipDeviceP2PAttributeNativeAtomicSupported", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaDevP2PAttrNativeAtomicSupported = 0x03) - - cuda2hipRename["cuDeviceGetP2PAttribute"] = {"hipDeviceGetP2PAttribute", CONV_DEV, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaDeviceGetP2PAttribute) + cuda2hipRename["CUdevice_P2PAttribute"] = {"hipDeviceP2PAttribute", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaDeviceP2PAttr) + // cuda2hipRename["CUdevice_P2PAttribute_enum"] = {"hipDeviceP2PAttribute", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_DEVICE_P2P_ATTRIBUTE_PERFORMANCE_RANK"] = {"hipDeviceP2PAttributePerformanceRank", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaDevP2PAttrPerformanceRank = 0x01) + cuda2hipRename["CU_DEVICE_P2P_ATTRIBUTE_ACCESS_SUPPORTED"] = {"hipDeviceP2PAttributeAccessSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaDevP2PAttrAccessSupported = 0x02) + cuda2hipRename["CU_DEVICE_P2P_ATTRIBUTE_NATIVE_ATOMIC_SUPPORTED"] = {"hipDeviceP2PAttributeNativeAtomicSupported", CONV_TYPE, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaDevP2PAttrNativeAtomicSupported = 0x03) // Events // pointer to CUevent_st @@ -993,6 +1000,24 @@ struct cuda2hipMap { cuda2hipRename["cuFuncSetSharedMemConfig"] = {"hipFuncSetSharedMemConfig", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuLaunchKernel"] = {"hipModuleLaunchKernel", CONV_MODULE, API_DRIVER}; + // Execution Control [DEPRECATED] + cuda2hipRename["cuFuncSetBlockShape"] = {"hipFuncSetBlockShape", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuFuncSetSharedSize"] = {"hipFuncSetSharedSize", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLaunch"] = {"hipLaunch", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaLaunch) + cuda2hipRename["cuLaunchGrid"] = {"hipLaunchGrid", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuLaunchGridAsync"] = {"hipLaunchGridAsync", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuParamSetf"] = {"hipParamSetf", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuParamSeti"] = {"hipParamSeti", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuParamSetSize"] = {"hipParamSetSize", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuParamSetSize"] = {"hipParamSetSize", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuParamSetv"] = {"hipParamSetv", CONV_MODULE, API_DRIVER, HIP_UNSUPPORTED}; + + // Occupancy + cuda2hipRename["cuOccupancyMaxActiveBlocksPerMultiprocessor"] = {"hipOccupancyMaxActiveBlocksPerMultiprocessor", CONV_OCCUPANCY, API_DRIVER}; // API_Runtime ANALOGUE (cudaOccupancyMaxActiveBlocksPerMultiprocessor) + cuda2hipRename["cuOccupancyMaxActiveBlocksPerMultiprocessorWithFlags"] = {"hipOccupancyMaxActiveBlocksPerMultiprocessorWithFlags", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags) + cuda2hipRename["cuOccupancyMaxPotentialBlockSize"] = {"hipOccupancyMaxPotentialBlockSize", CONV_OCCUPANCY, API_DRIVER}; // API_Runtime ANALOGUE (cudaOccupancyMaxPotentialBlockSize) + cuda2hipRename["cuOccupancyMaxPotentialBlockSizeWithFlags"] = {"hipOccupancyMaxPotentialBlockSizeWithFlags", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaOccupancyMaxPotentialBlockSizeWithFlags) + // Streams cuda2hipRename["cuStreamAddCallback"] = {"hipStreamAddCallback", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuStreamAttachMemAsync"] = {"hipStreamAttachMemAsync", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; @@ -1028,10 +1053,10 @@ struct cuda2hipMap { cuda2hipRename["cuMemcpy2DAsync"] = {"hipMemcpy2DAsync__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy2DAsync due to different signatures cuda2hipRename["cuMemcpy2DUnaligned"] = {"hipMemcpy2DUnaligned", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemcpy3D"] = {"hipMemcpy3D__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3D due to different signatures - cuda2hipRename["cuMemcpy3DAsync"] = {"hipMemcpy3DAsync__", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3DAsync due to different signatures - cuda2hipRename["cuMemcpy3DPeer"] = {"hipMemcpy3DPeer__", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3DPeer due to different signatures - cuda2hipRename["cuMemcpy3DPeerAsync"] = {"hipMemcpy3DPeerAsync__", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3DPeerAsync due to different signatures - cuda2hipRename["cuMemcpyAsync"] = {"hipMemcpyAsync__", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // Not equal to cudaMemcpyAsync due to different signatures + cuda2hipRename["cuMemcpy3DAsync"] = {"hipMemcpy3DAsync__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3DAsync due to different signatures + cuda2hipRename["cuMemcpy3DPeer"] = {"hipMemcpy3DPeer__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3DPeer due to different signatures + cuda2hipRename["cuMemcpy3DPeerAsync"] = {"hipMemcpy3DPeerAsync__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaMemcpy3DPeerAsync due to different signatures + cuda2hipRename["cuMemcpyAsync"] = {"hipMemcpyAsync__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaMemcpyAsync due to different signatures cuda2hipRename["cuMemcpyAtoA"] = {"hipMemcpyAtoA", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemcpyAtoD"] = {"hipMemcpyAtoD", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemcpyAtoH"] = {"hipMemcpyAtoH", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; @@ -1056,30 +1081,30 @@ struct cuda2hipMap { cuda2hipRename["cuMemHostGetFlags"] = {"hipMemHostGetFlags", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemHostRegister_v2"] = {"hipHostRegister", CONV_MEM, API_DRIVER}; // API_Runtime ANALOGUE (cudaHostAlloc) cuda2hipRename["cuMemHostUnregister"] = {"hipHostUnregister", CONV_MEM, API_DRIVER}; // API_Runtime ANALOGUE (cudaHostUnregister) - cuda2hipRename["cuMemsetD16_v2"] = {"hipMemsetD16", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD16Async"] = {"hipMemsetD16Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D16_v2"] = {"hipMemsetD2D16", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D16Async"] = {"hipMemsetD2D16Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D32_v2"] = {"hipMemsetD2D32", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D32Async"] = {"hipMemsetD2D32Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D8_v2"] = {"hipMemsetD2D8", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD2D8Async"] = {"hipMemsetD2D8Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD16_v2"] = {"hipMemsetD16", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD16Async"] = {"hipMemsetD16Async", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D16_v2"] = {"hipMemsetD2D16", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D16Async"] = {"hipMemsetD2D16Async", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D32_v2"] = {"hipMemsetD2D32", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D32Async"] = {"hipMemsetD2D32Async", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D8_v2"] = {"hipMemsetD2D8", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD2D8Async"] = {"hipMemsetD2D8Async", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuMemsetD32_v2"] = {"hipMemset", CONV_MEM, API_DRIVER}; // API_Runtime ANALOGUE (cudaMemset) cuda2hipRename["cuMemsetD32Async"] = {"hipMemsetAsync", CONV_MEM, API_DRIVER}; // API_Runtime ANALOGUE (cudaMemsetAsync) - cuda2hipRename["cuMemsetD8_v2"] = {"hipMemsetD8", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMemsetD8Async"] = {"hipMemsetD8Async", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMipmappedArrayCreate"] = {"hipMipmappedArrayCreate", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMipmappedArrayDestroy"] = {"hipMipmappedArrayDestroy", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cuMipmappedArrayGetLevel"] = {"hipMipmappedArrayGetLevel", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD8_v2"] = {"hipMemsetD8", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemsetD8Async"] = {"hipMemsetD8Async", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMipmappedArrayCreate"] = {"hipMipmappedArrayCreate", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMipmappedArrayDestroy"] = {"hipMipmappedArrayDestroy", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuMipmappedArrayGetLevel"] = {"hipMipmappedArrayGetLevel", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Unified Addressing - cuda2hipRename["cuMemPrefetchAsync"] = {"hipMemPrefetchAsync__", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // no API_Runtime ANALOGUE (cudaMemPrefetchAsync has different signature) - cuda2hipRename["cuMemAdvise"] = {"hipMemAdvise", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // API_Runtime ANALOGUE (cudaMemAdvise) - cuda2hipRename["cuMemRangeGetAttribute"] = {"hipMemRangeGetAttribute", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // API_Runtime ANALOGUE (cudaMemRangeGetAttribute) - cuda2hipRename["cuMemRangeGetAttributes"] = {"hipMemRangeGetAttributes", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // API_Runtime ANALOGUE (cudaMemRangeGetAttributes) - cuda2hipRename["cuPointerGetAttribute"] = {"hipPointerGetAttribute", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cuPointerGetAttributes"] = {"hipPointerGetAttributes", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cuPointerSetAttribute"] = {"hipPointerSetAttribute", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cuMemPrefetchAsync"] = {"hipMemPrefetchAsync__", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // no API_Runtime ANALOGUE (cudaMemPrefetchAsync has different signature) + cuda2hipRename["cuMemAdvise"] = {"hipMemAdvise", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // API_Runtime ANALOGUE (cudaMemAdvise) + cuda2hipRename["cuMemRangeGetAttribute"] = {"hipMemRangeGetAttribute", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // API_Runtime ANALOGUE (cudaMemRangeGetAttribute) + cuda2hipRename["cuMemRangeGetAttributes"] = {"hipMemRangeGetAttributes", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // API_Runtime ANALOGUE (cudaMemRangeGetAttributes) + cuda2hipRename["cuPointerGetAttribute"] = {"hipPointerGetAttribute", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuPointerGetAttributes"] = {"hipPointerGetAttributes", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuPointerSetAttribute"] = {"hipPointerSetAttribute", CONV_MEM, API_DRIVER, HIP_UNSUPPORTED}; // Texture Reference Mngmnt // Texture reference filtering modes @@ -1089,17 +1114,217 @@ struct cuda2hipMap { cuda2hipRename["CU_TR_FILTER_MODE_POINT"] = {"hipFilterModePoint", CONV_TEX, API_DRIVER}; // 0 // API_Runtime ANALOGUE (cudaFilterModePoint = 0) cuda2hipRename["CU_TR_FILTER_MODE_LINEAR"] = {"hipFilterModeLinear", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 1 // API_Runtime ANALOGUE (cudaFilterModeLinear = 1) - cuda2hipRename["cuTexRefSetBorderColor"] = {"hipTexRefSetBorderColor", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // no API_Runtime ANALOGUE - cuda2hipRename["cuTexRefGetBorderColor"] = {"hipTexRefGetBorderColor", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // no API_Runtime ANALOGUE + cuda2hipRename["cuTexRefGetAddress"] = {"hipTexRefGetAddress", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetAddressMode"] = {"hipTexRefGetAddressMode", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetArray"] = {"hipTexRefGetArray", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetBorderColor"] = {"hipTexRefGetBorderColor", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // no API_Runtime ANALOGUE + cuda2hipRename["cuTexRefGetFilterMode"] = {"hipTexRefGetFilterMode", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetFlags"] = {"hipTexRefGetFlags", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetFormat"] = {"hipTexRefGetFormat", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetMaxAnisotropy"] = {"hipTexRefGetMaxAnisotropy", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetMipmapFilterMode"] = {"hipTexRefGetMipmapFilterMode", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetMipmapLevelBias"] = {"hipTexRefGetMipmapLevelBias", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetMipmapLevelClamp"] = {"hipTexRefGetMipmapLevelClamp", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefGetMipmappedArray"] = {"hipTexRefGetMipmappedArray", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetAddress"] = {"hipTexRefSetAddress", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetAddress2D"] = {"hipTexRefSetAddress2D", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetAddressMode"] = {"hipTexRefSetAddressMode", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetArray"] = {"hipTexRefSetArray", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetBorderColor"] = {"hipTexRefSetBorderColor", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // [CUDA 8.0.44] // no API_Runtime ANALOGUE + cuda2hipRename["cuTexRefSetFilterMode"] = {"hipTexRefSetFilterMode", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetFlags"] = {"hipTexRefSetFlags", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetFormat"] = {"hipTexRefSetFormat", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetMaxAnisotropy"] = {"hipTexRefSetMaxAnisotropy", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetMipmapFilterMode"] = {"hipTexRefSetMipmapFilterMode", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetMipmapLevelBias"] = {"hipTexRefSetMipmapLevelBias", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetMipmapLevelClamp"] = {"hipTexRefSetMipmapLevelClamp", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefSetMipmappedArray"] = {"hipTexRefSetMipmappedArray", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + + // Texture Reference Mngmnt [DEPRECATED] + cuda2hipRename["cuTexRefCreate"] = {"hipTexRefCreate", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexRefDestroy"] = {"hipTexRefDestroy", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + + // Surface Reference Mngmnt + cuda2hipRename["cuSurfRefGetArray"] = {"hipSurfRefGetArray", CONV_SURFACE, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuSurfRefSetArray"] = {"hipSurfRefSetArray", CONV_SURFACE, API_DRIVER, HIP_UNSUPPORTED}; + + // Texture Object Mngmnt + cuda2hipRename["cuTexObjectCreate"] = {"hipTexObjectCreate", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexObjectDestroy"] = {"hipTexObjectDestroy", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexObjectGetResourceDesc"] = {"hipTexObjectGetResourceDesc", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexObjectGetResourceViewDesc"] = {"hipTexObjectGetResourceViewDesc", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuTexObjectGetTextureDesc"] = {"hipTexObjectGetTextureDesc", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + + // Surface Object Mngmnt + cuda2hipRename["cuSurfObjectCreate"] = {"hipSurfObjectCreate", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuSurfObjectDestroy"] = {"hipSurfObjectDestroy", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuSurfObjectGetResourceDesc"] = {"hipSurfObjectGetResourceDesc", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; + + // Graphics Interoperability + cuda2hipRename["cuGraphicsMapResources"] = {"hipGraphicsMapResources", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsMapResources) + cuda2hipRename["cuGraphicsResourceGetMappedMipmappedArray"] = {"hipGraphicsResourceGetMappedMipmappedArray", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsResourceGetMappedMipmappedArray) + cuda2hipRename["cuGraphicsResourceGetMappedPointer"] = {"hipGraphicsResourceGetMappedPointer", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsResourceGetMappedPointer) + cuda2hipRename["cuGraphicsResourceSetMapFlags"] = {"hipGraphicsResourceSetMapFlags", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsResourceSetMapFlags) + cuda2hipRename["cuGraphicsSubResourceGetMappedArray"] = {"hipGraphicsSubResourceGetMappedArray", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsSubResourceGetMappedArray) + cuda2hipRename["cuGraphicsUnmapResources"] = {"hipGraphicsUnmapResources", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsUnmapResources) + cuda2hipRename["cuGraphicsUnregisterResource"] = {"hipGraphicsUnregisterResource", CONV_GRAPHICS, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsUnregisterResource) // Profiler - // unsupported yet by HIP - cuda2hipRename["cuProfilerInitialize"] = {"hipProfilerInitialize", CONV_OTHER, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuProfilerInitialize"] = {"hipProfilerInitialize", CONV_OTHER, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaProfilerInitialize) + cuda2hipRename["cuProfilerStart"] = {"hipProfilerStart", CONV_OTHER, API_DRIVER}; // API_Runtime ANALOGUE (cudaProfilerStart) + cuda2hipRename["cuProfilerStop"] = {"hipProfilerStop", CONV_OTHER, API_DRIVER}; // API_Runtime ANALOGUE (cudaProfilerStop) - cuda2hipRename["cuProfilerStart"] = {"hipProfilerStart", CONV_OTHER, API_DRIVER}; - cuda2hipRename["cuProfilerStop"] = {"hipProfilerStop", CONV_OTHER, API_DRIVER}; + // OpenGL Interoperability + // enum CUGLDeviceList/CUGLDeviceList_enum + cuda2hipRename["CUGLDeviceList"] = {"hipGLDeviceList", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLDeviceList) + // cuda2hipRename["CUGLDeviceList_enum"] = {"hipGLDeviceList", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_GL_DEVICE_LIST_ALL"] = {"HIP_GL_DEVICE_LIST_ALL", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaGLDeviceListAll) + cuda2hipRename["CU_GL_DEVICE_LIST_CURRENT_FRAME"] = {"HIP_GL_DEVICE_LIST_CURRENT_FRAME", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaGLDeviceListCurrentFrame) + cuda2hipRename["CU_GL_DEVICE_LIST_NEXT_FRAME"] = {"HIP_GL_DEVICE_LIST_NEXT_FRAME", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaGLDeviceListNextFrame) - /////////////////////////////// CUDA RT API /////////////////////////////// + cuda2hipRename["cuGLGetDevices"] = {"hipGLGetDevices", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLGetDevices) + cuda2hipRename["cuGraphicsGLRegisterBuffer"] = {"hipGraphicsGLRegisterBuffer", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsGLRegisterBuffer) + cuda2hipRename["cuGraphicsGLRegisterImage"] = {"hipGraphicsGLRegisterImage", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsGLRegisterImage) + cuda2hipRename["cuWGLGetDevice"] = {"hipWGLGetDevice", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaWGLGetDevice) + + // OpenGL Interoperability [DEPRECATED] + // enum CUGLmap_flags/CUGLmap_flags_enum + cuda2hipRename["CUGLmap_flags"] = {"hipGLMapFlags", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLMapFlags) + // cuda2hipRename["CUGLmap_flags_enum"] = {"hipGLMapFlags", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_GL_MAP_RESOURCE_FLAGS_NONE"] = {"HIP_GL_MAP_RESOURCE_FLAGS_NONE", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaGLMapFlagsNone) + cuda2hipRename["CU_GL_MAP_RESOURCE_FLAGS_READ_ONLY"] = {"HIP_GL_MAP_RESOURCE_FLAGS_READ_ONLY", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaGLMapFlagsReadOnly) + cuda2hipRename["CU_GL_MAP_RESOURCE_FLAGS_WRITE_DISCARD"] = {"HIP_GL_MAP_RESOURCE_FLAGS_WRITE_DISCARD", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaGLMapFlagsWriteDiscard) + + cuda2hipRename["cuGLCtxCreate"] = {"hipGLCtxCreate", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuGLInit"] = {"hipGLInit", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuGLMapBufferObject"] = {"hipGLMapBufferObject", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaGLMapBufferObject due to different signatures + cuda2hipRename["cuGLMapBufferObjectAsync"] = {"hipGLMapBufferObjectAsync", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaGLMapBufferObjectAsync due to different signatures + cuda2hipRename["cuGLRegisterBufferObject"] = {"hipGLRegisterBufferObject", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLRegisterBufferObject) + cuda2hipRename["cuGLSetBufferObjectMapFlags"] = {"hipGLSetBufferObjectMapFlags", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLSetBufferObjectMapFlags) + cuda2hipRename["cuGLUnmapBufferObject"] = {"hipGLUnmapBufferObject", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLUnmapBufferObject) + cuda2hipRename["cuGLUnmapBufferObjectAsync"] = {"hipGLUnmapBufferObjectAsync", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLUnmapBufferObjectAsync) + cuda2hipRename["cuGLUnregisterBufferObject"] = {"hipGLUnregisterBufferObject", CONV_GL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGLUnregisterBufferObject) + + // Direct3D 9 Interoperability + // enum CUd3d9DeviceList/CUd3d9DeviceList_enum + cuda2hipRename["CUd3d9DeviceList"] = {"hipD3D9DeviceList", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9DeviceList) + // cuda2hipRename["CUd3d9DeviceList_enum"] = {"hipD3D9DeviceList", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D9_DEVICE_LIST_ALL"] = {"HIP_D3D9_DEVICE_LIST_ALL", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D9DeviceListAll) + cuda2hipRename["CU_D3D9_DEVICE_LIST_CURRENT_FRAME"] = {"HIP_D3D9_DEVICE_LIST_CURRENT_FRAME", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaD3D9DeviceListCurrentFrame) + cuda2hipRename["CU_D3D9_DEVICE_LIST_NEXT_FRAME"] = {"HIP_D3D9_DEVICE_LIST_NEXT_FRAME", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaD3D9DeviceListNextFrame) + + cuda2hipRename["cuD3D9CtxCreate"] = {"hipD3D9CtxCreate", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuD3D9CtxCreateOnDevice"] = {"hipD3D9CtxCreateOnDevice", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuD3D9GetDevice"] = {"hipD3D9GetDevice", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9GetDevice) + cuda2hipRename["cuD3D9GetDevices"] = {"hipD3D9GetDevices", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9GetDevices) + cuda2hipRename["cuD3D9GetDirect3DDevice"] = {"hipD3D9GetDirect3DDevice", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9GetDirect3DDevice) + cuda2hipRename["cuGraphicsD3D9RegisterResource"] = {"hipGraphicsD3D9RegisterResource", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsD3D9RegisterResource) + + // Direct3D 9 Interoperability [DEPRECATED] + // enum CUd3d9map_flags/CUd3d9map_flags_enum + cuda2hipRename["CUd3d9map_flags"] = {"hipD3D9MapFlags", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9MapFlags) + // cuda2hipRename["CUd3d9map_flags_enum"] = {"hipD3D9MapFlags", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D9_MAPRESOURCE_FLAGS_NONE"] = {"HIP_D3D9_MAPRESOURCE_FLAGS_NONE", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaD3D9MapFlagsNone) + cuda2hipRename["CU_D3D9_MAPRESOURCE_FLAGS_READONLY"] = {"HIP_D3D9_MAPRESOURCE_FLAGS_READONLY", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D9MapFlagsReadOnly) + cuda2hipRename["CU_D3D9_MAPRESOURCE_FLAGS_WRITEDISCARD"] = {"HIP_D3D9_MAPRESOURCE_FLAGS_WRITEDISCARD", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaD3D9MapFlagsWriteDiscard) + + // enum CUd3d9register_flags/CUd3d9register_flags_enum + cuda2hipRename["CUd3d9register_flags"] = {"hipD3D9RegisterFlags", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9RegisterFlags) + // cuda2hipRename["CUd3d9register_flags_enum"] = {"hipD3D9RegisterFlags", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D9_REGISTER_FLAGS_NONE"] = {"HIP_D3D9_REGISTER_FLAGS_NONE", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaD3D9RegisterFlagsNone) + cuda2hipRename["CU_D3D9_REGISTER_FLAGS_ARRAY"] = {"HIP_D3D9_REGISTER_FLAGS_ARRAY", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D9RegisterFlagsArray) + + cuda2hipRename["cuD3D9MapResources"] = {"hipD3D9MapResources", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9MapResources) + cuda2hipRename["cuD3D9RegisterResource"] = {"hipD3D9RegisterResource", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9RegisterResource) + cuda2hipRename["cuD3D9ResourceGetMappedArray"] = {"hipD3D9ResourceGetMappedArray", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9ResourceGetMappedArray) + cuda2hipRename["cuD3D9ResourceGetMappedPitch"] = {"hipD3D9ResourceGetMappedPitch", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9ResourceGetMappedPitch) + cuda2hipRename["cuD3D9ResourceGetMappedPointer"] = {"hipD3D9ResourceGetMappedPointer", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9ResourceGetMappedPointer) + cuda2hipRename["cuD3D9ResourceGetMappedSize"] = {"hipD3D9ResourceGetMappedSize", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9ResourceGetMappedSize) + cuda2hipRename["cuD3D9ResourceGetSurfaceDimensions"] = {"hipD3D9ResourceGetSurfaceDimensions", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9ResourceGetSurfaceDimensions) + cuda2hipRename["cuD3D9ResourceSetMapFlags"] = {"hipD3D9ResourceSetMapFlags", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9ResourceSetMapFlags) + cuda2hipRename["cuD3D9UnmapResources"] = {"hipD3D9UnmapResources", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9UnmapResources) + cuda2hipRename["cuD3D9UnregisterResource"] = {"hipD3D9UnregisterResource", CONV_D3D9, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D9UnregisterResource) + + // Direct3D 10 Interoperability + // enum CUd3d10DeviceList/CUd3d10DeviceList_enum + cuda2hipRename["CUd3d10DeviceList"] = {"hipd3d10DeviceList", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10DeviceList) + // cuda2hipRename["CUd3d10DeviceList_enum"] = {"hipD3D10DeviceList", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D10_DEVICE_LIST_ALL"] = {"HIP_D3D10_DEVICE_LIST_ALL", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D10DeviceListAll) + cuda2hipRename["CU_D3D10_DEVICE_LIST_CURRENT_FRAME"] = {"HIP_D3D10_DEVICE_LIST_CURRENT_FRAME", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaD3D10DeviceListCurrentFrame) + cuda2hipRename["CU_D3D10_DEVICE_LIST_NEXT_FRAME"] = {"HIP_D3D10_DEVICE_LIST_NEXT_FRAME", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaD3D10DeviceListNextFrame) + + cuda2hipRename["cuD3D10GetDevice"] = {"hipD3D10GetDevice", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10GetDevice) + cuda2hipRename["cuD3D10GetDevices"] = {"hipD3D10GetDevices", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10GetDevices) + cuda2hipRename["cuGraphicsD3D10RegisterResource"] = {"hipGraphicsD3D10RegisterResource", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsD3D10RegisterResource) + + // Direct3D 10 Interoperability [DEPRECATED] + // enum CUd3d10map_flags/CUd3d10map_flags_enum + cuda2hipRename["CUd3d10map_flags"] = {"hipD3D10MapFlags", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10MapFlags) + // cuda2hipRename["CUd3d10map_flags_enum"] = {"hipD3D10MapFlags", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D10_MAPRESOURCE_FLAGS_NONE"] = {"HIP_D3D10_MAPRESOURCE_FLAGS_NONE", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaD3D10MapFlagsNone) + cuda2hipRename["CU_D3D10_MAPRESOURCE_FLAGS_READONLY"] = {"HIP_D3D10_MAPRESOURCE_FLAGS_READONLY", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D10MapFlagsReadOnly) + cuda2hipRename["CU_D3D10_MAPRESOURCE_FLAGS_WRITEDISCARD"] = {"HIP_D3D10_MAPRESOURCE_FLAGS_WRITEDISCARD", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaD3D10MapFlagsWriteDiscard) + + // enum CUd3d10register_flags/CUd3d10register_flags_enum + cuda2hipRename["CUd3d10register_flags"] = {"hipD3D10RegisterFlags", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10RegisterFlags) + // cuda2hipRename["CUd3d10register_flags_enum"] = {"hipD3D10RegisterFlags", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D10_REGISTER_FLAGS_NONE"] = {"HIP_D3D10_REGISTER_FLAGS_NONE", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x00 // API_Runtime ANALOGUE (cudaD3D10RegisterFlagsNone) + cuda2hipRename["CU_D3D10_REGISTER_FLAGS_ARRAY"] = {"HIP_D3D10_REGISTER_FLAGS_ARRAY", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D10RegisterFlagsArray) + + cuda2hipRename["cuD3D10CtxCreate"] = {"hipD3D10CtxCreate", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuD3D10CtxCreateOnDevice"] = {"hipD3D10CtxCreateOnDevice", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuD3D10GetDirect3DDevice"] = {"hipD3D10GetDirect3DDevice", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10GetDirect3DDevice) + cuda2hipRename["cuD3D10MapResources"] = {"hipD3D10MapResources", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10MapResources) + cuda2hipRename["cuD3D10RegisterResource"] = {"hipD3D10RegisterResource", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10RegisterResource) + cuda2hipRename["cuD3D10ResourceGetMappedArray"] = {"hipD3D10ResourceGetMappedArray", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10ResourceGetMappedArray) + cuda2hipRename["cuD3D10ResourceGetMappedPitch"] = {"hipD3D10ResourceGetMappedPitch", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10ResourceGetMappedPitch) + cuda2hipRename["cuD3D10ResourceGetMappedPointer"] = {"hipD3D10ResourceGetMappedPointer", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10ResourceGetMappedPointer) + cuda2hipRename["cuD3D10ResourceGetMappedSize"] = {"hipD3D10ResourceGetMappedSize", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10ResourceGetMappedSize) + cuda2hipRename["cuD3D10ResourceGetSurfaceDimensions"] = {"hipD3D10ResourceGetSurfaceDimensions", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10ResourceGetSurfaceDimensions) + cuda2hipRename["cuD310ResourceSetMapFlags"] = {"hipD3D10ResourceSetMapFlags", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10ResourceSetMapFlags) + cuda2hipRename["cuD3D10UnmapResources"] = {"hipD3D10UnmapResources", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10UnmapResources) + cuda2hipRename["cuD3D10UnregisterResource"] = {"hipD3D10UnregisterResource", CONV_D3D10, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D10UnregisterResource) + + // Direct3D 11 Interoperability + // enum CUd3d11DeviceList/CUd3d11DeviceList_enum + cuda2hipRename["CUd3d11DeviceList"] = {"hipd3d11DeviceList", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D11DeviceList) + // cuda2hipRename["CUd3d11DeviceList_enum"] = {"hipD3D11DeviceList", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["CU_D3D11_DEVICE_LIST_ALL"] = {"HIP_D3D11_DEVICE_LIST_ALL", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // 0x01 // API_Runtime ANALOGUE (cudaD3D11DeviceListAll) + cuda2hipRename["CU_D3D11_DEVICE_LIST_CURRENT_FRAME"] = {"HIP_D3D11_DEVICE_LIST_CURRENT_FRAME", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // 0x02 // API_Runtime ANALOGUE (cudaD3D11DeviceListCurrentFrame) + cuda2hipRename["CU_D3D11_DEVICE_LIST_NEXT_FRAME"] = {"HIP_D3D11_DEVICE_LIST_NEXT_FRAME", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // 0x03 // API_Runtime ANALOGUE (cudaD3D11DeviceListNextFrame) + + cuda2hipRename["cuD3D11GetDevice"] = {"hipD3D11GetDevice", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D11GetDevice) + cuda2hipRename["cuD3D11GetDevices"] = {"hipD3D11GetDevices", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D11GetDevices) + cuda2hipRename["cuGraphicsD3D11RegisterResource"] = {"hipGraphicsD3D11RegisterResource", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsD3D11RegisterResource) + + // Direct3D 11 Interoperability [DEPRECATED] + cuda2hipRename["cuD3D11CtxCreate"] = {"hipD3D11CtxCreate", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuD3D11CtxCreateOnDevice"] = {"hipD3D11CtxCreateOnDevice", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuD3D11GetDirect3DDevice"] = {"hipD3D11GetDirect3DDevice", CONV_D3D11, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaD3D11GetDirect3DDevice) + + // VDPAU Interoperability + cuda2hipRename["cuGraphicsVDPAURegisterOutputSurface"] = {"hipGraphicsVDPAURegisterOutputSurface", CONV_VDPAU, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsVDPAURegisterOutputSurface) + cuda2hipRename["cuGraphicsVDPAURegisterVideoSurface"] = {"hipGraphicsVDPAURegisterVideoSurface", CONV_VDPAU, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsVDPAURegisterVideoSurface) + cuda2hipRename["cuVDPAUGetDevice"] = {"hipVDPAUGetDevice", CONV_VDPAU, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaVDPAUGetDevice) + cuda2hipRename["cuVDPAUCtxCreate"] = {"hipVDPAUCtxCreate", CONV_VDPAU, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + + // EGL Interoperability + cuda2hipRename["CUeglStreamConnection_st"] = {"hipEglStreamConnection", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEglStreamConnection) + cuda2hipRename["CUeglStreamConnection"] = {"hipEglStreamConnection", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEglStreamConnection) + + cuda2hipRename["cuEGLStreamConsumerAcquireFrame"] = {"hipEGLStreamConsumerAcquireFrame", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamConsumerAcquireFrame) + cuda2hipRename["cuEGLStreamConsumerConnect"] = {"hipEGLStreamConsumerConnect", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamConsumerConnect) + cuda2hipRename["cuEGLStreamConsumerConnectWithFlags"] = {"hipEGLStreamConsumerConnectWithFlags", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamConsumerConnectWithFlags) + cuda2hipRename["cuEGLStreamConsumerDisconnect"] = {"hipEGLStreamConsumerDisconnect", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // no API_Runtime ANALOGUE + cuda2hipRename["cuEGLStreamConsumerReleaseFrame"] = {"hipEGLStreamConsumerReleaseFrame", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamConsumerReleaseFrame) + cuda2hipRename["cuEGLStreamProducerConnect"] = {"hipEGLStreamProducerConnect", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamProducerConnect) + cuda2hipRename["cuEGLStreamProducerDisconnect"] = {"hipEGLStreamProducerDisconnect", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamProducerDisconnect) + cuda2hipRename["cuEGLStreamProducerPresentFrame"] = {"hipEGLStreamProducerPresentFrame", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamProducerPresentFrame) + cuda2hipRename["cuEGLStreamProducerReturnFrame"] = {"hipEGLStreamProducerReturnFrame", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaEGLStreamProducerReturnFrame) + cuda2hipRename["cuGraphicsEGLRegisterImage"] = {"hipGraphicsEGLRegisterImage", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsEGLRegisterImage) + cuda2hipRename["cuGraphicsResourceGetMappedEglFrame"] = {"hipGraphicsResourceGetMappedEglFrame", CONV_EGL, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaGraphicsResourceGetMappedEglFrame) + +/////////////////////////////// CUDA RT API /////////////////////////////// // Data types // unsupported yet by HIP [CUDA 8.0.44] cuda2hipRename["cudaDataType_t"] = {"hipDataType_t", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; @@ -1138,10 +1363,10 @@ struct cuda2hipMap { cuda2hipRename["cudaOccupancyDisableCachingOverride"] = {"hipOccupancyDisableCachingOverride", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE = 0x1) // Error API - cuda2hipRename["cudaGetLastError"] = {"hipGetLastError", CONV_ERR, API_RUNTIME}; - cuda2hipRename["cudaPeekAtLastError"] = {"hipPeekAtLastError", CONV_ERR, API_RUNTIME}; - cuda2hipRename["cudaGetErrorName"] = {"hipGetErrorName", CONV_ERR, API_RUNTIME}; - cuda2hipRename["cudaGetErrorString"] = {"hipGetErrorString", CONV_ERR, API_RUNTIME}; + cuda2hipRename["cudaGetLastError"] = {"hipGetLastError", CONV_ERROR, API_RUNTIME}; + cuda2hipRename["cudaPeekAtLastError"] = {"hipPeekAtLastError", CONV_ERROR, API_RUNTIME}; + cuda2hipRename["cudaGetErrorName"] = {"hipGetErrorName", CONV_ERROR, API_RUNTIME}; + cuda2hipRename["cudaGetErrorString"] = {"hipGetErrorString", CONV_ERROR, API_RUNTIME}; // Arrays cuda2hipRename["cudaArray"] = {"hipArray", CONV_MEM, API_RUNTIME}; @@ -1324,7 +1549,6 @@ struct cuda2hipMap { cuda2hipRename["cudaStream_t"] = {"hipStream_t", CONV_TYPE, API_RUNTIME}; cuda2hipRename["cudaStreamCreate"] = {"hipStreamCreate", CONV_STREAM, API_RUNTIME}; cuda2hipRename["cudaStreamCreateWithFlags"] = {"hipStreamCreateWithFlags", CONV_STREAM, API_RUNTIME}; - // unsupported yet by HIP cuda2hipRename["cudaStreamCreateWithPriority"] = {"hipStreamCreateWithPriority", CONV_STREAM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaStreamDestroy"] = {"hipStreamDestroy", CONV_STREAM, API_RUNTIME}; cuda2hipRename["cudaStreamWaitEvent"] = {"hipStreamWaitEvent", CONV_STREAM, API_RUNTIME}; @@ -1332,124 +1556,127 @@ struct cuda2hipMap { cuda2hipRename["cudaStreamGetFlags"] = {"hipStreamGetFlags", CONV_STREAM, API_RUNTIME}; cuda2hipRename["cudaStreamQuery"] = {"hipStreamQuery", CONV_STREAM, API_RUNTIME}; cuda2hipRename["cudaStreamAddCallback"] = {"hipStreamAddCallback", CONV_STREAM, API_RUNTIME}; - // unsupported yet by HIP cuda2hipRename["cudaStreamAttachMemAsync"] = {"hipStreamAttachMemAsync", CONV_STREAM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaStreamGetPriority"] = {"hipStreamGetPriority", CONV_STREAM, API_RUNTIME, HIP_UNSUPPORTED}; // Stream Flags - cuda2hipRename["cudaStreamDefault"] = {"hipStreamDefault", CONV_STREAM, API_RUNTIME}; - cuda2hipRename["cudaStreamNonBlocking"] = {"hipStreamNonBlocking", CONV_STREAM, API_RUNTIME}; + cuda2hipRename["cudaStreamDefault"] = {"hipStreamDefault", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaStreamNonBlocking"] = {"hipStreamNonBlocking", CONV_TYPE, API_RUNTIME}; // Other synchronization - cuda2hipRename["cudaDeviceSynchronize"] = {"hipDeviceSynchronize", CONV_DEV, API_RUNTIME}; - // translate deprecated cudaThreadSynchronize - cuda2hipRename["cudaThreadSynchronize"] = {"hipDeviceSynchronize", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceReset"] = {"hipDeviceReset", CONV_DEV, API_RUNTIME}; - // translate deprecated cudaThreadExit - cuda2hipRename["cudaThreadExit"] = {"hipDeviceReset", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaSetDevice"] = {"hipSetDevice", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaGetDevice"] = {"hipGetDevice", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaGetDeviceCount"] = {"hipGetDeviceCount", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaChooseDevice"] = {"hipChooseDevice", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaDeviceSynchronize"] = {"hipDeviceSynchronize", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaDeviceReset"] = {"hipDeviceReset", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaSetDevice"] = {"hipSetDevice", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaGetDevice"] = {"hipGetDevice", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaGetDeviceCount"] = {"hipGetDeviceCount", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaChooseDevice"] = {"hipChooseDevice", CONV_DEVICE, API_RUNTIME}; + + // Thread Management + cuda2hipRename["cudaThreadExit"] = {"hipDeviceReset", CONV_THREAD, API_RUNTIME}; + cuda2hipRename["cudaThreadGetCacheConfig"] = {"hipDeviceGetCacheConfig", CONV_THREAD, API_RUNTIME}; + cuda2hipRename["cudaThreadGetLimit"] = {"hipThreadGetLimit", CONV_THREAD, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaThreadSetCacheConfig"] = {"hipDeviceSetCacheConfig", CONV_THREAD, API_RUNTIME}; + cuda2hipRename["cudaThreadSetLimit"] = {"hipThreadSetLimit", CONV_THREAD, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaThreadSynchronize"] = {"hipDeviceSynchronize", CONV_THREAD, API_RUNTIME}; // Attributes - cuda2hipRename["cudaDeviceGetAttribute"] = {"hipDeviceGetAttribute", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaDeviceGetAttribute"] = {"hipDeviceGetAttribute", CONV_DEVICE, API_RUNTIME}; cuda2hipRename["cudaDeviceAttr"] = {"hipDeviceAttribute_t", CONV_TYPE, API_RUNTIME}; // API_DRIVER ANALOGUE (CUdevice_attribute) - cuda2hipRename["cudaDevAttrMaxThreadsPerBlock"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_DEV, API_RUNTIME}; // 1 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK = 1) - cuda2hipRename["cudaDevAttrMaxBlockDimX"] = {"hipDeviceAttributeMaxBlockDimX", CONV_DEV, API_RUNTIME}; // 2 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X = 2) - cuda2hipRename["cudaDevAttrMaxBlockDimY"] = {"hipDeviceAttributeMaxBlockDimY", CONV_DEV, API_RUNTIME}; // 3 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y = 3) - cuda2hipRename["cudaDevAttrMaxBlockDimZ"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_DEV, API_RUNTIME}; // 4 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z = 4) - cuda2hipRename["cudaDevAttrMaxGridDimX"] = {"hipDeviceAttributeMaxGridDimX", CONV_DEV, API_RUNTIME}; // 5 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 5) - cuda2hipRename["cudaDevAttrMaxGridDimY"] = {"hipDeviceAttributeMaxGridDimY", CONV_DEV, API_RUNTIME}; // 6 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 6) - cuda2hipRename["cudaDevAttrMaxGridDimZ"] = {"hipDeviceAttributeMaxGridDimZ", CONV_DEV, API_RUNTIME}; // 7 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 7) - cuda2hipRename["cudaDevAttrMaxSharedMemoryPerBlock"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_DEV, API_RUNTIME}; // 8 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK = 8) - cuda2hipRename["cudaDevAttrTotalConstantMemory"] = {"hipDeviceAttributeTotalConstantMemory", CONV_DEV, API_RUNTIME}; // 9 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY =9) - cuda2hipRename["cudaDevAttrWarpSize"] = {"hipDeviceAttributeWarpSize", CONV_DEV, API_RUNTIME}; // 10 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_WARP_SIZE = 10) - cuda2hipRename["cudaDevAttrMaxPitch"] = {"hipDeviceAttributeMaxPitch", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 11 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_PITCH = 11) - cuda2hipRename["cudaDevAttrMaxRegistersPerBlock"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_DEV, API_RUNTIME}; // 12 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK = 12) - cuda2hipRename["cudaDevAttrClockRate"] = {"hipDeviceAttributeClockRate", CONV_DEV, API_RUNTIME}; // 13 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CLOCK_RATE = 13) - cuda2hipRename["cudaDevAttrTextureAlignment"] = {"hipDeviceAttributeTextureAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 14 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT = 14) + cuda2hipRename["cudaDevAttrMaxThreadsPerBlock"] = {"hipDeviceAttributeMaxThreadsPerBlock", CONV_TYPE, API_RUNTIME}; // 1 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_BLOCK = 1) + cuda2hipRename["cudaDevAttrMaxBlockDimX"] = {"hipDeviceAttributeMaxBlockDimX", CONV_TYPE, API_RUNTIME}; // 2 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_X = 2) + cuda2hipRename["cudaDevAttrMaxBlockDimY"] = {"hipDeviceAttributeMaxBlockDimY", CONV_TYPE, API_RUNTIME}; // 3 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Y = 3) + cuda2hipRename["cudaDevAttrMaxBlockDimZ"] = {"hipDeviceAttributeMaxBlockDimZ", CONV_TYPE, API_RUNTIME}; // 4 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_BLOCK_DIM_Z = 4) + cuda2hipRename["cudaDevAttrMaxGridDimX"] = {"hipDeviceAttributeMaxGridDimX", CONV_TYPE, API_RUNTIME}; // 5 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 5) + cuda2hipRename["cudaDevAttrMaxGridDimY"] = {"hipDeviceAttributeMaxGridDimY", CONV_TYPE, API_RUNTIME}; // 6 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 6) + cuda2hipRename["cudaDevAttrMaxGridDimZ"] = {"hipDeviceAttributeMaxGridDimZ", CONV_TYPE, API_RUNTIME}; // 7 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_GRID_DIM_X = 7) + cuda2hipRename["cudaDevAttrMaxSharedMemoryPerBlock"] = {"hipDeviceAttributeMaxSharedMemoryPerBlock", CONV_TYPE, API_RUNTIME}; // 8 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_BLOCK = 8) + cuda2hipRename["cudaDevAttrTotalConstantMemory"] = {"hipDeviceAttributeTotalConstantMemory", CONV_TYPE, API_RUNTIME}; // 9 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TOTAL_CONSTANT_MEMORY =9) + cuda2hipRename["cudaDevAttrWarpSize"] = {"hipDeviceAttributeWarpSize", CONV_TYPE, API_RUNTIME}; // 10 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_WARP_SIZE = 10) + cuda2hipRename["cudaDevAttrMaxPitch"] = {"hipDeviceAttributeMaxPitch", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 11 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_PITCH = 11) + cuda2hipRename["cudaDevAttrMaxRegistersPerBlock"] = {"hipDeviceAttributeMaxRegistersPerBlock", CONV_TYPE, API_RUNTIME}; // 12 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_BLOCK = 12) + cuda2hipRename["cudaDevAttrClockRate"] = {"hipDeviceAttributeClockRate", CONV_TYPE, API_RUNTIME}; // 13 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CLOCK_RATE = 13) + cuda2hipRename["cudaDevAttrTextureAlignment"] = {"hipDeviceAttributeTextureAlignment", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 14 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT = 14) // Is not deprecated as CUDA Driver's API analogue CU_DEVICE_ATTRIBUTE_GPU_OVERLAP - cuda2hipRename["cudaDevAttrGpuOverlap"] = {"hipDeviceAttributeGpuOverlap", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 15 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GPU_OVERLAP = 15) - cuda2hipRename["cudaDevAttrMultiProcessorCount"] = {"hipDeviceAttributeMultiprocessorCount", CONV_DEV, API_RUNTIME}; // 16 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT = 16) - cuda2hipRename["cudaDevAttrKernelExecTimeout"] = {"hipDeviceAttributeKernelExecTimeout", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 17 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT = 17) - cuda2hipRename["cudaDevAttrIntegrated"] = {"hipDeviceAttributeIntegrated", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 18 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_INTEGRATED = 18) - cuda2hipRename["cudaDevAttrCanMapHostMemory"] = {"hipDeviceAttributeCanMapHostMemory", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 19 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY = 19) - cuda2hipRename["cudaDevAttrComputeMode"] = {"hipDeviceAttributeComputeMode", CONV_DEV, API_RUNTIME}; // 20 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_MODE = 20) - cuda2hipRename["cudaDevAttrMaxTexture1DWidth"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 21 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH = 21) - cuda2hipRename["cudaDevAttrMaxTexture2DWidth"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 22 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH = 22) - cuda2hipRename["cudaDevAttrMaxTexture2DHeight"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 23 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT = 23) - cuda2hipRename["cudaDevAttrMaxTexture3DWidth"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 24 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH = 24) - cuda2hipRename["cudaDevAttrMaxTexture3DHeight"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 25 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT = 25) - cuda2hipRename["cudaDevAttrMaxTexture3DDepth"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 26 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH = 26) - cuda2hipRename["cudaDevAttrMaxTexture2DLayeredWidth"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 27 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH = 27) - cuda2hipRename["cudaDevAttrMaxTexture2DLayeredHeight"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 28 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT = 28) - cuda2hipRename["cudaDevAttrMaxTexture2DLayeredLayers"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 29 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS = 29) - cuda2hipRename["cudaDevAttrSurfaceAlignment"] = {"hipDeviceAttributeSurfaceAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 30 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT = 30) - cuda2hipRename["cudaDevAttrConcurrentKernels"] = {"hipDeviceAttributeConcurrentKernels", CONV_DEV, API_RUNTIME}; // 31 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS = 31) - cuda2hipRename["cudaDevAttrEccEnabled"] = {"hipDeviceAttributeEccEnabled", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 32 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_ECC_ENABLED = 32) - cuda2hipRename["cudaDevAttrPciBusId"] = {"hipDeviceAttributePciBusId", CONV_DEV, API_RUNTIME}; // 33 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_BUS_ID = 33) - cuda2hipRename["cudaDevAttrPciDeviceId"] = {"hipDeviceAttributePciDeviceId", CONV_DEV, API_RUNTIME}; // 34 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID = 34) - cuda2hipRename["cudaDevAttrTccDriver"] = {"hipDeviceAttributeTccDriver", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 35 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TCC_DRIVER = 35) - cuda2hipRename["cudaDevAttrMemoryClockRate"] = {"hipDeviceAttributeMemoryClockRate", CONV_DEV, API_RUNTIME}; // 36 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE = 36) - cuda2hipRename["cudaDevAttrGlobalMemoryBusWidth"] = {"hipDeviceAttributeMemoryBusWidth", CONV_DEV, API_RUNTIME}; // 37 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH = 37) - cuda2hipRename["cudaDevAttrL2CacheSize"] = {"hipDeviceAttributeL2CacheSize", CONV_DEV, API_RUNTIME}; // 38 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE = 38) - cuda2hipRename["cudaDevAttrMaxThreadsPerMultiProcessor"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_DEV, API_RUNTIME}; // 39 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR = 39) - cuda2hipRename["cudaDevAttrAsyncEngineCount"] = {"hipDeviceAttributeAsyncEngineCount", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 40 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT = 40) - cuda2hipRename["cudaDevAttrUnifiedAddressing"] = {"hipDeviceAttributeUnifiedAddressing", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 41 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING = 41) - cuda2hipRename["cudaDevAttrMaxTexture1DLayeredWidth"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 42 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH = 42) - cuda2hipRename["cudaDevAttrMaxTexture1DLayeredLayers"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 43 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS = 43) + cuda2hipRename["cudaDevAttrGpuOverlap"] = {"hipDeviceAttributeGpuOverlap", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 15 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GPU_OVERLAP = 15) + cuda2hipRename["cudaDevAttrMultiProcessorCount"] = {"hipDeviceAttributeMultiprocessorCount", CONV_TYPE, API_RUNTIME}; // 16 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT = 16) + cuda2hipRename["cudaDevAttrKernelExecTimeout"] = {"hipDeviceAttributeKernelExecTimeout", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 17 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT = 17) + cuda2hipRename["cudaDevAttrIntegrated"] = {"hipDeviceAttributeIntegrated", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 18 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_INTEGRATED = 18) + cuda2hipRename["cudaDevAttrCanMapHostMemory"] = {"hipDeviceAttributeCanMapHostMemory", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 19 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY = 19) + cuda2hipRename["cudaDevAttrComputeMode"] = {"hipDeviceAttributeComputeMode", CONV_TYPE, API_RUNTIME}; // 20 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_MODE = 20) + cuda2hipRename["cudaDevAttrMaxTexture1DWidth"] = {"hipDeviceAttributeMaxTexture1DWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 21 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_WIDTH = 21) + cuda2hipRename["cudaDevAttrMaxTexture2DWidth"] = {"hipDeviceAttributeMaxTexture2DWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 22 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_WIDTH = 22) + cuda2hipRename["cudaDevAttrMaxTexture2DHeight"] = {"hipDeviceAttributeMaxTexture2DHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 23 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_HEIGHT = 23) + cuda2hipRename["cudaDevAttrMaxTexture3DWidth"] = {"hipDeviceAttributeMaxTexture3DWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 24 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH = 24) + cuda2hipRename["cudaDevAttrMaxTexture3DHeight"] = {"hipDeviceAttributeMaxTexture3DHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 25 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT = 25) + cuda2hipRename["cudaDevAttrMaxTexture3DDepth"] = {"hipDeviceAttributeMaxTexture3DDepth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 26 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH = 26) + cuda2hipRename["cudaDevAttrMaxTexture2DLayeredWidth"] = {"hipDeviceAttributeMaxTexture2DLayeredWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 27 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_WIDTH = 27) + cuda2hipRename["cudaDevAttrMaxTexture2DLayeredHeight"] = {"hipDeviceAttributeMaxTexture2DLayeredHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 28 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_HEIGHT = 28) + cuda2hipRename["cudaDevAttrMaxTexture2DLayeredLayers"] = {"hipDeviceAttributeMaxTexture2DLayeredLayers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 29 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LAYERED_LAYERS = 29) + cuda2hipRename["cudaDevAttrSurfaceAlignment"] = {"hipDeviceAttributeSurfaceAlignment", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 30 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_SURFACE_ALIGNMENT = 30) + cuda2hipRename["cudaDevAttrConcurrentKernels"] = {"hipDeviceAttributeConcurrentKernels", CONV_TYPE, API_RUNTIME}; // 31 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS = 31) + cuda2hipRename["cudaDevAttrEccEnabled"] = {"hipDeviceAttributeEccEnabled", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 32 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_ECC_ENABLED = 32) + cuda2hipRename["cudaDevAttrPciBusId"] = {"hipDeviceAttributePciBusId", CONV_TYPE, API_RUNTIME}; // 33 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_BUS_ID = 33) + cuda2hipRename["cudaDevAttrPciDeviceId"] = {"hipDeviceAttributePciDeviceId", CONV_TYPE, API_RUNTIME}; // 34 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID = 34) + cuda2hipRename["cudaDevAttrTccDriver"] = {"hipDeviceAttributeTccDriver", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 35 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TCC_DRIVER = 35) + cuda2hipRename["cudaDevAttrMemoryClockRate"] = {"hipDeviceAttributeMemoryClockRate", CONV_TYPE, API_RUNTIME}; // 36 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE = 36) + cuda2hipRename["cudaDevAttrGlobalMemoryBusWidth"] = {"hipDeviceAttributeMemoryBusWidth", CONV_TYPE, API_RUNTIME}; // 37 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH = 37) + cuda2hipRename["cudaDevAttrL2CacheSize"] = {"hipDeviceAttributeL2CacheSize", CONV_TYPE, API_RUNTIME}; // 38 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE = 38) + cuda2hipRename["cudaDevAttrMaxThreadsPerMultiProcessor"] = {"hipDeviceAttributeMaxThreadsPerMultiProcessor", CONV_TYPE, API_RUNTIME}; // 39 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_THREADS_PER_MULTIPROCESSOR = 39) + cuda2hipRename["cudaDevAttrAsyncEngineCount"] = {"hipDeviceAttributeAsyncEngineCount", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 40 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT = 40) + cuda2hipRename["cudaDevAttrUnifiedAddressing"] = {"hipDeviceAttributeUnifiedAddressing", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 41 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING = 41) + cuda2hipRename["cudaDevAttrMaxTexture1DLayeredWidth"] = {"hipDeviceAttributeMaxTexture1DLayeredWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 42 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_WIDTH = 42) + cuda2hipRename["cudaDevAttrMaxTexture1DLayeredLayers"] = {"hipDeviceAttributeMaxTexture1DLayeredLayers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 43 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LAYERED_LAYERS = 43) // 44 - no - cuda2hipRename["cudaDevAttrMaxTexture2DGatherWidth"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 45 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH = 45) - cuda2hipRename["cudaDevAttrMaxTexture2DGatherHeight"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 46 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT = 46) - cuda2hipRename["cudaDevAttrMaxTexture3DWidthAlt"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 47 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE = 47) - cuda2hipRename["cudaDevAttrMaxTexture3DHeightAlt"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 48 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE = 48) - cuda2hipRename["cudaDevAttrMaxTexture3DDepthAlt"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 49 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE = 49) - cuda2hipRename["cudaDevAttrPciDomainId"] = {"hipDeviceAttributePciDomainId", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 50 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID = 50) - cuda2hipRename["cudaDevAttrTexturePitchAlignment"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 51 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT = 51) - cuda2hipRename["cudaDevAttrMaxTextureCubemapWidth"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 52 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH = 52) - cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredWidth"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 53 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH = 53) - cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredLayers"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 54 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS = 54) - cuda2hipRename["cudaDevAttrMaxSurface1DWidth"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 55 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH = 55) - cuda2hipRename["cudaDevAttrMaxSurface2DWidth"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 56 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH = 56) - cuda2hipRename["cudaDevAttrMaxSurface2DHeight"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 57 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT = 57) - cuda2hipRename["cudaDevAttrMaxSurface3DWidth"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 58 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH = 58) - cuda2hipRename["cudaDevAttrMaxSurface3DHeight"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 59 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT = 59) - cuda2hipRename["cudaDevAttrMaxSurface3DDepth"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 60 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH = 60) - cuda2hipRename["cudaDevAttrMaxSurface1DLayeredWidth"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 61 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH = 61) - cuda2hipRename["cudaDevAttrMaxSurface1DLayeredLayers"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 62 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS = 62) - cuda2hipRename["cudaDevAttrMaxSurface2DLayeredWidth"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 63 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH = 63) - cuda2hipRename["cudaDevAttrMaxSurface2DLayeredHeight"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 64 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT = 64) - cuda2hipRename["cudaDevAttrMaxSurface2DLayeredLayers"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 65 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS = 65) - cuda2hipRename["cudaDevAttrMaxSurfaceCubemapWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 66 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH = 66) - cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 67 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH = 67) - cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredLayers"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 68 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS = 68) - cuda2hipRename["cudaDevAttrMaxTexture1DLinearWidth"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 69 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH = 69) - cuda2hipRename["cudaDevAttrMaxTexture2DLinearWidth"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 70 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH = 70) - cuda2hipRename["cudaDevAttrMaxTexture2DLinearHeight"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 71 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT = 71) - cuda2hipRename["cudaDevAttrMaxTexture2DLinearPitch"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 72 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH = 72) - cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 73 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH = 73) - cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedHeight"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 74 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT = 74) - cuda2hipRename["cudaDevAttrComputeCapabilityMajor"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_DEV, API_RUNTIME}; // 75 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR = 75) - cuda2hipRename["cudaDevAttrComputeCapabilityMinor"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_DEV, API_RUNTIME}; // 76 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR = 76) - cuda2hipRename["cudaDevAttrMaxTexture1DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 77 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH = 77) - cuda2hipRename["cudaDevAttrStreamPrioritiesSupported"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 78 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED = 78) - cuda2hipRename["cudaDevAttrGlobalL1CacheSupported"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 79 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED = 79) - cuda2hipRename["cudaDevAttrLocalL1CacheSupported"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 80 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED = 80) - cuda2hipRename["cudaDevAttrMaxSharedMemoryPerMultiprocessor"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_DEV, API_RUNTIME}; // 81 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR = 81) - cuda2hipRename["cudaDevAttrMaxRegistersPerMultiprocessor"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 82 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR = 82) - cuda2hipRename["cudaDevAttrManagedMemory"] = {"hipDeviceAttributeManagedMemory", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 83 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY = 83) - cuda2hipRename["cudaDevAttrIsMultiGpuBoard"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_DEV, API_RUNTIME}; // 84 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD = 84) - cuda2hipRename["cudaDevAttrMultiGpuBoardGroupID"] = {"hipDeviceAttributeMultiGpuBoardGroupID", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 85 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID = 85) + cuda2hipRename["cudaDevAttrMaxTexture2DGatherWidth"] = {"hipDeviceAttributeMaxTexture2DGatherWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 45 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_WIDTH = 45) + cuda2hipRename["cudaDevAttrMaxTexture2DGatherHeight"] = {"hipDeviceAttributeMaxTexture2DGatherHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 46 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_GATHER_HEIGHT = 46) + cuda2hipRename["cudaDevAttrMaxTexture3DWidthAlt"] = {"hipDeviceAttributeMaxTexture3DWidthAlternate", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 47 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_WIDTH_ALTERNATE = 47) + cuda2hipRename["cudaDevAttrMaxTexture3DHeightAlt"] = {"hipDeviceAttributeMaxTexture3DHeightAlternate", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 48 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_HEIGHT_ALTERNATE = 48) + cuda2hipRename["cudaDevAttrMaxTexture3DDepthAlt"] = {"hipDeviceAttributeMaxTexture3DDepthAlternate", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 49 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE3D_DEPTH_ALTERNATE = 49) + cuda2hipRename["cudaDevAttrPciDomainId"] = {"hipDeviceAttributePciDomainId", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 50 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID = 50) + cuda2hipRename["cudaDevAttrTexturePitchAlignment"] = {"hipDeviceAttributeTexturePitchAlignment", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 51 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT = 51) + cuda2hipRename["cudaDevAttrMaxTextureCubemapWidth"] = {"hipDeviceAttributeMaxTextureCubemapWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 52 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_WIDTH = 52) + cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredWidth"] = {"hipDeviceAttributeMaxTextureCubemapLayeredWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 53 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_WIDTH = 53) + cuda2hipRename["cudaDevAttrMaxTextureCubemapLayeredLayers"] = {"hipDeviceAttributeMaxTextureCubemapLayeredLayers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 54 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURECUBEMAP_LAYERED_LAYERS = 54) + cuda2hipRename["cudaDevAttrMaxSurface1DWidth"] = {"hipDeviceAttributeMaxSurface1DWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 55 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_WIDTH = 55) + cuda2hipRename["cudaDevAttrMaxSurface2DWidth"] = {"hipDeviceAttributeMaxSurface2DWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 56 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_WIDTH = 56) + cuda2hipRename["cudaDevAttrMaxSurface2DHeight"] = {"hipDeviceAttributeMaxSurface2DHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 57 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_HEIGHT = 57) + cuda2hipRename["cudaDevAttrMaxSurface3DWidth"] = {"hipDeviceAttributeMaxSurface3DWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 58 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_WIDTH = 58) + cuda2hipRename["cudaDevAttrMaxSurface3DHeight"] = {"hipDeviceAttributeMaxSurface3DHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 59 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_HEIGHT = 59) + cuda2hipRename["cudaDevAttrMaxSurface3DDepth"] = {"hipDeviceAttributeMaxSurface3DDepth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 60 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE3D_DEPTH = 60) + cuda2hipRename["cudaDevAttrMaxSurface1DLayeredWidth"] = {"hipDeviceAttributeMaxSurface1DLayeredWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 61 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_WIDTH = 61) + cuda2hipRename["cudaDevAttrMaxSurface1DLayeredLayers"] = {"hipDeviceAttributeMaxSurface1DLayeredLayers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 62 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE1D_LAYERED_LAYERS = 62) + cuda2hipRename["cudaDevAttrMaxSurface2DLayeredWidth"] = {"hipDeviceAttributeMaxSurface2DLayeredWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 63 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_WIDTH = 63) + cuda2hipRename["cudaDevAttrMaxSurface2DLayeredHeight"] = {"hipDeviceAttributeMaxSurface2DLayeredHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 64 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_HEIGHT = 64) + cuda2hipRename["cudaDevAttrMaxSurface2DLayeredLayers"] = {"hipDeviceAttributeMaxSurface2DLayeredLayers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 65 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACE2D_LAYERED_LAYERS = 65) + cuda2hipRename["cudaDevAttrMaxSurfaceCubemapWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 66 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_WIDTH = 66) + cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredWidth"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 67 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_WIDTH = 67) + cuda2hipRename["cudaDevAttrMaxSurfaceCubemapLayeredLayers"] = {"hipDeviceAttributeMaxSurfaceCubemapLayeredLayers", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 68 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_SURFACECUBEMAP_LAYERED_LAYERS = 68) + cuda2hipRename["cudaDevAttrMaxTexture1DLinearWidth"] = {"hipDeviceAttributeMaxTexture1DLinearWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 69 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_LINEAR_WIDTH = 69) + cuda2hipRename["cudaDevAttrMaxTexture2DLinearWidth"] = {"hipDeviceAttributeMaxTexture2DLinearWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 70 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_WIDTH = 70) + cuda2hipRename["cudaDevAttrMaxTexture2DLinearHeight"] = {"hipDeviceAttributeMaxTexture2DLinearHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 71 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_HEIGHT = 71) + cuda2hipRename["cudaDevAttrMaxTexture2DLinearPitch"] = {"hipDeviceAttributeMaxTexture2DLinearPitch", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 72 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_LINEAR_PITCH = 72) + cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture2DMipmappedWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 73 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_WIDTH = 73) + cuda2hipRename["cudaDevAttrMaxTexture2DMipmappedHeight"] = {"hipDeviceAttributeMaxTexture2DMipmappedHeight", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 74 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE2D_MIPMAPPED_HEIGHT = 74) + cuda2hipRename["cudaDevAttrComputeCapabilityMajor"] = {"hipDeviceAttributeComputeCapabilityMajor", CONV_TYPE, API_RUNTIME}; // 75 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR = 75) + cuda2hipRename["cudaDevAttrComputeCapabilityMinor"] = {"hipDeviceAttributeComputeCapabilityMinor", CONV_TYPE, API_RUNTIME}; // 76 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR = 76) + cuda2hipRename["cudaDevAttrMaxTexture1DMipmappedWidth"] = {"hipDeviceAttributeMaxTexture1DMipmappedWidth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 77 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAXIMUM_TEXTURE1D_MIPMAPPED_WIDTH = 77) + cuda2hipRename["cudaDevAttrStreamPrioritiesSupported"] = {"hipDeviceAttributeStreamPrioritiesSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 78 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_STREAM_PRIORITIES_SUPPORTED = 78) + cuda2hipRename["cudaDevAttrGlobalL1CacheSupported"] = {"hipDeviceAttributeGlobalL1CacheSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 79 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_GLOBAL_L1_CACHE_SUPPORTED = 79) + cuda2hipRename["cudaDevAttrLocalL1CacheSupported"] = {"hipDeviceAttributeLocalL1CacheSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 80 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_LOCAL_L1_CACHE_SUPPORTED = 80) + cuda2hipRename["cudaDevAttrMaxSharedMemoryPerMultiprocessor"] = {"hipDeviceAttributeMaxSharedMemoryPerMultiprocessor", CONV_TYPE, API_RUNTIME}; // 81 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_SHARED_MEMORY_PER_MULTIPROCESSOR = 81) + cuda2hipRename["cudaDevAttrMaxRegistersPerMultiprocessor"] = {"hipDeviceAttributeMaxRegistersPerMultiprocessor", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 82 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MAX_REGISTERS_PER_MULTIPROCESSOR = 82) + cuda2hipRename["cudaDevAttrManagedMemory"] = {"hipDeviceAttributeManagedMemory", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 83 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY = 83) + cuda2hipRename["cudaDevAttrIsMultiGpuBoard"] = {"hipDeviceAttributeIsMultiGpuBoard", CONV_TYPE, API_RUNTIME}; // 84 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD = 84) + cuda2hipRename["cudaDevAttrMultiGpuBoardGroupID"] = {"hipDeviceAttributeMultiGpuBoardGroupID", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 85 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID = 85) // unsupported yet by HIP [CUDA 8.0.44] - cuda2hipRename["cudaDevAttrHostNativeAtomicSupported"] = {"hipDeviceAttributeHostNativeAtomicSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 86 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_HOST_NATIVE_ATOMIC_SUPPORTED = 86) - cuda2hipRename["cudaDevAttrSingleToDoublePrecisionPerfRatio"] = {"hipDeviceAttributeSingleToDoublePrecisionPerfRatio", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 87 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_SINGLE_TO_DOUBLE_PRECISION_PERF_RATIO = 87) - cuda2hipRename["cudaDevAttrPageableMemoryAccess"] = {"hipDeviceAttributePageableMemoryAccess", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 88 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PAGEABLE_MEMORY_ACCESS = 88) - cuda2hipRename["cudaDevAttrConcurrentManagedAccess"] = {"hipDeviceAttributeConcurrentManagedAccess", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 89 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CONCURRENT_MANAGED_ACCESS = 89) - cuda2hipRename["cudaDevAttrComputePreemptionSupported"] = {"hipDeviceAttributeComputePreemptionSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 90 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_PREEMPTION_SUPPORTED = 90) - cuda2hipRename["cudaDevAttrCanUseHostPointerForRegisteredMem"] = {"hipDeviceAttributeCanUseHostPointerForRegisteredMem", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 91 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CAN_USE_HOST_POINTER_FOR_REGISTERED_MEM = 91) + cuda2hipRename["cudaDevAttrHostNativeAtomicSupported"] = {"hipDeviceAttributeHostNativeAtomicSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 86 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_HOST_NATIVE_ATOMIC_SUPPORTED = 86) + cuda2hipRename["cudaDevAttrSingleToDoublePrecisionPerfRatio"] = {"hipDeviceAttributeSingleToDoublePrecisionPerfRatio", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 87 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_SINGLE_TO_DOUBLE_PRECISION_PERF_RATIO = 87) + cuda2hipRename["cudaDevAttrPageableMemoryAccess"] = {"hipDeviceAttributePageableMemoryAccess", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 88 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_PAGEABLE_MEMORY_ACCESS = 88) + cuda2hipRename["cudaDevAttrConcurrentManagedAccess"] = {"hipDeviceAttributeConcurrentManagedAccess", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 89 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CONCURRENT_MANAGED_ACCESS = 89) + cuda2hipRename["cudaDevAttrComputePreemptionSupported"] = {"hipDeviceAttributeComputePreemptionSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 90 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_COMPUTE_PREEMPTION_SUPPORTED = 90) + cuda2hipRename["cudaDevAttrCanUseHostPointerForRegisteredMem"] = {"hipDeviceAttributeCanUseHostPointerForRegisteredMem", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 91 // API_DRIVER ANALOGUE (CU_DEVICE_ATTRIBUTE_CAN_USE_HOST_POINTER_FOR_REGISTERED_MEM = 91) // Pointer Attributes // struct cudaPointerAttributes @@ -1460,54 +1687,49 @@ struct cuda2hipMap { // Device cuda2hipRename["cudaDeviceProp"] = {"hipDeviceProp_t", CONV_TYPE, API_RUNTIME}; - cuda2hipRename["cudaGetDeviceProperties"] = {"hipGetDeviceProperties", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceGetPCIBusId"] = {"hipDeviceGetPCIBusId", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceGetByPCIBusId"] = {"hipDeviceGetByPCIBusId", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaGetDeviceProperties"] = {"hipGetDeviceProperties", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaDeviceGetPCIBusId"] = {"hipDeviceGetPCIBusId", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaDeviceGetByPCIBusId"] = {"hipDeviceGetByPCIBusId", CONV_DEVICE, API_RUNTIME}; // unsupported yet by HIP - cuda2hipRename["cudaDeviceGetStreamPriorityRange"] = {"hipDeviceGetStreamPriorityRange", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaSetValidDevices"] = {"hipSetValidDevices", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaDeviceGetStreamPriorityRange"] = {"hipDeviceGetStreamPriorityRange", CONV_DEVICE, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaSetValidDevices"] = {"hipSetValidDevices", CONV_DEVICE, API_RUNTIME, HIP_UNSUPPORTED}; // unsupported yet by HIP [CUDA 8.0.44] // P2P Attributes - cuda2hipRename["cudaDeviceP2PAttr"] = {"hipDeviceP2PAttribute", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (CUdevice_P2PAttribute) - cuda2hipRename["cudaDevP2PAttrPerformanceRank"] = {"hipDeviceP2PAttributePerformanceRank", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_DRIVER ANALOGUE (CU_DEVICE_P2P_ATTRIBUTE_PERFORMANCE_RANK = 0x01) - cuda2hipRename["cudaDevP2PAttrAccessSupported"] = {"hipDeviceP2PAttributeAccessSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_DRIVER ANALOGUE (CU_DEVICE_P2P_ATTRIBUTE_ACCESS_SUPPORTED = 0x02) - cuda2hipRename["cudaDevP2PAttrNativeAtomicSupported"] = {"hipDeviceP2PAttributeNativeAtomicSupported", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_DRIVER ANALOGUE (CU_DEVICE_P2P_ATTRIBUTE_NATIVE_ATOMIC_SUPPORTED = 0x03) + cuda2hipRename["cudaDeviceP2PAttr"] = {"hipDeviceP2PAttribute", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (CUdevice_P2PAttribute) + cuda2hipRename["cudaDevP2PAttrPerformanceRank"] = {"hipDeviceP2PAttributePerformanceRank", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_DRIVER ANALOGUE (CU_DEVICE_P2P_ATTRIBUTE_PERFORMANCE_RANK = 0x01) + cuda2hipRename["cudaDevP2PAttrAccessSupported"] = {"hipDeviceP2PAttributeAccessSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_DRIVER ANALOGUE (CU_DEVICE_P2P_ATTRIBUTE_ACCESS_SUPPORTED = 0x02) + cuda2hipRename["cudaDevP2PAttrNativeAtomicSupported"] = {"hipDeviceP2PAttributeNativeAtomicSupported", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_DRIVER ANALOGUE (CU_DEVICE_P2P_ATTRIBUTE_NATIVE_ATOMIC_SUPPORTED = 0x03) // [CUDA 8.0.44] - cuda2hipRename["cudaDeviceGetP2PAttribute"] = {"hipDeviceGetP2PAttribute", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (cuDeviceGetP2PAttribute) + cuda2hipRename["cudaDeviceGetP2PAttribute"] = {"hipDeviceGetP2PAttribute", CONV_DEVICE, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (cuDeviceGetP2PAttribute) // Compute mode - cuda2hipRename["cudaComputeMode"] = {"hipComputeMode", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (CUcomputemode) - cuda2hipRename["cudaComputeModeDefault"] = {"hipComputeModeDefault", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_DEFAULT = 0) - cuda2hipRename["cudaComputeModeExclusive"] = {"hipComputeModeExclusive", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_EXCLUSIVE = 1) - cuda2hipRename["cudaComputeModeProhibited"] = {"hipComputeModeProhibited", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_PROHIBITED = 2) - cuda2hipRename["cudaComputeModeExclusiveProcess"] = {"hipComputeModeExclusiveProcess", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; // 3 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_EXCLUSIVE_PROCESS = 3) + cuda2hipRename["cudaComputeMode"] = {"hipComputeMode", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // API_DRIVER ANALOGUE (CUcomputemode) + cuda2hipRename["cudaComputeModeDefault"] = {"hipComputeModeDefault", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_DEFAULT = 0) + cuda2hipRename["cudaComputeModeExclusive"] = {"hipComputeModeExclusive", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_EXCLUSIVE = 1) + cuda2hipRename["cudaComputeModeProhibited"] = {"hipComputeModeProhibited", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_PROHIBITED = 2) + cuda2hipRename["cudaComputeModeExclusiveProcess"] = {"hipComputeModeExclusiveProcess", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 3 // API_DRIVER ANALOGUE (CU_COMPUTEMODE_EXCLUSIVE_PROCESS = 3) // Device Flags - // unsupported yet by HIP - cuda2hipRename["cudaGetDeviceFlags"] = {"hipGetDeviceFlags", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaSetDeviceFlags"] = {"hipSetDeviceFlags", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceScheduleAuto"] = {"hipDeviceScheduleAuto", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceScheduleSpin"] = {"hipDeviceScheduleSpin", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceScheduleYield"] = {"hipDeviceScheduleYield", CONV_DEV, API_RUNTIME}; - // deprecated as of CUDA 4.0 and replaced with cudaDeviceScheduleBlockingSync - cuda2hipRename["cudaDeviceBlockingSync"] = {"hipDeviceScheduleBlockingSync", CONV_DEV, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaDeviceScheduleBlockingSync"] = {"hipDeviceScheduleBlockingSync", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceScheduleMask"] = {"hipDeviceScheduleMask", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaGetDeviceFlags"] = {"hipGetDeviceFlags", CONV_DEVICE, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaSetDeviceFlags"] = {"hipSetDeviceFlags", CONV_DEVICE, API_RUNTIME}; - cuda2hipRename["cudaDeviceMapHost"] = {"hipDeviceMapHost", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaDeviceScheduleAuto"] = {"hipDeviceScheduleAuto", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaDeviceScheduleSpin"] = {"hipDeviceScheduleSpin", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaDeviceScheduleYield"] = {"hipDeviceScheduleYield", CONV_TYPE, API_RUNTIME}; + // deprecated as of CUDA 4.0 and replaced with cudaDeviceScheduleBlockingSync + cuda2hipRename["cudaDeviceBlockingSync"] = {"hipDeviceScheduleBlockingSync", CONV_TYPE, API_RUNTIME}; // unsupported yet by HIP - cuda2hipRename["cudaDeviceLmemResizeToMax"] = {"hipDeviceLmemResizeToMax", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDeviceMask"] = {"hipDeviceMask", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaDeviceScheduleBlockingSync"] = {"hipDeviceScheduleBlockingSync", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaDeviceScheduleMask"] = {"hipDeviceScheduleMask", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; + + cuda2hipRename["cudaDeviceMapHost"] = {"hipDeviceMapHost", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaDeviceLmemResizeToMax"] = {"hipDeviceLmemResizeToMax", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaDeviceMask"] = {"hipDeviceMask", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // Cache config cuda2hipRename["cudaDeviceSetCacheConfig"] = {"hipDeviceSetCacheConfig", CONV_CACHE, API_RUNTIME}; - // translate deprecated - cuda2hipRename["cudaThreadSetCacheConfig"] = {"hipDeviceSetCacheConfig", CONV_CACHE, API_RUNTIME}; cuda2hipRename["cudaDeviceGetCacheConfig"] = {"hipDeviceGetCacheConfig", CONV_CACHE, API_RUNTIME}; - // translate deprecated - cuda2hipRename["cudaThreadGetCacheConfig"] = {"hipDeviceGetCacheConfig", CONV_CACHE, API_RUNTIME}; cuda2hipRename["cudaFuncSetCacheConfig"] = {"hipFuncSetCacheConfig", CONV_CACHE, API_RUNTIME}; // Execution control @@ -1533,25 +1755,25 @@ struct cuda2hipMap { cuda2hipRename["cudaLaunch"] = {"hipLaunch", CONV_EXEC, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaSetupArgument"] = {"hipSetupArgument", CONV_EXEC, API_RUNTIME, HIP_UNSUPPORTED}; - // Driver/Runtime - cuda2hipRename["cudaDriverGetVersion"] = {"hipDriverGetVersion", CONV_DRIVER, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaRuntimeGetVersion"] = {"hipRuntimeGetVersion", CONV_DEV, API_RUNTIME, HIP_UNSUPPORTED}; + // Version Management + cuda2hipRename["cudaDriverGetVersion"] = {"hipDriverGetVersion", CONV_VERSION, API_RUNTIME}; + cuda2hipRename["cudaRuntimeGetVersion"] = {"hipRuntimeGetVersion", CONV_VERSION, API_RUNTIME, HIP_UNSUPPORTED}; // Occupancy - cuda2hipRename["cudaOccupancyMaxPotentialBlockSize"] = {"hipOccupancyMaxPotentialBlockSize", CONV_OCCUPANCY, API_DRIVER}; + cuda2hipRename["cudaOccupancyMaxPotentialBlockSize"] = {"hipOccupancyMaxPotentialBlockSize", CONV_OCCUPANCY, API_RUNTIME}; // unsupported yet by HIP - cuda2hipRename["cudaOccupancyMaxPotentialBlockSizeWithFlags"] = {"hipOccupancyMaxPotentialBlockSizeWithFlags", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cudaOccupancyMaxActiveBlocksPerMultiprocessor"] = {"hipOccupancyMaxActiveBlocksPerMultiprocessor", CONV_OCCUPANCY, API_DRIVER}; + cuda2hipRename["cudaOccupancyMaxPotentialBlockSizeWithFlags"] = {"hipOccupancyMaxPotentialBlockSizeWithFlags", CONV_OCCUPANCY, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaOccupancyMaxActiveBlocksPerMultiprocessor"] = {"hipOccupancyMaxActiveBlocksPerMultiprocessor", CONV_OCCUPANCY, API_RUNTIME}; // unsupported yet by HIP - cuda2hipRename["cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags"] = {"hipOccupancyMaxActiveBlocksPerMultiprocessorWithFlags", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cudaOccupancyMaxPotentialBlockSizeVariableSMem"] = {"hipOccupancyMaxPotentialBlockSizeVariableSMem", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; - cuda2hipRename["cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags"] = {"hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags"] = {"hipOccupancyMaxActiveBlocksPerMultiprocessorWithFlags", CONV_OCCUPANCY, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaOccupancyMaxPotentialBlockSizeVariableSMem"] = {"hipOccupancyMaxPotentialBlockSizeVariableSMem", CONV_OCCUPANCY, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags"] = {"hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags", CONV_OCCUPANCY, API_RUNTIME, HIP_UNSUPPORTED}; // Peer2Peer - cuda2hipRename["cudaDeviceCanAccessPeer"] = {"hipDeviceCanAccessPeer", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceDisablePeerAccess"] = {"hipDeviceDisablePeerAccess", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceEnablePeerAccess"] = {"hipDeviceEnablePeerAccess", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaDeviceCanAccessPeer"] = {"hipDeviceCanAccessPeer", CONV_PEER, API_RUNTIME}; + cuda2hipRename["cudaDeviceDisablePeerAccess"] = {"hipDeviceDisablePeerAccess", CONV_PEER, API_RUNTIME}; + cuda2hipRename["cudaDeviceEnablePeerAccess"] = {"hipDeviceEnablePeerAccess", CONV_PEER, API_RUNTIME}; + cuda2hipRename["cudaMemcpyPeerAsync"] = {"hipMemcpyPeerAsync", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMemcpyPeer"] = {"hipMemcpyPeer", CONV_MEM, API_RUNTIME}; @@ -1559,17 +1781,16 @@ struct cuda2hipMap { cuda2hipRename["cudaIpcMemLazyEnablePeerAccess"] = {"hipIpcMemLazyEnablePeerAccess", CONV_TYPE, API_RUNTIME}; // 0x01 // API_Driver ANALOGUE (CU_IPC_MEM_LAZY_ENABLE_PEER_ACCESS = 0x1) // Shared memory - cuda2hipRename["cudaDeviceSetSharedMemConfig"] = {"hipDeviceSetSharedMemConfig", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaDeviceSetSharedMemConfig"] = {"hipDeviceSetSharedMemConfig", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaDeviceGetSharedMemConfig"] = {"hipDeviceGetSharedMemConfig", CONV_DEVICE, API_RUNTIME}; // translate deprecated - cuda2hipRename["cudaThreadSetSharedMemConfig"] = {"hipDeviceSetSharedMemConfig", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaDeviceGetSharedMemConfig"] = {"hipDeviceGetSharedMemConfig", CONV_DEV, API_RUNTIME}; - // translate deprecated - cuda2hipRename["cudaThreadGetSharedMemConfig"] = {"hipDeviceGetSharedMemConfig", CONV_DEV, API_RUNTIME}; + // cuda2hipRename["cudaThreadGetSharedMemConfig"] = {"hipDeviceGetSharedMemConfig", CONV_DEVICE, API_RUNTIME}; + // cuda2hipRename["cudaThreadSetSharedMemConfig"] = {"hipDeviceSetSharedMemConfig", CONV_DEVICE, API_RUNTIME}; - cuda2hipRename["cudaSharedMemConfig"] = {"hipSharedMemConfig", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaSharedMemBankSizeDefault"] = {"hipSharedMemBankSizeDefault", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaSharedMemBankSizeFourByte"] = {"hipSharedMemBankSizeFourByte", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaSharedMemBankSizeEightByte"] = {"hipSharedMemBankSizeEightByte", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaSharedMemConfig"] = {"hipSharedMemConfig", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaSharedMemBankSizeDefault"] = {"hipSharedMemBankSizeDefault", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaSharedMemBankSizeFourByte"] = {"hipSharedMemBankSizeFourByte", CONV_TYPE, API_RUNTIME}; + cuda2hipRename["cudaSharedMemBankSizeEightByte"] = {"hipSharedMemBankSizeEightByte", CONV_TYPE, API_RUNTIME}; // Limits cuda2hipRename["cudaLimit"] = {"hipLimit_t", CONV_TYPE, API_RUNTIME}; // API_Driver ANALOGUE (CUlimit) @@ -1579,14 +1800,12 @@ struct cuda2hipMap { cuda2hipRename["cudaLimitDevRuntimeSyncDepth"] = {"hipLimitDevRuntimeSyncDepth", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH = 0x03) cuda2hipRename["cudaLimitDevRuntimePendingLaunchCount"] = {"hipLimitDevRuntimePendingLaunchCount", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x04 // API_Driver ANALOGUE (CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT = 0x04) - cuda2hipRename["cudaDeviceGetLimit"] = {"hipDeviceGetLimit", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaDeviceGetLimit"] = {"hipDeviceGetLimit", CONV_DEVICE, API_RUNTIME}; // Profiler - // unsupported yet by HIP - cuda2hipRename["cudaProfilerInitialize"] = {"hipProfilerInitialize", CONV_OTHER, API_RUNTIME, HIP_UNSUPPORTED}; - - cuda2hipRename["cudaProfilerStart"] = {"hipProfilerStart", CONV_OTHER, API_RUNTIME}; - cuda2hipRename["cudaProfilerStop"] = {"hipProfilerStop", CONV_OTHER, API_RUNTIME}; + cuda2hipRename["cudaProfilerInitialize"] = {"hipProfilerInitialize", CONV_OTHER, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuProfilerInitialize) + cuda2hipRename["cudaProfilerStart"] = {"hipProfilerStart", CONV_OTHER, API_RUNTIME}; // API_Driver ANALOGUE (cuProfilerStart) + cuda2hipRename["cudaProfilerStop"] = {"hipProfilerStop", CONV_OTHER, API_RUNTIME}; // API_Driver ANALOGUE (cuProfilerStop) // unsupported yet by HIP cuda2hipRename["cudaOutputMode"] = {"hipOutputMode", CONV_OTHER, API_RUNTIME, HIP_UNSUPPORTED}; @@ -1630,6 +1849,8 @@ struct cuda2hipMap { cuda2hipRename["cudaResourceDesc"] = {"hipResourceDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaResourceViewDesc"] = {"hipResourceViewDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaTextureDesc"] = {"hipTextureDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["surfaceReference"] = {"hipSurfaceReference", CONV_SURFACE, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["textureReference"] = {"hipTextureReference", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // enums // enum cudaResourceType @@ -1673,7 +1894,7 @@ struct cuda2hipMap { cuda2hipRename["cudaResViewFormatSignedBlockCompressed4"] = {"hipResViewFormatSignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC4 = 0x1d) cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed5"] = {"hipResViewFormatUnsignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC5 = 0x1e) cuda2hipRename["cudaResViewFormatSignedBlockCompressed5"] = {"hipResViewFormatSignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC5 = 0x1f) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed6H"] = {"hipResViewFormatUnsignedBlockCompressed6H", CONV_TEX, API_DRIVER, HIP_UNSUPPORTED}; // 0x20 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC6H = 0x20) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed6H"] = {"hipResViewFormatUnsignedBlockCompressed6H", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x20 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC6H = 0x20) cuda2hipRename["cudaResViewFormatSignedBlockCompressed6H"] = {"hipResViewFormatSignedBlockCompressed6H", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x21 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC6H = 0x21) cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed7"] = {"hipResViewFormatUnsignedBlockCompressed7", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x22 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC7 = 0x22) @@ -1718,28 +1939,26 @@ struct cuda2hipMap { cuda2hipRename["cudaIpcMemHandle_st"] = {"hipIpcMemHandle_t", CONV_TYPE, API_RUNTIME}; // IPC functions - cuda2hipRename["cudaIpcCloseMemHandle"] = {"hipIpcCloseMemHandle", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaIpcGetEventHandle"] = {"hipIpcGetEventHandle", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaIpcGetMemHandle"] = {"hipIpcGetMemHandle", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaIpcOpenEventHandle"] = {"hipIpcOpenEventHandle", CONV_DEV, API_RUNTIME}; - cuda2hipRename["cudaIpcOpenMemHandle"] = {"hipIpcOpenMemHandle", CONV_DEV, API_RUNTIME}; + cuda2hipRename["cudaIpcCloseMemHandle"] = {"hipIpcCloseMemHandle", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaIpcGetEventHandle"] = {"hipIpcGetEventHandle", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaIpcGetMemHandle"] = {"hipIpcGetMemHandle", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaIpcOpenEventHandle"] = {"hipIpcOpenEventHandle", CONV_DEVICE, API_RUNTIME}; + cuda2hipRename["cudaIpcOpenMemHandle"] = {"hipIpcOpenMemHandle", CONV_DEVICE, API_RUNTIME}; // OpenGL Interoperability - // unsupported yet by HIP cuda2hipRename["cudaGLGetDevices"] = {"hipGLGetDevices", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaGraphicsGLRegisterBuffer"] = {"hipGraphicsGLRegisterBuffer", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaGraphicsGLRegisterImage"] = {"hipGraphicsGLRegisterImage", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaWGLGetDevice"] = {"hipWGLGetDevice", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // Graphics Interoperability - // unsupported yet by HIP - cuda2hipRename["cudaGraphicsMapResources"] = {"hipGraphicsMapResources", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsResourceGetMappedMipmappedArray"] = {"hipGraphicsResourceGetMappedMipmappedArray", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsResourceGetMappedPointer"] = {"hipGraphicsResourceGetMappedPointer", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsResourceSetMapFlags"] = {"hipGraphicsResourceSetMapFlags", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsSubResourceGetMappedArray"] = {"hipGraphicsSubResourceGetMappedArray", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsUnmapResources"] = {"hipGraphicsUnmapResources", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGraphicsUnregisterResource"] = {"hipGraphicsUnregisterResource", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaGraphicsMapResources"] = {"hipGraphicsMapResources", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsMapResources) + cuda2hipRename["cudaGraphicsResourceGetMappedMipmappedArray"] = {"hipGraphicsResourceGetMappedMipmappedArray", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsResourceGetMappedMipmappedArray) + cuda2hipRename["cudaGraphicsResourceGetMappedPointer"] = {"hipGraphicsResourceGetMappedPointer", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsResourceGetMappedPointer) + cuda2hipRename["cudaGraphicsResourceSetMapFlags"] = {"hipGraphicsResourceSetMapFlags", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsResourceSetMapFlags) + cuda2hipRename["cudaGraphicsSubResourceGetMappedArray"] = {"hipGraphicsSubResourceGetMappedArray", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsSubResourceGetMappedArray) + cuda2hipRename["cudaGraphicsUnmapResources"] = {"hipGraphicsUnmapResources", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsUnmapResources) + cuda2hipRename["cudaGraphicsUnregisterResource"] = {"hipGraphicsUnregisterResource", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsUnregisterResource) cuda2hipRename["cudaGraphicsCubeFace"] = {"hipGraphicsCubeFace", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaGraphicsCubeFacePositiveX"] = {"hipGraphicsCubeFacePositiveX", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; @@ -1763,6 +1982,142 @@ struct cuda2hipMap { cuda2hipRename["cudaGraphicsRegisterFlagsSurfaceLoadStore"] = {"hipGraphicsRegisterFlagsSurfaceLoadStore", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 4 // API_Driver ANALOGUE (CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST = 0x04) cuda2hipRename["cudaGraphicsRegisterFlagsTextureGather"] = {"hipGraphicsRegisterFlagsTextureGather", CONV_GRAPHICS, API_RUNTIME, HIP_UNSUPPORTED}; // 8 // API_Driver ANALOGUE (CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER = 0x08) + // OpenGL Interoperability + // enum cudaGLDeviceList + cuda2hipRename["cudaGLDeviceList"] = {"hipGLDeviceList", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUGLDeviceList) + cuda2hipRename["cudaGLDeviceListAll"] = {"HIP_GL_DEVICE_LIST_ALL", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_GL_DEVICE_LIST_ALL) + cuda2hipRename["cudaGLDeviceListCurrentFrame"] = {"HIP_GL_DEVICE_LIST_CURRENT_FRAME", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_GL_DEVICE_LIST_CURRENT_FRAME) + cuda2hipRename["cudaGLDeviceListNextFrame"] = {"HIP_GL_DEVICE_LIST_NEXT_FRAME", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_GL_DEVICE_LIST_NEXT_FRAME) + + cuda2hipRename["cudaGLGetDevices"] = {"hipGLGetDevices", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGLGetDevices) + cuda2hipRename["cudaGraphicsGLRegisterBuffer"] = {"hipGraphicsGLRegisterBuffer", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsGLRegisterBuffer) + cuda2hipRename["cudaGraphicsGLRegisterImage"] = {"hipGraphicsGLRegisterImage", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsGLRegisterImage) + cuda2hipRename["cudaWGLGetDevice"] = {"hipWGLGetDevice", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuWGLGetDevice) + + // OpenGL Interoperability [DEPRECATED] + // enum cudaGLMapFlags + cuda2hipRename["cudaGLMapFlags"] = {"hipGLMapFlags", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUGLmap_flags) + cuda2hipRename["cudaGLMapFlagsNone"] = {"HIP_GL_MAP_RESOURCE_FLAGS_NONE", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_GL_MAP_RESOURCE_FLAGS_NONE) + cuda2hipRename["cudaGLMapFlagsReadOnly"] = {"HIP_GL_MAP_RESOURCE_FLAGS_READ_ONLY", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_GL_MAP_RESOURCE_FLAGS_READ_ONLY) + cuda2hipRename["cudaGLMapFlagsWriteDiscard"] = {"HIP_GL_MAP_RESOURCE_FLAGS_WRITE_DISCARD", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_GL_MAP_RESOURCE_FLAGS_WRITE_DISCARD) + + cuda2hipRename["cudaGLMapBufferObject"] = {"hipGLMapBufferObject__", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // Not equal to cuGLMapBufferObject due to different signatures + cuda2hipRename["cudaGLMapBufferObjectAsync"] = {"hipGLMapBufferObjectAsync__", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // Not equal to cuGLMapBufferObjectAsync due to different signatures + cuda2hipRename["cudaGLRegisterBufferObject"] = {"hipGLRegisterBufferObject", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGLRegisterBufferObject) + cuda2hipRename["cudaGLSetBufferObjectMapFlags"] = {"hipGLSetBufferObjectMapFlags", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGLSetBufferObjectMapFlags) + cuda2hipRename["cudaGLSetGLDevice"] = {"hipGLSetGLDevice", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // no API_Driver ANALOGUE + cuda2hipRename["cudaGLUnmapBufferObject"] = {"hipGLUnmapBufferObject", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGLUnmapBufferObject) + cuda2hipRename["cudaGLUnmapBufferObjectAsync"] = {"hipGLUnmapBufferObjectAsync", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGLUnmapBufferObjectAsync) + cuda2hipRename["cudaGLUnregisterBufferObject"] = {"hipGLUnregisterBufferObject", CONV_GL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGLUnregisterBufferObject) + + // Direct3D 9 Interoperability + // enum CUd3d9DeviceList + cuda2hipRename["cudaD3D9DeviceList"] = {"hipD3D9DeviceList", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d9DeviceList) + cuda2hipRename["cudaD3D9DeviceListAll"] = {"HIP_D3D9_DEVICE_LIST_ALL", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D9_DEVICE_LIST_ALL) + cuda2hipRename["cudaD3D9DeviceListCurrentFrame"] = {"HIP_D3D9_DEVICE_LIST_CURRENT_FRAME", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_D3D9_DEVICE_LIST_CURRENT_FRAME) + cuda2hipRename["cudaD3D9DeviceListNextFrame"] = {"HIP_D3D9_DEVICE_LIST_NEXT_FRAME", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 3 // API_Driver ANALOGUE (CU_D3D9_DEVICE_LIST_NEXT_FRAME) + + cuda2hipRename["cudaD3D9GetDevice"] = {"hipD3D9GetDevice", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9GetDevice) + cuda2hipRename["cudaD3D9GetDevices"] = {"hipD3D9GetDevices", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9GetDevices) + cuda2hipRename["cudaD3D9GetDirect3DDevice"] = {"hipD3D9GetDirect3DDevice", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9GetDirect3DDevice) + cuda2hipRename["cudaD3D9SetDirect3DDevice"] = {"hipD3D9SetDirect3DDevice", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // no API_Driver ANALOGUE + cuda2hipRename["cudaGraphicsD3D9RegisterResource"] = {"hipGraphicsD3D9RegisterResource", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsD3D9RegisterResource) + + // Direct3D 9 Interoperability [DEPRECATED] + // enum cudaD3D9MapFlags + cuda2hipRename["cudaD3D9MapFlags"] = {"hipD3D9MapFlags", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d9map_flags) + cuda2hipRename["cudaD3D9MapFlagsNone"] = {"HIP_D3D9_MAPRESOURCE_FLAGS_NONE", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_Driver ANALOGUE (CU_D3D9_MAPRESOURCE_FLAGS_NONE) + cuda2hipRename["cudaD3D9MapFlagsReadOnly"] = {"HIP_D3D9_MAPRESOURCE_FLAGS_READONLY", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D9_MAPRESOURCE_FLAGS_READONLY) + cuda2hipRename["cudaD3D9MapFlagsWriteDiscard"] = {"HIP_D3D9_MAPRESOURCE_FLAGS_WRITEDISCARD", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_D3D9_MAPRESOURCE_FLAGS_WRITEDISCARD) + + // enum cudaD3D9RegisterFlags + cuda2hipRename["cudaD3D9RegisterFlags"] = {"hipD3D9RegisterFlags", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d9Register_flags) + cuda2hipRename["cudaD3D9RegisterFlagsNone"] = {"HIP_D3D9_REGISTER_FLAGS_NONE", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_Driver ANALOGUE (CU_D3D9_REGISTER_FLAGS_NONE) + cuda2hipRename["cudaD3D9RegisterFlagsArray"] = {"HIP_D3D9_REGISTER_FLAGS_ARRAY", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D9_REGISTER_FLAGS_ARRAY) + + cuda2hipRename["cudaD3D9MapResources"] = {"hipD3D9MapResources", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9MapResources) + cuda2hipRename["cudaD3D9RegisterResource"] = {"hipD3D9RegisterResource", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9RegisterResource) + cuda2hipRename["cudaD3D9ResourceGetMappedArray"] = {"hipD3D9ResourceGetMappedArray", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9ResourceGetMappedArray) + cuda2hipRename["cudaD3D9ResourceGetMappedPitch"] = {"hipD3D9ResourceGetMappedPitch", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cudaD3D9ResourceGetMappedPitch) + cuda2hipRename["cudaD3D9ResourceGetMappedPointer"] = {"hipD3D9ResourceGetMappedPointer", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9ResourceGetMappedPointer) + cuda2hipRename["cudaD3D9ResourceGetMappedSize"] = {"hipD3D9ResourceGetMappedSize", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9ResourceGetMappedSize) + cuda2hipRename["cudaD3D9ResourceGetSurfaceDimensions"] = {"hipD3D9ResourceGetSurfaceDimensions", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9ResourceGetSurfaceDimensions) + cuda2hipRename["cudaD3D9ResourceSetMapFlags"] = {"hipD3D9ResourceSetMapFlags", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9ResourceSetMapFlags) + cuda2hipRename["cudaD3D9UnmapResources"] = {"hipD3D9UnmapResources", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9UnmapResources) + cuda2hipRename["cudaD3D9UnregisterResource"] = {"hipD3D9UnregisterResource", CONV_D3D9, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D9UnregisterResource) + + // Direct3D 10 Interoperability + // enum cudaD3D10DeviceList + cuda2hipRename["cudaD3D10DeviceList"] = {"hipd3d10DeviceList", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d10DeviceList) + cuda2hipRename["cudaD3D10DeviceListAll"] = {"HIP_D3D10_DEVICE_LIST_ALL", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D10_DEVICE_LIST_ALL) + cuda2hipRename["cudaD3D10DeviceListCurrentFrame"] = {"HIP_D3D10_DEVICE_LIST_CURRENT_FRAME", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_D3D10_DEVICE_LIST_CURRENT_FRAME) + cuda2hipRename["cudaD3D10DeviceListNextFrame"] = {"HIP_D3D10_DEVICE_LIST_NEXT_FRAME", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 3 // API_Driver ANALOGUE (CU_D3D10_DEVICE_LIST_NEXT_FRAME) + + cuda2hipRename["cudaD3D10GetDevice"] = {"hipD3D10GetDevice", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10GetDevice) + cuda2hipRename["cudaD3D10GetDevices"] = {"hipD3D10GetDevices", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10GetDevices) + cuda2hipRename["cudaGraphicsD3D10RegisterResource"] = {"hipGraphicsD3D10RegisterResource", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsD3D10RegisterResource) + + // Direct3D 10 Interoperability [DEPRECATED] + // enum cudaD3D10MapFlags + cuda2hipRename["cudaD3D10MapFlags"] = {"hipD3D10MapFlags", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d10map_flags) + cuda2hipRename["cudaD3D10MapFlagsNone"] = {"HIP_D3D10_MAPRESOURCE_FLAGS_NONE", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_Driver ANALOGUE (CU_D3D10_MAPRESOURCE_FLAGS_NONE) + cuda2hipRename["cudaD3D10MapFlagsReadOnly"] = {"HIP_D3D10_MAPRESOURCE_FLAGS_READONLY", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D10_MAPRESOURCE_FLAGS_READONLY) + cuda2hipRename["cudaD3D10MapFlagsWriteDiscard"] = {"HIP_D3D10_MAPRESOURCE_FLAGS_WRITEDISCARD", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_D3D10_MAPRESOURCE_FLAGS_WRITEDISCARD) + + // enum cudaD3D10RegisterFlags + cuda2hipRename["cudaD3D10RegisterFlags"] = {"hipD3D10RegisterFlags", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d10Register_flags) + cuda2hipRename["cudaD3D10RegisterFlagsNone"] = {"HIP_D3D10_REGISTER_FLAGS_NONE", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 0 // API_Driver ANALOGUE (CU_D3D10_REGISTER_FLAGS_NONE) + cuda2hipRename["cudaD3D10RegisterFlagsArray"] = {"HIP_D3D10_REGISTER_FLAGS_ARRAY", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D10_REGISTER_FLAGS_ARRAY) + + cuda2hipRename["cudaD3D10GetDirect3DDevice"] = {"hipD3D10GetDirect3DDevice", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cudaD3D10GetDirect3DDevice) + cuda2hipRename["cudaD3D10MapResources"] = {"hipD3D10MapResources", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10MapResources) + cuda2hipRename["cudaD3D10RegisterResource"] = {"hipD3D10RegisterResource", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10RegisterResource) + cuda2hipRename["cudaD3D10ResourceGetMappedArray"] = {"hipD3D10ResourceGetMappedArray", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10ResourceGetMappedArray) + cuda2hipRename["cudaD3D10ResourceGetMappedPitch"] = {"hipD3D10ResourceGetMappedPitch", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cudaD3D10ResourceGetMappedPitch) + cuda2hipRename["cudaD3D10ResourceGetMappedPointer"] = {"hipD3D10ResourceGetMappedPointer", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10ResourceGetMappedPointer) + cuda2hipRename["cudaD3D10ResourceGetMappedSize"] = {"hipD3D10ResourceGetMappedSize", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10ResourceGetMappedSize) + cuda2hipRename["cudaD3D10ResourceGetSurfaceDimensions"] = {"hipD3D10ResourceGetSurfaceDimensions", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10ResourceGetSurfaceDimensions) + cuda2hipRename["cudaD3D10ResourceSetMapFlags"] = {"hipD3D10ResourceSetMapFlags", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10ResourceSetMapFlags) + cuda2hipRename["cudaD3D10SetDirect3DDevice"] = {"hipD3D10SetDirect3DDevice", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // no API_Driver ANALOGUE + cuda2hipRename["cudaD3D10UnmapResources"] = {"hipD3D10UnmapResources", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10UnmapResources) + cuda2hipRename["cudaD3D10UnregisterResource"] = {"hipD3D10UnregisterResource", CONV_D3D10, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D10UnregisterResource) + + // Direct3D 11 Interoperability + // enum cudaD3D11DeviceList + cuda2hipRename["cudaD3D11DeviceList"] = {"hipd3d11DeviceList", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUd3d11DeviceList) + cuda2hipRename["cudaD3D11DeviceListAll"] = {"HIP_D3D11_DEVICE_LIST_ALL", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_Driver ANALOGUE (CU_D3D11_DEVICE_LIST_ALL) + cuda2hipRename["cudaD3D11DeviceListCurrentFrame"] = {"HIP_D3D11_DEVICE_LIST_CURRENT_FRAME", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // 2 // API_Driver ANALOGUE (CU_D3D11_DEVICE_LIST_CURRENT_FRAME) + cuda2hipRename["cudaD3D11DeviceListNextFrame"] = {"HIP_D3D11_DEVICE_LIST_NEXT_FRAME", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // 3 // API_Driver ANALOGUE (CU_D3D11_DEVICE_LIST_NEXT_FRAME) + + cuda2hipRename["cudaD3D11GetDevice"] = {"hipD3D11GetDevice", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D11GetDevice) + cuda2hipRename["cudaD3D11GetDevices"] = {"hipD3D11GetDevices", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D11GetDevices) + cuda2hipRename["cudaGraphicsD3D11RegisterResource"] = {"hipGraphicsD3D11RegisterResource", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsD3D11RegisterResource) + + // Direct3D 11 Interoperability [DEPRECATED] + cuda2hipRename["cudaD3D11GetDevice"] = {"hipD3D11GetDevice", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D11GetDevice) + cuda2hipRename["cudaD3D11GetDevices"] = {"hipD3D11GetDevices", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuD3D11GetDevices) + cuda2hipRename["cudaGraphicsD3D11RegisterResource"] = {"hipGraphicsD3D11RegisterResource", CONV_D3D11, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsD3D11RegisterResource) + + // VDPAU Interoperability + cuda2hipRename["cudaGraphicsVDPAURegisterOutputSurface"] = {"hipGraphicsVDPAURegisterOutputSurface", CONV_VDPAU, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsVDPAURegisterOutputSurface) + cuda2hipRename["cudaGraphicsVDPAURegisterVideoSurface"] = {"hipGraphicsVDPAURegisterVideoSurface", CONV_VDPAU, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsVDPAURegisterVideoSurface) + cuda2hipRename["cudaVDPAUGetDevice"] = {"hipVDPAUGetDevice", CONV_VDPAU, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuVDPAUGetDevice) + cuda2hipRename["cudaVDPAUSetVDPAUDevice"] = {"hipVDPAUSetDevice", CONV_VDPAU, API_RUNTIME, HIP_UNSUPPORTED}; // no API_Driver ANALOGUE + + // EGL Interoperability + cuda2hipRename["cudaEglStreamConnection"] = {"hipEglStreamConnection", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUeglStreamConnection) + + cuda2hipRename["cudaEGLStreamConsumerAcquireFrame"] = {"hipEGLStreamConsumerAcquireFrame", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamConsumerAcquireFrame) + cuda2hipRename["cudaEGLStreamConsumerConnect"] = {"hipEGLStreamConsumerConnect", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamConsumerConnect) + cuda2hipRename["cudaEGLStreamConsumerConnectWithFlags"] = {"hipEGLStreamConsumerConnectWithFlags", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamConsumerConnectWithFlags) + cuda2hipRename["cudaEGLStreamConsumerReleaseFrame"] = {"hipEGLStreamConsumerReleaseFrame", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamConsumerReleaseFrame) + cuda2hipRename["cudaEGLStreamProducerConnect"] = {"hipEGLStreamProducerConnect", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamProducerConnect) + cuda2hipRename["cudaEGLStreamProducerDisconnect"] = {"hipEGLStreamProducerDisconnect", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamProducerDisconnect) + cuda2hipRename["cudaEGLStreamProducerPresentFrame"] = {"hipEGLStreamProducerPresentFrame", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamProducerPresentFrame) + cuda2hipRename["cudaEGLStreamProducerReturnFrame"] = {"hipEGLStreamProducerReturnFrame", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuEGLStreamProducerReturnFrame) + cuda2hipRename["cudaGraphicsEGLRegisterImage"] = {"hipGraphicsEGLRegisterImage", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsEGLRegisterImage) + cuda2hipRename["cudaGraphicsResourceGetMappedEglFrame"] = {"hipGraphicsResourceGetMappedEglFrame", CONV_EGL, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (cuGraphicsResourceGetMappedEglFrame) + //---------------------------------------BLAS-------------------------------------// // Blas types cuda2hipRename["cublasHandle_t"] = {"hipblasHandle_t", CONV_TYPE, API_BLAS}; From e63c8494fda0011d1085ea1ebc651e792ec1a523 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Wed, 26 Jul 2017 13:51:32 +0530 Subject: [PATCH 042/108] [cmake] No longer rebuild cmake cache on each build Cmake cache was being rebuilt on each build. This was being done to update HIP_VERSION, HCC_VERSION, .hipInfo and .hipVersion. However, rebuilding cache also re-runs HIT parser which is slow. Removing the cache rebuild should speed up the build. But user needs to explicitly rebuild the cache in case HIP_VERSION or HCC_VERSION changes by calling "make rebuild_cache" Change-Id: Ia5476eb7105aa614239c4dc7968c37f5e6cb0b29 --- CMakeLists.txt | 10 ---------- 1 file changed, 10 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index 7d4039b4b4..1f246b05ec 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -141,9 +141,6 @@ add_to_config(_buildInfo COMPILE_HIP_ATP_MARKER) ############################# # Build steps ############################# -# Rebuild cmake cache updates .hipInfo and .hipVersion -add_custom_target(update_build_and_version_info COMMAND make rebuild_cache) - # Build clang hipify if enabled add_subdirectory(hipify-clang) @@ -212,13 +209,6 @@ if(HIP_PLATFORM STREQUAL "hcc") endforeach() target_link_libraries(hip_hcc INTERFACE hcc::hccrt;hcc::hc_am) - # Generate hcc_version.txt - add_custom_target(query_hcc_version COMMAND ${HCC_HOME}/bin/hcc --version > ${PROJECT_BINARY_DIR}/hcc_version.tmp) - add_custom_target(check_hcc_version COMMAND ${CMAKE_COMMAND} -E copy_if_different ${PROJECT_BINARY_DIR}/hcc_version.tmp ${PROJECT_BINARY_DIR}/hcc_version.txt DEPENDS query_hcc_version) - set_source_files_properties(${PROJECT_BINARY_DIR}/hcc_version.txt PROPERTIES GENERATED TRUE) - set_source_files_properties(${SOURCE_FILES_RUNTIME} ${SOURCE_FILES_DEVICE} PROPERTIES OBJECT_DEPENDS ${PROJECT_BINARY_DIR}/hcc_version.txt) - add_dependencies(hip_hcc check_hcc_version update_build_and_version_info) - # Generate .hipInfo file(WRITE "${PROJECT_BINARY_DIR}/.hipInfo" ${_buildInfo}) endif() From 368aebc4bcea8da0bb42d8ea6a812c3780941569 Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Wed, 26 Jul 2017 19:29:50 +0300 Subject: [PATCH 043/108] [HIPIFY] CUDA Stream Callback support --- docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md | 2 +- docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md | 1 + hipify-clang/src/Cuda2Hip.cpp | 4 +++- 3 files changed, 5 insertions(+), 2 deletions(-) diff --git a/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md index bd6634dbe5..9a4700b19c 100644 --- a/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Driver_API_functions_supported_by_HIP.md @@ -599,7 +599,7 @@ | **CUDA** | **HIP** | |-----------------------------------------------------------|-------------------------------| -| `cuStreamAddCallback` | | +| `cuStreamAddCallback` | `hipStreamAddCallback` | | `cuStreamAttachMemAsync` | | | `cuStreamCreate` | | | `cuStreamCreateWithPriority` | | diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index b4c004e2dc..6627f5fd19 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -741,6 +741,7 @@ | 0x01 |*`cudaCSV`* | | | typedef | `cudaOutputMode_t` | | | typedef | `cudaStream_t` | `hipStream_t` | +| typedef | `cudaStreamCallback_t` | `hipStreamCallback_t` | | typedef | `cudaSurfaceObject_t` | | | typedef | `cudaTextureObject_t` | | | typedef | `CUuuid_stcudaUUID_t` | | diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index 138c3434a6..191cb8e56c 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -1019,7 +1019,7 @@ struct cuda2hipMap { cuda2hipRename["cuOccupancyMaxPotentialBlockSizeWithFlags"] = {"hipOccupancyMaxPotentialBlockSizeWithFlags", CONV_OCCUPANCY, API_DRIVER, HIP_UNSUPPORTED}; // API_Runtime ANALOGUE (cudaOccupancyMaxPotentialBlockSizeWithFlags) // Streams - cuda2hipRename["cuStreamAddCallback"] = {"hipStreamAddCallback", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; + cuda2hipRename["cuStreamAddCallback"] = {"hipStreamAddCallback", CONV_STREAM, API_DRIVER}; cuda2hipRename["cuStreamAttachMemAsync"] = {"hipStreamAttachMemAsync", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; cuda2hipRename["cuStreamCreate"] = {"hipStreamCreate__", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; // Not equal to cudaStreamCreate due to different signatures cuda2hipRename["cuStreamCreateWithPriority"] = {"hipStreamCreateWithPriority", CONV_STREAM, API_DRIVER, HIP_UNSUPPORTED}; @@ -1362,6 +1362,8 @@ struct cuda2hipMap { cuda2hipRename["cudaOccupancyDefault"] = {"hipOccupancyDefault", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_OCCUPANCY_DEFAULT = 0x0) cuda2hipRename["cudaOccupancyDisableCachingOverride"] = {"hipOccupancyDisableCachingOverride", CONV_TYPE, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_OCCUPANCY_DISABLE_CACHING_OVERRIDE = 0x1) + cuda2hipRename["cudaStreamCallback_t"] = {"hipStreamCallback_t", CONV_TYPE, API_RUNTIME}; + // Error API cuda2hipRename["cudaGetLastError"] = {"hipGetLastError", CONV_ERROR, API_RUNTIME}; cuda2hipRename["cudaPeekAtLastError"] = {"hipPeekAtLastError", CONV_ERROR, API_RUNTIME}; From d9d9a8620fabe5e286d02804a013b1f0b4842e28 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Sat, 22 Jul 2017 22:08:08 -0500 Subject: [PATCH 044/108] Update release notes. --- RELEASE.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/RELEASE.md b/RELEASE.md index d426f8563c..98c076dc6e 100644 --- a/RELEASE.md +++ b/RELEASE.md @@ -5,6 +5,10 @@ We have attempted to document known bugs and limitations - in particular the [HI =================================================================================================== +Upcoming: +- HIP development now fully open and on GitHub. Developers should submit pull requests. + + ## Revision History: =================================================================================================== From 8b8e97ff2824b146e9846261d6e367b415603cd8 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Wed, 26 Jul 2017 18:40:37 -0500 Subject: [PATCH 045/108] Make host memory allocations coherent by default. Associated change is to optimize event recording so it uses agent-scope releaes (since it was only using system-scope release to support non-coherent host mem). Flags and environment variables exist to obtain previous behavior if desired. Options are documented in new performance guide. --- README.md | 1 + docs/markdown/hip_performance.md | 39 ----------- docs/markdown/hip_programming_guide.md | 89 ++++++++++++++++++++++++++ src/hip_device.cpp | 1 + src/hip_hcc.cpp | 6 +- src/hip_hcc_internal.h | 2 +- src/hip_memory.cpp | 2 +- 7 files changed, 96 insertions(+), 44 deletions(-) delete mode 100644 docs/markdown/hip_performance.md create mode 100644 docs/markdown/hip_programming_guide.md diff --git a/README.md b/README.md index 568b105049..4f7f1a0123 100644 --- a/README.md +++ b/README.md @@ -32,6 +32,7 @@ HIP releases are typically of two types. The tag naming convention is different - [HIP Runtime API (Doxygen)](http://rocm-developer-tools.github.io/HIP) - [HIP Porting Guide](docs/markdown/hip_porting_guide.md) - [HIP Porting Driver Guide](docs/markdown/hip_porting_driver_api.md) +- [HIP Programming Guide](docs/markdown/hip_programming_guide.md) - [HIP Profiling ](docs/markdown/hip_profiling.md) - [HIP Debugging](docs/markdown/hip_debugging.md) - [HIP Terminology](docs/markdown/hip_terms.md) (including Rosetta Stone of GPU computing terms across CUDA/HIP/HC/AMP/OpenL) diff --git a/docs/markdown/hip_performance.md b/docs/markdown/hip_performance.md deleted file mode 100644 index 67a2f88b58..0000000000 --- a/docs/markdown/hip_performance.md +++ /dev/null @@ -1,39 +0,0 @@ -# HIP Performance Optimizations - -Please note that this document lists possible ways for experimenting with HIP stack to gain performance. Performance may vary from platform to platform. - -### Unpinned Memory Transfer Optimizations - -#### On Small BAR Setup - -There are two possible ways to transfer data from host-to-device (H2D) and device-to-host(D2H) - * Using Staging Buffers - * Using PinInPlace - -#### On Large BAR Setup - -There are three possible ways to transfer data from host-to-device (H2D) - * Using Staging Buffers - * Using PinInPlace - * Direct Memcpy - - And there are two possible ways to transfer data from device-to-host (D2H) - * Using Staging Buffers - * Using PinInPlace - -Some GPUs may not be able to directly access host memory, and in these cases we need to -stage the copy through an optimized pinned staging buffer, to implement H2D and D2H copies.The copy is broken into buffer-sized chunks to limit the size of the buffer and also to provide better performance by overlapping the CPU copies with the DMA copies. - -PinInPlace is another algorithm which pins the host memory "in-place", and copies it with the DMA engine. - -By default staging buffers are used for unpinned memory transfers. Environment variables allow control over the unpinned copy algorithm and parameters: - -- HIP_PININPLACE - This environment variable forces the use of PinInPlace logic for all unpinned memory copies - -- HIP_OPTIMAL_MEM_TRANSFER- This environment variable enables a hybrid memory copy logic based on thresholds. These thresholds can be managed with following environment variables: - - HIP_H2D_MEM_TRANSFER_THRESHOLD_STAGING_OR_PININPLACE - Threshold in bytes for H2D copy. For sizes smaller than threshold staging buffers logic would be used else PinInPlace logic. - - HIP_H2D_MEM_TRANSFER_THRESHOLD_DIRECT_OR_STAGING - Threshold in bytes for H2D copy. For sizes smaller than threshold direct copy logic would be used else staging buffers logic. - - HIP_D2H_MEM_TRANSFER_THRESHOLD - Threshold in bytes for D2H copy. For sizes smaller than threshold staging buffer logic would be used else PinInPlace logic. - - - diff --git a/docs/markdown/hip_programming_guide.md b/docs/markdown/hip_programming_guide.md new file mode 100644 index 0000000000..5d0c1f2497 --- /dev/null +++ b/docs/markdown/hip_programming_guide.md @@ -0,0 +1,89 @@ +# HIP Programming Guide + +## Host Memory + +### Introduction +hipHostMemory allocates pinned host memory which is mapped into the address space of all GPUs in the system. +There are two use cases for this host memory: +- Faster HostToDevice and DeviceToHost Data Transfers: +The runtime tracks the hipHostMalloc allocations and can avoid some of the setup required for regular unpinned memory. For exact measurements on a specific system, experiment with --unpinned and --pinned switches for the hipBusBandwidth tool. +- Zero-Copy GPU Access: +GPU can directly access the host memory over the CPU/GPU interconnect, without need to copy the data. This avoids the need for the copy, but during the kernel access each memory access must traverse the interconnect, which can be tens of times slower than accessing the GPU's local device memory. Zero-copy memory can be a good choice when the memory accesses are infrequent (perhaps only once). Zero-copy memory is typically "Coherent" and thus not cached by the GPU but this can be overridden if desired and is explained in more detail below. + +### Memory allocation flags +hipHostMalloc always sets the hipHostMallocPortable and hipHostMallocMapped flags. Both usage models described above use the same allocation flags, and the difference is in how the surrounding code uses the host memory. +See the hipHostMalloc API for more information. + + +### Coherency Controls +ROCm defines two coherency options for host memory: +- Coherent memory : Supports fine-grain synchronization while the kernel is running.  For example, a kernel can perform atomic operations that are visible to the host CPU or to other (peer) GPUs.  Synchronization instructions include threadfence_system and C++11-style atomic operations.   However, coherent memory cannot be cached by the GPU and thus may have lower performance. +- Non-coherent memory : Can be cached by GPU, but cannot support synchronization while the kernel is running.  Non-coherent memory can be optionally synchronized only at command (end-of-kernel or copy command) boundaries.  This memory is appropriate for high-performance access when fine-grain synchronization is not required. + +IP provides the developer with controls to select which type of memory is used via allocation flags passed to hipHostMalloc and the HIP_HOST_COHERENT environment variable: +- hipHostllocCoherent=0, hipHostMallocNonCoherent=0: Use HIP_HOST_COHERENT environment variable: + - If HIP_HOST_COHERENT is 1 or undefined, the host memory allocation is coherent. + - If host memory is `defined and 0: the host memory allocation is non-coherent. +- hipHostMallocCoherent=1, hipHostMallocNonCoherent=0: The host memory allocation will be coherent.  HIP_HOST_COHERENT env variable is ignored. +- hipHostMallocCoherent=0, hipHostMallocNonCoherent=1: The host memory allocation will be non-coherent.  HIP_HOST_COHERENT env variable is ignored. +- hipHostMallocCoherent=1, hipHostMallocNonCoherent=1: Illegal. + + +### Visibility of Zero-Copy Host Memory +Coherent host memory is automatically visible at synchronization points. +Non-coherent + +| HIP API | Synchronization Effect | Fence | Coherent Host Memory Visibiity | Non-Coherent Host Memory Visibility| +| --- | --- | --- | --- | --- | +| hipStreamSynchronize | host waits for all commands in the specified stream to complete | system-scope release | yes | yes | +| hipDeviceSynchronize | host waits for all commands in all streams on the specified device to complete | system-scope release | yes | yes | +| hipEventSynchronize | host waits for the specified event to complete | device-scope release | yes | depends - see below| +| hipStreamWaitEvent | stream waits for the specified event to complete | none | yes | no | + + +### hipEventSynchronize +Developers can control the release scope for hipEvents: +- By default, the GPU performs a device-scope acquire and release operation with each recorded event.  This will make host and device memory visible to other commands executing on the same device.  + +A stronger system-level fence can be specified when the event is created with hipEventCreateWithFlags: +- hipEventReleaseToSystem : Perform a system-scope release operation when the event is recorded.  This will make both Coherent and Non-Coherent host memory visible to other agents in the system, but may involve heavyweight operations such as cache flushing.  Coherent memory will typically use lighter-weight in-kernel synchronization mechanisms such as an atomic operation and thus does not need to use hipEventReleaseToSystem. + +### Summary and Recommendations: + +- Coherent host memory is the default and is the easiest to use since the memory is visible to the CPU at typical synchronization points. This memory allows in-kernel synchronization commands such as threadfence_system to work transparently. +- HIP/ROCm also supports the ability to cache host memory in the GPU using the "Non-Coherent" host memory allocations. This can provide performance benefit, but care must be taken to use the correct synchronization. + + +## Unpinned Memory Transfer Optimizations +Please note that this document lists possible ways for experimenting with HIP stack to gain performance. Performance may vary from platform to platform. + +### On Small BAR Setup + +There are two possible ways to transfer data from host-to-device (H2D) and device-to-host(D2H) + * Using Staging Buffers + * Using PinInPlace + +### On Large BAR Setup + +There are three possible ways to transfer data from host-to-device (H2D) + * Using Staging Buffers + * Using PinInPlace + * Direct Memcpy + + And there are two possible ways to transfer data from device-to-host (D2H) + * Using Staging Buffers + * Using PinInPlace + +Some GPUs may not be able to directly access host memory, and in these cases we need to +stage the copy through an optimized pinned staging buffer, to implement H2D and D2H copies.The copy is broken into buffer-sized chunks to limit the size of the buffer and also to provide better performance by overlapping the CPU copies with the DMA copies. + +PinInPlace is another algorithm which pins the host memory "in-place", and copies it with the DMA engine. + +By default staging buffers are used for unpinned memory transfers. Environment variables allow control over the unpinned copy algorithm and parameters: + +- HIP_PININPLACE - This environment variable forces the use of PinInPlace logic for all unpinned memory copies + +- HIP_OPTIMAL_MEM_TRANSFER- This environment variable enables a hybrid memory copy logic based on thresholds. These thresholds can be managed with following environment variables: + - HIP_H2D_MEM_TRANSFER_THRESHOLD_STAGING_OR_PININPLACE - Threshold in bytes for H2D copy. For sizes smaller than threshold staging buffers logic would be used else PinInPlace logic. + - HIP_H2D_MEM_TRANSFER_THRESHOLD_DIRECT_OR_STAGING - Threshold in bytes for H2D copy. For sizes smaller than threshold direct copy logic would be used else staging buffers logic. + - HIP_D2H_MEM_TRANSFER_THRESHOLD - Threshold in bytes for D2H copy. For sizes smaller than threshold staging buffer logic would be used else PinInPlace logic. diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 9086cd8012..7e59e37cad 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -180,6 +180,7 @@ hipError_t hipDeviceReset(void) return ihipLogStatus(hipSuccess); } + hipError_t ihipDeviceSetState(void) { hipError_t e = hipErrorInvalidContext; diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 9d431df2dd..d4f7723dbc 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -83,8 +83,8 @@ int HIP_HIDDEN_FREE_MEM = 256; int HIP_FORCE_SYNC_COPY = 0; // TODO - set these to 0 and 1 -int HIP_EVENT_SYS_RELEASE=1; -int HIP_COHERENT_HOST_ALLOC = 0; +int HIP_EVENT_SYS_RELEASE=0; +int HIP_HOST_COHERENT = 1; // TODO - set to 0 once we resolve stability. // USE_ HIP_SYNC_HOST_ALLOC @@ -1226,7 +1226,7 @@ void HipReadEnv() READ_ENV_I(release, HIP_SYNC_NULL_STREAM, 0, "Synchronize on host for null stream submissions"); - READ_ENV_I(release, HIP_COHERENT_HOST_ALLOC, 0, "If set, all host memory will be allocated as fine-grained system memory. This allows threadfence_system to work but prevents host memory from being cached on GPU which may have performance impact."); + READ_ENV_I(release, HIP_HOST_COHERENT, 0, "If set, all host memory will be allocated as fine-grained system memory. This allows threadfence_system to work but prevents host memory from being cached on GPU which may have performance impact."); READ_ENV_I(release, HCC_OPT_FLUSH, 0, "When set, use agent-scope fence operations rather than system-scope fence operationsflush when possible. This flag controls both HIP and HCC behavior."); diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index 1499406780..88c7eedda0 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -59,7 +59,7 @@ extern int HIP_STREAM_SIGNALS; /* number of signals to allocate at stream creat extern int HIP_VISIBLE_DEVICES; /* Contains a comma-separated sequence of GPU identifiers */ extern int HIP_FORCE_P2P_HOST; -extern int HIP_COHERENT_HOST_ALLOC; +extern int HIP_HOST_COHERENT; extern int HIP_HIDDEN_FREE_MEM; //--- diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 6728ae42d3..fc7ca323e0 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -293,7 +293,7 @@ hipError_t hipHostMalloc(void** ptr, size_t sizeBytes, unsigned int flags) amFlags = amHostNonCoherent; } else { // depends on env variables: - amFlags = HIP_COHERENT_HOST_ALLOC ? amHostCoherent : amHostNonCoherent; + amFlags = HIP_HOST_COHERENT ? amHostCoherent : amHostNonCoherent; } From ffcf5d32a07e517c35dfd317f35bfb9b8ebe76c1 Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Thu, 27 Jul 2017 18:39:40 +0300 Subject: [PATCH 046/108] [HIPIFY] CUDA RT Textures support --- ..._Runtime_API_functions_supported_by_HIP.md | 150 ++++++++-------- hipify-clang/src/Cuda2Hip.cpp | 167 +++++++++--------- 2 files changed, 157 insertions(+), 160 deletions(-) diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index 6627f5fd19..dde2a04b22 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -125,7 +125,7 @@ | `cudaHostUnregister` | `hipHostUnregister` | | `cudaMalloc` | `hipMalloc` | | `cudaMalloc3D` | | -| `cudaMalloc3DArray` | | +| `cudaMalloc3DArray` | `hipMalloc2DArray` | | `cudaMallocArray` | `hipMallocArray` | | `cudaMallocHost` | `hipHostMalloc` | | `cudaMallocManaged` | | @@ -140,7 +140,7 @@ | `cudaMemcpy2DFromArrayAsync` | | | `cudaMemcpy2DToArray` | `hipMemcpy2DToArray` | | `cudaMemcpy2DToArrayAsync` | | -| `cudaMemcpy3D` | | +| `cudaMemcpy3D` | `hipMemcpy3D` | | `cudaMemcpy3DAsync` | | | `cudaMemcpy3DPeer` | | | `cudaMemcpy3DPeerAsync` | | @@ -157,14 +157,14 @@ | `cudaMemcpyToSymbol` | `hipMemcpyToSymbol` | | `cudaMemcpyToSymbolAsync` | `hipMemcpyToSymbolAsync` | | `cudaMemset` | `hipMemset` | -| `cudaMemset2D` | | +| `cudaMemset2D` | `hipMemset2D` | | `cudaMemset2DAsync` | | | `cudaMemset3D` | | | `cudaMemset3DAsync` | | | `cudaMemsetAsync` | `hipMemsetAsync` | -| `make_cudaExtent` | | -| `make_cudaPitchedPtr` | | -| `make_cudaPos` | | +| `make_cudaExtent` | `make_hipExtent` | +| `make_cudaPitchedPtr` | `make_hipPitchedPtr` | +| `make_cudaPos` | `make_hipPos` | ## **10. Unified Addressing** @@ -308,13 +308,13 @@ | **CUDA** | **HIP** | |-----------------------------------------------------------|-------------------------------| | `cudaBindTexture` | | -| `cudaBindTexture2D` | | -| `cudaBindTextureToArray` | | -| `cudaBindTextureToMipmappedArray` | | +| `cudaBindTexture2D` | `hipBindTexture2D` | +| `cudaBindTextureToArray` | `hipBindTextureToArray` | +| `cudaBindTextureToMipmappedArray` | `hipBindTextureToMipmappedArray` | | `cudaCreateChannelDesc` | | | `cudaGetChannelDesc` | | -| `cudaGetTextureAlignmentOffset` | | -| `cudaGetTextureReference` | | +| `cudaGetTextureAlignmentOffset` | `hipGetTextureAlignmentOffset`| +| `cudaGetTextureReference` | `hipGetTextureReference` | | `cudaUnbindTexture` | | ## **24. Surface Reference Management** @@ -328,11 +328,11 @@ | **CUDA** | **HIP** | |-----------------------------------------------------------|-------------------------------| -| `cudaCreateTextureObject` | | -| `cudaDestroyTextureObject` | | -| `cudaGetTextureObjectResourceDesc` | | -| `cudaGetTextureObjectResourceViewDesc` | | -| `cudaGetTextureObjectTextureDesc` | | +| `cudaCreateTextureObject` |`hipCreateTextureObject` | +| `cudaDestroyTextureObject` |`hipDestroyTextureObject` | +| `cudaGetTextureObjectResourceDesc` |`hipGetTextureObjectResourceDesc` | +| `cudaGetTextureObjectResourceViewDesc` |`hipGetTextureObjectResourceViewDesc` | +| `cudaGetTextureObjectTextureDesc` |`hipGetTextureObjectTextureDesc` | ## **26. Surface Object Management** @@ -399,18 +399,18 @@ |-------------:|-----------------------------------------------|------------------------------------------------------| | struct | `cudaChannelFormatDesc` | `hipChannelFormatDesc` | | struct | `cudaDeviceProp` | `hipDeviceProp_t` | -| struct | `cudaExtent` | | +| struct | `cudaExtent` | `hipExtent` | | struct | `cudaFuncAttributes` | | | struct | `cudaIpcEventHandle_t` | `hipIpcEventHandle_t` | | struct | `cudaIpcMemHandle_t` | `hipIpcMemHandle_t` | -| struct | `cudaMemcpy3DParms` | | +| struct | `cudaMemcpy3DParms` | `hipMemcpy3DParms` | | struct | `cudaMemcpy3DPeerParms` | | -| struct | `cudaPitchedPtr` | | +| struct | `cudaPitchedPtr` | `hipPitchedPtr` | | struct | `cudaPointerAttributes` | `hipPointerAttribute_t` | -| struct | `cudaPos` | | -| struct | `cudaResourceDesc` | | -| struct | `cudaResourceViewDesc` | | -| struct | `cudaTextureDesc` | | +| struct | `cudaPos` | `hipPos` | +| struct | `cudaResourceDesc` | `hipResourceDesc` | +| struct | `cudaResourceViewDesc` | `hipResourceViewDesc` | +| struct | `cudaTextureDesc` | `hipTextureDesc` | | struct | `surfaceReference` | | | struct | `textureReference` | `textureReference` | | enum |***`cudaChannelFormatKind`*** |***`hipChannelFormatKind`*** | @@ -630,47 +630,47 @@ | enum |***`cudaMemoryType`*** |***`hipMemoryType`*** | | 1 |*`cudaMemoryTypeHost`* |*`hipMemoryTypeHost`* | | 2 |*`cudaMemoryTypeDevice`* |*`hipMemoryTypeDevice`* | -| enum |***`cudaResourceType`*** | | -| 0 |*`cudaResourceTypeArray`* | | -| 1 |*`cudaResourceTypeMipmappedArray`* | | -| 2 |*`cudaResourceTypeLinear`* | | -| 3 |*`cudaResourceTypePitch2D`* | | -| enum |***`cudaResourceViewFormat`*** | | -| 0x00 |*`cudaResViewFormatNone`* | | -| 0x01 |*`cudaResViewFormatUnsignedChar1`* | | -| 0x02 |*`cudaResViewFormatUnsignedChar2`* | | -| 0x03 |*`cudaResViewFormatUnsignedChar4`* | | -| 0x04 |*`cudaResViewFormatSignedChar1`* | | -| 0x05 |*`cudaResViewFormatSignedChar2`* | | -| 0x06 |*`cudaResViewFormatSignedChar4`* | | -| 0x07 |*`cudaResViewFormatUnsignedShort1`* | | -| 0x08 |*`cudaResViewFormatUnsignedShort2`* | | -| 0x09 |*`cudaResViewFormatUnsignedShort4`* | | -| 0x0a |*`cudaResViewFormatSignedShort1`* | | -| 0x0b |*`cudaResViewFormatSignedShort2`* | | -| 0x0c |*`cudaResViewFormatSignedShort4`* | | -| 0x0d |*`cudaResViewFormatUnsignedInt1`* | | -| 0x0e |*`cudaResViewFormatUnsignedInt2`* | | -| 0x0f |*`cudaResViewFormatUnsignedInt4`* | | -| 0x10 |*`cudaResViewFormatSignedInt1`* | | -| 0x11 |*`cudaResViewFormatSignedInt2`* | | -| 0x12 |*`cudaResViewFormatSignedInt4`* | | -| 0x13 |*`cudaResViewFormatHalf1`* | | -| 0x14 |*`cudaResViewFormatHalf2`* | | -| 0x15 |*`cudaResViewFormatHalf4`* | | -| 0x16 |*`cudaResViewFormatFloat1`* | | -| 0x17 |*`cudaResViewFormatFloat2`* | | -| 0x18 |*`cudaResViewFormatFloat4`* | | -| 0x19 |*`cudaResViewFormatUnsignedBlockCompressed1`* | | -| 0x1a |*`cudaResViewFormatUnsignedBlockCompressed2`* | | -| 0x1b |*`cudaResViewFormatUnsignedBlockCompressed3`* | | -| 0x1c |*`cudaResViewFormatUnsignedBlockCompressed4`* | | -| 0x1d |*`cudaResViewFormatSignedBlockCompressed4`* | | -| 0x1e |*`cudaResViewFormatUnsignedBlockCompressed5`* | | -| 0x1f |*`cudaResViewFormatSignedBlockCompressed5`* | | -| 0x20 |*`cudaResViewFormatUnsignedBlockCompressed6H`* | | -| 0x21 |*`cudaResViewFormatSignedBlockCompressed6H`* | | -| 0x22 |*`cudaResViewFormatUnsignedBlockCompressed7`* | | +| enum |***`cudaResourceType`*** |***`hipResourceType`*** | +| 0 |*`cudaResourceTypeArray`* |*`hipResourceTypeArray`* | +| 1 |*`cudaResourceTypeMipmappedArray`* |*`hipResourceTypeMipmappedArray`* | +| 2 |*`cudaResourceTypeLinear`* |*`hipResourceTypeLinear`* | +| 3 |*`cudaResourceTypePitch2D`* |*`hipResourceTypePitch2D`* | +| enum |***`cudaResourceViewFormat`*** |***`hipResourceViewFormat`*** | +| 0x00 |*`cudaResViewFormatNone`* |*`hipResViewFormatNone`* | +| 0x01 |*`cudaResViewFormatUnsignedChar1`* |*`hipResViewFormatUnsignedChar1`* | +| 0x02 |*`cudaResViewFormatUnsignedChar2`* |*`hipResViewFormatUnsignedChar2`* | +| 0x03 |*`cudaResViewFormatUnsignedChar4`* |*`hipResViewFormatUnsignedChar4`* | +| 0x04 |*`cudaResViewFormatSignedChar1`* |*`hipResViewFormatSignedChar1`* | +| 0x05 |*`cudaResViewFormatSignedChar2`* |*`hipResViewFormatSignedChar2`* | +| 0x06 |*`cudaResViewFormatSignedChar4`* |*`hipResViewFormatSignedChar4`* | +| 0x07 |*`cudaResViewFormatUnsignedShort1`* |*`hipResViewFormatUnsignedShort1`* | +| 0x08 |*`cudaResViewFormatUnsignedShort2`* |*`hipResViewFormatUnsignedShort2`* | +| 0x09 |*`cudaResViewFormatUnsignedShort4`* |*`hipResViewFormatUnsignedShort4`* | +| 0x0a |*`cudaResViewFormatSignedShort1`* |*`hipResViewFormatSignedShort1`* | +| 0x0b |*`cudaResViewFormatSignedShort2`* |*`hipResViewFormatSignedShort2`* | +| 0x0c |*`cudaResViewFormatSignedShort4`* |*`hipResViewFormatSignedShort4`* | +| 0x0d |*`cudaResViewFormatUnsignedInt1`* |*`hipResViewFormatUnsignedInt1`* | +| 0x0e |*`cudaResViewFormatUnsignedInt2`* |*`hipResViewFormatUnsignedInt2`* | +| 0x0f |*`cudaResViewFormatUnsignedInt4`* |*`hipResViewFormatUnsignedInt4`* | +| 0x10 |*`cudaResViewFormatSignedInt1`* |*`hipResViewFormatSignedInt1`* | +| 0x11 |*`cudaResViewFormatSignedInt2`* |*`hipResViewFormatSignedInt2`* | +| 0x12 |*`cudaResViewFormatSignedInt4`* |*`hipResViewFormatSignedInt4`* | +| 0x13 |*`cudaResViewFormatHalf1`* |*`hipResViewFormatHalf1`* | +| 0x14 |*`cudaResViewFormatHalf2`* |*`hipResViewFormatHalf2`* | +| 0x15 |*`cudaResViewFormatHalf4`* |*`hipResViewFormatHalf4`* | +| 0x16 |*`cudaResViewFormatFloat1`* |*`hipResViewFormatFloat1`* | +| 0x17 |*`cudaResViewFormatFloat2`* |*`hipResViewFormatFloat2`* | +| 0x18 |*`cudaResViewFormatFloat4`* |*`hipResViewFormatFloat4`* | +| 0x19 |*`cudaResViewFormatUnsignedBlockCompressed1`* |*`hipResViewFormatUnsignedBlockCompressed1`* | +| 0x1a |*`cudaResViewFormatUnsignedBlockCompressed2`* |*`hipResViewFormatUnsignedBlockCompressed2`* | +| 0x1b |*`cudaResViewFormatUnsignedBlockCompressed3`* |*`hipResViewFormatUnsignedBlockCompressed3`* | +| 0x1c |*`cudaResViewFormatUnsignedBlockCompressed4`* |*`hipResViewFormatUnsignedBlockCompressed4`* | +| 0x1d |*`cudaResViewFormatSignedBlockCompressed4`* |*`hipResViewFormatSignedBlockCompressed4`* | +| 0x1e |*`cudaResViewFormatUnsignedBlockCompressed5`* |*`hipResViewFormatUnsignedBlockCompressed5`* | +| 0x1f |*`cudaResViewFormatSignedBlockCompressed5`* |*`hipResViewFormatSignedBlockCompressed5`* | +| 0x20 |*`cudaResViewFormatUnsignedBlockCompressed6H`* |*`hipResViewFormatUnsignedBlockCompressed6H`* | +| 0x21 |*`cudaResViewFormatSignedBlockCompressed6H`* |*`hipResViewFormatSignedBlockCompressed6H`* | +| 0x22 |*`cudaResViewFormatUnsignedBlockCompressed7`* |*`hipResViewFormatUnsignedBlockCompressed7`* | | enum |***`cudaSharedMemConfig`*** |***`hipSharedMemConfig`*** | | 0 |*`cudaSharedMemBankSizeDefault`* |*`hipSharedMemBankSizeDefault`* | | 1 |*`cudaSharedMemBankSizeFourByte`* |*`hipSharedMemBankSizeFourByte`* | @@ -682,17 +682,17 @@ | enum |***`cudaSurfaceFormatMode`*** | | | 0 |*`cudaFormatModeForced`* | | | 1 |*`cudaFormatModeAuto`* | | -| enum |***`cudaTextureAddressMode`*** | | -| 0 |*`cudaAddressModeWrap`* | | -| 1 |*`cudaAddressModeClamp`* | | -| 2 |*`cudaAddressModeMirror`* | | -| 3 |*`cudaAddressModeBorder`* | | +| enum |***`cudaTextureAddressMode`*** |***`hipTextureAddressMode`*** | +| 0 |*`cudaAddressModeWrap`* |*`hipAddressModeWrap`* | +| 1 |*`cudaAddressModeClamp`* |*`hipAddressModeClamp`* | +| 2 |*`cudaAddressModeMirror`* |*`hipAddressModeMirror`* | +| 3 |*`cudaAddressModeBorder`* |*`hipAddressModeBorder`* | | enum |***`cudaTextureFilterMode`*** |***`hipTextureFilterMode`*** | | 0 |*`cudaFilterModePoint`* |*`hipFilterModePoint`* | -| 1 |*`cudaFilterModeLinear`* | | +| 1 |*`cudaFilterModeLinear`* |*`hipFilterModeLinear`* | | enum |***`cudaTextureReadMode`*** |***`hipTextureReadMode`*** | | 0 |*`cudaReadModeElementType`* |*`hipReadModeElementType`* | -| 1 |*`cudaReadModeNormalizedFloat`* | | +| 1 |*`cudaReadModeNormalizedFloat`* |*`hipReadModeNormalizedFloat`* | | enum |***`cudaGLDeviceList`*** | | | 0x01 |*`cudaGLDeviceListAll`* | | | 0x02 |*`cudaGLDeviceListCurrentFrame`* | | @@ -728,14 +728,14 @@ | 2 |*`cudaD3D11DeviceListCurrentFrame`* | | | 3 |*`cudaD3D11DeviceListNextFrame`* | | | struct | `cudaArray` | `hipArray` | -| typedef | `cudaArray_t` | `hipArray *` | -| typedef | `cudaArray_const_t` | `const hipArray *` | +| typedef | `cudaArray_t` | `hipArray_t` | +| typedef | `cudaArray_const_t` | `hipArray_const_t` | | enum | `cudaError` | `hipError_t` | | typedef | `cudaError_t` | `hipError_t` | | typedef | `cudaEvent_t` | `hipEvent_t` | | typedef | `cudaGraphicsResource_t` | | -| typedef | `cudaMipmappedArray_t` | | -| typedef | `cudaMipmappedArray_const_t` | | +| typedef | `cudaMipmappedArray_t` | `hipMipmappedArray_t` | +| typedef | `cudaMipmappedArray_const_t` | `hipMipmappedArray_const_t` | | enum |***`cudaOutputMode`*** | | | 0x00 |*`cudaKeyValuePair`* | | | 0x01 |*`cudaCSV`* | | diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index 191cb8e56c..c36ca5369a 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -1373,17 +1373,15 @@ struct cuda2hipMap { // Arrays cuda2hipRename["cudaArray"] = {"hipArray", CONV_MEM, API_RUNTIME}; // typedef struct cudaArray *cudaArray_t; - cuda2hipRename["cudaArray_t"] = {"hipArray *", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaArray_t"] = {"hipArray_t", CONV_MEM, API_RUNTIME}; // typedef const struct cudaArray *cudaArray_const_t; - cuda2hipRename["cudaArray_const_t"] = {"const hipArray *", CONV_MEM, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaMipmappedArray_t"] = {"hipMipmappedArray *", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaMipmappedArray_const_t"] = {"const hipMipmappedArray *", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaArray_const_t"] = {"hipArray_const_t", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaMipmappedArray_t"] = {"hipMipmappedArray_t", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaMipmappedArray_const_t"] = {"hipMipmappedArray_const_t", CONV_MEM, API_RUNTIME}; // memcpy // memcpy structs - // unsupported yet by HIP - cuda2hipRename["cudaMemcpy3DParms"] = {"hipMemcpy3DParms", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaMemcpy3DParms"] = {"hipMemcpy3DParms", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMemcpy3DPeerParms"] = {"hipMemcpy3DPeerParms", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; // memcpy functions @@ -1400,7 +1398,7 @@ struct cuda2hipMap { cuda2hipRename["cudaMemcpy2DFromArray"] = {"hipMemcpy2DFromArray", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemcpy2DFromArrayAsync"] = {"hipMemcpy2DFromArrayAsync", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemcpy2DToArrayAsync"] = {"hipMemcpy2DToArrayAsync", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaMemcpy3D"] = {"hipMemcpy3D", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaMemcpy3D"] = {"hipMemcpy3D", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMemcpy3DAsync"] = {"hipMemcpy3DAsync", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemcpy3DPeer"] = {"hipMemcpy3DPeer", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemcpy3DPeerAsync"] = {"hipMemcpy3DPeerAsync", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; @@ -1439,8 +1437,8 @@ struct cuda2hipMap { // memset cuda2hipRename["cudaMemset"] = {"hipMemset", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMemsetAsync"] = {"hipMemsetAsync", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaMemset2D"] = {"hipMemset2D", CONV_MEM, API_RUNTIME}; // unsupported yet by HIP - cuda2hipRename["cudaMemset2D"] = {"hipMemset2D", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemset2DAsync"] = {"hipMemset2DAsync", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemset3D"] = {"hipMemset3D", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMemset3DAsync"] = {"hipMemset3DAsync", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; @@ -1459,9 +1457,8 @@ struct cuda2hipMap { cuda2hipRename["cudaMalloc"] = {"hipMalloc", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMallocHost"] = {"hipHostMalloc", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMallocArray"] = {"hipMallocArray", CONV_MEM, API_RUNTIME}; - // unsupported yet by HIP cuda2hipRename["cudaMalloc3D"] = {"hipMalloc3D", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaMalloc3DArray"] = {"hipMalloc3DArray", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaMalloc3DArray"] = {"hipMalloc3DArray", CONV_MEM, API_RUNTIME}; cuda2hipRename["cudaMallocManaged"] = {"hipMallocManaged", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMallocMipmappedArray"] = {"hipMallocMipmappedArray", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; cuda2hipRename["cudaMallocPitch"] = {"hipMallocPitch", CONV_MEM, API_RUNTIME}; @@ -1480,14 +1477,13 @@ struct cuda2hipMap { cuda2hipRename["cudaMemoryTypeDevice"] = {"hipMemoryTypeDevice", CONV_MEM, API_RUNTIME}; // make memory functions - // unsupported yet by HIP - cuda2hipRename["make_cudaExtent"] = {"make_hipExtent", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["make_cudaPitchedPtr"] = {"make_hipPitchedPtr", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["make_cudaPos"] = {"make_hipPos", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["make_cudaExtent"] = {"make_hipExtent", CONV_MEM, API_RUNTIME}; + cuda2hipRename["make_cudaPitchedPtr"] = {"make_hipPitchedPtr", CONV_MEM, API_RUNTIME}; + cuda2hipRename["make_cudaPos"] = {"make_hipPos", CONV_MEM, API_RUNTIME}; - cuda2hipRename["cudaExtent"] = {"hipExtent", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaPitchedPtr"] = {"hipPitchedPtr", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaPos"] = {"hipPos", CONV_MEM, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaExtent"] = {"hipExtent", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaPitchedPtr"] = {"hipPitchedPtr", CONV_MEM, API_RUNTIME}; + cuda2hipRename["cudaPos"] = {"hipPos", CONV_MEM, API_RUNTIME}; // Host Malloc Flags cuda2hipRename["cudaHostAllocDefault"] = {"hipHostMallocDefault", CONV_MEM, API_RUNTIME}; @@ -1818,21 +1814,19 @@ struct cuda2hipMap { // enums cuda2hipRename["cudaTextureReadMode"] = {"hipTextureReadMode", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaReadModeElementType"] = {"hipReadModeElementType", CONV_TEX, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaReadModeNormalizedFloat"] = {"hipReadModeNormalizedFloat", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaReadModeNormalizedFloat"] = {"hipReadModeNormalizedFloat", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaTextureFilterMode"] = {"hipTextureFilterMode", CONV_TEX, API_RUNTIME}; // API_DRIVER ANALOGUE (CUfilter_mode) cuda2hipRename["cudaFilterModePoint"] = {"hipFilterModePoint", CONV_TEX, API_RUNTIME}; // 0 // API_DRIVER ANALOGUE (CU_TR_FILTER_MODE_POINT = 0) - cuda2hipRename["cudaFilterModeLinear"] = {"hipFilterModeLinear", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 1 // API_DRIVER ANALOGUE (CU_TR_FILTER_MODE_POINT = 1) + cuda2hipRename["cudaFilterModeLinear"] = {"hipFilterModeLinear", CONV_TEX, API_RUNTIME}; // 1 // API_DRIVER ANALOGUE (CU_TR_FILTER_MODE_POINT = 1) cuda2hipRename["cudaBindTexture"] = {"hipBindTexture", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaUnbindTexture"] = {"hipUnbindTexture", CONV_TEX, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaBindTexture2D"] = {"hipBindTexture2D", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaBindTextureToArray"] = {"hipBindTextureToArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaBindTextureToMipmappedArray"] = {"hipBindTextureToMipmappedArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureAlignmentOffset"] = {"hipGetTextureAlignmentOffset", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureReference"] = {"hipGetTextureReference", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaBindTexture2D"] = {"hipBindTexture2D", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaBindTextureToArray"] = {"hipBindTextureToArray", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaBindTextureToMipmappedArray"] = {"hipBindTextureToMipmappedArray", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaGetTextureAlignmentOffset"] = {"hipGetTextureAlignmentOffset", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaGetTextureReference"] = {"hipGetTextureReference", CONV_TEX, API_RUNTIME}; // Channel cuda2hipRename["cudaChannelFormatKind"] = {"hipChannelFormatKind", CONV_TEX, API_RUNTIME}; @@ -1841,77 +1835,80 @@ struct cuda2hipMap { cuda2hipRename["cudaChannelFormatKindFloat"] = {"hipChannelFormatKindFloat", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaChannelFormatKindNone"] = {"hipChannelFormatKindNone", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaChannelFormatDesc"] = {"hipChannelFormatDesc", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaCreateChannelDesc"] = {"hipCreateChannelDesc", CONV_TEX, API_RUNTIME}; - // unsupported yet by HIP - cuda2hipRename["cudaGetChannelDesc"] = {"hipGetChannelDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaGetChannelDesc"] = {"hipGetChannelDesc", CONV_TEX, API_RUNTIME}; // Texture Object Management // structs - // unsupported yet by HIP - cuda2hipRename["cudaResourceDesc"] = {"hipResourceDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaResourceViewDesc"] = {"hipResourceViewDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaTextureDesc"] = {"hipTextureDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaResourceDesc"] = {"hipResourceDesc", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaResourceViewDesc"] = {"hipResourceViewDesc", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaTextureDesc"] = {"hipTextureDesc", CONV_TEX, API_RUNTIME}; cuda2hipRename["surfaceReference"] = {"hipSurfaceReference", CONV_SURFACE, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["textureReference"] = {"hipTextureReference", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + // Leaved unchanged + // cuda2hipRename["textureReference"] = {"textureReference", CONV_TEX, API_RUNTIME}; + + // typedefs + cuda2hipRename["cudaTextureObject_t"] = {"hipTextureObject_t", CONV_TEX, API_RUNTIME}; // enums // enum cudaResourceType - cuda2hipRename["cudaResourceType"] = {"hipResourceType", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUresourcetype) - cuda2hipRename["cudaResourceTypeArray"] = {"hipResourceTypeArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_ARRAY = 0x00) - cuda2hipRename["cudaResourceTypeMipmappedArray"] = {"hipResourceTypeMipmappedArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_MIPMAPPED_ARRAY = 0x01) - cuda2hipRename["cudaResourceTypeLinear"] = {"hipResourceTypeLinear", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_LINEAR = 0x02) - cuda2hipRename["cudaResourceTypePitch2D"] = {"hipResourceTypePitch2D", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_PITCH2D = 0x03) + cuda2hipRename["cudaResourceType"] = {"hipResourceType", CONV_TEX, API_RUNTIME}; // API_Driver ANALOGUE (CUresourcetype) + cuda2hipRename["cudaResourceTypeArray"] = {"hipResourceTypeArray", CONV_TEX, API_RUNTIME}; // 0x00 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_ARRAY = 0x00) + cuda2hipRename["cudaResourceTypeMipmappedArray"] = {"hipResourceTypeMipmappedArray", CONV_TEX, API_RUNTIME}; // 0x01 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_MIPMAPPED_ARRAY = 0x01) + cuda2hipRename["cudaResourceTypeLinear"] = {"hipResourceTypeLinear", CONV_TEX, API_RUNTIME}; // 0x02 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_LINEAR = 0x02) + cuda2hipRename["cudaResourceTypePitch2D"] = {"hipResourceTypePitch2D", CONV_TEX, API_RUNTIME}; // 0x03 // API_Driver ANALOGUE (CU_RESOURCE_TYPE_PITCH2D = 0x03) // enum cudaResourceViewFormat - cuda2hipRename["cudaResourceViewFormat"] = {"hipResourceViewFormat", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // API_Driver ANALOGUE (CUresourceViewFormat) - cuda2hipRename["cudaResViewFormatNone"] = {"hipResViewFormatNone", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x00 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_NONE = 0x00) - cuda2hipRename["cudaResViewFormatUnsignedChar1"] = {"hipResViewFormatUnsignedChar1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x01 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X8 = 0x01) - cuda2hipRename["cudaResViewFormatUnsignedChar2"] = {"hipResViewFormatUnsignedChar2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x02 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X8 = 0x02) - cuda2hipRename["cudaResViewFormatUnsignedChar4"] = {"hipResViewFormatUnsignedChar4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x03 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X8 = 0x03) - cuda2hipRename["cudaResViewFormatSignedChar1"] = {"hipResViewFormatSignedChar1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x04 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X8 = 0x04) - cuda2hipRename["cudaResViewFormatSignedChar2"] = {"hipResViewFormatSignedChar2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x05 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X8 = 0x05) - cuda2hipRename["cudaResViewFormatSignedChar4"] = {"hipResViewFormatSignedChar4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x06 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X8 = 0x06) - cuda2hipRename["cudaResViewFormatUnsignedShort1"] = {"hipResViewFormatUnsignedShort1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x07 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X16 = 0x07) - cuda2hipRename["cudaResViewFormatUnsignedShort2"] = {"hipResViewFormatUnsignedShort2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x08 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X16 = 0x08) - cuda2hipRename["cudaResViewFormatUnsignedShort4"] = {"hipResViewFormatUnsignedShort4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x09 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X16 = 0x09) - cuda2hipRename["cudaResViewFormatSignedShort1"] = {"hipResViewFormatSignedShort1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0a // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X16 = 0x0a) - cuda2hipRename["cudaResViewFormatSignedShort2"] = {"hipResViewFormatSignedShort2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0b // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X16 = 0x0b) - cuda2hipRename["cudaResViewFormatSignedShort4"] = {"hipResViewFormatSignedShort4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0c // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X16 = 0x0c) - cuda2hipRename["cudaResViewFormatUnsignedInt1"] = {"hipResViewFormatUnsignedInt1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X32 = 0x0d) - cuda2hipRename["cudaResViewFormatUnsignedInt2"] = {"hipResViewFormatUnsignedInt2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X32 = 0x0e) - cuda2hipRename["cudaResViewFormatUnsignedInt4"] = {"hipResViewFormatUnsignedInt4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x0f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X32 = 0x0f) - cuda2hipRename["cudaResViewFormatSignedInt1"] = {"hipResViewFormatSignedInt1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x10 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X32 = 0x10) - cuda2hipRename["cudaResViewFormatSignedInt2"] = {"hipResViewFormatSignedInt2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x11 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X32 = 0x11) - cuda2hipRename["cudaResViewFormatSignedInt4"] = {"hipResViewFormatSignedInt4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x12 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X32 = 0x12) - cuda2hipRename["cudaResViewFormatHalf1"] = {"hipResViewFormatHalf1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x13 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_1X16 = 0x13) - cuda2hipRename["cudaResViewFormatHalf2"] = {"hipResViewFormatHalf2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x14 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_2X16 = 0x14) - cuda2hipRename["cudaResViewFormatHalf4"] = {"hipResViewFormatHalf4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x15 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_4X16 = 0x15) - cuda2hipRename["cudaResViewFormatFloat1"] = {"hipResViewFormatFloat1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x16 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_1X32 = 0x16) - cuda2hipRename["cudaResViewFormatFloat2"] = {"hipResViewFormatFloat2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x17 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_2X32 = 0x17) - cuda2hipRename["cudaResViewFormatFloat4"] = {"hipResViewFormatFloat4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x18 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_4X32 = 0x18) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed1"] = {"hipResViewFormatUnsignedBlockCompressed1", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x19 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC1 = 0x19) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed2"] = {"hipResViewFormatUnsignedBlockCompressed2", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1a // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC2 = 0x1a) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed3"] = {"hipResViewFormatUnsignedBlockCompressed3", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1b // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC3 = 0x1b) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed4"] = {"hipResViewFormatUnsignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1c // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC4 = 0x1c) - cuda2hipRename["cudaResViewFormatSignedBlockCompressed4"] = {"hipResViewFormatSignedBlockCompressed4", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC4 = 0x1d) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed5"] = {"hipResViewFormatUnsignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC5 = 0x1e) - cuda2hipRename["cudaResViewFormatSignedBlockCompressed5"] = {"hipResViewFormatSignedBlockCompressed5", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x1f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC5 = 0x1f) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed6H"] = {"hipResViewFormatUnsignedBlockCompressed6H", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x20 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC6H = 0x20) - cuda2hipRename["cudaResViewFormatSignedBlockCompressed6H"] = {"hipResViewFormatSignedBlockCompressed6H", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x21 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC6H = 0x21) - cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed7"] = {"hipResViewFormatUnsignedBlockCompressed7", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // 0x22 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC7 = 0x22) + cuda2hipRename["cudaResourceViewFormat"] = {"hipResourceViewFormat", CONV_TEX, API_RUNTIME}; // API_Driver ANALOGUE (CUresourceViewFormat) + cuda2hipRename["cudaResViewFormatNone"] = {"hipResViewFormatNone", CONV_TEX, API_RUNTIME}; // 0x00 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_NONE = 0x00) + cuda2hipRename["cudaResViewFormatUnsignedChar1"] = {"hipResViewFormatUnsignedChar1", CONV_TEX, API_RUNTIME}; // 0x01 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X8 = 0x01) + cuda2hipRename["cudaResViewFormatUnsignedChar2"] = {"hipResViewFormatUnsignedChar2", CONV_TEX, API_RUNTIME}; // 0x02 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X8 = 0x02) + cuda2hipRename["cudaResViewFormatUnsignedChar4"] = {"hipResViewFormatUnsignedChar4", CONV_TEX, API_RUNTIME}; // 0x03 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X8 = 0x03) + cuda2hipRename["cudaResViewFormatSignedChar1"] = {"hipResViewFormatSignedChar1", CONV_TEX, API_RUNTIME}; // 0x04 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X8 = 0x04) + cuda2hipRename["cudaResViewFormatSignedChar2"] = {"hipResViewFormatSignedChar2", CONV_TEX, API_RUNTIME}; // 0x05 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X8 = 0x05) + cuda2hipRename["cudaResViewFormatSignedChar4"] = {"hipResViewFormatSignedChar4", CONV_TEX, API_RUNTIME}; // 0x06 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X8 = 0x06) + cuda2hipRename["cudaResViewFormatUnsignedShort1"] = {"hipResViewFormatUnsignedShort1", CONV_TEX, API_RUNTIME}; // 0x07 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X16 = 0x07) + cuda2hipRename["cudaResViewFormatUnsignedShort2"] = {"hipResViewFormatUnsignedShort2", CONV_TEX, API_RUNTIME}; // 0x08 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X16 = 0x08) + cuda2hipRename["cudaResViewFormatUnsignedShort4"] = {"hipResViewFormatUnsignedShort4", CONV_TEX, API_RUNTIME}; // 0x09 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X16 = 0x09) + cuda2hipRename["cudaResViewFormatSignedShort1"] = {"hipResViewFormatSignedShort1", CONV_TEX, API_RUNTIME}; // 0x0a // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X16 = 0x0a) + cuda2hipRename["cudaResViewFormatSignedShort2"] = {"hipResViewFormatSignedShort2", CONV_TEX, API_RUNTIME}; // 0x0b // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X16 = 0x0b) + cuda2hipRename["cudaResViewFormatSignedShort4"] = {"hipResViewFormatSignedShort4", CONV_TEX, API_RUNTIME}; // 0x0c // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X16 = 0x0c) + cuda2hipRename["cudaResViewFormatUnsignedInt1"] = {"hipResViewFormatUnsignedInt1", CONV_TEX, API_RUNTIME}; // 0x0d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_1X32 = 0x0d) + cuda2hipRename["cudaResViewFormatUnsignedInt2"] = {"hipResViewFormatUnsignedInt2", CONV_TEX, API_RUNTIME}; // 0x0e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_2X32 = 0x0e) + cuda2hipRename["cudaResViewFormatUnsignedInt4"] = {"hipResViewFormatUnsignedInt4", CONV_TEX, API_RUNTIME}; // 0x0f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UINT_4X32 = 0x0f) + cuda2hipRename["cudaResViewFormatSignedInt1"] = {"hipResViewFormatSignedInt1", CONV_TEX, API_RUNTIME}; // 0x10 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_1X32 = 0x10) + cuda2hipRename["cudaResViewFormatSignedInt2"] = {"hipResViewFormatSignedInt2", CONV_TEX, API_RUNTIME}; // 0x11 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_2X32 = 0x11) + cuda2hipRename["cudaResViewFormatSignedInt4"] = {"hipResViewFormatSignedInt4", CONV_TEX, API_RUNTIME}; // 0x12 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SINT_4X32 = 0x12) + cuda2hipRename["cudaResViewFormatHalf1"] = {"hipResViewFormatHalf1", CONV_TEX, API_RUNTIME}; // 0x13 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_1X16 = 0x13) + cuda2hipRename["cudaResViewFormatHalf2"] = {"hipResViewFormatHalf2", CONV_TEX, API_RUNTIME}; // 0x14 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_2X16 = 0x14) + cuda2hipRename["cudaResViewFormatHalf4"] = {"hipResViewFormatHalf4", CONV_TEX, API_RUNTIME}; // 0x15 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_4X16 = 0x15) + cuda2hipRename["cudaResViewFormatFloat1"] = {"hipResViewFormatFloat1", CONV_TEX, API_RUNTIME}; // 0x16 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_1X32 = 0x16) + cuda2hipRename["cudaResViewFormatFloat2"] = {"hipResViewFormatFloat2", CONV_TEX, API_RUNTIME}; // 0x17 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_2X32 = 0x17) + cuda2hipRename["cudaResViewFormatFloat4"] = {"hipResViewFormatFloat4", CONV_TEX, API_RUNTIME}; // 0x18 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_FLOAT_4X32 = 0x18) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed1"] = {"hipResViewFormatUnsignedBlockCompressed1", CONV_TEX, API_RUNTIME}; // 0x19 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC1 = 0x19) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed2"] = {"hipResViewFormatUnsignedBlockCompressed2", CONV_TEX, API_RUNTIME}; // 0x1a // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC2 = 0x1a) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed3"] = {"hipResViewFormatUnsignedBlockCompressed3", CONV_TEX, API_RUNTIME}; // 0x1b // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC3 = 0x1b) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed4"] = {"hipResViewFormatUnsignedBlockCompressed4", CONV_TEX, API_RUNTIME}; // 0x1c // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC4 = 0x1c) + cuda2hipRename["cudaResViewFormatSignedBlockCompressed4"] = {"hipResViewFormatSignedBlockCompressed4", CONV_TEX, API_RUNTIME}; // 0x1d // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC4 = 0x1d) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed5"] = {"hipResViewFormatUnsignedBlockCompressed5", CONV_TEX, API_RUNTIME}; // 0x1e // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC5 = 0x1e) + cuda2hipRename["cudaResViewFormatSignedBlockCompressed5"] = {"hipResViewFormatSignedBlockCompressed5", CONV_TEX, API_RUNTIME}; // 0x1f // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC5 = 0x1f) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed6H"] = {"hipResViewFormatUnsignedBlockCompressed6H", CONV_TEX, API_RUNTIME}; // 0x20 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC6H = 0x20) + cuda2hipRename["cudaResViewFormatSignedBlockCompressed6H"] = {"hipResViewFormatSignedBlockCompressed6H", CONV_TEX, API_RUNTIME}; // 0x21 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_SIGNED_BC6H = 0x21) + cuda2hipRename["cudaResViewFormatUnsignedBlockCompressed7"] = {"hipResViewFormatUnsignedBlockCompressed7", CONV_TEX, API_RUNTIME}; // 0x22 // API_Driver ANALOGUE (CU_RES_VIEW_FORMAT_UNSIGNED_BC7 = 0x22) - cuda2hipRename["cudaTextureAddressMode"] = {"hipTextureAddressMode", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaAddressModeWrap"] = {"hipAddressModeWrap", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaAddressModeClamp"] = {"hipAddressModeClamp", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaAddressModeMirror"] = {"hipAddressModeMirror", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaAddressModeBorder"] = {"hipAddressModeBorder", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaTextureAddressMode"] = {"hipTextureAddressMode", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaAddressModeWrap"] = {"hipAddressModeWrap", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaAddressModeClamp"] = {"hipAddressModeClamp", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaAddressModeMirror"] = {"hipAddressModeMirror", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaAddressModeBorder"] = {"hipAddressModeBorder", CONV_TEX, API_RUNTIME}; // functions - cuda2hipRename["cudaCreateTextureObject"] = {"hipCreateTextureObject", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaDestroyTextureObject"] = {"hipDestroyTextureObject", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureObjectResourceDesc"] = {"hipGetTextureObjectResourceDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureObjectResourceViewDesc"] = {"hipGetTextureObjectResourceViewDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureObjectTextureDesc"] = {"hipGetTextureObjectTextureDesc", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaCreateTextureObject"] = {"hipCreateTextureObject", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaDestroyTextureObject"] = {"hipDestroyTextureObject", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaGetTextureObjectResourceDesc"] = {"hipGetTextureObjectResourceDesc", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaGetTextureObjectResourceViewDesc"] = {"hipGetTextureObjectResourceViewDesc", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaGetTextureObjectTextureDesc"] = {"hipGetTextureObjectTextureDesc", CONV_TEX, API_RUNTIME}; // Surface Reference Management // unsupported yet by HIP From 0aaa800109265757752ec75587ccaaf2f0b9e8c6 Mon Sep 17 00:00:00 2001 From: Yifan Sun Date: Thu, 27 Jul 2017 11:46:18 -0400 Subject: [PATCH 047/108] Allow hip cmake to link libraries --- cmake/FindHIP.cmake | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cmake/FindHIP.cmake b/cmake/FindHIP.cmake index 5a5813ba0d..6668e60332 100644 --- a/cmake/FindHIP.cmake +++ b/cmake/FindHIP.cmake @@ -520,7 +520,7 @@ macro(HIP_ADD_EXECUTABLE hip_target) if("x${HCC_HOME}" STREQUAL "x") set(HCC_HOME "/opt/rocm/hcc") endif() - set(CMAKE_HIP_LINK_EXECUTABLE "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HCC_HOME} -o ") + set(CMAKE_HIP_LINK_EXECUTABLE "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HCC_HOME} -o ") add_executable(${hip_target} ${_cmake_options} ${_generated_files} ${_sources}) set_target_properties(${hip_target} PROPERTIES LINKER_LANGUAGE HIP) endmacro() From 77fb9893b4fb1108fdd71281a5090e25cb51630a Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Thu, 27 Jul 2017 11:11:54 -0500 Subject: [PATCH 048/108] Set HIP_SYNC_NULL_STREAM=0. Optimizes null stream synchronization so it uses GPU-side dependency resolution. Requires HCC __hcc_workweek__ > 17300. --- src/hip_hcc.cpp | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 9d431df2dd..26799a9b11 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -86,13 +86,18 @@ int HIP_FORCE_SYNC_COPY = 0; int HIP_EVENT_SYS_RELEASE=1; int HIP_COHERENT_HOST_ALLOC = 0; -// TODO - set to 0 once we resolve stability. -// USE_ HIP_SYNC_HOST_ALLOC int HIP_SYNC_HOST_ALLOC = 1; + +#if (__hcc_workweek__ >= 17300) +// Make sure we have required bug fix in HCC +// Perform resolution on the GPU: // Chicken bit to sync on host to implement null stream. // If 0, null stream synchronization is performed on the GPU +int HIP_SYNC_NULL_STREAM = 0; +#else int HIP_SYNC_NULL_STREAM = 1; +#endif // HIP needs to change some behavior based on HCC_OPT_FLUSH : #if (__hcc_workweek__ >= 17296) From ed0d6cf7454d792d3d2edd7ac72dcee07b5dc974 Mon Sep 17 00:00:00 2001 From: sunway513 Date: Thu, 27 Jul 2017 17:41:49 +0000 Subject: [PATCH 049/108] Change to use hipDeviceGetPCIBusId api for hipEnvVarDriver test --- tests/src/hipEnvVar.cpp | 15 ++++++------- tests/src/hipEnvVarDriver.cpp | 42 +++++++++++++++++------------------ 2 files changed, 27 insertions(+), 30 deletions(-) diff --git a/tests/src/hipEnvVar.cpp b/tests/src/hipEnvVar.cpp index d5942518e0..ff68103816 100644 --- a/tests/src/hipEnvVar.cpp +++ b/tests/src/hipEnvVar.cpp @@ -109,15 +109,14 @@ int main(int argc, char **argv) std::cout << devCount << std::endl; } if (retDevInfo) { - hipSetDevice(device); - hipDeviceProp_t devProp; + hipDevice_t deviceT; + hipDeviceGet(&deviceT, device); - hipGetDeviceProperties(&devProp, device); - if (devProp.major < 1) { - printf("%d does not support HIP\n", device); - return -1; - } - std::cout << devProp.pciBusID << std::endl; + char pciBusId[100]; + memset(pciBusId,0,100); + hipDeviceGetPCIBusId(pciBusId,100,deviceT); + + cout< devPCINum; + char pciBusID[100]; //collect the device pci bus ID for all devices int totalDeviceNum = getDeviceNumber(); std::cout << "The total number of available devices is " << totalDeviceNum<< std::endl <<"Valid index range is 0 - "< devPCINum; for (int i = 0; i < totalDeviceNum ; i++) { - devPCINum.push_back(getDevicePCIBusNum(i)); + getDevicePCIBusNum(i, pciBusID); + devPCINum.push_back(pciBusID); std::cout <<"The collected device PCI Bus ID of Device "< Date: Fri, 28 Jul 2017 01:38:50 +0300 Subject: [PATCH 050/108] Update CUDA_Runtime_API_functions_supported_by_HIP.md --- docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index dde2a04b22..b5db3fc705 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -310,7 +310,7 @@ | `cudaBindTexture` | | | `cudaBindTexture2D` | `hipBindTexture2D` | | `cudaBindTextureToArray` | `hipBindTextureToArray` | -| `cudaBindTextureToMipmappedArray` | `hipBindTextureToMipmappedArray` | +| `cudaBindTextureToMipmappedArray` | | | `cudaCreateChannelDesc` | | | `cudaGetChannelDesc` | | | `cudaGetTextureAlignmentOffset` | `hipGetTextureAlignmentOffset`| From bcca169631646b39857789db2a2a2f3622bf4e0f Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Fri, 28 Jul 2017 01:41:14 +0300 Subject: [PATCH 051/108] Update CUDA_Runtime_API_functions_supported_by_HIP.md --- docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index b5db3fc705..cb8d359853 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -313,8 +313,8 @@ | `cudaBindTextureToMipmappedArray` | | | `cudaCreateChannelDesc` | | | `cudaGetChannelDesc` | | -| `cudaGetTextureAlignmentOffset` | `hipGetTextureAlignmentOffset`| -| `cudaGetTextureReference` | `hipGetTextureReference` | +| `cudaGetTextureAlignmentOffset` | | +| `cudaGetTextureReference` | | | `cudaUnbindTexture` | | ## **24. Surface Reference Management** From d614f53ad71740efa93781956963c87df6136390 Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Fri, 28 Jul 2017 01:45:24 +0300 Subject: [PATCH 052/108] Update Cuda2Hip.cpp --- hipify-clang/src/Cuda2Hip.cpp | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index c36ca5369a..79be71967c 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -1824,9 +1824,9 @@ struct cuda2hipMap { cuda2hipRename["cudaUnbindTexture"] = {"hipUnbindTexture", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaBindTexture2D"] = {"hipBindTexture2D", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaBindTextureToArray"] = {"hipBindTextureToArray", CONV_TEX, API_RUNTIME}; - cuda2hipRename["cudaBindTextureToMipmappedArray"] = {"hipBindTextureToMipmappedArray", CONV_TEX, API_RUNTIME}; - cuda2hipRename["cudaGetTextureAlignmentOffset"] = {"hipGetTextureAlignmentOffset", CONV_TEX, API_RUNTIME}; - cuda2hipRename["cudaGetTextureReference"] = {"hipGetTextureReference", CONV_TEX, API_RUNTIME}; + cuda2hipRename["cudaBindTextureToMipmappedArray"] = {"hipBindTextureToMipmappedArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaGetTextureAlignmentOffset"] = {"hipGetTextureAlignmentOffset", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaGetTextureReference"] = {"hipGetTextureReference", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; // Channel cuda2hipRename["cudaChannelFormatKind"] = {"hipChannelFormatKind", CONV_TEX, API_RUNTIME}; From 3a4dfc0f852f1f84a8b8fed093229fc7a2dd82aa Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Thu, 27 Jul 2017 22:00:15 -0500 Subject: [PATCH 053/108] Pass kernel name to HCC dispatch_hsa_kernel, for debug/profile --- src/hip_module.cpp | 3 --- 1 file changed, 3 deletions(-) diff --git a/src/hip_module.cpp b/src/hip_module.cpp index b8c032da27..5dad689c69 100644 --- a/src/hip_module.cpp +++ b/src/hip_module.cpp @@ -452,10 +452,7 @@ hipError_t ihipModuleLaunchKernel(hipFunction_t f, lp.av->dispatch_hsa_kernel(&aql, config[1] /* kernarg*/, kernArgSize, (startEvent || stopEvent) ? &cf : nullptr -#define USE_NAMED_KERNEL 0 -#if USE_NAMED_KERNEL , f->_name.c_str() -#endif ); From 4980a6d3ab39f2449d69974438efd15d01af09f5 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Thu, 27 Jul 2017 23:00:58 -0500 Subject: [PATCH 054/108] Add workweek check to make sure we have a new enough compiler --- src/hip_module.cpp | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/hip_module.cpp b/src/hip_module.cpp index 5dad689c69..10fcd15fdc 100644 --- a/src/hip_module.cpp +++ b/src/hip_module.cpp @@ -452,7 +452,9 @@ hipError_t ihipModuleLaunchKernel(hipFunction_t f, lp.av->dispatch_hsa_kernel(&aql, config[1] /* kernarg*/, kernArgSize, (startEvent || stopEvent) ? &cf : nullptr +#if (__hcc_workweek__ >= 17300) , f->_name.c_str() +#endif ); From 830e85060b451f343519c6d245e49d1c060076aa Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Fri, 28 Jul 2017 16:18:15 +0530 Subject: [PATCH 055/108] [hipcc] Cleanup amdgpu target logic Existing logic has a bug. If user specifies targetA via commandline options, while enumerator returns targetB, hipcc will create a fatbin containing targets targetA and targetB. enumerator should only be used when no target is specified by user (commandline or env var). Change-Id: I6da857f86860c0e671b5988cd858644a08f723b9 --- bin/hipcc | 133 ++++++++++++++++++++++++++++++++---------------------- 1 file changed, 80 insertions(+), 53 deletions(-) diff --git a/bin/hipcc b/bin/hipcc index 8a1874a36a..b40b4ee855 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -75,6 +75,7 @@ $target_gfx801 = 0; $target_gfx802 = 0; $target_gfx803 = 0; $target_gfx900 = 0; +$default_amdgpu_target = 1; if ($HIP_PLATFORM eq "hcc") { $HSA_PATH=$ENV{'HSA_PATH'} // "/opt/rocm/hsa"; @@ -103,29 +104,6 @@ if ($HIP_PLATFORM eq "hcc") { $HIPLDFLAGS = `${HCC_HOME}/bin/hcc-config --ldflags`; - $ROCM_AGENT_ENUM = "${ROCM_PATH}/bin/rocm_agent_enumerator"; - - my $myAgents = `${ROCM_AGENT_ENUM} -t GPU`; - my @agentsLine = split('\n', $myAgents); - - foreach my $val (@agentsLine) { - if($val eq "gfx701") { - $target_gfx701 = 1; - } - if($val eq "gfx801") { - $target_gfx801 = 1; - } - if($val eq "gfx802") { - $target_gfx802 = 1; - } - if($val eq "gfx803") { - $target_gfx803 = 1; - } - if($val eq "gfx900") { - $target_gfx900 = 1; - } - } - #### GCC system includes workaround #### $HCC_WA_FLAGS = " "; if ($HCC_VERSION_MAJOR eq 1) { @@ -278,25 +256,32 @@ foreach $arg (@ARGV) $HIPCXXFLAGS .= " -stdlib=libc++"; $setStdLib = 1; } + + # TODO: Add support for comma separated list like HCC_AMDGPU_TARGET if($arg eq '--amdgpu-target=gfx701') { $target_gfx701 = 1; + $default_amdgpu_target = 0; } if($arg eq '--amdgpu-target=gfx801') { $target_gfx801 = 1; + $default_amdgpu_target = 0; } if($arg eq '--amdgpu-target=gfx802') { $target_gfx802 = 1; + $default_amdgpu_target = 0; } if($arg eq '--amdgpu-target=gfx803') { $target_gfx803 = 1; + $default_amdgpu_target = 0; } if($arg eq '--amdgpu-target=gfx900') { $target_gfx900 = 1; + $default_amdgpu_target = 0; } if(($trimarg eq '-stdlib=libstdc++') and ($setStdLib eq 0)) @@ -358,38 +343,80 @@ foreach $arg (@ARGV) } $toolArgs .= " $arg" unless $swallowArg; } -if(defined $ENV{HCC_AMDGPU_TARGET}) -{ - foreach my $target (split(/,/, $ENV{HCC_AMDGPU_TARGET})) - { - if($target eq 'gfx701') - { - $target_gfx701 = 1; - } - if($target eq 'gfx801') - { - $target_gfx801 = 1; - } - if($target eq 'gfx802') - { - $target_gfx802 = 1; - } - if($target eq 'gfx803') - { - $target_gfx803 = 1; - } - if($target eq 'gfx900') - { - $target_gfx900 = 1; - } - } -} -if ($target_gfx701 eq 0 and $target_gfx801 eq 0 and $target_gfx802 eq 0 and $target_gfx803 eq 0 and $target_gfx900 eq 0) -{ - $target_gfx803 = 1; -} if($HIP_PLATFORM eq "hcc"){ + # No AMDGPU target specified at commandline + if($default_amdgpu_target eq 1) + { + # Look for HCC_AMDGPU_TARGET + if(defined $ENV{HCC_AMDGPU_TARGET}) + { + foreach my $target (split(/,/, $ENV{HCC_AMDGPU_TARGET})) + { + if($target eq 'gfx701') + { + $target_gfx701 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx801') + { + $target_gfx801 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx802') + { + $target_gfx802 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx803') + { + $target_gfx803 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx900') + { + $target_gfx900 = 1; + $default_amdgpu_target = 0; + } + } + } + # Else try using rocm_agent_enumerator + else + { + $ROCM_AGENT_ENUM = "${ROCM_PATH}/bin/rocm_agent_enumerator"; + + my $myAgents = `${ROCM_AGENT_ENUM} -t GPU`; + my @agentsLine = split('\n', $myAgents); + + foreach my $val (@agentsLine) { + if($val eq "gfx701") { + $target_gfx701 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx801") { + $target_gfx801 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx802") { + $target_gfx802 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx803") { + $target_gfx803 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx900") { + $target_gfx900 = 1; + $default_amdgpu_target = 0; + } + } + } + } + # All options including enumerator failed! Fallback to gfx803 for now + if ($default_amdgpu_target eq 1) + { + $target_gfx803 = 1; + } $ENV{HCC_EXTRA_LIBRARIES}="$HIP_PATH/lib/hip_hc.ll\n"; From 0053e4db0341d970674f559d5a1467a4baa1ae11 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Sat, 29 Jul 2017 06:50:56 +0530 Subject: [PATCH 056/108] fix hipMemcpy2DAsync --- src/hip_memory.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 5f20fcb24b..9bb31d2d78 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -879,7 +879,7 @@ hipError_t hipMemcpy2D(void* dst, size_t dpitch, const void* src, size_t spitch, return ihipLogStatus(e); } -hipError_t hipMemcpy2D(void* dst, size_t dpitch, const void* src, size_t spitch, +hipError_t hipMemcpy2DAsync(void* dst, size_t dpitch, const void* src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind, hipStream_t stream) { HIP_INIT_SPECIAL_API((TRACE_MCMD), dst, dpitch, src, spitch, width, height, kind, stream); if(width > dpitch || width > spitch) From ecbb494a5861debf0bb1cc9ad8cd6197bfd81fef Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Fri, 28 Jul 2017 22:13:43 -0500 Subject: [PATCH 057/108] Some fixes to tracing. --- src/grid_launch.cpp | 4 ++-- src/hip_device.cpp | 6 ++++-- src/hip_hcc.cpp | 11 +++++------ 3 files changed, 11 insertions(+), 10 deletions(-) diff --git a/src/grid_launch.cpp b/src/grid_launch.cpp index f3b28c5f60..fd5c2a1573 100644 --- a/src/grid_launch.cpp +++ b/src/grid_launch.cpp @@ -72,8 +72,8 @@ namespace hip_impl } if (COMPILE_HIP_DB && HIP_TRACE_API) { - std::cerr << API_COLOR << os.str() << API_COLOR_END - << std::endl; + std::string fullStr; + recordApiTrace(&fullStr, os.str()); } } } diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 7e59e37cad..5ff6dbf04d 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -358,7 +358,8 @@ hipError_t hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device hipError_t hipDeviceGetName(char *name,int len,hipDevice_t device) { - HIP_INIT_API(name,len, device); + // Cast to void* here to avoid printing garbage in debug modes. + HIP_INIT_API((void*)name,len, device); hipError_t e = hipSuccess; auto deviceHandle = ihipGetDevice(device); int nameLen = strlen(deviceHandle->_props.name); @@ -369,7 +370,8 @@ hipError_t hipDeviceGetName(char *name,int len,hipDevice_t device) hipError_t hipDeviceGetPCIBusId (char *pciBusId,int len, int device) { - HIP_INIT_API(pciBusId, len, device); + // Cast to void* here to avoid printing garbage in debug modes. + HIP_INIT_API((void*)pciBusId, len, device); hipError_t e = hipErrorInvalidValue; int deviceCount = 0; ihipGetDeviceCount( &deviceCount ); diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index d4f7723dbc..d080e34993 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -1424,9 +1424,7 @@ void ihipPrintKernelLaunch(const char *kernelName, const grid_launch_parm *lp, c { if ((HIP_TRACE_API & (1<grid_dim @@ -1434,6 +1432,11 @@ void ihipPrintKernelLaunch(const char *kernelName, const grid_launch_parm *lp, c << " sharedMem:+" << lp->dynamic_group_mem_bytes << " " << *stream; + if (COMPILE_HIP_DB && HIP_TRACE_API) { + std::string fullStr; + recordApiTrace(&fullStr, os.str()); + } + if (HIP_PROFILE_API == 0x1) { std::string shortAtpString("hipLaunchKernel:"); shortAtpString += kernelName; @@ -1441,10 +1444,6 @@ void ihipPrintKernelLaunch(const char *kernelName, const grid_launch_parm *lp, c } else if (HIP_PROFILE_API == 0x2) { MARKER_BEGIN(os.str().c_str(), "HIP"); } - - if (COMPILE_HIP_DB && HIP_TRACE_API) { - std::cerr << API_COLOR << os.str() << API_COLOR_END << std::endl; - } } } From 85ff6e3ef4525497c5203e58b61c2d8be6978694 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Mon, 31 Jul 2017 10:11:19 +0530 Subject: [PATCH 058/108] [hipcc] Fix amdgpu target selection logic - Refactoring introduced a bug when user does not specify any target via --amdgpu-target, but has an invalid target specified in HCC_AMDGPU_TARGET. In this case the selection logic was defaulting to gf803. - Removed defaulting to any specific target if rocm_agent_enumerator fails. hipcc will report this and die if linking was required. Change-Id: I76131867049fef92331807dd19a926406dcc1d02 --- bin/hipcc | 122 ++++++++++++++++++++++++++---------------------------- 1 file changed, 59 insertions(+), 63 deletions(-) diff --git a/bin/hipcc b/bin/hipcc index b40b4ee855..624777ab25 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -345,77 +345,73 @@ foreach $arg (@ARGV) } if($HIP_PLATFORM eq "hcc"){ - # No AMDGPU target specified at commandline - if($default_amdgpu_target eq 1) + # No AMDGPU target specified at commandline. So look for HCC_AMDGPU_TARGET + if($default_amdgpu_target eq 1 and defined $ENV{HCC_AMDGPU_TARGET}) { - # Look for HCC_AMDGPU_TARGET - if(defined $ENV{HCC_AMDGPU_TARGET}) + foreach my $target (split(/,/, $ENV{HCC_AMDGPU_TARGET})) { - foreach my $target (split(/,/, $ENV{HCC_AMDGPU_TARGET})) + if($target eq 'gfx701') { - if($target eq 'gfx701') - { - $target_gfx701 = 1; - $default_amdgpu_target = 0; - } - if($target eq 'gfx801') - { - $target_gfx801 = 1; - $default_amdgpu_target = 0; - } - if($target eq 'gfx802') - { - $target_gfx802 = 1; - $default_amdgpu_target = 0; - } - if($target eq 'gfx803') - { - $target_gfx803 = 1; - $default_amdgpu_target = 0; - } - if($target eq 'gfx900') - { - $target_gfx900 = 1; - $default_amdgpu_target = 0; - } + $target_gfx701 = 1; + $default_amdgpu_target = 0; } - } - # Else try using rocm_agent_enumerator - else - { - $ROCM_AGENT_ENUM = "${ROCM_PATH}/bin/rocm_agent_enumerator"; - - my $myAgents = `${ROCM_AGENT_ENUM} -t GPU`; - my @agentsLine = split('\n', $myAgents); - - foreach my $val (@agentsLine) { - if($val eq "gfx701") { - $target_gfx701 = 1; - $default_amdgpu_target = 0; - } - if($val eq "gfx801") { - $target_gfx801 = 1; - $default_amdgpu_target = 0; - } - if($val eq "gfx802") { - $target_gfx802 = 1; - $default_amdgpu_target = 0; - } - if($val eq "gfx803") { - $target_gfx803 = 1; - $default_amdgpu_target = 0; - } - if($val eq "gfx900") { - $target_gfx900 = 1; - $default_amdgpu_target = 0; - } + if($target eq 'gfx801') + { + $target_gfx801 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx802') + { + $target_gfx802 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx803') + { + $target_gfx803 = 1; + $default_amdgpu_target = 0; + } + if($target eq 'gfx900') + { + $target_gfx900 = 1; + $default_amdgpu_target = 0; } } } - # All options including enumerator failed! Fallback to gfx803 for now - if ($default_amdgpu_target eq 1) + # Else try using rocm_agent_enumerator + if($default_amdgpu_target eq 1) { - $target_gfx803 = 1; + $ROCM_AGENT_ENUM = "${ROCM_PATH}/bin/rocm_agent_enumerator"; + + my $myAgents = `${ROCM_AGENT_ENUM} -t GPU`; + my @agentsLine = split('\n', $myAgents); + + foreach my $val (@agentsLine) { + if($val eq "gfx701") { + $target_gfx701 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx801") { + $target_gfx801 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx802") { + $target_gfx802 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx803") { + $target_gfx803 = 1; + $default_amdgpu_target = 0; + } + if($val eq "gfx900") { + $target_gfx900 = 1; + $default_amdgpu_target = 0; + } + } + } + # rocm_agent_enumerator failed! Throw an error and die if linking is required + if ($default_amdgpu_target eq 1 and $compileOnly eq 0) + { + print "No valid AMD GPU target was either specified or found. Please specify a valid target using --amdgpu-target=" and die(); } $ENV{HCC_EXTRA_LIBRARIES}="$HIP_PATH/lib/hip_hc.ll\n"; From e9c995959b5be31d6377538ce13a8d04810e70a5 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Mon, 31 Jul 2017 14:57:32 +0530 Subject: [PATCH 059/108] [doc] Add details of HIP build time dependencies to INSTALL.md Change-Id: Iee87d1345aba716d0ce49f7afb45ba073a16be88 --- INSTALL.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/INSTALL.md b/INSTALL.md index dc9ae41b9c..eb219e19f7 100644 --- a/INSTALL.md +++ b/INSTALL.md @@ -59,7 +59,8 @@ HIP source code is available and the project can be built from source on the HCC 1. Follow the above steps to install and validate the binary packages. 2. Download HIP source code (from the [GitHub repot](https://github.com/ROCm-Developer-Tools/HIP).) -3. Build and install HIP (This is the simple version assuming default paths ; see below for additional options.) +3. Install HIP build-time dependencies using ```sudo apt-get install libelf-dev```. +4. Build and install HIP (This is the simple version assuming default paths ; see below for additional options.) ``` cd HIP mkdir build From 96e79f97e2132d15f8ba0de1444ad0f656456294 Mon Sep 17 00:00:00 2001 From: Siu Chi Chan Date: Wed, 2 Aug 2017 08:50:18 +0000 Subject: [PATCH 060/108] implement __threadfence_system --- include/hip/hcc_detail/hip_runtime.h | 2 +- src/device_util.cpp | 2 +- .../src/deviceLib/hip_threadfence_system.cpp | 128 ++++++++++++++++++ 3 files changed, 130 insertions(+), 2 deletions(-) create mode 100644 tests/src/deviceLib/hip_threadfence_system.cpp diff --git a/include/hip/hcc_detail/hip_runtime.h b/include/hip/hcc_detail/hip_runtime.h index fecf79c496..1092e22c47 100644 --- a/include/hip/hcc_detail/hip_runtime.h +++ b/include/hip/hcc_detail/hip_runtime.h @@ -130,7 +130,7 @@ extern int HIP_TRACE_API; #define __HIP_ARCH_HAS_WARP_FUNNEL_SHIFT__ (0) //sync -#define __HIP_ARCH_HAS_THREAD_FENCE_SYSTEM__ (0) +#define __HIP_ARCH_HAS_THREAD_FENCE_SYSTEM__ (1) #define __HIP_ARCH_HAS_SYNC_THREAD_EXT__ (0) // misc diff --git a/src/device_util.cpp b/src/device_util.cpp index 8bba36c9c7..6afc797ec6 100644 --- a/src/device_util.cpp +++ b/src/device_util.cpp @@ -1264,7 +1264,7 @@ __device__ double __hip_fast_dsqrt_rz(double x) { } __device__ void __threadfence_system(void){ - // no-op + std::atomic_thread_fence(std::memory_order_seq_cst); } double __hip_host_j0(double x) diff --git a/tests/src/deviceLib/hip_threadfence_system.cpp b/tests/src/deviceLib/hip_threadfence_system.cpp new file mode 100644 index 0000000000..2ff3505c15 --- /dev/null +++ b/tests/src/deviceLib/hip_threadfence_system.cpp @@ -0,0 +1,128 @@ +/* +Copyright (c) 2015-2016 Advanced Micro Devices, Inc. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN +THE SOFTWARE. +*/ + +/* HIT_START + * BUILD: %t %s ../test_common.cpp + * RUN: %t + * HIT_END + */ + +#include +#include +#include +#include "hip/hip_runtime.h" +#include "hip/device_functions.h" +#include "test_common.h" + + +#define HIP_ASSERT(x) (assert((x)==hipSuccess)) + +__host__ void fence_system() { + std::atomic_thread_fence(std::memory_order_seq_cst); +} + +__device__ void fence_system() { + __threadfence_system(); +} + +__host__ __device__ void round_robin(const int id, const int num_dev, const int num_iter, volatile int* data, volatile int* flag) { + for (int i = 0; i < num_iter; i++) { + while(*flag%num_dev != id) + fence_system(); // invalid the cache for read + + (*data)++; + fence_system(); // make sure the store to data is sequenced before the store to flag + (*flag)++; + fence_system(); // invalid the cache to flush out flag + } +} + +__global__ void gpu_round_robin(const int id, const int num_dev, const int num_iter, volatile int* data, volatile int* flag) { + round_robin(id, num_dev, num_iter, data, flag); +} + +int main() { + + int num_gpus = 0; + HIP_ASSERT(hipGetDeviceCount(&num_gpus)); + if (num_gpus == 0) { + passed(); + return 0; + } + + volatile int* data; + HIP_ASSERT(hipHostMalloc(&data, sizeof(int), hipHostMallocCoherent)); + constexpr int init_data = 1000; + *data = init_data; + + volatile int* flag; + HIP_ASSERT(hipHostMalloc(&flag, sizeof(int), hipHostMallocCoherent)); + *flag = 0; + + // number of rounds per device + constexpr int num_iter = 1000; + + // one CPU thread + 1 kernel/GPU + const int num_dev = num_gpus + 1; + + int next_id = 0; + std::vector threads; + + // create a CPU thread for the round_robin + threads.push_back(std::thread(round_robin, next_id++, num_dev, num_iter, data, flag)); + + // run one thread per GPU + dim3 dim_block(1,1,1); + dim3 dim_grid(1,1,1); + + // launch one kernel per device for the round robin + for (; next_id < num_dev; ++next_id) { + threads.push_back(std::thread([=]() { + HIP_ASSERT(hipSetDevice(next_id-1)); + hipLaunchKernelGGL(gpu_round_robin, dim_grid, dim_block, 0, 0x0 + , next_id, num_dev, num_iter, data, flag); + HIP_ASSERT(hipDeviceSynchronize()); + })); + } + + for (auto& t : threads) { + t.join(); + } + + int expected_data = init_data + num_dev * num_iter; + int expected_flag = num_dev * num_iter; + + bool passed = *data == expected_data + && *flag == expected_flag; + + HIP_ASSERT(hipHostFree((void*)data)); + HIP_ASSERT(hipHostFree((void*)flag)); + + if (passed) { + passed(); + } + else { + failed("Failed Verification!\n"); + } + + return 0; +} From 2f617bac6d6bdd55ae49e3861395b981607d39ac Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Wed, 2 Aug 2017 14:47:13 +0530 Subject: [PATCH 061/108] Disable failing test p2p_copy_coherency --- tests/src/runtimeApi/memory/p2p_copy_coherency.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/src/runtimeApi/memory/p2p_copy_coherency.cpp b/tests/src/runtimeApi/memory/p2p_copy_coherency.cpp index 9fadebea1e..0c2c387c2f 100644 --- a/tests/src/runtimeApi/memory/p2p_copy_coherency.cpp +++ b/tests/src/runtimeApi/memory/p2p_copy_coherency.cpp @@ -24,7 +24,7 @@ THE SOFTWARE. /* HIT_START * BUILD: %t %s ../../test_common.cpp NVCC_OPTIONS --std=c++11 - * RUN: %t + * RUN: %t EXCLUDE_HIP_PLATFORM all * HIT_END */ From 3e84cf4aba89ce89aaf5b6134544c68427c7c527 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Tue, 8 Aug 2017 07:02:22 +0530 Subject: [PATCH 062/108] Updated context management logic: 1) hipSetDevice sets a flag so that next call to hipCtxGetCurrent returns primary context on current device 2) hipCtxGetCurrent returns primary context on current device if TLS context stack is empty 3) hipCtxPopCurrent falls back to primary context on current device as default 4) hipCtxPushCurrent, hipCtxSetCurrent and hipCtxCreate reset the flag set in hipSetDevice --- src/hip_context.cpp | 46 ++++++++++++++++++++---------------------- src/hip_device.cpp | 1 + src/hip_hcc_internal.h | 1 + 3 files changed, 24 insertions(+), 24 deletions(-) diff --git a/src/hip_context.cpp b/src/hip_context.cpp index 11ef6d6da5..69d75e7f31 100644 --- a/src/hip_context.cpp +++ b/src/hip_context.cpp @@ -31,6 +31,7 @@ THE SOFTWARE. // Stack of contexts thread_local std::stack tls_ctxStack; +thread_local bool tls_getPrimaryCtx = true; void ihipCtxStackUpdate() { @@ -65,6 +66,7 @@ hipError_t hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device) *ctx = ictx; ihipSetTlsDefaultCtx(*ctx); tls_ctxStack.push(*ctx); + tls_getPrimaryCtx = false; deviceCrit->addContext(ictx); } @@ -93,8 +95,7 @@ hipError_t hipDriverGetVersion(int *driverVersion) hipError_t e = hipSuccess; if (driverVersion) { *driverVersion = 4; - } - else { + } else { e = hipErrorInvalidValue; } @@ -107,8 +108,7 @@ hipError_t hipRuntimeGetVersion(int *runtimeVersion) hipError_t e = hipSuccess; if (runtimeVersion) { *runtimeVersion = HIP_VERSION_PATCH; - } - else { + } else { e = hipErrorInvalidValue; } @@ -124,9 +124,7 @@ hipError_t hipCtxDestroy(hipCtx_t ctx) if(primaryCtx== ctx) { e = hipErrorInvalidValue; - } - else - { + } else { if(currentCtx == ctx) { //need to destroy the ctx associated with calling thread tls_ctxStack.pop(); @@ -146,19 +144,21 @@ hipError_t hipCtxPopCurrent(hipCtx_t* ctx) { HIP_INIT_API(ctx); hipError_t e = hipSuccess; - ihipCtx_t* tempCtx; - *ctx = ihipGetTlsDefaultCtx(); + ihipCtx_t* currentCtx = ihipGetTlsDefaultCtx(); + auto deviceHandle = currentCtx->getDevice(); + *ctx = currentCtx; + if(!tls_ctxStack.empty()) { tls_ctxStack.pop(); } + if(!tls_ctxStack.empty()) { - tempCtx= tls_ctxStack.top(); - } - else { - tempCtx = nullptr; + currentCtx= tls_ctxStack.top(); + } else { + currentCtx = deviceHandle->_primaryCtx; } - ihipSetTlsDefaultCtx(tempCtx); //TOD0 - Shall check for NULL? + ihipSetTlsDefaultCtx(currentCtx); //TOD0 - Shall check for NULL? return ihipLogStatus(e); } @@ -169,8 +169,8 @@ hipError_t hipCtxPushCurrent(hipCtx_t ctx) if(ctx != NULL) { //TODO- is this check needed? ihipSetTlsDefaultCtx(ctx); tls_ctxStack.push(ctx); - } - else { + tls_getPrimaryCtx = false; + } else { e = hipErrorInvalidContext; } return ihipLogStatus(e); @@ -180,12 +180,11 @@ hipError_t hipCtxGetCurrent(hipCtx_t* ctx) { HIP_INIT_API(ctx); hipError_t e = hipSuccess; - if(!tls_ctxStack.empty()) { + if((tls_getPrimaryCtx) || tls_ctxStack.empty()) { + *ctx = ihipGetTlsDefaultCtx(); + } else { *ctx= tls_ctxStack.top(); } - else { - *ctx = NULL; - } return ihipLogStatus(e); } @@ -195,10 +194,10 @@ hipError_t hipCtxSetCurrent(hipCtx_t ctx) hipError_t e = hipSuccess; if(ctx == NULL) { tls_ctxStack.pop(); - } - else { + } else { ihipSetTlsDefaultCtx(ctx); tls_ctxStack.push(ctx); + tls_getPrimaryCtx = false; } return ihipLogStatus(e); } @@ -213,8 +212,7 @@ hipError_t hipCtxGetDevice(hipDevice_t *device) if(ctx == nullptr) { e = hipErrorInvalidContext; // TODO *device = nullptr; - } - else { + } else { auto deviceHandle = ctx->getDevice(); *device = deviceHandle->_deviceId; } diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 5ff6dbf04d..1800c9369c 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -146,6 +146,7 @@ hipError_t hipSetDevice(int deviceId) return ihipLogStatus(hipErrorInvalidDevice); } else { ihipSetTlsDefaultCtx(ihipGetPrimaryCtx(deviceId)); + tls_getPrimaryCtx = true; return ihipLogStatus(hipSuccess); } } diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index 88c7eedda0..4cb85ffc19 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -114,6 +114,7 @@ private: //Extern tls extern thread_local hipError_t tls_lastHipError; extern thread_local TidInfo tls_tidInfo; +extern thread_local bool tls_getPrimaryCtx; extern std::vector g_dbStartTriggers; extern std::vector g_dbStopTriggers; From eb25bd72dbd60ad94d6d6e4224f6d20890f91ca3 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Tue, 8 Aug 2017 09:13:12 +0530 Subject: [PATCH 063/108] docs update for memcpytosymbol support --- docs/markdown/hip_faq.md | 1 - docs/markdown/hip_porting_guide.md | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/docs/markdown/hip_faq.md b/docs/markdown/hip_faq.md index ddf70f2875..b92ae512c5 100644 --- a/docs/markdown/hip_faq.md +++ b/docs/markdown/hip_faq.md @@ -56,7 +56,6 @@ At a high-level, the following features are not supported: - CUDA Driver API - CUDA IPC Functions (Under Development) - CUDA array, mipmappedArray and pitched memory -- MemcpyToSymbol functions - Queue priority controls See the [API Support Table](CUDA_Runtime_API_functions_supported_by_HIP.md) for more detailed information. diff --git a/docs/markdown/hip_porting_guide.md b/docs/markdown/hip_porting_guide.md index 84887fd512..12ec931f2a 100644 --- a/docs/markdown/hip_porting_guide.md +++ b/docs/markdown/hip_porting_guide.md @@ -405,7 +405,7 @@ Code should not assume a warp size of 32 or 64. See [Warp Cross-Lane Functions] ## memcpyToSymbol -HIP support for hipMemCpyToSymbol is complete. This feature allows a kernel +HIP support for hipMemcpyToSymbol is complete. This feature allows a kernel to define a device-side data symbol which can be accessed on the host side. The symbol can be in __constant or device space. From 8d6acec1351e4155fdea32960e38b93e49d7456c Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Tue, 8 Aug 2017 10:15:32 +0530 Subject: [PATCH 064/108] Bump min hcc_workweek required for named kernel dispatch to 17312 Change-Id: I8c7b58306b279ed113d03260e4bc6086bb8b4e68 --- src/hip_module.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hip_module.cpp b/src/hip_module.cpp index 10fcd15fdc..74b0092b84 100644 --- a/src/hip_module.cpp +++ b/src/hip_module.cpp @@ -452,7 +452,7 @@ hipError_t ihipModuleLaunchKernel(hipFunction_t f, lp.av->dispatch_hsa_kernel(&aql, config[1] /* kernarg*/, kernArgSize, (startEvent || stopEvent) ? &cf : nullptr -#if (__hcc_workweek__ >= 17300) +#if (__hcc_workweek__ > 17312) , f->_name.c_str() #endif ); From 4264a4ce563f9e6b6f27b44583ce1dc31284cb40 Mon Sep 17 00:00:00 2001 From: Weixing Zhang Date: Thu, 3 Aug 2017 00:29:10 -0400 Subject: [PATCH 065/108] [HIP Texture] The GPU virtual address for texture memory needs to be aligned. In hcc_am, a bigger buffer will be allocated for alignment purpose and _unalignedDevicePointer is added in struct AmPointerInfo for original allocated address. --- src/hip_hcc.cpp | 14 ++--- src/hip_memory.cpp | 141 +++++++++++++++++++++++++++++++++++++++------ 2 files changed, 129 insertions(+), 26 deletions(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index b89099ea0d..07d58599cf 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -1866,8 +1866,8 @@ void ihipStream_t::locked_copySync(void* dst, const void* src, size_t sizeBytes, } hc::accelerator acc; - hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); - hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo dstPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo srcPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); bool dstTracked = getTailoredPtrInfo(&dstPtrInfo, dst, sizeBytes); bool srcTracked = getTailoredPtrInfo(&srcPtrInfo, src, sizeBytes); @@ -1902,7 +1902,7 @@ void ihipStream_t::locked_copySync(void* dst, const void* src, size_t sizeBytes, } void ihipStream_t::addSymbolPtrToTracker(hc::accelerator& acc, void* ptr, size_t sizeBytes) { - hc::AmPointerInfo ptrInfo(NULL, ptr, sizeBytes, acc, true, false); + hc::AmPointerInfo ptrInfo(NULL, ptr, ptr, sizeBytes, acc, true, false); hc::am_memtracker_add(ptr, ptrInfo); } @@ -1926,7 +1926,7 @@ void ihipStream_t::lockedSymbolCopyAsync(hc::accelerator &acc, void* dst, void* { // TODO - review - this looks broken , should not be adding pointers to tracker dynamically: if(kind == hipMemcpyHostToDevice) { - hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo srcPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); bool srcTracked = (hc::am_memtracker_getinfo(&srcPtrInfo, src) == AM_SUCCESS); if(srcTracked) { addSymbolPtrToTracker(acc, dst, sizeBytes); @@ -1938,7 +1938,7 @@ void ihipStream_t::lockedSymbolCopyAsync(hc::accelerator &acc, void* dst, void* } } if(kind == hipMemcpyDeviceToHost) { - hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo dstPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); bool dstTracked = (hc::am_memtracker_getinfo(&dstPtrInfo, dst) == AM_SUCCESS); if(dstTracked) { addSymbolPtrToTracker(acc, src, sizeBytes); @@ -1977,8 +1977,8 @@ void ihipStream_t::locked_copyAsync(void* dst, const void* src, size_t sizeBytes } else { hc::accelerator acc; - hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); - hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo dstPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo srcPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); bool dstTracked = getTailoredPtrInfo(&dstPtrInfo, dst, sizeBytes); bool srcTracked = getTailoredPtrInfo(&srcPtrInfo, src, sizeBytes); diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 9bb31d2d78..bf51748070 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -101,14 +101,18 @@ int sharePtr(void *ptr, ihipCtx_t *ctx, bool shareWithAll, unsigned hipFlags) // Allocate a new pointer with am_alloc and share with all valid peers. // Returns null-ptr if a memory error occurs (either allocation or sharing) -void * allocAndSharePtr(const char *msg, size_t sizeBytes, ihipCtx_t *ctx, bool shareWithAll, unsigned amFlags, unsigned hipFlags) +void * allocAndSharePtr(const char *msg, size_t sizeBytes, ihipCtx_t *ctx, bool shareWithAll, unsigned amFlags, unsigned hipFlags, size_t alignment) { void *ptr = nullptr; auto device = ctx->getWriteableDevice(); - ptr = hc::am_alloc(sizeBytes, device->_acc, amFlags); + if (alignment != 0) { + ptr = hc::am_aligned_alloc(sizeBytes, device->_acc, amFlags, alignment); + } else { + ptr = hc::am_alloc(sizeBytes, device->_acc, amFlags); + } tprintf(DB_MEM, " alloc %s ptr:%p-%p size:%zu on dev:%d\n", msg, ptr, static_cast(ptr)+sizeBytes, sizeBytes, device->_deviceId); @@ -145,7 +149,7 @@ hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void hipError_t e = hipSuccess; hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if (status == AM_SUCCESS) { @@ -195,7 +199,7 @@ hipError_t hipHostGetDevicePointer(void **devicePointer, void *hostPointer, unsi e = hipErrorInvalidValue; } else { hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, hostPointer); if (status == AM_SUCCESS) { *devicePointer = static_cast(amPointerInfo._devicePointer) + (static_cast(hostPointer) - static_cast(amPointerInfo._hostPointer)) ; @@ -226,7 +230,7 @@ hipError_t hipMalloc(void** ptr, size_t sizeBytes) } else { auto device = ctx->getWriteableDevice(); - *ptr = hip_internal::allocAndSharePtr("device_mem", sizeBytes, ctx, false/*shareWithAll*/, 0/*amFlags*/, 0/*hipFlags*/); + *ptr = hip_internal::allocAndSharePtr("device_mem", sizeBytes, ctx, false/*shareWithAll*/, 0/*amFlags*/, 0/*hipFlags*/, 0); if(sizeBytes && (*ptr == NULL)){ hip_status = hipErrorMemoryAllocation; @@ -294,7 +298,7 @@ hipError_t hipHostMalloc(void** ptr, size_t sizeBytes, unsigned int flags) *ptr = hip_internal::allocAndSharePtr((amFlags & amHostCoherent) ? "finegrained_host":"pinned_host", - sizeBytes, ctx, (trueFlags & hipHostMallocPortable) /*shareWithAll*/, amFlags, flags); + sizeBytes, ctx, (trueFlags & hipHostMallocPortable) /*shareWithAll*/, amFlags, flags, 0); if(sizeBytes && (*ptr == NULL)){ hip_status = hipErrorMemoryAllocation; @@ -340,10 +344,29 @@ hipError_t hipMallocPitch(void** ptr, size_t* pitch, size_t width, size_t height //err = hipMalloc(ptr, (*pitch)*height); if (ctx) { - auto device = ctx->getWriteableDevice(); + hc::accelerator acc = ctx->getDevice()->_acc; + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + size_t allocGranularity = 0; + hsa_amd_memory_pool_t *allocRegion = static_cast(acc.get_hsa_am_region()); + hsa_amd_memory_pool_get_info(*allocRegion, HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_GRANULE, &allocGranularity); + + hsa_ext_image_descriptor_t imageDescriptor; + imageDescriptor.width = *pitch; + imageDescriptor.height = height; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + imageDescriptor.format.channel_order = HSA_EXT_IMAGE_CHANNEL_ORDER_R; + imageDescriptor.format.channel_type = HSA_EXT_IMAGE_CHANNEL_TYPE_UNSIGNED_INT32; + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + hsa_ext_image_data_info_t imageInfo; + hsa_status_t status = hsa_ext_image_data_get_info(*agent, &imageDescriptor, permission, &imageInfo); + size_t alignment = imageInfo.alignment <= allocGranularity ? 0 : imageInfo.alignment; const unsigned am_flags = 0; - *ptr = hip_internal::allocAndSharePtr("device_pitch", sizeBytes, ctx, false/*shareWithAll*/, am_flags, 0); + *ptr = hip_internal::allocAndSharePtr("device_pitch", sizeBytes, ctx, false/*shareWithAll*/, am_flags, 0, alignment); if (sizeBytes && (*ptr == NULL)) { hip_status = hipErrorMemoryAllocation; @@ -363,6 +386,11 @@ hipChannelFormatDesc hipCreateChannelDesc(int x, int y, int z, int w, hipChannel return cd; } +extern void getChannelOrderAndType(const hipChannelFormatDesc& desc, + enum hipTextureReadMode readMode, + hsa_ext_image_channel_order_t& channelOrder, + hsa_ext_image_channel_type_t& channelType); + hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, size_t width, size_t height, unsigned int flags) { @@ -403,7 +431,43 @@ hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, hip_status = hipErrorUnknown; break; } - *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false/*shareWithAll*/, am_flags, 0); + hc::accelerator acc = ctx->getDevice()->_acc; + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + size_t allocGranularity = 0; + hsa_amd_memory_pool_t *allocRegion = static_cast(acc.get_hsa_am_region()); + hsa_amd_memory_pool_get_info(*allocRegion, HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_GRANULE, &allocGranularity); + + hsa_ext_image_descriptor_t imageDescriptor; + + imageDescriptor.width = width; + imageDescriptor.height = height; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + switch (flags) { + case hipArrayLayered: + case hipArrayCubemap: + case hipArraySurfaceLoadStore: + case hipArrayTextureGather: + assert(0); + break; + case hipArrayDefault: + default: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; + break; + } + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + getChannelOrderAndType(*desc, hipReadModeElementType, channelOrder, channelType); + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + hsa_ext_image_data_info_t imageInfo; + hsa_status_t status = hsa_ext_image_data_get_info(*agent, &imageDescriptor, permission, &imageInfo); + size_t alignment = imageInfo.alignment <= allocGranularity ? 0 : imageInfo.alignment; + + *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false/*shareWithAll*/, am_flags, 0, alignment); if (size && (*ptr == NULL)) { hip_status = hipErrorMemoryAllocation; } @@ -457,7 +521,47 @@ hipError_t hipMalloc3DArray(hipArray_t *array, hip_status = hipErrorUnknown; break; } - *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false, am_flags, 0); + + hc::accelerator acc = ctx->getDevice()->_acc; + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + + size_t allocGranularity = 0; + hsa_amd_memory_pool_t *allocRegion = static_cast(acc.get_hsa_am_region()); + hsa_amd_memory_pool_get_info(*allocRegion, HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_GRANULE, &allocGranularity); + + hsa_ext_image_descriptor_t imageDescriptor; + imageDescriptor.width = extent.width; + imageDescriptor.height = extent.height; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + switch (flags) { + case hipArrayLayered: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2DA; + imageDescriptor.array_size = extent.depth; + break; + case hipArraySurfaceLoadStore: + case hipArrayTextureGather: + case hipArrayDefault: + assert(0); + break; + case hipArrayCubemap: + default: + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_3D; + imageDescriptor.depth = extent.depth; + break; + } + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + getChannelOrderAndType(*desc, hipReadModeElementType, channelOrder, channelType); + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; + + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + hsa_ext_image_data_info_t imageInfo; + hsa_status_t status = hsa_ext_image_data_get_info(*agent, &imageDescriptor, permission, &imageInfo); + size_t alignment = imageInfo.alignment <= allocGranularity ? 0 : imageInfo.alignment; + + *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false, am_flags, 0, alignment); if (size && (*ptr == NULL)) { hip_status = hipErrorMemoryAllocation; @@ -467,7 +571,6 @@ hipError_t hipMalloc3DArray(hipArray_t *array, hip_status = hipErrorMemoryAllocation; } - //return ihipLogStatus(hip_status); return hip_status; } @@ -478,7 +581,7 @@ hipError_t hipHostGetFlags(unsigned int* flagsPtr, void* hostPtr) hipError_t hip_status = hipSuccess; hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, hostPtr); if(status == AM_SUCCESS){ *flagsPtr = amPointerInfo._appAllocationFlags; @@ -509,7 +612,7 @@ hipError_t hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags) } hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t am_status = hc::am_memtracker_getinfo(&amPointerInfo, hostPtr); if(am_status == AM_SUCCESS){ @@ -1309,7 +1412,7 @@ hipError_t hipMemPtrGetInfo(void *ptr, size_t *size) if(ptr != nullptr && size != nullptr){ hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if(status == AM_SUCCESS){ *size = amPointerInfo._sizeBytes; @@ -1334,7 +1437,7 @@ hipError_t hipFree(void* ptr) if (ptr) { hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if(status == AM_SUCCESS){ if(amPointerInfo._hostPointer == NULL){ @@ -1362,7 +1465,7 @@ hipError_t hipHostFree(void* ptr) hipError_t hipStatus = hipErrorInvalidValue; if (ptr) { hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if(status == AM_SUCCESS){ if(amPointerInfo._hostPointer == ptr){ @@ -1396,7 +1499,7 @@ hipError_t hipFreeArray(hipArray* array) if(array->data) { hc::accelerator acc; - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, array->data); if(status == AM_SUCCESS){ if(amPointerInfo._hostPointer == NULL){ @@ -1414,7 +1517,7 @@ hipError_t hipMemGetAddressRange ( hipDeviceptr_t* pbase, size_t* psize, hipDevi HIP_INIT_API ( pbase , psize , dptr ); hipError_t hipStatus = hipSuccess; hc::accelerator acc; - hc::AmPointerInfo amPointerInfo( NULL , NULL , 0 , acc , 0 , 0 ); + hc::AmPointerInfo amPointerInfo( NULL , NULL , NULL, 0 , acc , 0 , 0 ); am_status_t status = hc::am_memtracker_getinfo( &amPointerInfo , dptr ); if (status == AM_SUCCESS) { *pbase = amPointerInfo._devicePointer; @@ -1437,7 +1540,7 @@ hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t* handle, void* devPtr){ if((handle == NULL) || (devPtr == NULL)) { hipStatus = hipErrorInvalidResourceHandle; } else { - hc::AmPointerInfo amPointerInfo( NULL , NULL , 0 , acc , 0 , 0 ); + hc::AmPointerInfo amPointerInfo( NULL , NULL , NULL, 0 , acc , 0 , 0 ); am_status_t status = hc::am_memtracker_getinfo( &amPointerInfo , devPtr ); if (status == AM_SUCCESS) { psize = (size_t)amPointerInfo._sizeBytes; From 69e431f03588a78673b3979f252cf393038476c0 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Thu, 10 Aug 2017 12:05:58 +0530 Subject: [PATCH 066/108] [dtests] Disable hip_threadfence_system.tst on nvcc The test needs some more modifications before it is ready for building on nvcc path. Disabling it for now so that the pull request can be merged. Change-Id: I19a894fcda7b2159f86a4e4e95a409c5655d3760 --- tests/src/deviceLib/hip_threadfence_system.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/src/deviceLib/hip_threadfence_system.cpp b/tests/src/deviceLib/hip_threadfence_system.cpp index 2ff3505c15..673cb4507a 100644 --- a/tests/src/deviceLib/hip_threadfence_system.cpp +++ b/tests/src/deviceLib/hip_threadfence_system.cpp @@ -21,7 +21,7 @@ THE SOFTWARE. */ /* HIT_START - * BUILD: %t %s ../test_common.cpp + * BUILD: %t %s ../test_common.cpp EXCLUDE_HIP_PLATFORM nvcc * RUN: %t * HIT_END */ From 9a04c5cdd419b30b68970f0164d579eed627bd3b Mon Sep 17 00:00:00 2001 From: Siu Chi Chan Date: Thu, 10 Aug 2017 17:58:36 +0000 Subject: [PATCH 067/108] add C++11 compilation flags and minor bug fixes --- tests/src/deviceLib/hip_threadfence_system.cpp | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/tests/src/deviceLib/hip_threadfence_system.cpp b/tests/src/deviceLib/hip_threadfence_system.cpp index 673cb4507a..be59d56ab1 100644 --- a/tests/src/deviceLib/hip_threadfence_system.cpp +++ b/tests/src/deviceLib/hip_threadfence_system.cpp @@ -21,11 +21,13 @@ THE SOFTWARE. */ /* HIT_START - * BUILD: %t %s ../test_common.cpp EXCLUDE_HIP_PLATFORM nvcc + * BUILD: %t %s ../test_common.cpp -std=c++11 * RUN: %t * HIT_END */ +#include +#include #include #include #include @@ -33,15 +35,14 @@ THE SOFTWARE. #include "hip/device_functions.h" #include "test_common.h" - #define HIP_ASSERT(x) (assert((x)==hipSuccess)) -__host__ void fence_system() { - std::atomic_thread_fence(std::memory_order_seq_cst); -} - -__device__ void fence_system() { +__host__ __device__ void fence_system() { +#ifdef __HIP_DEVICE_COMPILE__ __threadfence_system(); +#else + std::atomic_thread_fence(std::memory_order_seq_cst); +#endif } __host__ __device__ void round_robin(const int id, const int num_dev, const int num_iter, volatile int* data, volatile int* flag) { From 4bcb3372b52f150cfead9ca9c82be54a8b8bcbc8 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Fri, 11 Aug 2017 14:49:17 +0530 Subject: [PATCH 068/108] [dtests] Fix hip_threadfence_system test's BUILD command Change-Id: I037c523369e4bddd7a386ffe117fe97a178900fd --- tests/src/deviceLib/hip_threadfence_system.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/src/deviceLib/hip_threadfence_system.cpp b/tests/src/deviceLib/hip_threadfence_system.cpp index be59d56ab1..31f6db1e17 100644 --- a/tests/src/deviceLib/hip_threadfence_system.cpp +++ b/tests/src/deviceLib/hip_threadfence_system.cpp @@ -21,7 +21,7 @@ THE SOFTWARE. */ /* HIT_START - * BUILD: %t %s ../test_common.cpp -std=c++11 + * BUILD: %t %s ../test_common.cpp NVCC_OPTIONS -std=c++11 * RUN: %t * HIT_END */ From ec3db6791e01b76cae504cc2d99f7fb6c05b860e Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Fri, 11 Aug 2017 15:57:20 +0530 Subject: [PATCH 069/108] [ci] Continue building remaining dtests on failure Change-Id: Iab7fd53517e0b66dd5f71efb9e65bd2e9c91355b --- Jenkinsfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Jenkinsfile b/Jenkinsfile index 526a825c65..8aba810dc0 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -118,7 +118,7 @@ def docker_build_inside_image( def build_image, String inside_args, String platf set -x cd ${build_dir_rel} make install -j\$(nproc) - make build_tests -j\$(nproc) + make build_tests -i -j\$(nproc) make test """ // If unit tests output a junit or xunit file in the future, jenkins can parse that file From e40047f2bf698515e4b75e21d81cd0d57c462f81 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Tue, 15 Aug 2017 15:51:38 +0530 Subject: [PATCH 070/108] [texture] guard new HCC APIs under workweek Change-Id: I4f60a64fb0b0496ca1eb01ffe6ddda121c25d976 --- src/hip_hcc.cpp | 22 ++++++++++++++++++++++ src/hip_memory.cpp | 45 ++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 66 insertions(+), 1 deletion(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 07d58599cf..f7082e7e9e 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -1866,8 +1866,13 @@ void ihipStream_t::locked_copySync(void* dst, const void* src, size_t sizeBytes, } hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo dstPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); hc::AmPointerInfo srcPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); +#endif bool dstTracked = getTailoredPtrInfo(&dstPtrInfo, dst, sizeBytes); bool srcTracked = getTailoredPtrInfo(&srcPtrInfo, src, sizeBytes); @@ -1902,7 +1907,11 @@ void ihipStream_t::locked_copySync(void* dst, const void* src, size_t sizeBytes, } void ihipStream_t::addSymbolPtrToTracker(hc::accelerator& acc, void* ptr, size_t sizeBytes) { +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo ptrInfo(NULL, ptr, ptr, sizeBytes, acc, true, false); +#else + hc::AmPointerInfo ptrInfo(NULL, ptr, sizeBytes, acc, true, false); +#endif hc::am_memtracker_add(ptr, ptrInfo); } @@ -1926,7 +1935,11 @@ void ihipStream_t::lockedSymbolCopyAsync(hc::accelerator &acc, void* dst, void* { // TODO - review - this looks broken , should not be adding pointers to tracker dynamically: if(kind == hipMemcpyHostToDevice) { +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo srcPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); +#endif bool srcTracked = (hc::am_memtracker_getinfo(&srcPtrInfo, src) == AM_SUCCESS); if(srcTracked) { addSymbolPtrToTracker(acc, dst, sizeBytes); @@ -1938,7 +1951,11 @@ void ihipStream_t::lockedSymbolCopyAsync(hc::accelerator &acc, void* dst, void* } } if(kind == hipMemcpyDeviceToHost) { +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo dstPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); +#endif bool dstTracked = (hc::am_memtracker_getinfo(&dstPtrInfo, dst) == AM_SUCCESS); if(dstTracked) { addSymbolPtrToTracker(acc, src, sizeBytes); @@ -1977,8 +1994,13 @@ void ihipStream_t::locked_copyAsync(void* dst, const void* src, size_t sizeBytes } else { hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo dstPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); hc::AmPointerInfo srcPtrInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo dstPtrInfo(NULL, NULL, 0, acc, 0, 0); + hc::AmPointerInfo srcPtrInfo(NULL, NULL, 0, acc, 0, 0); +#endif bool dstTracked = getTailoredPtrInfo(&dstPtrInfo, dst, sizeBytes); bool srcTracked = getTailoredPtrInfo(&srcPtrInfo, src, sizeBytes); diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index bf51748070..6db3d82376 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -108,9 +108,12 @@ void * allocAndSharePtr(const char *msg, size_t sizeBytes, ihipCtx_t *ctx, bool auto device = ctx->getWriteableDevice(); +#if (__hcc_workweek__ >= 17332) if (alignment != 0) { ptr = hc::am_aligned_alloc(sizeBytes, device->_acc, amFlags, alignment); - } else { + } else +#endif + { ptr = hc::am_alloc(sizeBytes, device->_acc, amFlags); } tprintf(DB_MEM, " alloc %s ptr:%p-%p size:%zu on dev:%d\n", @@ -149,7 +152,11 @@ hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void hipError_t e = hipSuccess; hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if (status == AM_SUCCESS) { @@ -199,7 +206,11 @@ hipError_t hipHostGetDevicePointer(void **devicePointer, void *hostPointer, unsi e = hipErrorInvalidValue; } else { hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, hostPointer); if (status == AM_SUCCESS) { *devicePointer = static_cast(amPointerInfo._devicePointer) + (static_cast(hostPointer) - static_cast(amPointerInfo._hostPointer)) ; @@ -581,7 +592,11 @@ hipError_t hipHostGetFlags(unsigned int* flagsPtr, void* hostPtr) hipError_t hip_status = hipSuccess; hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, hostPtr); if(status == AM_SUCCESS){ *flagsPtr = amPointerInfo._appAllocationFlags; @@ -612,7 +627,11 @@ hipError_t hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags) } hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t am_status = hc::am_memtracker_getinfo(&amPointerInfo, hostPtr); if(am_status == AM_SUCCESS){ @@ -1412,7 +1431,11 @@ hipError_t hipMemPtrGetInfo(void *ptr, size_t *size) if(ptr != nullptr && size != nullptr){ hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if(status == AM_SUCCESS){ *size = amPointerInfo._sizeBytes; @@ -1437,7 +1460,11 @@ hipError_t hipFree(void* ptr) if (ptr) { hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if(status == AM_SUCCESS){ if(amPointerInfo._hostPointer == NULL){ @@ -1465,7 +1492,11 @@ hipError_t hipHostFree(void* ptr) hipError_t hipStatus = hipErrorInvalidValue; if (ptr) { hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); if(status == AM_SUCCESS){ if(amPointerInfo._hostPointer == ptr){ @@ -1499,7 +1530,11 @@ hipError_t hipFreeArray(hipArray* array) if(array->data) { hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, array->data); if(status == AM_SUCCESS){ if(amPointerInfo._hostPointer == NULL){ @@ -1517,7 +1552,11 @@ hipError_t hipMemGetAddressRange ( hipDeviceptr_t* pbase, size_t* psize, hipDevi HIP_INIT_API ( pbase , psize , dptr ); hipError_t hipStatus = hipSuccess; hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo( NULL , NULL , NULL, 0 , acc , 0 , 0 ); +#else + hc::AmPointerInfo amPointerInfo( NULL , NULL, 0 , acc , 0 , 0 ); +#endif am_status_t status = hc::am_memtracker_getinfo( &amPointerInfo , dptr ); if (status == AM_SUCCESS) { *pbase = amPointerInfo._devicePointer; @@ -1540,7 +1579,11 @@ hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t* handle, void* devPtr){ if((handle == NULL) || (devPtr == NULL)) { hipStatus = hipErrorInvalidResourceHandle; } else { +#if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo( NULL , NULL , NULL, 0 , acc , 0 , 0 ); +#else + hc::AmPointerInfo amPointerInfo( NULL , NULL , 0 , acc , 0 , 0 ); +#endif am_status_t status = hc::am_memtracker_getinfo( &amPointerInfo , devPtr ); if (status == AM_SUCCESS) { psize = (size_t)amPointerInfo._sizeBytes; From 512214e930f59c5907984a9054ce89634959827a Mon Sep 17 00:00:00 2001 From: wsttiger Date: Tue, 15 Aug 2017 13:22:20 -0500 Subject: [PATCH 071/108] Add missing CUDA tags --- bin/hipify-perl | 9 ++++++--- include/hip/hip_runtime_api.h | 1 + 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/bin/hipify-perl b/bin/hipify-perl index 4b6e6e89b9..a37f728bd1 100755 --- a/bin/hipify-perl +++ b/bin/hipify-perl @@ -247,6 +247,7 @@ while (@ARGV) { $ft{'err'} += s/\bcudaErrorInvalidValue\b/hipErrorInvalidValue/g; $ft{'err'} += s/\bcudaErrorInvalidResourceHandle\b/hipErrorInvalidResourceHandle/g; $ft{'err'} += s/\bcudaErrorInvalidDevice\b/hipErrorInvalidDevice/g; + $ft{'err'} += s/\bcudaErrorInvalidDevicePointer\b/hipErrorInvalidDevicePointer/g; $ft{'err'} += s/\bcudaErrorNoDevice\b/hipErrorNoDevice/g; $ft{'err'} += s/\bcudaErrorNotReady\b/hipErrorNotReady/g; $ft{'err'} += s/\bcudaErrorUnknown\b/hipErrorUnknown/g; @@ -280,6 +281,7 @@ while (@ARGV) { $ft{'mem'} += s/\bcudaMemcpyKind\b/hipMemcpyKind/g; $ft{'mem'} += s/\bcudaPointerAttributes\b/hipPointerAttribute_t/g; + $ft{'mem'} += s/\bcudaPointerGetAttributes\b/hipPointerGetAttribute_t/g; $ft{'mem'} += s/\bcudaMemcpy2D\b/hipMemcpy2D/g; $ft{'mem'} += s/\bcudaMemcpy2DToArray\b/hipMemcpy2DToArray/g; @@ -339,6 +341,7 @@ while (@ARGV) { $ft{'event'} += s/\bcudaEventElapsedTime\b/hipEventElapsedTime/g; $ft{'event'} += s/\bcudaEventSynchronize\b/hipEventSynchronize/g; $ft{'event'} += s/\bcudaEventDisableTiming\b/hipEventDisableTiming/g; + $ft{'event'} += s/\bcudaEventQuery\b/hipEventQuery/g; #-------- # Streams @@ -489,15 +492,15 @@ while (@ARGV) { my $kernelName; # Handle the <>> syntax: - $k += s/(\w+)\s*(<.*>)?\s*<<<\s*(.+)\s*,\s*(.+)\s*,\s*(.+)\s*,\s*(.+)\s*>>>([\s*\\]*)\(/hipLaunchKernel(HIP_KERNEL_NAME($1$2), dim3($3), dim3($4), $5, $6, /g; + $k += s/(\w+)\s*(<.*>)?\s*<<<\s*(.+)\s*,\s*(.+)\s*,\s*(.+)\s*,\s*(.+)\s*>>>([\s*\\]*)\(/hipLaunchKernelGGL(($1$2), dim3($3), dim3($4), $5, $6, /g; $kernelName = $1 if $k; # Handle the <>> syntax: - $k += s/(\w+)\s*(<.*>)?\s*<<<\s*(.+)\s*,\s*(.+)\s*,\s*(.+)\s*>>>([\s*\\]*)\(/hipLaunchKernel(HIP_KERNEL_NAME($1$2), dim3($3), dim3($4), $5, 0, /g; + $k += s/(\w+)\s*(<.*>)?\s*<<<\s*(.+)\s*,\s*(.+)\s*,\s*(.+)\s*>>>([\s*\\]*)\(/hipLaunchKernelGGL(($1$2), dim3($3), dim3($4), $5, 0, /g; $kernelName = $1 if $k; # Handle the <>> syntax: - $k += s/(\w+)\s*(<.*>)?\s*<<<\s*(.+)\s*,\s*(.+)\s*>>>([\s\\]*)\(/hipLaunchKernel(HIP_KERNEL_NAME($1$2), dim3($3), dim3($4), 0, 0, /g; + $k += s/(\w+)\s*(<.*>)?\s*<<<\s*(.+)\s*,\s*(.+)\s*>>>([\s\\]*)\(/hipLaunchKernelGGL(($1$2), dim3($3), dim3($4), 0, 0, /g; $kernelName = $1 if $k; $ft{'kern'} += $k; diff --git a/include/hip/hip_runtime_api.h b/include/hip/hip_runtime_api.h index 0cdace0e99..97733964bf 100644 --- a/include/hip/hip_runtime_api.h +++ b/include/hip/hip_runtime_api.h @@ -161,6 +161,7 @@ typedef enum hipError_t { hipErrorProfilerNotInitialized = 6, hipErrorProfilerAlreadyStarted = 7, hipErrorProfilerAlreadyStopped = 8, + hipErrorInsufficientDriver = 35, hipErrorInvalidImage = 200, hipErrorInvalidContext = 201, ///< Produced when input context is invalid. hipErrorContextAlreadyCurrent = 202, From b6a3d8c946ed324c0e20eeec41241aaf2084f337 Mon Sep 17 00:00:00 2001 From: wsttiger Date: Tue, 15 Aug 2017 13:23:09 -0500 Subject: [PATCH 072/108] Modify header such that it's compatible with strict C compiler --- include/hip/nvcc_detail/hip_runtime_api.h | 56 +++++++++++++---------- 1 file changed, 31 insertions(+), 25 deletions(-) diff --git a/include/hip/nvcc_detail/hip_runtime_api.h b/include/hip/nvcc_detail/hip_runtime_api.h index 502d205ac4..090ceb6339 100644 --- a/include/hip/nvcc_detail/hip_runtime_api.h +++ b/include/hip/nvcc_detail/hip_runtime_api.h @@ -31,6 +31,13 @@ THE SOFTWARE. extern "C" { #endif +#ifdef __cplusplus + #define __dparm(x) \ + = x +#else + #define __dparm(x) +#endif + //TODO -move to include/hip_runtime_api.h as a common implementation. /** * Memory copy types @@ -102,8 +109,8 @@ typedef cudaEvent_t hipEvent_t; typedef cudaStream_t hipStream_t; typedef cudaIpcEventHandle_t hipIpcEventHandle_t; typedef cudaIpcMemHandle_t hipIpcMemHandle_t; -typedef cudaLimit hipLimit_t; -typedef cudaFuncCache hipFuncCache_t; +typedef enum cudaLimit hipLimit_t; +typedef enum cudaFuncCache hipFuncCache_t; typedef CUcontext hipCtx_t; typedef CUsharedconfig hipSharedMemConfig; typedef CUfunc_cache hipFuncCache; @@ -112,10 +119,10 @@ typedef CUdevice hipDevice_t; typedef CUmodule hipModule_t; typedef CUfunction hipFunction_t; typedef CUdeviceptr hipDeviceptr_t; -typedef cudaChannelFormatKind hipChannelFormatKind; -typedef cudaChannelFormatDesc hipChannelFormatDesc; -typedef cudaTextureReadMode hipTextureReadMode; -typedef cudaArray hipArray; +typedef enum cudaChannelFormatKind hipChannelFormatKind; +typedef struct cudaChannelFormatDesc hipChannelFormatDesc; +typedef enum cudaTextureReadMode hipTextureReadMode; +typedef struct cudaArray hipArray; // Flags that can be used with hipStreamCreateWithFlags #define hipStreamDefault cudaStreamDefault @@ -187,7 +194,7 @@ switch(hError) { } } -inline static cudaMemcpyKind hipMemcpyKindToCudaMemcpyKind(hipMemcpyKind kind) { +inline static enum cudaMemcpyKind hipMemcpyKindToCudaMemcpyKind(hipMemcpyKind kind) { switch(kind) { case hipMemcpyHostToHost: return cudaMemcpyHostToHost; @@ -250,7 +257,7 @@ inline static hipError_t hipHostMalloc(void** ptr, size_t size, unsigned int fla return hipCUDAErrorTohipError(cudaHostAlloc(ptr, size, flags)); } -inline static hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, size_t width, size_t height, unsigned int flags) { +inline static hipError_t hipMallocArray(hipArray** array, const struct hipChannelFormatDesc* desc, size_t width, size_t height, unsigned int flags) { return hipCUDAErrorTohipError(cudaMallocArray(array, desc, width, height, flags)); } @@ -289,8 +296,8 @@ inline static hipError_t hipSetDevice(int device) { inline static hipError_t hipChooseDevice( int* device, const hipDeviceProp_t* prop ) { - cudaDeviceProp cdprop; - memset(&cdprop,0x0,sizeof(cudaDeviceProp)); + struct cudaDeviceProp cdprop; + memset(&cdprop,0x0,sizeof(struct cudaDeviceProp)); cdprop.major= prop->major; cdprop.minor = prop->minor; cdprop.totalGlobalMem = prop->totalGlobalMem ; @@ -351,25 +358,24 @@ inline static hipError_t hipMemcpy(void* dst, const void* src, size_t sizeBytes, } -inline static hipError_t hipMemcpyAsync(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind copyKind, hipStream_t stream=0) { - return hipCUDAErrorTohipError(cudaMemcpyAsync(dst, src, sizeBytes, hipMemcpyKindToCudaMemcpyKind(copyKind), stream)); +inline static hipError_t hipMemcpyAsync(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind copyKind, hipStream_t stream __dparm(0)) { + return hipCUDAErrorTohipError(cudaMemcpyAsync(dst, src, sizeBytes, hipMemcpyKindToCudaMemcpyKind(copyKind), 0)); } - -inline static hipError_t hipMemcpyToSymbol(const void* symbol, const void* src, size_t sizeBytes, size_t offset = 0, hipMemcpyKind copyType = hipMemcpyHostToDevice) { +inline static hipError_t hipMemcpyToSymbol(const void* symbol, const void* src, size_t sizeBytes, size_t offset __dparm(0), hipMemcpyKind copyType __dparm(hipMemcpyHostToDevice)) { return hipCUDAErrorTohipError(cudaMemcpyToSymbol(symbol, src, sizeBytes, offset, hipMemcpyKindToCudaMemcpyKind(copyType))); } -inline static hipError_t hipMemcpyToSymbolAsync(const void* symbol, const void* src, size_t sizeBytes, size_t offset, hipMemcpyKind copyType, hipStream_t stream = 0) { +inline static hipError_t hipMemcpyToSymbolAsync(const void* symbol, const void* src, size_t sizeBytes, size_t offset, hipMemcpyKind copyType, hipStream_t stream __dparm(0)) { return hipCUDAErrorTohipError(cudaMemcpyToSymbolAsync(symbol, src, sizeBytes, offset, hipMemcpyKindToCudaMemcpyKind(copyType), stream)); } -inline static hipError_t hipMemcpyFromSymbol(void *dst, const void* symbolName, size_t sizeBytes, size_t offset = 0, hipMemcpyKind kind = hipMemcpyDeviceToHost) +inline static hipError_t hipMemcpyFromSymbol(void *dst, const void* symbolName, size_t sizeBytes, size_t offset __dparm(0), hipMemcpyKind kind __dparm(hipMemcpyDeviceToHost)) { return hipCUDAErrorTohipError(cudaMemcpyFromSymbol(dst, symbolName, sizeBytes, offset, hipMemcpyKindToCudaMemcpyKind(kind))); } -inline static hipError_t hipMemcpyFromSymbolAsync(void *dst, const void* symbolName, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream = 0) +inline static hipError_t hipMemcpyFromSymbolAsync(void *dst, const void* symbolName, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream __dparm(0)) { return hipCUDAErrorTohipError(cudaMemcpyFromSymbolAsync(dst, symbolName, sizeBytes, offset, hipMemcpyKindToCudaMemcpyKind(kind), stream)); } @@ -438,7 +444,7 @@ inline static hipError_t hipMemset(void* devPtr,int value, size_t count) { return hipCUDAErrorTohipError(cudaMemset(devPtr, value, count)); } -inline static hipError_t hipMemsetAsync(void* devPtr,int value, size_t count, hipStream_t stream = 0) { +inline static hipError_t hipMemsetAsync(void* devPtr,int value, size_t count, hipStream_t stream __dparm(0)) { return hipCUDAErrorTohipError(cudaMemsetAsync(devPtr, value, count, stream)); } @@ -449,7 +455,7 @@ inline static hipError_t hipMemsetD8(hipDeviceptr_t dest, unsigned char value, inline static hipError_t hipGetDeviceProperties(hipDeviceProp_t *p_prop, int device) { - cudaDeviceProp cdprop; + struct cudaDeviceProp cdprop; cudaError_t cerror; cerror = cudaGetDeviceProperties(&cdprop,device); strncpy(p_prop->name,cdprop.name, 256); @@ -510,7 +516,7 @@ inline static hipError_t hipGetDeviceProperties(hipDeviceProp_t *p_prop, int dev inline static hipError_t hipDeviceGetAttribute(int* pi, hipDeviceAttribute_t attr, int device) { - cudaDeviceAttr cdattr; + enum cudaDeviceAttr cdattr; cudaError_t cerror; switch (attr) { @@ -586,7 +592,7 @@ inline static hipError_t hipOccupancyMaxActiveBlocksPerMultiprocessor( } inline static hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, void* ptr){ - cudaPointerAttributes cPA; + struct cudaPointerAttributes cPA; hipError_t err = hipCUDAErrorTohipError(cudaPointerGetAttributes(&cPA, ptr)); if(err == hipSuccess){ switch (cPA.memoryType){ @@ -617,7 +623,7 @@ inline static hipError_t hipEventCreate( hipEvent_t* event) return hipCUDAErrorTohipError(cudaEventCreate(event)); } -inline static hipError_t hipEventRecord( hipEvent_t event, hipStream_t stream = NULL) +inline static hipError_t hipEventRecord( hipEvent_t event, hipStream_t stream __dparm(NULL)) { return hipCUDAErrorTohipError(cudaEventRecord(event,stream)); } @@ -750,18 +756,18 @@ inline static hipError_t hipMemcpyPeer ( void* dst, int dstDevice, const void* return hipCUDAErrorTohipError(cudaMemcpyPeer(dst, dstDevice, src, srcDevice, count)); } -inline static hipError_t hipMemcpyPeerAsync ( void* dst, int dstDevice, const void* src, int srcDevice, size_t count, hipStream_t stream=0 ) +inline static hipError_t hipMemcpyPeerAsync ( void* dst, int dstDevice, const void* src, int srcDevice, size_t count, hipStream_t stream __dparm(0)) { return hipCUDAErrorTohipError(cudaMemcpyPeerAsync(dst, dstDevice, src, srcDevice, count, stream)); } // Profile APIs: -inline hipError_t hipProfilerStart() +inline static hipError_t hipProfilerStart() { return hipCUDAErrorTohipError(cudaProfilerStart()); } -inline hipError_t hipProfilerStop() +inline static hipError_t hipProfilerStop() { return hipCUDAErrorTohipError(cudaProfilerStop()); } From e3161bb40ef9dad304586ddb51a4c9d475cb37d7 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Wed, 16 Aug 2017 03:50:04 +0000 Subject: [PATCH 073/108] Add some new HIP_TRACE_API options. --- src/hip_device.cpp | 2 +- src/hip_event.cpp | 6 +++--- src/hip_hcc_internal.h | 10 ++++++---- src/hip_stream.cpp | 5 +++-- 4 files changed, 13 insertions(+), 10 deletions(-) diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 1800c9369c..3322690731 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -153,7 +153,7 @@ hipError_t hipSetDevice(int deviceId) hipError_t hipDeviceSynchronize(void) { - HIP_INIT_API(); + HIP_INIT_SPECIAL_API(TRACE_SYNC); return ihipLogStatus(ihipSynchronize()); } diff --git a/src/hip_event.cpp b/src/hip_event.cpp index 71da9fdc30..3a8f1ab611 100644 --- a/src/hip_event.cpp +++ b/src/hip_event.cpp @@ -127,7 +127,7 @@ hipError_t hipEventCreate(hipEvent_t* event) hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream) { - HIP_INIT_API(event, stream); + HIP_INIT_SPECIAL_API(TRACE_QUERY, event, stream); if (event && event->_state != hipEventStatusUnitialized) { stream = ihipSyncAndResolveStream(stream); @@ -178,7 +178,7 @@ hipError_t hipEventDestroy(hipEvent_t event) hipError_t hipEventSynchronize(hipEvent_t event) { - HIP_INIT_API(event); + HIP_INIT_SPECIAL_API(TRACE_SYNC, event); if (event) { if (event->_state == hipEventStatusUnitialized) { @@ -257,7 +257,7 @@ hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop) hipError_t hipEventQuery(hipEvent_t event) { - HIP_INIT_API(event); + HIP_INIT_SPECIAL_API(TRACE_QUERY, event); if ((event->_state == hipEventStatusRecording) && (!event->_marker.is_ready())) { return ihipLogStatus(hipErrorNotReady); diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index 4cb85ffc19..af5e7a121b 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -191,10 +191,12 @@ extern const char *API_COLOR_END; //--- //HIP Trace modes - use with HIP_TRACE_API=... -#define TRACE_ALL 0 // 0x1 -#define TRACE_KCMD 1 // 0x2, kernel command -#define TRACE_MCMD 2 // 0x4, memory command -#define TRACE_MEM 3 // 0x8, memory allocation or deallocation. +#define TRACE_ALL 0 // 0x01 +#define TRACE_KCMD 1 // 0x02, kernel command +#define TRACE_MCMD 2 // 0x04, memory command +#define TRACE_MEM 3 // 0x08, memory allocation or deallocation. +#define TRACE_SYNC 4 // 0x10, synchronization (host or hipStreamWaitEvent) +#define TRACE_QUERY 5 // 0x20, hipEventRecord, hipEventQuery, hipStreamQuery //--- diff --git a/src/hip_stream.cpp b/src/hip_stream.cpp index 40aade28b9..f0e9283201 100644 --- a/src/hip_stream.cpp +++ b/src/hip_stream.cpp @@ -84,7 +84,7 @@ hipError_t hipStreamCreate(hipStream_t *stream) hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags) { - HIP_INIT_API(stream, event, flags); + HIP_INIT_SPECIAL_API(TRACE_SYNC, stream, event, flags); hipError_t e = hipSuccess; @@ -114,7 +114,7 @@ hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int //--- hipError_t hipStreamQuery(hipStream_t stream) { - HIP_INIT_API(stream); + HIP_INIT_SPECIAL_API(TRACE_QUERY, stream); // Use default stream if 0 specified: if (stream == hipStreamNull) { @@ -140,6 +140,7 @@ hipError_t hipStreamQuery(hipStream_t stream) hipError_t hipStreamSynchronize(hipStream_t stream) { HIP_INIT_API(stream); + HIP_INIT_SPECIAL_API(TRACE_SYNC, stream); hipError_t e = hipSuccess; From c11b8737b92ab9cb11b1ecce09e4574fd5ffa97b Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Wed, 16 Aug 2017 10:57:07 +0530 Subject: [PATCH 074/108] [ci] Archive packages based on hcc-ctu instead of hcc-1.6 Change-Id: Ie1ea6411ed19aae87b979fa001db3e8927fc11f6 --- Jenkinsfile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 8aba810dc0..af80a2c18b 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -140,8 +140,8 @@ def docker_build_inside_image( def build_image, String inside_args, String platf // No matter the base platform, all packages have the same name // Only upload 1 set of packages, so we don't have a race condition uploading packages - // I arbitrarily pick hcc-1.6 as the most stable - if( platform.toLowerCase( ).startsWith( 'hcc-1.6' ) ) + // Using hcc-ctu since that is what most people want + if( platform.toLowerCase( ).startsWith( 'hcc-ctu' ) ) { archiveArtifacts artifacts: "${build_dir_rel}/*.deb", fingerprint: true archiveArtifacts artifacts: "${build_dir_rel}/*.rpm", fingerprint: true From 8c94bc00099147a9e941ad2f999e97eac0ac174f Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Wed, 16 Aug 2017 11:18:51 +0530 Subject: [PATCH 075/108] Bump HIP version to 1.3 and update release notes Change-Id: I53f89efca03472cb906bd56200a2bf2996453915 --- RELEASE.md | 15 ++++++++++++--- bin/hipconfig | 2 +- 2 files changed, 13 insertions(+), 4 deletions(-) diff --git a/RELEASE.md b/RELEASE.md index 98c076dc6e..452ac54ee1 100644 --- a/RELEASE.md +++ b/RELEASE.md @@ -5,12 +5,21 @@ We have attempted to document known bugs and limitations - in particular the [HI =================================================================================================== -Upcoming: -- HIP development now fully open and on GitHub. Developers should submit pull requests. - ## Revision History: +=================================================================================================== +Release: 1.3 +Date: 2017.08.16 +- hipcc now auto-detects amdgcn arch. No need to specify the arch when building for same system. +- HIP texture support +- Implemented __threadfence_support +- Improvements in HIP context management logic +- Bug fixes in several APIs including hipDeviceGetPCIBusId, hipEventDestroy, hipMemcpy2DAsync +- Updates to hipify-clang and documentation +- HIP development now fully open and on GitHub. Developers should submit pull requests. + + =================================================================================================== Release: 1.2 Date: 2017.06.29 diff --git a/bin/hipconfig b/bin/hipconfig index 39fdab5a99..c74d757fb5 100755 --- a/bin/hipconfig +++ b/bin/hipconfig @@ -1,7 +1,7 @@ #!/usr/bin/perl -w $HIP_BASE_VERSION_MAJOR = "1"; -$HIP_BASE_VERSION_MINOR = "2"; +$HIP_BASE_VERSION_MINOR = "3"; # Need perl > 5.10 to use logic-defined or use 5.006; use v5.10.1; From e87f7a4d00bd141993c37caebb3f82f6c2f124ad Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Wed, 26 Jul 2017 18:52:53 -0500 Subject: [PATCH 076/108] Refactor dispatch latency test and fix several bugs. --- .../hipDispatchLatency/ResultDatabase.cpp | 29 ++- .../hipDispatchLatency/hipDispatchLatency.cpp | 219 +++++++++--------- 2 files changed, 136 insertions(+), 112 deletions(-) diff --git a/samples/1_Utils/hipDispatchLatency/ResultDatabase.cpp b/samples/1_Utils/hipDispatchLatency/ResultDatabase.cpp index d207154e39..f6f2fab709 100644 --- a/samples/1_Utils/hipDispatchLatency/ResultDatabase.cpp +++ b/samples/1_Utils/hipDispatchLatency/ResultDatabase.cpp @@ -7,16 +7,23 @@ using namespace std; +#define SORT_BY_NAME 0 +#define SORT_RETAIN_ATTS_ORDER 1 + + bool ResultDatabase::Result::operator<(const Result &rhs) const { if (test < rhs.test) return true; if (test > rhs.test) return false; +#if (SORT_RETAIN_ATTS_ORDER == 0) + // For ties, sort by the value of the attribute: if (atts < rhs.atts) return true; if (atts > rhs.atts) return false; +#endif return false; // less-operator returns false on equal } @@ -189,7 +196,10 @@ void ResultDatabase::AddResult(const string &test_orig, void ResultDatabase::DumpDetailed(ostream &out) { vector sorted(results); - sort(sorted.begin(), sorted.end()); + +#if SORT_BY_NAME + stable_sort(sorted.begin(), sorted.end()); +#endif const int testNameW = 24 ; const int attW = 12; @@ -283,12 +293,15 @@ void ResultDatabase::DumpDetailed(ostream &out) void ResultDatabase::DumpSummary(ostream &out) { vector sorted(results); - sort(sorted.begin(), sorted.end()); - const int testNameW = 24 ; +#if SORT_BY_NAME + stable_sort(sorted.begin(), sorted.end()); +#endif + + const int testNameW = 32 ; const int attW = 12; const int fieldW = 9; - out << std::fixed << right << std::setprecision(4); + out << std::fixed << right << std::setprecision(2); // TODO: in big parallel runs, the "trials" are the procs // and we really don't want to print them all out.... @@ -334,8 +347,8 @@ void ResultDatabase::DumpSummary(ostream &out) } if (0) { out << endl - << "Note: results marked with (*) had missing values such as" << endl - << "might occur with a mixture of architectural capabilities." << endl; + << "Note: results marked with (*) had missing values such as" << endl + << "might occur with a mixture of architectural capabilities." << endl; } } @@ -381,7 +394,9 @@ void ResultDatabase::DumpCsv(string fileName) bool emptyFile; vector sorted(results); - sort(sorted.begin(), sorted.end()); +#if SORT_BY_NAME + stable_sort(sorted.begin(), sorted.end()); +#endif //Check to see if the file is empty - if so, add the headers emptyFile = this->IsFileEmpty(fileName); diff --git a/samples/1_Utils/hipDispatchLatency/hipDispatchLatency.cpp b/samples/1_Utils/hipDispatchLatency/hipDispatchLatency.cpp index b343386b5c..2a4f6ff649 100644 --- a/samples/1_Utils/hipDispatchLatency/hipDispatchLatency.cpp +++ b/samples/1_Utils/hipDispatchLatency/hipDispatchLatency.cpp @@ -25,15 +25,27 @@ THE SOFTWARE. #include #include"ResultDatabase.h" -#define check(msg, status) \ -if(status != hipSuccess){ \ - printf("%s failed.\n",#msg); \ - exit(1); \ +#define PRINT_PROGRESS 0 + +#define check(cmd) \ +{\ + hipError_t status = cmd;\ + if(status != hipSuccess){ \ + printf("error: '%s'(%d) from %s at %s:%d\n", \ + hipGetErrorString(status), status, #cmd,\ + __FILE__, __LINE__); \ + abort(); \ + }\ } #define LEN 1024*1024 -#define SIZE LEN * sizeof(float) -#define ITER 10120 + +#define NUM_GROUPS 1 +#define GROUP_SIZE 64 +#define TEST_ITERS 20 +#define DISPATCHES_PER_TEST 100 + +const unsigned p_tests = 0xfffffff; // HCC optimizes away fully NULL kernel calls, so run one that is nearly null: @@ -44,115 +56,112 @@ __global__ void NearlyNull(hipLaunchParm lp, float* Ad){ } +ResultDatabase resultDB; + + +void stopTest(hipEvent_t start, hipEvent_t stop, const char *msg, int iters) +{ + float mS = 0; + check(hipEventRecord(stop)); + check(hipDeviceSynchronize()); + check(hipEventElapsedTime(&mS, start, stop)); + resultDB.AddResult(std::string(msg), "", "uS", mS*1000/iters); + if (PRINT_PROGRESS & 0x1 ) { + std::cout<< msg <<"\t\t"< Date: Tue, 22 Aug 2017 14:05:44 +0530 Subject: [PATCH 077/108] Device count check on input device id --- src/hip_device.cpp | 38 +++++++++++++++++++++++++++----------- 1 file changed, 27 insertions(+), 11 deletions(-) diff --git a/src/hip_device.cpp b/src/hip_device.cpp index 3322690731..ff511b5509 100644 --- a/src/hip_device.cpp +++ b/src/hip_device.cpp @@ -274,6 +274,9 @@ hipError_t ihipDeviceGetAttribute(int* pi, hipDeviceAttribute_t attr, int device hipError_t hipDeviceGetAttribute(int* pi, hipDeviceAttribute_t attr, int device) { HIP_INIT_API(pi, attr, device); + if ((device < 0) || (device >= g_deviceCnt)) { + return ihipLogStatus(hipErrorInvalidDevice); + } return ihipLogStatus(ihipDeviceGetAttribute(pi,attr,device)); } @@ -300,6 +303,9 @@ hipError_t ihipGetDeviceProperties(hipDeviceProp_t* props, int device) hipError_t hipGetDeviceProperties(hipDeviceProp_t* props, int device) { HIP_INIT_API(props, device); + if ((device < 0) || (device >= g_deviceCnt)) { + return ihipLogStatus(hipErrorInvalidDevice); + } return ihipLogStatus(ihipGetDeviceProperties(props, device)); } @@ -352,8 +358,12 @@ hipError_t hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device { HIP_INIT_API(major,minor, device); hipError_t e = hipSuccess; - e = ihipDeviceGetAttribute(major, hipDeviceAttributeComputeCapabilityMajor, device); - e = ihipDeviceGetAttribute(minor, hipDeviceAttributeComputeCapabilityMinor, device); + if ((device < 0) || (device >= g_deviceCnt)) { + e = hipErrorInvalidDevice; + } else { + e = ihipDeviceGetAttribute(major, hipDeviceAttributeComputeCapabilityMajor, device); + e = ihipDeviceGetAttribute(minor, hipDeviceAttributeComputeCapabilityMinor, device); + } return ihipLogStatus(e); } @@ -362,10 +372,14 @@ hipError_t hipDeviceGetName(char *name,int len,hipDevice_t device) // Cast to void* here to avoid printing garbage in debug modes. HIP_INIT_API((void*)name,len, device); hipError_t e = hipSuccess; - auto deviceHandle = ihipGetDevice(device); - int nameLen = strlen(deviceHandle->_props.name); - if(nameLen <= len) - memcpy(name,deviceHandle->_props.name,nameLen); + if ((device < 0) || (device >= g_deviceCnt)) { + e = hipErrorInvalidDevice; + } else { + auto deviceHandle = ihipGetDevice(device); + int nameLen = strlen(deviceHandle->_props.name); + if(nameLen <= len) + memcpy(name,deviceHandle->_props.name,nameLen); + } return ihipLogStatus(e); } @@ -374,9 +388,7 @@ hipError_t hipDeviceGetPCIBusId (char *pciBusId,int len, int device) // Cast to void* here to avoid printing garbage in debug modes. HIP_INIT_API((void*)pciBusId, len, device); hipError_t e = hipErrorInvalidValue; - int deviceCount = 0; - ihipGetDeviceCount( &deviceCount ); - if((device > deviceCount) || (device < 0)) { + if ((device < 0) || (device >= g_deviceCnt)) { e = hipErrorInvalidDevice; } else { if((pciBusId != nullptr) && (len > 0)) { @@ -394,8 +406,12 @@ hipError_t hipDeviceTotalMem (size_t *bytes,hipDevice_t device) { HIP_INIT_API(bytes, device); hipError_t e = hipSuccess; - auto deviceHandle = ihipGetDevice(device); - *bytes= deviceHandle->_props.totalGlobalMem; + if ((device < 0) || (device >= g_deviceCnt)) { + e = hipErrorInvalidDevice; + } else { + auto deviceHandle = ihipGetDevice(device); + *bytes= deviceHandle->_props.totalGlobalMem; + } return ihipLogStatus(e); } From 9441f3a50e5b7fbd9541b61cf05b67ff8a42d07a Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Tue, 22 Aug 2017 15:41:16 +0530 Subject: [PATCH 078/108] PCI Bus ID related dtests --- .../device/hipDeviceGetByPCIBusId.cpp | 56 +++++++++++++++++++ .../device/hipDeviceGetPCIBusId.cpp | 51 +++++++++++++++++ 2 files changed, 107 insertions(+) create mode 100644 tests/src/runtimeApi/device/hipDeviceGetByPCIBusId.cpp create mode 100644 tests/src/runtimeApi/device/hipDeviceGetPCIBusId.cpp diff --git a/tests/src/runtimeApi/device/hipDeviceGetByPCIBusId.cpp b/tests/src/runtimeApi/device/hipDeviceGetByPCIBusId.cpp new file mode 100644 index 0000000000..448c9d3730 --- /dev/null +++ b/tests/src/runtimeApi/device/hipDeviceGetByPCIBusId.cpp @@ -0,0 +1,56 @@ +/* +Copyright (c) 2015-2017 Advanced Micro Devices, Inc. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN +THE SOFTWARE. +*/ + +/* HIT_START + * BUILD: %t %s ../../test_common.cpp + * RUN: %t + * HIT_END + */ + +#include +#include "hip/hip_runtime.h" +#include "test_common.h" + +int main( void ) { + char pciBusId[13]; + int deviceCount = 0; + HIPCHECK(hipGetDeviceCount(&deviceCount)); + HIPASSERT(deviceCount != 0); + for(int i=0; i< deviceCount;i++) { + int pciBusID = -1; + int pciDeviceID = -1; + int pciDomainID = -1; + int tempPciBusId = -1; + int tempDeviceId = -1; + HIPCHECK(hipDeviceGetPCIBusId ( &pciBusId[0], 13, i )); + sscanf (pciBusId,"%04x:%02x:%02x",&pciDomainID,&pciBusID,&pciDeviceID); + HIPCHECK(hipDeviceGetAttribute(&tempPciBusId,hipDeviceAttributePciBusId , i)); + if(pciBusID != tempPciBusId) { + exit(EXIT_FAILURE); + } + HIPCHECK(hipDeviceGetByPCIBusId ( &tempDeviceId, pciBusId )); + if(tempDeviceId != i) { + exit(EXIT_FAILURE); + } + } + passed(); +} diff --git a/tests/src/runtimeApi/device/hipDeviceGetPCIBusId.cpp b/tests/src/runtimeApi/device/hipDeviceGetPCIBusId.cpp new file mode 100644 index 0000000000..7fb0340a48 --- /dev/null +++ b/tests/src/runtimeApi/device/hipDeviceGetPCIBusId.cpp @@ -0,0 +1,51 @@ +/* +Copyright (c) 2015-2017 Advanced Micro Devices, Inc. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN +THE SOFTWARE. +*/ + +/* HIT_START + * BUILD: %t %s ../../test_common.cpp + * RUN: %t + * HIT_END + */ + +#include +#include "hip/hip_runtime.h" +#include "test_common.h" + +int main( void ) { + char pciBusId[13]; + int deviceCount = 0; + HIPCHECK(hipGetDeviceCount(&deviceCount)); + HIPASSERT(deviceCount != 0); + for(int i=0; i< deviceCount;i++) { + int pciBusID = -1; + int pciDeviceID = -1; + int pciDomainID = -1; + int tempPciBusId = -1; + HIPCHECK(hipDeviceGetPCIBusId ( &pciBusId[0], 13, i )); + sscanf (pciBusId,"%04x:%02x:%02x",&pciDomainID,&pciBusID,&pciDeviceID); + HIPCHECK(hipDeviceGetAttribute(&tempPciBusId,hipDeviceAttributePciBusId , i)); + if(pciBusID != tempPciBusId) { + exit(EXIT_FAILURE); + } + } + passed(); +} From 5c09f35b6520c275eab6af7abc1ac147f45dbd8f Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Thu, 24 Aug 2017 15:05:45 +0530 Subject: [PATCH 079/108] Add a basic build+install script Change-Id: Ida0702057fda3cbbe9b4674e07aecacb1004fc90 --- install.sh | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 48 insertions(+) create mode 100755 install.sh diff --git a/install.sh b/install.sh new file mode 100755 index 0000000000..f8ad640798 --- /dev/null +++ b/install.sh @@ -0,0 +1,48 @@ +#!/bin/bash + +BUILD_ROOT="$( mktemp -d )" +SRC_ROOT="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )" +WORKING_DIR=$PWD +DASH_JAY="-j $(getconf _NPROCESSORS_ONLN)" + +err() { + echo "${1-Died}." >&2 +} + +die() { + err "$1" + exit 1 +} + +pushd () { + command pushd "$@" > /dev/null +} + +popd () { + command popd "$@" > /dev/null +} + +function setupENV() +{ + sudo apt-get update + sudo apt-get install dpkg-dev rpm doxygen libelf-dev +} + +function buildHIP() +{ + pushd $BUILD_ROOT + cmake $SRC_ROOT -DCMAKE_BUILD_TYPE=Release -DCOMPILE_HIP_ATP_MARKER=1 + make $DASH_JAY + make package + rename -v 's/([a-z0-9_.\-]).deb/$1-amd64.deb/' *.deb;rename -v 's/([a-z0-9_.\-]).rpm/$1.x86_64.rpm/' *.rpm + cp hip_*.deb $WORKING_DIR + sudo dpkg -i hip_base*.deb hip_hcc*.deb hip_sample*.deb hip_doc*.deb + popd + rm -rf $BUILD_ROOT +} + +echo "Preparing build environment" +setupENV || die "setupENV failed" +echo "Building and installing HIP packages" +buildHIP || die "buildHIP failed" +echo "Finished building HIP packages" From 4b98572f4725a90901024d3c635608787707ab91 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Thu, 24 Aug 2017 15:51:58 +0530 Subject: [PATCH 080/108] Enable hipStreamSync2 and record_event dtests --- tests/src/runtimeApi/event/record_event.cpp | 2 +- tests/src/runtimeApi/stream/hipStreamSync2.cpp | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/src/runtimeApi/event/record_event.cpp b/tests/src/runtimeApi/event/record_event.cpp index c117f28f5c..bd8a3ada8e 100644 --- a/tests/src/runtimeApi/event/record_event.cpp +++ b/tests/src/runtimeApi/event/record_event.cpp @@ -19,7 +19,7 @@ THE SOFTWARE. /* HIT_START * BUILD: %t %s ../../test_common.cpp - * RUN: %t EXCLUDE_HIP_PLATFORM hcc + * RUN: %t * HIT_END */ diff --git a/tests/src/runtimeApi/stream/hipStreamSync2.cpp b/tests/src/runtimeApi/stream/hipStreamSync2.cpp index 70b4e559fb..c6a58ce7d4 100644 --- a/tests/src/runtimeApi/stream/hipStreamSync2.cpp +++ b/tests/src/runtimeApi/stream/hipStreamSync2.cpp @@ -19,7 +19,7 @@ THE SOFTWARE. /* HIT_START * BUILD: %t %s ../../test_common.cpp - * RUN: %t EXCLUDE_HIP_PLATFORM hcc + * RUN: %t * HIT_END */ From 38c9febe21ef60beb20f080e0bd19e13d3b2c182 Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Fri, 25 Aug 2017 08:46:34 +0530 Subject: [PATCH 081/108] Null check on input pointer arguments --- src/hip_memory.cpp | 77 +++++++++++++++++++++++----------------------- 1 file changed, 38 insertions(+), 39 deletions(-) diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 6db3d82376..4b3a4fcb12 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -150,45 +150,45 @@ hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void HIP_INIT_API(attributes, ptr); hipError_t e = hipSuccess; - - hc::accelerator acc; -#if (__hcc_workweek__ >= 17332) - hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); -#else - hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); -#endif - am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); - if (status == AM_SUCCESS) { - - attributes->memoryType = amPointerInfo._isInDeviceMem ? hipMemoryTypeDevice: hipMemoryTypeHost; - attributes->hostPointer = amPointerInfo._hostPointer; - attributes->devicePointer = amPointerInfo._devicePointer; - attributes->isManaged = 0; - if(attributes->memoryType == hipMemoryTypeHost){ - attributes->hostPointer = (void*)ptr; - } - if(attributes->memoryType == hipMemoryTypeDevice){ - attributes->devicePointer = (void*)ptr; - } - attributes->allocationFlags = amPointerInfo._appAllocationFlags; - attributes->device = amPointerInfo._appId; - - if (attributes->device < 0) { - e = hipErrorInvalidDevice; - } - - + if((attributes == nullptr) || (ptr == nullptr)) { + e = hipErrorInvalidValue; } else { - attributes->memoryType = hipMemoryTypeDevice; - attributes->hostPointer = 0; - attributes->devicePointer = 0; - attributes->device = -1; - attributes->isManaged = 0; - attributes->allocationFlags = 0; + hc::accelerator acc; +#if (__hcc_workweek__ >= 17332) + hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); +#else + hc::AmPointerInfo amPointerInfo(NULL, NULL, 0, acc, 0, 0); +#endif + am_status_t status = hc::am_memtracker_getinfo(&amPointerInfo, ptr); + if (status == AM_SUCCESS) { - e = hipErrorUnknown; // TODO - should be hipErrorInvalidValue ? + attributes->memoryType = amPointerInfo._isInDeviceMem ? hipMemoryTypeDevice: hipMemoryTypeHost; + attributes->hostPointer = amPointerInfo._hostPointer; + attributes->devicePointer = amPointerInfo._devicePointer; + attributes->isManaged = 0; + if(attributes->memoryType == hipMemoryTypeHost){ + attributes->hostPointer = (void*)ptr; + } + if(attributes->memoryType == hipMemoryTypeDevice){ + attributes->devicePointer = (void*)ptr; + } + attributes->allocationFlags = amPointerInfo._appAllocationFlags; + attributes->device = amPointerInfo._appId; + + if (attributes->device < 0) { + e = hipErrorInvalidDevice; + } + } else { + attributes->memoryType = hipMemoryTypeDevice; + attributes->hostPointer = 0; + attributes->devicePointer = 0; + attributes->device = -1; + attributes->isManaged = 0; + attributes->allocationFlags = 0; + + e = hipErrorUnknown; // TODO - should be hipErrorInvalidValue ? + } } - return ihipLogStatus(e); } @@ -199,13 +199,12 @@ hipError_t hipHostGetDevicePointer(void **devicePointer, void *hostPointer, unsi hipError_t e = hipSuccess; - *devicePointer = NULL; - // Flags must be 0: - if (flags != 0) { + if ((flags != 0) || (devicePointer == nullptr) || (hostPointer == nullptr)){ e = hipErrorInvalidValue; } else { hc::accelerator acc; + *devicePointer = NULL; #if (__hcc_workweek__ >= 17332) hc::AmPointerInfo amPointerInfo(NULL, NULL, NULL, 0, acc, 0, 0); #else From d50fb898bfc2f8bcfe9c895897529f6aa17a1e0d Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Mon, 28 Aug 2017 13:58:18 -0500 Subject: [PATCH 082/108] Update the nvidia volume driver version Move deleting docker images into the try clause, don't treat failure to delete images as a build failure --- Jenkinsfile | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index af80a2c18b..ee3a301dd3 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -119,7 +119,7 @@ def docker_build_inside_image( def build_image, String inside_args, String platf cd ${build_dir_rel} make install -j\$(nproc) make build_tests -i -j\$(nproc) - make test + make test """ // If unit tests output a junit or xunit file in the future, jenkins can parse that file // to display test results on the dashboard @@ -194,15 +194,15 @@ def docker_upload_artifactory( String hcc_ver, String from_image, String source_ hip_install_image.push( 'latest' ) } } + + // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" } catch( err ) { currentBuild.result = 'SUCCESS' } - - // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable - // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" } } @@ -288,10 +288,10 @@ nvcc: String nvcc_ver = 'nvcc-8.0' String from_image = 'nvidia/cuda:8.0-devel' - // This unfortunately hardcodes the driver version nvidia_driver_375.66 in the volume mount. Research if a way + // This unfortunately hardcodes the driver version nvidia_driver_375.74 in the volume mount. Research if a way // exists to get volume driver to customize the volume names to leave out driver version String inside_args = '''--device=/dev/nvidiactl --device=/dev/nvidia0 --device=/dev/nvidia-uvm --device=/dev/nvidia-uvm-tools - --volume-driver=nvidia-docker --volume=nvidia_driver_375.66:/usr/local/nvidia:ro'''; + --volume-driver=nvidia-docker --volume=nvidia_driver_375.74:/usr/local/nvidia:ro'''; // Checkout source code, dependencies and version files String source_hip_rel = checkout_and_version( nvcc_ver ) From 1a9e58893232a5dcbec58d3dec986b9985d8d64e Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Mon, 28 Aug 2017 15:44:38 -0500 Subject: [PATCH 083/108] Update docs. - Review unsupported comments and refresh. - Review experimental tag in API. --- docs/markdown/hip_faq.md | 32 +++++++++++------------- docs/markdown/hip_kernel_language.md | 4 +-- include/hip/hcc_detail/hip_runtime_api.h | 26 ++++++++----------- 3 files changed, 27 insertions(+), 35 deletions(-) diff --git a/docs/markdown/hip_faq.md b/docs/markdown/hip_faq.md index b92ae512c5..c311aae320 100644 --- a/docs/markdown/hip_faq.md +++ b/docs/markdown/hip_faq.md @@ -48,12 +48,11 @@ The HIP API documentation describes each API and its limitations, if any, compar ### What is not supported? #### Runtime/Driver API features -At a high-level, the following features are not supported: -- Textures +)t a high-level, the following features are not supported: +- Textures (partial support available) - Dynamic parallelism (CUDA 5.0) - Managed memory (CUDA 6.5) - Graphics interoperability with OpenGL or Direct3D -- CUDA Driver API - CUDA IPC Functions (Under Development) - CUDA array, mipmappedArray and pitched memory - Queue priority controls @@ -61,17 +60,12 @@ At a high-level, the following features are not supported: See the [API Support Table](CUDA_Runtime_API_functions_supported_by_HIP.md) for more detailed information. #### Kernel language features -- Device-side dynamic memory allocations (malloc, free, new, delete) (CUDA 4.0) +- C++-style device-side dynamic memory allocations (free, new, delete) (CUDA 4.0) - Virtual functions, indirect functions and try/catch (CUDA 4.0) - `__prof_trigger` - PTX assembly (CUDA 4.0). HCC supports inline GCN assembly. - Several kernel features are under development. See the [HIP Kernel Language](hip_kernel_language.md) for more information. These include: - printf - - assert - - `__restrict__` - - `__threadfence*_`, `__syncthreads*` - - Unbounded loop unroll - ### Is HIP a drop-in replacement for CUDA? @@ -100,18 +94,20 @@ However, we can provide a rough summary of the features included in each CUDA SD - Per-thread-streams (under development) - C++11 (HCC supports all of C++11, all of C++14 and some C++17 features) - CUDA 7.5 - - float16 + - float16 (supported) - CUDA 8.0 - - TBD. + - Page Migration including cudaMemAdvise, cudaMemPrefetch, other cudaMem* APIs(not supported) + ### What libraries does HIP support? -HIP includes growing support for the 4 key math libraries using hcBlas, hcFft, hcrng and hcsparse. -These offer pointer-based memory interfaces (as opposed to opaque buffers) and can be easily interfaced with other HCC applications. Developers should use conditional compilation if portability to nvcc systems is desired - using calls to cu* routines on one path and hc* routines on the other. +HIP includes growing support for the 4 key math libraries using hcBlas, hcFft, hcrng and hcsparse, as well as MIOpen for machine intelligence applications. +These offer pointer-based memory interfaces (as opposed to opaque buffers) and can be easily interfaced with other HIP applications. +The hip interfaces support both ROCm and CUDA paths, with familiar library interfaces. -- [hcblas](https://bitbucket.org/multicoreware/hcblas) -- [hcfft](https://bitbucket.org/multicoreware/hcfft) -- [hcsparse](https://bitbucket.org/multicoreware/hcsparse) -- [hcrng](https://bitbucket.org/multicoreware/hcrng) +- [hipBlas](https://github.com/ROCmSoftwarePlatform/hipBLAS), which utilizes [rocBlas](https://github.com/ROCmSoftwarePlatform/rocBLAS). +- [hipfft](https://github.com/ROCmSoftwarePlatform/hcFFT) +- [hipsparse](https://github.com/ROCmSoftwarePlatform/hcSPARSE) +- [hiprng](https://github.com/ROCmSoftwarePlatform/hcrng) Additionally, some of the cublas routines are automatically converted to hipblas equivalents by the hipify-clang tool. These APIs use cublas or hcblas depending on the platform, and replace the need to use conditional compilation. @@ -218,7 +214,7 @@ If platform portability is important, use #ifdef __HIP_PLATFORM_NVCC__ to guard ### On HCC, can I use HC functionality with HIP? Yes. The code can include hc.hpp and use HC functions inside the kernel. A typical use-case is to use AMD-specific hardware features such as the permute, swizzle, or DPP operations. -The "-stdlib=libc++" must be passed to hipcc in order to compile hc.hpp. See the 'bit_extract' sample for an example. +See the 'bit_extract' sample for an example. Also these functions can be used to extract HCC accelerator and accelerator_view structures from the HIP deviceId and hipStream_t: hipHccGetAccelerator(int deviceId, hc::accelerator *acc); diff --git a/docs/markdown/hip_kernel_language.md b/docs/markdown/hip_kernel_language.md index cfa5d0f871..a4f4d5d47f 100644 --- a/docs/markdown/hip_kernel_language.md +++ b/docs/markdown/hip_kernel_language.md @@ -167,7 +167,7 @@ The `__shared__` keyword is supported. Managed memory, including the `__managed__` keyword, are not supported in HIP. ### `__restrict__` -The `__restrict__` keyword tells the compiler that the associated memory pointer will not alias with any other pointer in the kernel or function. This feature can help the compiler generate better code. In most cases, all pointer arguments must use this keyword to realize the benefit. hcc support for the `__restrict__` qualifier on kernel arguments is under development. +The `__restrict__` keyword tells the compiler that the associated memory pointer will not alias with any other pointer in the kernel or function. This feature can help the compiler generate better code. In most cases, all pointer arguments must use this keyword to realize the benefit. ## Built-In Variables @@ -603,6 +603,7 @@ The Cuda `__prof_trigger()` instruction is not supported. ## Assert The assert function is under development. +HIP does support an "abort" call which will terminate the process execution from inside the kernel. ## Printf @@ -690,7 +691,6 @@ for (int i=0; i<16; i++) ... ``` -Unbounded loop unroll is under development on HCC compiler. ``` #pragma unroll /* hint to compiler to completely unroll next loop. */ for (int i=0; i<16; i++) ... diff --git a/include/hip/hcc_detail/hip_runtime_api.h b/include/hip/hcc_detail/hip_runtime_api.h index e28dbab82c..03be587b0d 100644 --- a/include/hip/hcc_detail/hip_runtime_api.h +++ b/include/hip/hcc_detail/hip_runtime_api.h @@ -171,7 +171,7 @@ typedef enum hipJitOption { /** - * @warning On AMD devices and recent Nvidia devices, these hints and controls are ignored. + * @warning On AMD devices and some Nvidia devices, these hints and controls are ignored. */ typedef enum hipFuncCache_t { hipFuncCachePreferNone, ///< no preference for shared memory or L1 (default) @@ -182,7 +182,7 @@ typedef enum hipFuncCache_t { /** - * @warning On AMD devices and recent Nvidia devices, these hints and controls are ignored. + * @warning On AMD devices and some Nvidia devices, these hints and controls are ignored. */ typedef enum hipSharedMemConfig { hipSharedMemBankSizeDefault, ///< The compiler selects a device-specific value for the banking. @@ -364,7 +364,7 @@ hipError_t hipGetDeviceProperties(hipDeviceProp_t* prop, int deviceId); * @param [in] cacheConfig * * @returns #hipSuccess, #hipErrorInitializationError - * Note: AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. + * Note: AMD devices and some Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. * */ hipError_t hipDeviceSetCacheConfig ( hipFuncCache_t cacheConfig ); @@ -376,7 +376,7 @@ hipError_t hipDeviceSetCacheConfig ( hipFuncCache_t cacheConfig ); * @param [in] cacheConfig * * @returns #hipSuccess, #hipErrorInitializationError - * Note: AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. + * Note: AMD devices and some Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. * */ hipError_t hipDeviceGetCacheConfig ( hipFuncCache_t *cacheConfig ); @@ -400,7 +400,7 @@ hipError_t hipDeviceGetLimit(size_t *pValue, enum hipLimit_t limit); * @param [in] config; * * @returns #hipSuccess, #hipErrorInitializationError - * Note: AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. + * Note: AMD devices and some Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. * */ hipError_t hipFuncSetCacheConfig (const void* func, hipFuncCache_t config ); @@ -412,7 +412,7 @@ hipError_t hipFuncSetCacheConfig (const void* func, hipFuncCache_t config ); * * @returns #hipSuccess, #hipErrorInvalidValue, #hipErrorInitializationError * - * Note: AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. + * Note: AMD devices and some Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. * */ hipError_t hipDeviceGetSharedMemConfig ( hipSharedMemConfig * pConfig ); @@ -425,7 +425,7 @@ hipError_t hipDeviceGetSharedMemConfig ( hipSharedMemConfig * pConfig ); * * @returns #hipSuccess, #hipErrorInvalidValue, #hipErrorInitializationError * - * Note: AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. + * Note: AMD devices and some Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. * */ hipError_t hipDeviceSetSharedMemConfig ( hipSharedMemConfig config ); @@ -1449,7 +1449,6 @@ hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p); * * @returns #hipSuccess, * @returns #hipErrorInvalidDevice if deviceId or peerDeviceId are not valid devices - * @warning PeerToPeer support is experimental. */ hipError_t hipDeviceCanAccessPeer (int* canAccessPeer, int deviceId, int peerDeviceId); @@ -1467,7 +1466,6 @@ hipError_t hipDeviceCanAccessPeer (int* canAccessPeer, int deviceId, int peerDev * * Returns #hipSuccess, #hipErrorInvalidDevice, #hipErrorInvalidValue, * @returns #hipErrorPeerAccessAlreadyEnabled if peer access is already enabled for this device. - * @warning PeerToPeer support is experimental. */ hipError_t hipDeviceEnablePeerAccess (int peerDeviceId, unsigned int flags); @@ -1480,7 +1478,6 @@ hipError_t hipDeviceEnablePeerAccess (int peerDeviceId, unsigned int flags); * @param [in] peerDeviceId * * @returns #hipSuccess, #hipErrorPeerAccessNotEnabled - * @warning PeerToPeer support is experimental. */ hipError_t hipDeviceDisablePeerAccess (int peerDeviceId); @@ -1512,7 +1509,6 @@ hipError_t hipMemGetAddressRange ( hipDeviceptr_t* pbase, size_t* psize, hipDevi * @param [in] sizeBytes - Size of memory copy in bytes * * @returns #hipSuccess, #hipErrorInvalidValue, #hipErrorInvalidDevice - * @warning PeerToPeer support is experimental. */ hipError_t hipMemcpyPeer (void* dst, int dstDeviceId, const void* src, int srcDeviceId, size_t sizeBytes); @@ -1671,7 +1667,7 @@ hipError_t hipCtxGetApiVersion (hipCtx_t ctx,int *apiVersion); * * @return #hipSuccess * - * @warning AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. + * @warning AMD devices and some Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. * * @see hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice */ @@ -1684,7 +1680,7 @@ hipError_t hipCtxGetCacheConfig ( hipFuncCache_t *cacheConfig ); * * @return #hipSuccess * - * @warning AMD devices and recent Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. + * @warning AMD devices and some Nvidia GPUS do not support reconfigurable cache. This hint is ignored on those architectures. * * @see hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice */ @@ -1697,7 +1693,7 @@ hipError_t hipCtxSetCacheConfig ( hipFuncCache_t cacheConfig ); * * @return #hipSuccess * - * @warning AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. + * @warning AMD devices and some Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. * * @see hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice */ @@ -1710,7 +1706,7 @@ hipError_t hipCtxSetSharedMemConfig ( hipSharedMemConfig config ); * * @return #hipSuccess * - * @warning AMD devices and recent Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. + * @warning AMD devices and some Nvidia GPUS do not support shared cache banking, and the hint is ignored on those architectures. * * @see hipCtxCreate, hipCtxDestroy, hipCtxGetFlags, hipCtxPopCurrent, hipCtxGetCurrent, hipCtxSetCurrent, hipCtxPushCurrent, hipCtxSetCacheConfig, hipCtxSynchronize, hipCtxGetDevice */ From 6ff74d0e977a403a7469e5c023c67a53b9c7878d Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Sat, 26 Aug 2017 14:39:14 +0000 Subject: [PATCH 084/108] Lock streams when waiting on event completion or querying event safety. --- src/hip_event.cpp | 23 ++++++++++++++++------- src/hip_hcc.cpp | 26 +++++++++++++++++++++++--- src/hip_hcc_internal.h | 15 ++++++++++++--- src/hip_stream.cpp | 2 +- 4 files changed, 52 insertions(+), 14 deletions(-) diff --git a/src/hip_event.cpp b/src/hip_event.cpp index 3a8f1ab611..d1ee37a45e 100644 --- a/src/hip_event.cpp +++ b/src/hip_event.cpp @@ -55,13 +55,13 @@ void ihipEvent_t::attachToCompletionFuture(const hc::completion_future *cf, void ihipEvent_t::refereshEventStatus() { - bool isReady0 = _marker.is_ready(); + bool isReady0 = locked_isReady(); bool isReady1; int val = 0; if (_state == hipEventStatusRecording) { // TODO - use completion-future functions to obtain ticks and timestamps: hsa_signal_t *sig = static_cast (_marker.get_native_handle()); - isReady1 = _marker.is_ready(); + isReady1 = locked_isReady(); if (sig) { val = hsa_signal_load_acquire(*sig); if (val == 0) { @@ -86,6 +86,17 @@ void ihipEvent_t::refereshEventStatus() } +bool ihipEvent_t::locked_isReady() +{ + return _stream->locked_eventIsReady(this); +} + +void ihipEvent_t::locked_waitComplete(hc::hcWaitMode waitMode) +{ + return _stream->locked_eventWaitComplete(this, waitMode); +} + + hipError_t ihipEventCreate(hipEvent_t* event, unsigned flags) { hipError_t e = hipSuccess; @@ -127,7 +138,7 @@ hipError_t hipEventCreate(hipEvent_t* event) hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream) { - HIP_INIT_SPECIAL_API(TRACE_QUERY, event, stream); + HIP_INIT_SPECIAL_API(TRACE_SYNC, event, stream); if (event && event->_state != hipEventStatusUnitialized) { stream = ihipSyncAndResolveStream(stream); @@ -192,9 +203,7 @@ hipError_t hipEventSynchronize(hipEvent_t event) ctx->locked_syncDefaultStream(true, true); return ihipLogStatus(hipSuccess); } else { - event->_marker.wait((event->_flags & hipEventBlockingSync) ? hc::hcWaitModeBlocked : hc::hcWaitModeActive); - - assert (event->_marker.is_ready()); + event->locked_waitComplete((event->_flags & hipEventBlockingSync) ? hc::hcWaitModeBlocked : hc::hcWaitModeActive); return ihipLogStatus(hipSuccess); } @@ -259,7 +268,7 @@ hipError_t hipEventQuery(hipEvent_t event) { HIP_INIT_SPECIAL_API(TRACE_QUERY, event); - if ((event->_state == hipEventStatusRecording) && (!event->_marker.is_ready())) { + if ((event->_state == hipEventStatusRecording) && !event->locked_isReady()) { return ihipLogStatus(hipErrorNotReady); } else { return ihipLogStatus(hipSuccess); diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index f7082e7e9e..dc72714e3e 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -328,14 +328,34 @@ void ihipStream_t::locked_wait() // Causes current stream to wait for specified event to complete: // Note this does not provide any kind of host serialization. -void ihipStream_t::locked_waitEvent(hipEvent_t event) +void ihipStream_t::locked_streamWaitEvent(hipEvent_t event) { LockedAccessor_StreamCrit_t crit(_criticalData); - crit->_av.create_blocking_marker(event->_marker, hc::accelerator_scope); + crit->_av.create_blocking_marker(event->marker(), hc::accelerator_scope); } + +// Causes current stream to wait for specified event to complete: +// Note this does not provide any kind of host serialization. +bool ihipStream_t::locked_eventIsReady(hipEvent_t event) +{ + // Event query that returns "Complete" may cause HCC to manipulate + // internal queue state so lock the stream's queue here. + LockedAccessor_StreamCrit_t crit(_criticalData); + + return (event->marker().is_ready()); +} + +void ihipStream_t::locked_eventWaitComplete(hipEvent_t event, hc::hcWaitMode waitMode) +{ + LockedAccessor_StreamCrit_t crit(_criticalData); + + event->marker().wait(waitMode); +} + + // Create a marker in this stream. // Save state in the event so it can track the status of the event. void ihipStream_t::locked_recordEvent(hipEvent_t event) @@ -354,7 +374,7 @@ void ihipStream_t::locked_recordEvent(hipEvent_t event) scopeFlag = HIP_EVENT_SYS_RELEASE ? hc::system_scope : hc::accelerator_scope; } - event->_marker = crit->_av.create_marker(scopeFlag); + event->marker(crit->_av.create_marker(scopeFlag)); }; //============================================================================= diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index af5e7a121b..56ca37b3e2 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -517,9 +517,12 @@ public: hc::accelerator_view* locked_getAv() { LockedAccessor_StreamCrit_t crit(_criticalData); return &(crit->_av); }; - void locked_waitEvent(hipEvent_t event); + void locked_streamWaitEvent(hipEvent_t event); void locked_recordEvent(hipEvent_t event); + bool locked_eventIsReady(hipEvent_t event); + void locked_eventWaitComplete(hipEvent_t event, hc::hcWaitMode waitMode); + ihipStreamCritical_t &criticalData() { return _criticalData; }; //--- @@ -608,18 +611,24 @@ public: ihipEvent_t(unsigned flags); void attachToCompletionFuture(const hc::completion_future *cf, hipStream_t stream, ihipEventType_t eventType); void refereshEventStatus(); + hc::completion_future & marker() { return _marker; } + void marker(hc::completion_future cf) { _marker = cf; }; + + bool locked_isReady(); + void locked_waitComplete(hc::hcWaitMode waitMode); + uint64_t timestamp() const { return _timestamp; } ; ihipEventType_t type() const { return _type; }; public: hipEventStatus_t _state; - hipStream_t _stream; // Stream where the event is recorded, or NULL if all streams. + hipStream_t _stream; // Stream where the event is recorded. Null stream is resolved to actual stream when recorded unsigned _flags; - hc::completion_future _marker; private: + hc::completion_future _marker; ihipEventType_t _type; uint64_t _timestamp; // store timestamp, may be set on host or by marker. friend hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream); diff --git a/src/hip_stream.cpp b/src/hip_stream.cpp index f0e9283201..51aeb01412 100644 --- a/src/hip_stream.cpp +++ b/src/hip_stream.cpp @@ -96,7 +96,7 @@ hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int if (stream != hipStreamNull) { // This will user create_blocking_marker to wait on the specified queue. - stream->locked_waitEvent(event); + stream->locked_streamWaitEvent(event); } else { // TODO-hcc Convert to use create_blocking_marker(...) functionality. From 882dab4536736945e574344674bc83f890116ae4 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Wed, 30 Aug 2017 19:56:47 +0000 Subject: [PATCH 085/108] Refactor hipStreamWaitEvent - Null streams use same flow as non-null. - Add HIP_SYNC_STREAM_WAIT - Resolve null stream. --- src/hip_hcc.cpp | 3 ++ src/hip_hcc_internal.h | 1 + src/hip_stream.cpp | 15 +++---- .../runtimeApi/stream/hipStreamWaitEvent.cpp | 40 +++++++++++++++++-- 4 files changed, 46 insertions(+), 13 deletions(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index dc72714e3e..7c9f0f966f 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -88,6 +88,8 @@ int HIP_HOST_COHERENT = 1; int HIP_SYNC_HOST_ALLOC = 1; +int HIP_SYNC_STREAM_WAIT = 0; + #if (__hcc_workweek__ >= 17300) // Make sure we have required bug fix in HCC @@ -1249,6 +1251,7 @@ void HipReadEnv() READ_ENV_I(release, HIP_SYNC_HOST_ALLOC, 0, "Sync before and after all host memory allocations. May help stability"); READ_ENV_I(release, HIP_SYNC_NULL_STREAM, 0, "Synchronize on host for null stream submissions"); + READ_ENV_I(release, HIP_SYNC_STREAM_WAIT, 0, "hipStreamWaitEvent will synchronize to host"); READ_ENV_I(release, HIP_HOST_COHERENT, 0, "If set, all host memory will be allocated as fine-grained system memory. This allows threadfence_system to work but prevents host memory from being cached on GPU which may have performance impact."); diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index 56ca37b3e2..af855861c6 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -65,6 +65,7 @@ extern int HIP_HIDDEN_FREE_MEM; //--- // Chicken bits for disabling functionality to work around potential issues: extern int HIP_SYNC_HOST_ALLOC; +extern int HIP_SYNC_STREAM_WAIT; extern int HIP_SYNC_NULL_STREAM; diff --git a/src/hip_stream.cpp b/src/hip_stream.cpp index 51aeb01412..0b645ec658 100644 --- a/src/hip_stream.cpp +++ b/src/hip_stream.cpp @@ -93,18 +93,15 @@ hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int } else if (event->_state != hipEventStatusUnitialized) { - if (stream != hipStreamNull) { - + if (HIP_SYNC_STREAM_WAIT || (HIP_SYNC_NULL_STREAM && (stream == 0))) { + // conservative wait on host for the specified event to complete: + event->locked_waitComplete((event->_flags & hipEventBlockingSync) ? hc::hcWaitModeBlocked : hc::hcWaitModeActive); + } else { + stream = ihipSyncAndResolveStream(stream); // This will user create_blocking_marker to wait on the specified queue. stream->locked_streamWaitEvent(event); - - } else { - // TODO-hcc Convert to use create_blocking_marker(...) functionality. - // Currently we have a super-conservative version of this - block on host, and drain the queue. - // This should create a barrier packet in the target queue. - // TODO-HIP_SYNC_NULL_STREAM - stream->locked_wait(); } + } // else event not recorded, return immediately and don't create marker. return ihipLogStatus(e); diff --git a/tests/src/runtimeApi/stream/hipStreamWaitEvent.cpp b/tests/src/runtimeApi/stream/hipStreamWaitEvent.cpp index 9bbd43828c..f5b1b79550 100644 --- a/tests/src/runtimeApi/stream/hipStreamWaitEvent.cpp +++ b/tests/src/runtimeApi/stream/hipStreamWaitEvent.cpp @@ -446,9 +446,41 @@ int main(int argc, char *argv[]) if (p_tests & 0x1000) { - printf ("==> Test 0x1000 try null stream\n"); - hipStreamQuery(0/* try null stream*/); + printf ("==> Test 0x1000 simple null stream tests\n"); + // try some null stream: + hipStreamQuery(0); + + + hipStream_t s1; + hipEvent_t e1; + + { + // stream null waits on event in s1 stream: + HIPCHECK(hipStreamCreate(&s1)); + HIPCHECK(hipEventCreate(&e1)); + + HIPCHECK(hipEventRecord(e1, s1)) + + HIPCHECK(hipStreamWaitEvent(hipStream_t(0), e1, 0/*flags*/)); + + HIPCHECK(hipStreamDestroy(s1)); + HIPCHECK(hipEventDestroy(e1)); + } + + { + // stream s1 waits on event in null stream: + HIPCHECK(hipStreamCreate(&s1)); + HIPCHECK(hipEventCreate(&e1)); + + HIPCHECK(hipEventRecord(e1, hipStream_t(0))) + + HIPCHECK(hipStreamWaitEvent(s1, e1, 0/*flags*/)); + + HIPCHECK(hipStreamDestroy(s1)); + HIPCHECK(hipEventDestroy(e1)); + } + } @@ -471,8 +503,8 @@ int main(int argc, char *argv[]) } - { - printf ("test: alternating memcpy/count-reverse followed by event\n"); + if (p_tests & 0x4000 ) { + printf ("test: %x alternating memcpy/count-reverse followed by event\n", p_tests); RUN_SYNC_TEST(0x4000, streamersDev0, sync_queryAllUntilComplete(streamersDev0), true); RUN_SYNC_TEST(0x8000, streamersDev0, sync_streamWaitEvent(streamersDev0.back()->event(), 0, sideStreams[0], false), true); } From 4ac6d643c12cf0da8a07726c52d2d0eb72ecd781 Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Thu, 31 Aug 2017 02:50:30 +0000 Subject: [PATCH 086/108] hipStreamQuery uses av::is_empty. Add HIP_FORCE_NULL_STREAM. --- src/hip_hcc.cpp | 4 ++++ src/hip_hcc_internal.h | 1 + src/hip_stream.cpp | 38 ++++++++++++++++++++++---------------- 3 files changed, 27 insertions(+), 16 deletions(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 7c9f0f966f..860c6f87a8 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -90,6 +90,8 @@ int HIP_SYNC_HOST_ALLOC = 1; int HIP_SYNC_STREAM_WAIT = 0; +int HIP_FORCE_NULL_STREAM=0; + #if (__hcc_workweek__ >= 17300) // Make sure we have required bug fix in HCC @@ -1251,6 +1253,8 @@ void HipReadEnv() READ_ENV_I(release, HIP_SYNC_HOST_ALLOC, 0, "Sync before and after all host memory allocations. May help stability"); READ_ENV_I(release, HIP_SYNC_NULL_STREAM, 0, "Synchronize on host for null stream submissions"); + READ_ENV_I(release, HIP_FORCE_NULL_STREAM, 0, "Force all stream allocations to secretly return the null stream"); + READ_ENV_I(release, HIP_SYNC_STREAM_WAIT, 0, "hipStreamWaitEvent will synchronize to host"); diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index af855861c6..88717e48bf 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -68,6 +68,7 @@ extern int HIP_SYNC_HOST_ALLOC; extern int HIP_SYNC_STREAM_WAIT; extern int HIP_SYNC_NULL_STREAM; +extern int HIP_FORCE_NULL_STREAM; // TODO - remove when this is standard behavior. extern int HCC_OPT_FLUSH; diff --git a/src/hip_stream.cpp b/src/hip_stream.cpp index 0b645ec658..7dd6efd39c 100644 --- a/src/hip_stream.cpp +++ b/src/hip_stream.cpp @@ -38,21 +38,26 @@ hipError_t ihipStreamCreate(hipStream_t *stream, unsigned int flags) hipError_t e = hipSuccess; if (ctx) { - hc::accelerator acc = ctx->getWriteableDevice()->_acc; - // TODO - se try-catch loop to detect memory exception? - // - //Note this is an execute_in_order queue, so all kernels submitted will atuomatically wait for prev to complete: - //This matches CUDA stream behavior: + if (HIP_FORCE_NULL_STREAM) { + *stream = 0; + } else { + hc::accelerator acc = ctx->getWriteableDevice()->_acc; - { - // Obtain mutex access to the device critical data, release by destructor - LockedAccessor_CtxCrit_t ctxCrit(ctx->criticalData()); + // TODO - se try-catch loop to detect memory exception? + // + //Note this is an execute_in_order queue, so all kernels submitted will atuomatically wait for prev to complete: + //This matches CUDA stream behavior: - auto istream = new ihipStream_t(ctx, acc.create_view(), flags); + { + // Obtain mutex access to the device critical data, release by destructor + LockedAccessor_CtxCrit_t ctxCrit(ctx->criticalData()); - ctxCrit->addStream(istream); - *stream = istream; + auto istream = new ihipStream_t(ctx, acc.create_view(), flags); + + ctxCrit->addStream(istream); + *stream = istream; + } } tprintf(DB_SYNC, "hipStreamCreate, %s\n", ToString(*stream).c_str()); @@ -119,15 +124,14 @@ hipError_t hipStreamQuery(hipStream_t stream) stream = device->_defaultStream; } - int pendingOps = 0; + bool isEmpty = 0; { LockedAccessor_StreamCrit_t crit(stream->_criticalData); - pendingOps = crit->_av.get_pending_async_ops(); + isEmpty = crit->_av.get_is_empty(); } - - hipError_t e = (pendingOps > 0) ? hipErrorNotReady : hipSuccess; + hipError_t e = isEmpty ? hipSuccess : hipErrorNotReady ; return ihipLogStatus(e); } @@ -167,7 +171,9 @@ hipError_t hipStreamDestroy(hipStream_t stream) //--- Drain the stream: if (stream == NULL) { - e = hipErrorInvalidResourceHandle; // TODO - review - what happens if try to destroy null stream + if (!HIP_FORCE_NULL_STREAM) { + e = hipErrorInvalidResourceHandle; + } } else { stream->locked_wait(); From 40ff42e555e5c6c3eb7f54b7db8237f267b3fced Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Thu, 7 Sep 2017 20:45:27 +0530 Subject: [PATCH 087/108] Device side assert support --- include/hip/hcc_detail/hip_runtime.h | 2 +- include/hip/nvcc_detail/hip_runtime.h | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/include/hip/hcc_detail/hip_runtime.h b/include/hip/hcc_detail/hip_runtime.h index 1092e22c47..2386ea08cb 100644 --- a/include/hip/hcc_detail/hip_runtime.h +++ b/include/hip/hcc_detail/hip_runtime.h @@ -99,7 +99,7 @@ extern int HIP_TRACE_API; // TODO-HCC add a dummy implementation of assert, need to replace with a proper kernel exit call. #if __HIP_DEVICE_COMPILE__ == 1 #undef assert - #define assert(COND) { if (COND) {} } + #define assert(COND) { if (!COND) {abort();} } #endif diff --git a/include/hip/nvcc_detail/hip_runtime.h b/include/hip/nvcc_detail/hip_runtime.h index 8c08f3d151..fca9ab6e39 100644 --- a/include/hip/nvcc_detail/hip_runtime.h +++ b/include/hip/nvcc_detail/hip_runtime.h @@ -111,6 +111,8 @@ kernelName<<>>(__VA_ARGS__);\ #ifdef __HIP_DEVICE_COMPILE__ #define abort() {asm("trap;");} +#undef assert +#define assert(COND) { if (!COND) {abort();} } #endif #endif From edd0b360b24f56e1d7b2ebfda48a0b95ab0a176f Mon Sep 17 00:00:00 2001 From: Aaron Enye Shi Date: Fri, 8 Sep 2017 11:46:10 -0400 Subject: [PATCH 088/108] [CentOS7] Fix HIP sample hipInfo and other tests Cent OS 7 has conflicts between its libc++ and libstdc++. Seems that we cannot use libstdc++ on Cent OS. This is related to SWDEV-131972 [ROCm CQE][Cent OS 7][G] Building any HIP sample giving an fatal error: 'bits/c++config.h'. --- bin/hipcc | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/bin/hipcc b/bin/hipcc index 624777ab25..5869eeda0d 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -106,6 +106,7 @@ if ($HIP_PLATFORM eq "hcc") { #### GCC system includes workaround #### $HCC_WA_FLAGS = " "; + $HOST_OSNAME= `cat /etc/os-release | grep "^ID\=" | cut -d= -f2 | tr -d '\n'`; if ($HCC_VERSION_MAJOR eq 1) { my $GCC_CUR_VER = `gcc -dumpversion`; my $GPP_CUR_VER = `g++ -dumpversion`; @@ -116,7 +117,8 @@ if ($HIP_PLATFORM eq "hcc") { # Only include the libstdc++ headers and libraries flags explicitly if the g++ is older than version 5. # That's because HCC already uses libstdc++ by default if a newer g++/libstdc++ is available - if (${GCC_CUR_VER} eq ${GPP_CUR_VER} and $GPP_VER_FIELDS[0] < 5) { + # Cent OS 7 cannot use libstdc++ for compilation, defaults to libc++ + if (${GCC_CUR_VER} eq ${GPP_CUR_VER} and $GPP_VER_FIELDS[0] < 5 and ($HOST_OSNAME ne "\"centos\"")) { $HCC_WA_FLAGS .= " -stdlib=libstdc++ -I/usr/include/x86_64-linux-gnu -I/usr/include/x86_64-linux-gnu/c++/${GCC_CUR_VER} -I/usr/include/c++/${GCC_CUR_VER} "; # Add C++ libs for GCC. $HIPLDFLAGS .= " -lstdc++"; @@ -124,7 +126,6 @@ if ($HIP_PLATFORM eq "hcc") { } # Force -stdlib=libc++ on UB14.04 - $HOST_OSNAME= `cat /etc/os-release | grep "^ID\=" | cut -d= -f2 | tr -d '\n'`; $HOST_OSVER= `cat /etc/os-release | grep "^VERSION_ID\=" | cut -d= -f2 | tr -d '\n'`; if ($HOST_OSNAME eq "ubuntu" and $HOST_OSVER eq "\"14.04\"") { $HIPCXXFLAGS .= " -stdlib=libc++"; From 2a915b279068ba59c944d510886d5335047b0bbc Mon Sep 17 00:00:00 2001 From: Rahul Garg Date: Tue, 12 Sep 2017 21:52:11 +0530 Subject: [PATCH 089/108] Fixed hipMallocArray for 1D cases --- src/hip_memory.cpp | 132 +++++++++++++++++++++++---------------------- 1 file changed, 69 insertions(+), 63 deletions(-) diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 4b3a4fcb12..6149271b8d 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -407,83 +407,89 @@ hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, HIP_INIT_SPECIAL_API((TRACE_MEM), array, desc, width, height, flags); HIP_SET_DEVICE(); hipError_t hip_status = hipSuccess; + if(width > 0) { + auto ctx = ihipGetTlsDefaultCtx(); - auto ctx = ihipGetTlsDefaultCtx(); + *array = (hipArray*)malloc(sizeof(hipArray)); + array[0]->type = flags; + array[0]->width = width; + array[0]->height = height; + array[0]->depth = 1; + array[0]->desc = *desc; - *array = (hipArray*)malloc(sizeof(hipArray)); - array[0]->type = flags; - array[0]->width = width; - array[0]->height = height; - array[0]->depth = 1; - array[0]->desc = *desc; + void ** ptr = &array[0]->data; - void ** ptr = &array[0]->data; + if (ctx) { + const unsigned am_flags = 0; + size_t size = width; + if(height > 0) { + size = size * height; + } - if (ctx) { - const unsigned am_flags = 0; - const size_t size = width*height; + size_t allocSize = 0; + switch(desc->f) { + case hipChannelFormatKindSigned: + allocSize = size * sizeof(int); + break; + case hipChannelFormatKindUnsigned: + allocSize = size * sizeof(unsigned int); + break; + case hipChannelFormatKindFloat: + allocSize = size * sizeof(float); + break; + case hipChannelFormatKindNone: + allocSize = size * sizeof(size_t); + break; + default: + hip_status = hipErrorUnknown; + break; + } + hc::accelerator acc = ctx->getDevice()->_acc; + hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); - size_t allocSize = 0; - switch(desc->f) { - case hipChannelFormatKindSigned: - allocSize = size * sizeof(int); - break; - case hipChannelFormatKindUnsigned: - allocSize = size * sizeof(unsigned int); - break; - case hipChannelFormatKindFloat: - allocSize = size * sizeof(float); - break; - case hipChannelFormatKindNone: - allocSize = size * sizeof(size_t); + size_t allocGranularity = 0; + hsa_amd_memory_pool_t *allocRegion = static_cast(acc.get_hsa_am_region()); + hsa_amd_memory_pool_get_info(*allocRegion, HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_GRANULE, &allocGranularity); + + hsa_ext_image_descriptor_t imageDescriptor; + + imageDescriptor.width = width; + imageDescriptor.height = height; + imageDescriptor.depth = 0; + imageDescriptor.array_size = 0; + switch (flags) { + case hipArrayLayered: + case hipArrayCubemap: + case hipArraySurfaceLoadStore: + case hipArrayTextureGather: + assert(0); break; + case hipArrayDefault: default: - hip_status = hipErrorUnknown; + imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; break; - } - hc::accelerator acc = ctx->getDevice()->_acc; - hsa_agent_t* agent =static_cast(acc.get_hsa_agent()); + } + hsa_ext_image_channel_order_t channelOrder; + hsa_ext_image_channel_type_t channelType; + getChannelOrderAndType(*desc, hipReadModeElementType, channelOrder, channelType); + imageDescriptor.format.channel_order = channelOrder; + imageDescriptor.format.channel_type = channelType; - size_t allocGranularity = 0; - hsa_amd_memory_pool_t *allocRegion = static_cast(acc.get_hsa_am_region()); - hsa_amd_memory_pool_get_info(*allocRegion, HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_GRANULE, &allocGranularity); + hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; + hsa_ext_image_data_info_t imageInfo; + hsa_status_t status = hsa_ext_image_data_get_info(*agent, &imageDescriptor, permission, &imageInfo); + size_t alignment = imageInfo.alignment <= allocGranularity ? 0 : imageInfo.alignment; - hsa_ext_image_descriptor_t imageDescriptor; + *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false/*shareWithAll*/, am_flags, 0, alignment); + if (size && (*ptr == NULL)) { + hip_status = hipErrorMemoryAllocation; + } - imageDescriptor.width = width; - imageDescriptor.height = height; - imageDescriptor.depth = 0; - imageDescriptor.array_size = 0; - switch (flags) { - case hipArrayLayered: - case hipArrayCubemap: - case hipArraySurfaceLoadStore: - case hipArrayTextureGather: - assert(0); - break; - case hipArrayDefault: - default: - imageDescriptor.geometry = HSA_EXT_IMAGE_GEOMETRY_2D; - break; - } - hsa_ext_image_channel_order_t channelOrder; - hsa_ext_image_channel_type_t channelType; - getChannelOrderAndType(*desc, hipReadModeElementType, channelOrder, channelType); - imageDescriptor.format.channel_order = channelOrder; - imageDescriptor.format.channel_type = channelType; - - hsa_access_permission_t permission = HSA_ACCESS_PERMISSION_RW; - hsa_ext_image_data_info_t imageInfo; - hsa_status_t status = hsa_ext_image_data_get_info(*agent, &imageDescriptor, permission, &imageInfo); - size_t alignment = imageInfo.alignment <= allocGranularity ? 0 : imageInfo.alignment; - - *ptr = hip_internal::allocAndSharePtr("device_array", allocSize, ctx, false/*shareWithAll*/, am_flags, 0, alignment); - if (size && (*ptr == NULL)) { + } else { hip_status = hipErrorMemoryAllocation; } - } else { - hip_status = hipErrorMemoryAllocation; + hip_status = hipErrorInvalidValue; } return ihipLogStatus(hip_status); From c7746b12acc0f4e57415f99220daed9d835fe2fe Mon Sep 17 00:00:00 2001 From: Aaron Enye Shi Date: Tue, 12 Sep 2017 16:59:43 -0400 Subject: [PATCH 090/108] Must explicitly add libc++ and libc++abi --- bin/hipcc | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/bin/hipcc b/bin/hipcc index 5869eeda0d..d9aee09693 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -131,6 +131,11 @@ if ($HIP_PLATFORM eq "hcc") { $HIPCXXFLAGS .= " -stdlib=libc++"; $setStdLib = 1; } + if ($HOST_OSNAME eq "\"centos\"" and $HOST_OSVER eq "\"7\"") { + $HIPCXXFLAGS .= " -stdlib=libc++"; + $HIPLDFLAGS .= " -lc++ -lc++abi"; + $setStdLib = 1; + } $HIPCXXFLAGS .= " -I$HIP_PATH/include/hip/hcc_detail/cuda"; $HIPCXXFLAGS .= " -I$HSA_PATH/include"; From b340f4a6ed5232ebbd03b0b4c43d3f4729f1e8a0 Mon Sep 17 00:00:00 2001 From: Aaron Enye Shi Date: Tue, 12 Sep 2017 17:50:44 -0400 Subject: [PATCH 091/108] Update hipcc --- bin/hipcc | 1 + 1 file changed, 1 insertion(+) diff --git a/bin/hipcc b/bin/hipcc index d9aee09693..30ed8145bd 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -131,6 +131,7 @@ if ($HIP_PLATFORM eq "hcc") { $HIPCXXFLAGS .= " -stdlib=libc++"; $setStdLib = 1; } + # Force -stdlib=libc++ on Cent OS 7 if ($HOST_OSNAME eq "\"centos\"" and $HOST_OSVER eq "\"7\"") { $HIPCXXFLAGS .= " -stdlib=libc++"; $HIPLDFLAGS .= " -lc++ -lc++abi"; From e43802035f44dd7b32b9bcc43e60c503777b9a2a Mon Sep 17 00:00:00 2001 From: Sandeep Kumar Date: Wed, 13 Sep 2017 12:57:37 +0530 Subject: [PATCH 092/108] Add more info for inline asm in hip kernel guide and cookbook readme --- docs/markdown/hip_kernel_language.md | 12 +++++++++++- samples/2_Cookbook/10_inline_asm/Readme.md | 15 ++++++++++++++- 2 files changed, 25 insertions(+), 2 deletions(-) diff --git a/docs/markdown/hip_kernel_language.md b/docs/markdown/hip_kernel_language.md index a4f4d5d47f..094d7531e8 100644 --- a/docs/markdown/hip_kernel_language.md +++ b/docs/markdown/hip_kernel_language.md @@ -699,8 +699,18 @@ for (int i=0; i<16; i++) ... ## In-Line Assembly -In-line assembly, including in-line PTX, in-line HSAIL and in-line GCN ISA, is not supported. Users who need these features should employ conditional compilation to provide different functionally equivalent implementations on each target platform. +GCN ISA In-line assembly, is supported. For example: +``` +asm volatile ("v_mac_f32_e32 %0, %2, %3" : "=v" (out[i]) : "0"(out[i]), "v" (a), "v" (in[i])); +``` + +We insert the GCN isa into the kernel using `asm()` Assembler statement. +`volatile` keyword is used so that the optimizers must not change the number of volatile operations or change their order of execution relative to other volatile operations. +`v_mac_f32_e32` is the GCN instruction, for more information please refer - [AMD GCN3 ISA architecture manual](http://gpuopen.com/compute-product/amd-gcn3-isa-architecture-manual/) +Index for the respective operand in the ordered fashion is provided by `%` followed by position in the list of operands +`"v"` is the constraint code (for target-specific AMDGPU) for 32-bit VGPR register, for more info please refer - [Supported Constraint Code List for AMDGPU](https://llvm.org/docs/LangRef.html#supported-constraint-code-list) +Output Constraints are specified by an `"="` prefix as shown above ("=v"). This indicate that assemby will write to this operand, and the operand will then be made available as a return value of the asm expression. Input constraints do not have a prefix - just the constraint code. The constraint string of `"0"` says to use the assigned register for output as an input as well (it being the 0'th constraint). ## C++ Support The following C++ features are not supported: diff --git a/samples/2_Cookbook/10_inline_asm/Readme.md b/samples/2_Cookbook/10_inline_asm/Readme.md index 0e64fe9c6e..7d0301bc74 100644 --- a/samples/2_Cookbook/10_inline_asm/Readme.md +++ b/samples/2_Cookbook/10_inline_asm/Readme.md @@ -27,10 +27,23 @@ We will be using the Simple Matrix Transpose application from the our very first ## asm() Assembler statement -We insert the GCN isa into the kernel using asm() Assembler statement. In the same sourcecode, we used for MatrixTranspose. We'll add the following: +In the same sourcecode, we used for MatrixTranspose. We'll add the following: ` asm volatile ("v_mov_b32_e32 %0, %1" : "=v" (out[x*width + y]) : "v" (in[y*width + x])); ` +GCN ISA In-line assembly, is supported. For example: + +``` +asm volatile ("v_mac_f32_e32 %0, %2, %3" : "=v" (out[i]) : "0"(out[i]), "v" (a), "v" (in[i])); +``` + +We insert the GCN isa into the kernel using `asm()` Assembler statement. +`volatile` keyword is used so that the optimizers must not change the number of volatile operations or change their order of execution relative to other volatile operations. +`v_mac_f32_e32` is the GCN instruction, for more information please refer - [AMD GCN3 ISA architecture manual](http://gpuopen.com/compute-product/amd-gcn3-isa-architecture-manual/) +Index for the respective operand in the ordered fashion is provided by `%` followed by position in the list of operands +`"v"` is the constraint code (for target-specific AMDGPU) for 32-bit VGPR register, for more info please refer - [Supported Constraint Code List for AMDGPU](https://llvm.org/docs/LangRef.html#supported-constraint-code-list) +Output Constraints are specified by an `"="` prefix as shown above ("=v"). This indicate that assemby will write to this operand, and the operand will then be made available as a return value of the asm expression. Input constraints do not have a prefix - just the constraint code. The constraint string of `"0"` says to use the assigned register for output as an input as well (it being the 0'th constraint). + ## How to build and run: Use the make command and execute it using ./exe Use hipcc to build the application, which is using hcc on AMD and nvcc on nvidia. From a6028ade1eea47717e0074beac7d5d83a21a6d3b Mon Sep 17 00:00:00 2001 From: Aaron Enye Shi Date: Wed, 13 Sep 2017 14:21:43 -0400 Subject: [PATCH 093/108] Added support for RHEL 7.4 --- bin/hipcc | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/bin/hipcc b/bin/hipcc index 30ed8145bd..19c957757c 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -117,8 +117,8 @@ if ($HIP_PLATFORM eq "hcc") { # Only include the libstdc++ headers and libraries flags explicitly if the g++ is older than version 5. # That's because HCC already uses libstdc++ by default if a newer g++/libstdc++ is available - # Cent OS 7 cannot use libstdc++ for compilation, defaults to libc++ - if (${GCC_CUR_VER} eq ${GPP_CUR_VER} and $GPP_VER_FIELDS[0] < 5 and ($HOST_OSNAME ne "\"centos\"")) { + # Cent OS 7 and RHEL 7.4 cannot use libstdc++ for compilation, default to libc++ + if (${GCC_CUR_VER} eq ${GPP_CUR_VER} and $GPP_VER_FIELDS[0] < 5 and ($HOST_OSNAME ne "\"centos\"") and ($HOST_OSNAME ne "\"rhel\"")) { $HCC_WA_FLAGS .= " -stdlib=libstdc++ -I/usr/include/x86_64-linux-gnu -I/usr/include/x86_64-linux-gnu/c++/${GCC_CUR_VER} -I/usr/include/c++/${GCC_CUR_VER} "; # Add C++ libs for GCC. $HIPLDFLAGS .= " -lstdc++"; @@ -131,8 +131,8 @@ if ($HIP_PLATFORM eq "hcc") { $HIPCXXFLAGS .= " -stdlib=libc++"; $setStdLib = 1; } - # Force -stdlib=libc++ on Cent OS 7 - if ($HOST_OSNAME eq "\"centos\"" and $HOST_OSVER eq "\"7\"") { + # Force -stdlib=libc++ on Cent OS 7 and RHEL 7.4 + if (($HOST_OSNAME eq "\"centos\"" and $HOST_OSVER eq "\"7\"") or ($HOST_OSNAME eq "\"rhel\"" and $HOST_OSVER eq "\"7.4\"")) { $HIPCXXFLAGS .= " -stdlib=libc++"; $HIPLDFLAGS .= " -lc++ -lc++abi"; $setStdLib = 1; From cea80cd8b3cfaba4c7a2d43b47264eb3e28f954b Mon Sep 17 00:00:00 2001 From: Ben Sander Date: Wed, 13 Sep 2017 23:31:48 +0000 Subject: [PATCH 094/108] Add HIP_INIT_ALLOC to init allocated memory. --- src/hip_hcc.cpp | 5 +++-- src/hip_hcc_internal.h | 1 + src/hip_memory.cpp | 6 ++++++ 3 files changed, 10 insertions(+), 2 deletions(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index dc72714e3e..5e297465ec 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -88,6 +88,8 @@ int HIP_HOST_COHERENT = 1; int HIP_SYNC_HOST_ALLOC = 1; +int HIP_INIT_ALLOC=-1; + #if (__hcc_workweek__ >= 17300) // Make sure we have required bug fix in HCC @@ -267,8 +269,6 @@ ihipStream_t::ihipStream_t(ihipCtx_t *ctx, hc::accelerator_view av, unsigned int case hipDeviceScheduleBlockingSync : _scheduleMode = Yield; break; default:_scheduleMode = Auto; }; - - }; @@ -1248,6 +1248,7 @@ void HipReadEnv() READ_ENV_I(release, HIP_FAIL_SOC, 0, "Fault on Sub-Optimal-Copy, rather than use a slower but functional implementation. Bit 0x1=Fail on async copy with unpinned memory. Bit 0x2=Fail peer copy rather than use staging buffer copy"); READ_ENV_I(release, HIP_SYNC_HOST_ALLOC, 0, "Sync before and after all host memory allocations. May help stability"); + READ_ENV_I(release, HIP_INIT_ALLOC, 0, "If not -1, initialize allocated memory to specified byte"); READ_ENV_I(release, HIP_SYNC_NULL_STREAM, 0, "Synchronize on host for null stream submissions"); diff --git a/src/hip_hcc_internal.h b/src/hip_hcc_internal.h index 56ca37b3e2..159a447081 100644 --- a/src/hip_hcc_internal.h +++ b/src/hip_hcc_internal.h @@ -67,6 +67,7 @@ extern int HIP_HIDDEN_FREE_MEM; extern int HIP_SYNC_HOST_ALLOC; extern int HIP_SYNC_NULL_STREAM; +extern int HIP_INIT_ALLOC; // TODO - remove when this is standard behavior. extern int HCC_OPT_FLUSH; diff --git a/src/hip_memory.cpp b/src/hip_memory.cpp index 4b3a4fcb12..b2f795ba75 100644 --- a/src/hip_memory.cpp +++ b/src/hip_memory.cpp @@ -119,6 +119,11 @@ void * allocAndSharePtr(const char *msg, size_t sizeBytes, ihipCtx_t *ctx, bool tprintf(DB_MEM, " alloc %s ptr:%p-%p size:%zu on dev:%d\n", msg, ptr, static_cast(ptr)+sizeBytes, sizeBytes, device->_deviceId); + if (HIP_INIT_ALLOC != -1) { + // TODO , dont' call HIP API directly here: + hipMemset(ptr, HIP_INIT_ALLOC, sizeBytes); + } + if (ptr != nullptr) { int r = sharePtr(ptr, ctx, shareWithAll, hipFlags); if (r != 0) { @@ -1180,6 +1185,7 @@ ihipMemsetKernel(hipStream_t stream, } + // TODO-sync: function is async unless target is pinned host memory - then these are fully sync. hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t stream ) { From 3e69b2ed5a4256a012269ba898c127c68a4450d0 Mon Sep 17 00:00:00 2001 From: sriharikarnam Date: Thu, 14 Sep 2017 16:03:19 +0530 Subject: [PATCH 095/108] Add hipDevice[Get/Set]SharedMemConfig API on nvcc details --- include/hip/nvcc_detail/hip_runtime_api.h | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) diff --git a/include/hip/nvcc_detail/hip_runtime_api.h b/include/hip/nvcc_detail/hip_runtime_api.h index 502d205ac4..cd8b2c17a0 100644 --- a/include/hip/nvcc_detail/hip_runtime_api.h +++ b/include/hip/nvcc_detail/hip_runtime_api.h @@ -105,7 +105,7 @@ typedef cudaIpcMemHandle_t hipIpcMemHandle_t; typedef cudaLimit hipLimit_t; typedef cudaFuncCache hipFuncCache_t; typedef CUcontext hipCtx_t; -typedef CUsharedconfig hipSharedMemConfig; +typedef cudaSharedMemConfig hipSharedMemConfig; typedef CUfunc_cache hipFuncCache; typedef CUjit_option hipJitOption; typedef CUdevice hipDevice_t; @@ -124,6 +124,11 @@ typedef cudaArray hipArray; //typedef cudaChannelFormatDesc hipChannelFormatDesc; #define hipChannelFormatDesc cudaChannelFormatDesc +//adding code for hipmemSharedConfig +#define hipSharedMemBankSizeDefault cudaSharedMemBankSizeDefault +#define hipSharedMemBankSizeFourByte cudaSharedMemBankSizeFourByte +#define hipSharedMemBankSizeEightByte cudaSharedMemBankSizeEightByte + inline static hipError_t hipCUDAErrorTohipError(cudaError_t cuError) { switch(cuError) { case cudaSuccess : return hipSuccess; @@ -833,12 +838,12 @@ inline static hipError_t hipCtxSetCacheConfig (hipFuncCache cacheConfig) inline static hipError_t hipCtxSetSharedMemConfig (hipSharedMemConfig config) { - return hipCUResultTohipError(cuCtxSetSharedMemConfig(config)); + return hipCUResultTohipError(cuCtxSetSharedMemConfig((CUsharedconfig)config)); } inline static hipError_t hipCtxGetSharedMemConfig ( hipSharedMemConfig * pConfig ) { - return hipCUResultTohipError(cuCtxGetSharedMemConfig(pConfig)); + return hipCUResultTohipError(cuCtxGetSharedMemConfig((CUsharedconfig *)pConfig)); } inline static hipError_t hipCtxSynchronize ( void ) @@ -881,6 +886,16 @@ inline static hipError_t hipDeviceGetByPCIBusId(int* device, const char *pciBusI return hipCUDAErrorTohipError(cudaDeviceGetByPCIBusId(device, pciBusId)); } +inline static hipError_t hipDeviceGetSharedMemConfig(hipSharedMemConfig *config) +{ + return hipCUDAErrorTohipError(cudaDeviceGetSharedMemConfig(config)); +} + +inline static hipError_t hipDeviceSetSharedMemConfig(hipSharedMemConfig config) +{ + return hipCUDAErrorTohipError(cudaDeviceSetSharedMemConfig(config)); +} + inline static hipError_t hipDeviceGetLimit(size_t *pValue, hipLimit_t limit) { return hipCUDAErrorTohipError(cudaDeviceGetLimit(pValue, limit)); From c3c94dcd8de503599287727cb28d799300060e61 Mon Sep 17 00:00:00 2001 From: Maneesh Gupta Date: Thu, 14 Sep 2017 16:19:57 +0530 Subject: [PATCH 096/108] Fix typo in nvcc_detail/hip_runtime_api.h --- include/hip/nvcc_detail/hip_runtime_api.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/hip/nvcc_detail/hip_runtime_api.h b/include/hip/nvcc_detail/hip_runtime_api.h index 090ceb6339..e5c12f6038 100644 --- a/include/hip/nvcc_detail/hip_runtime_api.h +++ b/include/hip/nvcc_detail/hip_runtime_api.h @@ -359,7 +359,7 @@ inline static hipError_t hipMemcpy(void* dst, const void* src, size_t sizeBytes, inline static hipError_t hipMemcpyAsync(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind copyKind, hipStream_t stream __dparm(0)) { - return hipCUDAErrorTohipError(cudaMemcpyAsync(dst, src, sizeBytes, hipMemcpyKindToCudaMemcpyKind(copyKind), 0)); + return hipCUDAErrorTohipError(cudaMemcpyAsync(dst, src, sizeBytes, hipMemcpyKindToCudaMemcpyKind(copyKind), stream)); } inline static hipError_t hipMemcpyToSymbol(const void* symbol, const void* src, size_t sizeBytes, size_t offset __dparm(0), hipMemcpyKind copyType __dparm(hipMemcpyHostToDevice)) { From ffe21c9d751e85c0d1cfdb4f0c6c8a7840bdb8fb Mon Sep 17 00:00:00 2001 From: Aaron Enye Shi Date: Thu, 14 Sep 2017 10:55:38 -0400 Subject: [PATCH 097/108] hcc-config already sets -lc++ -lc++abi flags Since hcc-config will set the flags for CentOS7 and RHEL 7.X, we don't need to set flags here. --- bin/hipcc | 6 ------ 1 file changed, 6 deletions(-) diff --git a/bin/hipcc b/bin/hipcc index 19c957757c..8f18ab0d8e 100755 --- a/bin/hipcc +++ b/bin/hipcc @@ -131,12 +131,6 @@ if ($HIP_PLATFORM eq "hcc") { $HIPCXXFLAGS .= " -stdlib=libc++"; $setStdLib = 1; } - # Force -stdlib=libc++ on Cent OS 7 and RHEL 7.4 - if (($HOST_OSNAME eq "\"centos\"" and $HOST_OSVER eq "\"7\"") or ($HOST_OSNAME eq "\"rhel\"" and $HOST_OSVER eq "\"7.4\"")) { - $HIPCXXFLAGS .= " -stdlib=libc++"; - $HIPLDFLAGS .= " -lc++ -lc++abi"; - $setStdLib = 1; - } $HIPCXXFLAGS .= " -I$HIP_PATH/include/hip/hcc_detail/cuda"; $HIPCXXFLAGS .= " -I$HSA_PATH/include"; From 7565d8ca8cd3bdccb25e1a4506df8c39c0d36d5a Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Thu, 10 Aug 2017 12:08:46 -0500 Subject: [PATCH 098/108] Adding support to upload images to docker-hub --- Jenkinsfile | 32 +++++++++++++++++++++++++------- 1 file changed, 25 insertions(+), 7 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index ee3a301dd3..280ea9b5ad 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -156,7 +156,7 @@ def docker_build_inside_image( def build_image, String inside_args, String platf //////////////////////////////////////////////////////////////////////// // This builds a fresh docker image FROM a clean base image, with no build dependencies included // Uploads the new docker image to internal artifactory -def docker_upload_artifactory( String hcc_ver, String from_image, String source_hip_rel, String build_dir_rel ) +def docker_upload_install_image( String hcc_ver, String from_image, String source_hip_rel, String build_dir_rel ) { def hip_install_image = null String image_name = "hip-${hcc_ver}-ubuntu-16.04" @@ -194,15 +194,33 @@ def docker_upload_artifactory( String hcc_ver, String from_image, String source_ hip_install_image.push( 'latest' ) } } - - // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable - // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" } catch( err ) { currentBuild.result = 'SUCCESS' } + + // Do not treat failures to push to docker-hub as a build fail + try + { + // Only push changes to the master branch to docker-hub + if( env.BRANCH_NAME.toLowerCase( ).startsWith( 'docker' ) ) + { + docker.withRegistry('https://hub.docker.com', 'docker-hub-cred' ) + { + hip_install_image.push( "${env.BUILD_NUMBER}" ) + hip_install_image.push( 'latest' ) + } + } + } + catch( err ) + { + currentBuild.result = 'SUCCESS' + } + + // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" } } @@ -242,7 +260,7 @@ parallel hcc_ctu: docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) // After a successful build, upload a docker image of the results - docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + docker_upload_install_image( hcc_ver, from_image, source_hip_rel, build_hip_rel ) } }, hcc_1_6: @@ -276,7 +294,7 @@ hcc_1_6: docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) // After a successful build, upload a docker image of the results - docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + docker_upload_install_image( hcc_ver, from_image, source_hip_rel, build_hip_rel ) } }, nvcc: From a31cd5cf18b9df667c89db0069d88c078fa25ae0 Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Thu, 10 Aug 2017 14:52:09 -0500 Subject: [PATCH 099/108] Adding docker_upload_dockerhub Refactored when docker images get deleted to the end of the pipeline --- Jenkinsfile | 43 ++++++++++++++++++++++++++++++++++--------- 1 file changed, 34 insertions(+), 9 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 280ea9b5ad..7bb6fc8756 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -156,7 +156,7 @@ def docker_build_inside_image( def build_image, String inside_args, String platf //////////////////////////////////////////////////////////////////////// // This builds a fresh docker image FROM a clean base image, with no build dependencies included // Uploads the new docker image to internal artifactory -def docker_upload_install_image( String hcc_ver, String from_image, String source_hip_rel, String build_dir_rel ) +def docker_upload_artifactory( String hcc_ver, String from_image, String source_hip_rel, String build_dir_rel ) { def hip_install_image = null String image_name = "hip-${hcc_ver}-ubuntu-16.04" @@ -173,7 +173,6 @@ def docker_upload_install_image( String hcc_ver, String from_image, String sourc // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 // hip_install_image = docker.build( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}", "--pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" ) - // The --build-arg REPO_RADEON= is a temporary fix to get around a DNS issue with our build machines // JENKINS-44836 workaround by using a bash script instead of docker.build() sh "docker build -t ${artifactory_org}/${image_name}:${env.BUILD_NUMBER} --pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" hip_install_image = docker.image( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}" ) @@ -199,6 +198,21 @@ def docker_upload_install_image( String hcc_ver, String from_image, String sourc { currentBuild.result = 'SUCCESS' } + } + + return hip_install_image +} + +//////////////////////////////////////////////////////////////////////// +// This builds a fresh docker image FROM a clean base image, with no build dependencies included +// Uploads the new docker image to internal artifactory +def docker_upload_dockerhub( def hip_install_image, String hcc_ver ) +{ + String image_name = "hip-${hcc_ver}-ubuntu-16.04" + + stage( 'docker-hub' ) + { + docker_hub_image = docker.tag( "rocm/${image_name}:${env.BUILD_NUMBER}" ) // Do not treat failures to push to docker-hub as a build fail try @@ -208,8 +222,8 @@ def docker_upload_install_image( String hcc_ver, String from_image, String sourc { docker.withRegistry('https://hub.docker.com', 'docker-hub-cred' ) { - hip_install_image.push( "${env.BUILD_NUMBER}" ) - hip_install_image.push( 'latest' ) + docker_hub_image.push( "${env.BUILD_NUMBER}" ) + docker_hub_image.push( 'latest' ) } } } @@ -218,12 +232,22 @@ def docker_upload_install_image( String hcc_ver, String from_image, String sourc currentBuild.result = 'SUCCESS' } - // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable - // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" } } +// Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable +def docker_clean_images( String hcc_ver ) +{ + String image_name = "hip-${hcc_ver}-ubuntu-16.04" + String artifactory_org = env.JOB_NAME.toLowerCase( ) + + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" + + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"rocm/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" +} + //////////////////////////////////////////////////////////////////////// // -- MAIN // Following this line is the start of MAIN of this Jenkinsfile @@ -260,7 +284,8 @@ parallel hcc_ctu: docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) // After a successful build, upload a docker image of the results - docker_upload_install_image( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + hip_install_image = docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + docker_upload_dockerhub( hip_install_image, hcc_ver ) } }, hcc_1_6: @@ -294,7 +319,7 @@ hcc_1_6: docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) // After a successful build, upload a docker image of the results - docker_upload_install_image( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) } }, nvcc: From 0e4467513954545db06492781ddaaac1fb0c99c6 Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Thu, 10 Aug 2017 16:13:57 -0500 Subject: [PATCH 100/108] Adding docker_clean_images --- Jenkinsfile | 155 ++++++++++++++++++++++++++++------------------------ 1 file changed, 84 insertions(+), 71 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 7bb6fc8756..efcb0b955b 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -140,7 +140,6 @@ def docker_build_inside_image( def build_image, String inside_args, String platf // No matter the base platform, all packages have the same name // Only upload 1 set of packages, so we don't have a race condition uploading packages - // Using hcc-ctu since that is what most people want if( platform.toLowerCase( ).startsWith( 'hcc-ctu' ) ) { archiveArtifacts artifacts: "${build_dir_rel}/*.deb", fingerprint: true @@ -174,8 +173,8 @@ def docker_upload_artifactory( String hcc_ver, String from_image, String source_ // hip_install_image = docker.build( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}", "--pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" ) // JENKINS-44836 workaround by using a bash script instead of docker.build() - sh "docker build -t ${artifactory_org}/${image_name}:${env.BUILD_NUMBER} --pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" - hip_install_image = docker.image( "${artifactory_org}/${image_name}:${env.BUILD_NUMBER}" ) + sh "docker build -t ${artifactory_org}/${image_name} --pull -f ${build_dir_rel}/dockerfile-hip-ubuntu-16.04 --build-arg base_image=${from_image} ${build_dir_rel}" + hip_install_image = docker.image( "${artifactory_org}/${image_name}" ) // The connection to artifactory can fail sometimes, but this should not be treated as a build fail try @@ -204,31 +203,43 @@ def docker_upload_artifactory( String hcc_ver, String from_image, String source_ } //////////////////////////////////////////////////////////////////////// -// This builds a fresh docker image FROM a clean base image, with no build dependencies included -// Uploads the new docker image to internal artifactory +// Uploads the new docker image to the public docker-hub def docker_upload_dockerhub( def hip_install_image, String hcc_ver ) { String image_name = "hip-${hcc_ver}-ubuntu-16.04" + String artifactory_org = env.JOB_NAME.toLowerCase( ) stage( 'docker-hub' ) { - docker_hub_image = docker.tag( "rocm/${image_name}:${env.BUILD_NUMBER}" ) - // Do not treat failures to push to docker-hub as a build fail try { // Only push changes to the master branch to docker-hub if( env.BRANCH_NAME.toLowerCase( ).startsWith( 'docker' ) ) { - docker.withRegistry('https://hub.docker.com', 'docker-hub-cred' ) + println "inside startswith" + + sh """#!/usr/bin/env bash + set -x + echo inside sh + docker tag ${artifactory_org}/${image_name} rocm/${image_name} + """ + println "after sh" + + hip_install_image = docker.image( "rocm/${image_name}" ) + println "after docker.image" + + docker.withRegistry('https://registry.hub.docker.com', 'docker-hub-cred' ) { - docker_hub_image.push( "${env.BUILD_NUMBER}" ) - docker_hub_image.push( 'latest' ) + println "docker.withRegistry" + hip_install_image.push( "${env.BUILD_NUMBER}" ) + hip_install_image.push( 'latest' ) } } } catch( err ) { + println "err: " + err.toString() currentBuild.result = 'SUCCESS' } @@ -253,43 +264,43 @@ def docker_clean_images( String hcc_ver ) // Following this line is the start of MAIN of this Jenkinsfile String build_config = 'Release' -parallel hcc_ctu: -{ - node('docker && rocm && gfx803') - { - String hcc_ver = 'hcc-ctu' - String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/clang_tot_upgrade/hcc-lc-ubuntu-16.04:latest' - String inside_args = '--device=/dev/kfd' +// parallel hcc_ctu: +// { +// node('docker && rocm && gfx803') +// { +// String hcc_ver = 'hcc-ctu' +// String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/clang_tot_upgrade/hcc-lc-ubuntu-16.04:latest' +// String inside_args = '--device=/dev/kfd' - // Checkout source code, dependencies and version files - String source_hip_rel = checkout_and_version( hcc_ver ) +// // Checkout source code, dependencies and version files +// String source_hip_rel = checkout_and_version( hcc_ver ) - // Create/reuse a docker image that represents the hip build environment - def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) +// // Create/reuse a docker image that represents the hip build environment +// def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) - // Print system information for the log - hip_build_image.inside( inside_args ) - { - sh """#!/usr/bin/env bash - set -x - /opt/rocm/bin/rocm_agent_enumerator -t ALL - /opt/rocm/bin/hcc --version - """ - } +// // Print system information for the log +// hip_build_image.inside( inside_args ) +// { +// sh """#!/usr/bin/env bash +// set -x +// /opt/rocm/bin/rocm_agent_enumerator -t ALL +// /opt/rocm/bin/hcc --version +// """ +// } - // Conctruct a binary directory path based on build config - String build_hip_rel = build_directory_rel( build_config ); +// // Conctruct a binary directory path based on build config +// String build_hip_rel = build_directory_rel( build_config ); - // Build hip inside of the build environment - docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) +// // Build hip inside of the build environment +// docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) - // After a successful build, upload a docker image of the results - hip_install_image = docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) - docker_upload_dockerhub( hip_install_image, hcc_ver ) - } -}, -hcc_1_6: -{ +// // After a successful build, upload a docker image of the results +// hip_install_image = docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) +// docker_upload_dockerhub( hip_install_image, hcc_ver ) +// } +// }, +// hcc_1_6: +// { node('docker && rocm && gfx803') { String hcc_ver = 'hcc-1.6' @@ -319,45 +330,47 @@ hcc_1_6: docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) // After a successful build, upload a docker image of the results - docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + hip_install_image = docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) + docker_upload_dockerhub( hip_install_image, hcc_ver ) + docker_clean_images( hcc_ver ) } -}, -nvcc: -{ - node('docker && cuda') - { - //////////////////////////////////////////////////////////////////////// - // Block of string constants customizing behavior for cuda - String nvcc_ver = 'nvcc-8.0' - String from_image = 'nvidia/cuda:8.0-devel' +// }, +// nvcc: +// { +// node('docker && cuda') +// { +// //////////////////////////////////////////////////////////////////////// +// // Block of string constants customizing behavior for cuda +// String nvcc_ver = 'nvcc-8.0' +// String from_image = 'nvidia/cuda:8.0-devel' // This unfortunately hardcodes the driver version nvidia_driver_375.74 in the volume mount. Research if a way // exists to get volume driver to customize the volume names to leave out driver version String inside_args = '''--device=/dev/nvidiactl --device=/dev/nvidia0 --device=/dev/nvidia-uvm --device=/dev/nvidia-uvm-tools --volume-driver=nvidia-docker --volume=nvidia_driver_375.74:/usr/local/nvidia:ro'''; - // Checkout source code, dependencies and version files - String source_hip_rel = checkout_and_version( nvcc_ver ) +// // Checkout source code, dependencies and version files +// String source_hip_rel = checkout_and_version( nvcc_ver ) - // We pull public nvidia images - def hip_build_image = docker_build_image( nvcc_ver, source_hip_rel, from_image ) +// // We pull public nvidia images +// def hip_build_image = docker_build_image( nvcc_ver, source_hip_rel, from_image ) - // Print system information for the log - hip_build_image.inside( inside_args ) - { - sh """#!/usr/bin/env bash - set -x - nvidia-smi - nvcc --version - """ - } +// // Print system information for the log +// hip_build_image.inside( inside_args ) +// { +// sh """#!/usr/bin/env bash +// set -x +// nvidia-smi +// nvcc --version +// """ +// } - // Conctruct a binary directory path based on build config - String build_hip_rel = build_directory_rel( build_config ); +// // Conctruct a binary directory path based on build config +// String build_hip_rel = build_directory_rel( build_config ); - // Build hip inside of the build environment - docker_build_inside_image( hip_build_image, inside_args, nvcc_ver, "-DHIP_NVCC_FLAGS=--Wno-deprecated-gpu-targets", build_config, source_hip_rel, build_hip_rel ) +// // Build hip inside of the build environment +// docker_build_inside_image( hip_build_image, inside_args, nvcc_ver, "-DHIP_NVCC_FLAGS=--Wno-deprecated-gpu-targets", build_config, source_hip_rel, build_hip_rel ) - // Not pushing an Nvidia based HiP to artifactory at this time - } -} +// // Not pushing an Nvidia based HiP to artifactory at this time +// } +// } From eecc936e5445b98acf362e1f36343ccbf6d422ea Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Fri, 11 Aug 2017 14:14:27 -0500 Subject: [PATCH 101/108] Adding support to take parameters from upstream hcc HCC build can start passing in build parameters into hip build --- Jenkinsfile | 214 ++++++++++++++++++++++++++++++---------------------- 1 file changed, 122 insertions(+), 92 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index efcb0b955b..392c018259 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -10,6 +10,45 @@ properties([buildDiscarder(logRotator( [$class: 'CopyArtifactPermissionProperty', projectNames: '*'] ]) +//////////////////////////////////////////////////////////////////////// +// -- Test & Bootstrapping code + +node('docker && rocm') +{ + String hcc_ver = 'hcc-1.6' + String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/roc-1.6.x/hcc-lc-ubuntu-16.04:latest' + String inside_args = '--device=/dev/kfd' + + stage( 'parameters' ) + { + cleanWs( ) + + println "upstream_hcc: ${params.upstream_hcc}" + if( params.upstream_hcc ) + { + step([$class: 'CopyArtifact', filter: 'archive/**/*.deb, docker/dockerfile-*', + fingerprintArtifacts: true, projectName: "${params.upstream_hcc}", flatten: true, + selector: [$class: 'TriggeredBuildSelector', allowUpstreamDependencies: false, fallbackToLastSuccessful: false, upstreamFilterStrategy: 'UseGlobalSetting'], + target: 'integration-testing' ]) + + // // The following that copies from workspace apparently copies from the PREVIOUS COMPLETED build, so not as handy + // step( [$class: 'CopyArtifact', filter: '**', fingerprintArtifacts: true, flatten: true, + // projectName: "${params.upstream_hcc}", selector: [$class: 'WorkspaceSelector'], target: 'integration-testing'] ) + } + else + { + println "upstream_hcc: tested false" + } + + sh """#!/usr/bin/env bash + set -x + ls -Rlah + """ + } +} + +return + //////////////////////////////////////////////////////////////////////// // -- AUXILLARY HELPER FUNCTIONS @@ -155,11 +194,10 @@ def docker_build_inside_image( def build_image, String inside_args, String platf //////////////////////////////////////////////////////////////////////// // This builds a fresh docker image FROM a clean base image, with no build dependencies included // Uploads the new docker image to internal artifactory -def docker_upload_artifactory( String hcc_ver, String from_image, String source_hip_rel, String build_dir_rel ) +String docker_upload_artifactory( String hcc_ver, String artifactory_org, String from_image, String source_hip_rel, String build_dir_rel ) { def hip_install_image = null String image_name = "hip-${hcc_ver}-ubuntu-16.04" - String artifactory_org = env.JOB_NAME.toLowerCase( ) stage( 'artifactory' ) { @@ -199,16 +237,13 @@ def docker_upload_artifactory( String hcc_ver, String from_image, String source_ } } - return hip_install_image + return image_name } //////////////////////////////////////////////////////////////////////// // Uploads the new docker image to the public docker-hub -def docker_upload_dockerhub( def hip_install_image, String hcc_ver ) +def docker_upload_dockerhub( String artifactory_org, String image_name ) { - String image_name = "hip-${hcc_ver}-ubuntu-16.04" - String artifactory_org = env.JOB_NAME.toLowerCase( ) - stage( 'docker-hub' ) { // Do not treat failures to push to docker-hub as a build fail @@ -217,90 +252,86 @@ def docker_upload_dockerhub( def hip_install_image, String hcc_ver ) // Only push changes to the master branch to docker-hub if( env.BRANCH_NAME.toLowerCase( ).startsWith( 'docker' ) ) { - println "inside startswith" - sh """#!/usr/bin/env bash set -x echo inside sh docker tag ${artifactory_org}/${image_name} rocm/${image_name} """ - println "after sh" - hip_install_image = docker.image( "rocm/${image_name}" ) - println "after docker.image" + docker_hub_image = docker.image( "rocm/${image_name}" ) - docker.withRegistry('https://registry.hub.docker.com', 'docker-hub-cred' ) - { - println "docker.withRegistry" - hip_install_image.push( "${env.BUILD_NUMBER}" ) - hip_install_image.push( 'latest' ) - } + // docker.withRegistry('https://registry.hub.docker.com', 'docker-hub-cred' ) + // { + // docker_hub_image.push( "${env.BUILD_NUMBER}" ) + // docker_hub_image.push( 'latest' ) + // } } } catch( err ) { - println "err: " + err.toString() currentBuild.result = 'SUCCESS' } - } } // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable -def docker_clean_images( String hcc_ver ) +def docker_clean_images( String artifactory_org, String image_name ) { - String image_name = "hip-${hcc_ver}-ubuntu-16.04" - String artifactory_org = env.JOB_NAME.toLowerCase( ) + // Check if any images exist first, the script returns a 0 for success, indicating grep found images + def docker_images = sh( script: "docker images | grep \"${artifactory_org}/${image_name}\"", returnStatus: true ) - // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" - - // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"rocm/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" + if( docker_images == 0 ) + { + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" + } } //////////////////////////////////////////////////////////////////////// // -- MAIN // Following this line is the start of MAIN of this Jenkinsfile String build_config = 'Release' +String job_name = env.JOB_NAME.toLowerCase( ) -// parallel hcc_ctu: -// { -// node('docker && rocm && gfx803') -// { -// String hcc_ver = 'hcc-ctu' -// String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/clang_tot_upgrade/hcc-lc-ubuntu-16.04:latest' -// String inside_args = '--device=/dev/kfd' +parallel hcc_ctu: +{ + node('docker && rocm && gfx803') + { + String hcc_ver = 'hcc-ctu' + String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/clang_tot_upgrade/hcc-lc-ubuntu-16.04:latest' + String inside_args = '--device=/dev/kfd' -// // Checkout source code, dependencies and version files -// String source_hip_rel = checkout_and_version( hcc_ver ) + // Checkout source code, dependencies and version files + String source_hip_rel = checkout_and_version( hcc_ver ) -// // Create/reuse a docker image that represents the hip build environment -// def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) + // Create/reuse a docker image that represents the hip build environment + def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) -// // Print system information for the log -// hip_build_image.inside( inside_args ) -// { -// sh """#!/usr/bin/env bash -// set -x -// /opt/rocm/bin/rocm_agent_enumerator -t ALL -// /opt/rocm/bin/hcc --version -// """ -// } + // Print system information for the log + hip_build_image.inside( inside_args ) + { + sh """#!/usr/bin/env bash + set -x + /opt/rocm/bin/rocm_agent_enumerator -t ALL + /opt/rocm/bin/hcc --version + """ + } -// // Conctruct a binary directory path based on build config -// String build_hip_rel = build_directory_rel( build_config ); + // Conctruct a binary directory path based on build config + String build_hip_rel = build_directory_rel( build_config ); -// // Build hip inside of the build environment -// docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) + // Build hip inside of the build environment + docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) -// // After a successful build, upload a docker image of the results -// hip_install_image = docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) -// docker_upload_dockerhub( hip_install_image, hcc_ver ) -// } -// }, -// hcc_1_6: -// { + // After a successful build, upload a docker image of the results + hip_install_image = docker_upload_artifactory( hcc_ver, job_name, from_image, source_hip_rel, build_hip_rel ) + docker_upload_dockerhub( job_name, hip_image_name ) + docker_clean_images( job_name, hip_image_name ) + docker_clean_images( 'rocm', hip_image_name ) + } +}, +hcc_1_6: +{ node('docker && rocm && gfx803') { String hcc_ver = 'hcc-1.6' @@ -329,48 +360,47 @@ String build_config = 'Release' // Build hip inside of the build environment docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) - // After a successful build, upload a docker image of the results - hip_install_image = docker_upload_artifactory( hcc_ver, from_image, source_hip_rel, build_hip_rel ) - docker_upload_dockerhub( hip_install_image, hcc_ver ) - docker_clean_images( hcc_ver ) + // Not pushing hip-hcc-1.6 builds at this time + hip_image_name = docker_upload_artifactory( hcc_ver, job_name, from_image, source_hip_rel, build_hip_rel ) + docker_clean_images( job_name, hip_image_name ) } -// }, -// nvcc: -// { -// node('docker && cuda') -// { -// //////////////////////////////////////////////////////////////////////// -// // Block of string constants customizing behavior for cuda -// String nvcc_ver = 'nvcc-8.0' -// String from_image = 'nvidia/cuda:8.0-devel' +}, +nvcc: +{ + node('docker && cuda') + { + //////////////////////////////////////////////////////////////////////// + // Block of string constants customizing behavior for cuda + String nvcc_ver = 'nvcc-8.0' + String from_image = 'nvidia/cuda:8.0-devel' // This unfortunately hardcodes the driver version nvidia_driver_375.74 in the volume mount. Research if a way // exists to get volume driver to customize the volume names to leave out driver version String inside_args = '''--device=/dev/nvidiactl --device=/dev/nvidia0 --device=/dev/nvidia-uvm --device=/dev/nvidia-uvm-tools --volume-driver=nvidia-docker --volume=nvidia_driver_375.74:/usr/local/nvidia:ro'''; -// // Checkout source code, dependencies and version files -// String source_hip_rel = checkout_and_version( nvcc_ver ) + // Checkout source code, dependencies and version files + String source_hip_rel = checkout_and_version( nvcc_ver ) -// // We pull public nvidia images -// def hip_build_image = docker_build_image( nvcc_ver, source_hip_rel, from_image ) + // We pull public nvidia images + def hip_build_image = docker_build_image( nvcc_ver, source_hip_rel, from_image ) -// // Print system information for the log -// hip_build_image.inside( inside_args ) -// { -// sh """#!/usr/bin/env bash -// set -x -// nvidia-smi -// nvcc --version -// """ -// } + // Print system information for the log + hip_build_image.inside( inside_args ) + { + sh """#!/usr/bin/env bash + set -x + nvidia-smi + nvcc --version + """ + } -// // Conctruct a binary directory path based on build config -// String build_hip_rel = build_directory_rel( build_config ); + // Conctruct a binary directory path based on build config + String build_hip_rel = build_directory_rel( build_config ); -// // Build hip inside of the build environment -// docker_build_inside_image( hip_build_image, inside_args, nvcc_ver, "-DHIP_NVCC_FLAGS=--Wno-deprecated-gpu-targets", build_config, source_hip_rel, build_hip_rel ) + // Build hip inside of the build environment + docker_build_inside_image( hip_build_image, inside_args, nvcc_ver, "-DHIP_NVCC_FLAGS=--Wno-deprecated-gpu-targets", build_config, source_hip_rel, build_hip_rel ) -// // Not pushing an Nvidia based HiP to artifactory at this time -// } -// } + // Not pushing an Nvidia based HiP to artifactory at this time + } +} From 0bd6d0c9e4697df2812f3eec4549fa54a5c37040 Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Tue, 15 Aug 2017 17:53:21 -0500 Subject: [PATCH 102/108] Made function hcc_integration_testing --- Jenkinsfile | 163 +++++++++++++++++++++++++++++++++++----------------- 1 file changed, 111 insertions(+), 52 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 392c018259..12d76803b8 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -7,50 +7,36 @@ properties([buildDiscarder(logRotator( daysToKeepStr: '', numToKeepStr: '10')), disableConcurrentBuilds(), + // parameters([string(name: 'sample_string', defaultValue: '', description: 'description of a sample string')]), [$class: 'CopyArtifactPermissionProperty', projectNames: '*'] ]) //////////////////////////////////////////////////////////////////////// -// -- Test & Bootstrapping code - -node('docker && rocm') -{ - String hcc_ver = 'hcc-1.6' - String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/roc-1.6.x/hcc-lc-ubuntu-16.04:latest' - String inside_args = '--device=/dev/kfd' - - stage( 'parameters' ) - { - cleanWs( ) - - println "upstream_hcc: ${params.upstream_hcc}" - if( params.upstream_hcc ) - { - step([$class: 'CopyArtifact', filter: 'archive/**/*.deb, docker/dockerfile-*', - fingerprintArtifacts: true, projectName: "${params.upstream_hcc}", flatten: true, - selector: [$class: 'TriggeredBuildSelector', allowUpstreamDependencies: false, fallbackToLastSuccessful: false, upstreamFilterStrategy: 'UseGlobalSetting'], - target: 'integration-testing' ]) - - // // The following that copies from workspace apparently copies from the PREVIOUS COMPLETED build, so not as handy - // step( [$class: 'CopyArtifact', filter: '**', fingerprintArtifacts: true, flatten: true, - // projectName: "${params.upstream_hcc}", selector: [$class: 'WorkspaceSelector'], target: 'integration-testing'] ) - } - else - { - println "upstream_hcc: tested false" - } - - sh """#!/usr/bin/env bash - set -x - ls -Rlah - """ - } -} - -return +// -- AUXILLARY HELPER FUNCTIONS //////////////////////////////////////////////////////////////////////// -// -- AUXILLARY HELPER FUNCTIONS +// Return build number of upstream job +@NonCPS +int get_upstream_build_num( ) +{ + def upstream_cause = currentBuild.rawBuild.getCause( hudson.model.Cause$UpstreamCause ) + if( upstream_cause == null) + return 0 + + return upstream_cause.getUpstreamBuild() +} + +//////////////////////////////////////////////////////////////////////// +// Return project name of upstream job +@NonCPS +String get_upstream_build_project( ) +{ + def upstream_cause = currentBuild.rawBuild.getCause( hudson.model.Cause$UpstreamCause ) + if( upstream_cause == null) + return null + + return upstream_cause.getUpstreamProject() +} //////////////////////////////////////////////////////////////////////// // Construct the relative path of the build directory @@ -99,12 +85,10 @@ String checkout_and_version( String platform ) //////////////////////////////////////////////////////////////////////// // This creates the docker image that we use to build the project in // The docker images contains all dependencies, including OS platform, to build -def docker_build_image( String platform, String source_hip_rel, String from_image ) +def docker_build_image( String platform, String org, String optional_build_parm, String source_hip_rel, String from_image ) { - String project = "hip" - String build_type_name = "build-ubuntu-16.04" - String dockerfile_name = "dockerfile-${build_type_name}" - String build_image_name = "${build_type_name}" + String build_image_name = "build-ubuntu-16.04" + String dockerfile_name = "dockerfile-build-ubuntu-16.04" def build_image = null stage("${platform} build image") @@ -115,11 +99,11 @@ def docker_build_image( String platform, String source_hip_rel, String from_imag // Docker 17.05 introduced the ability to use ARG values in FROM statements // Docker inspect failing on FROM statements with ARG https://issues.jenkins-ci.org/browse/JENKINS-44836 - //build_image = docker.build( "${project}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) + // build_image = docker.build( "${org}/${build_image_name}:latest", "--pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." ) // JENKINS-44836 workaround by using a bash script instead of docker.build() - sh "docker build -t ${project}/${build_image_name}:latest --pull -f docker/${dockerfile_name} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." - build_image = docker.image( "${project}/${build_image_name}:latest" ) + sh "docker build -t ${org}/${build_image_name}:latest -f docker/${dockerfile_name} ${optional_build_parm} --build-arg user_uid=${user_uid} --build-arg base_image=${from_image} ." + build_image = docker.image( "${org}/${build_image_name}:latest" ) } } @@ -275,24 +259,99 @@ def docker_upload_dockerhub( String artifactory_org, String image_name ) } // Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable -def docker_clean_images( String artifactory_org, String image_name ) +def docker_clean_images( String org, String image_name ) { // Check if any images exist first, the script returns a 0 for success, indicating grep found images - def docker_images = sh( script: "docker images | grep \"${artifactory_org}/${image_name}\"", returnStatus: true ) + def docker_images = sh( script: "docker images | grep \"${org}/${image_name}\"", returnStatus: true ) if( docker_images == 0 ) { // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"${artifactory_org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" + sh "docker images | grep \"${org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" } } +//////////////////////////////////////////////////////////////////////// +// hcc_integration_testing +// This function is sets up compilation and testing of HiP on a compiler downloaded from an upstream build +// Integration testing is centered around docker and constructing clean test environments every time + +// NOTES: I have implemeneted integration testing 3 different ways, and I've come to the conclusion nothing is perfect +// 1. I've tried having HCC push the test compiler to artifactory, and having HiP download the test docker image from artifactory +// a. The act of uploading and downloading images from artifactory takes minutes +// b. There is no good way of deleting images from a repository. You have to use an arcane CURL command and I don't know how +// to keep the password secret. These test integration images are meant to be ephemeral. +// 2. I tried 'docker save' to export a docker image into a tarball, and transfering the image through 'copy artifacts plugin' +// a. The HCC docker image uncompressed is over 1GB +// b. Compressing the docker image takes even longer than uploading the image to artifactory +// 3. Download the HCC .deb and dockerfile through 'copy artifacts plugin'. Create a new HCC image on the fly +// a. There is inefficency in building a new ubuntu image and installing HCC twice (once in HCC build, once here) +// b. This solution doesn't scale when we start testing downstream libraries + +// I've implemented solution #3 above, probably transitioning to #2 down the line (probably without compression) +String hcc_integration_testing( String inside_args, String job, String build_config ) +{ + // Attempt to make unique docker image names for each build, to support concurrent builds + // Mangle docker org name with upstream build info + String testing_org_name = 'hcc-test-' + get_upstream_build_project( ).replaceAll('/','-') + '-' + get_upstream_build_num( ) + + // Tag image name with this build number + String hcc_test_image_name = "hcc:${env.BUILD_NUMBER}" + + def hip_integration_image = null + + dir( 'integration-testing' ) + { + deleteDir( ) + + // This invokes 'copy artifact plugin' to copy archived files from upstream build + step([$class: 'CopyArtifact', filter: 'archive/**/*.deb, docker/dockerfile-*', + fingerprintArtifacts: true, projectName: get_upstream_build_project( ), flatten: true, + selector: [$class: 'TriggeredBuildSelector', allowUpstreamDependencies: false, fallbackToLastSuccessful: false, upstreamFilterStrategy: 'UseGlobalSetting'], + target: '.' ]) +// // The following 'copy artifact' is supposed to copy direct from workspace, but it doesn't seem to work across machines +// step( [$class: 'CopyArtifact', filter: '**', fingerprintArtifacts: true, flatten: true, +// projectName: "${params.upstream_hcc}", selector: [$class: 'WorkspaceSelector'], target: 'integration-testing'] ) + + docker.build( "${testing_org_name}/${hcc_test_image_name}", "-f dockerfile-hcc-lc-ubuntu-16.04 ." ) + } + + // Checkout source code, dependencies and version files + String source_hip_rel = checkout_and_version( job ) + + // Conctruct a binary directory path based on build config + String build_hip_rel = build_directory_rel( build_config ); + + // Build hip inside of the build environment + hip_integration_image = docker_build_image( job, testing_org_name, '', source_hip_rel, "${testing_org_name}/${hcc_test_image_name}" ) + + docker_build_inside_image( hip_integration_image, inside_args, job, '', build_config, source_hip_rel, build_hip_rel ) + + docker_clean_images( testing_org_name, '*' ) +} + //////////////////////////////////////////////////////////////////////// // -- MAIN // Following this line is the start of MAIN of this Jenkinsfile String build_config = 'Release' String job_name = env.JOB_NAME.toLowerCase( ) +// Integration testing is a special path which implies testing of an upsteam build of hcc, +// but does not need testing across older builds of hcc or cuda. This is more of a compiler +// hcc unit test +if( params.hcc_integration_test ) +{ + println "HCC integration testing" + + node('docker && rocm') + { + hcc_integration_testing( '--device=/dev/kfd', 'hcc-ctu', build_config ) + } + + return +} + +// The following launches 3 builds in parallel: hcc-ctu, hcc-1.6 and cuda parallel hcc_ctu: { node('docker && rocm && gfx803') @@ -305,7 +364,7 @@ parallel hcc_ctu: String source_hip_rel = checkout_and_version( hcc_ver ) // Create/reuse a docker image that represents the hip build environment - def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) + def hip_build_image = docker_build_image( hcc_ver, 'hip', ' --pull', source_hip_rel, from_image ) // Print system information for the log hip_build_image.inside( inside_args ) @@ -342,7 +401,7 @@ hcc_1_6: String source_hip_rel = checkout_and_version( hcc_ver ) // Create/reuse a docker image that represents the hip build environment - def hip_build_image = docker_build_image( hcc_ver, source_hip_rel, from_image ) + def hip_build_image = docker_build_image( hcc_ver, 'hip', ' --pull', source_hip_rel, from_image ) // Print system information for the log hip_build_image.inside( inside_args ) @@ -383,7 +442,7 @@ nvcc: String source_hip_rel = checkout_and_version( nvcc_ver ) // We pull public nvidia images - def hip_build_image = docker_build_image( nvcc_ver, source_hip_rel, from_image ) + def hip_build_image = docker_build_image( nvcc_ver, 'hip', ' --pull', source_hip_rel, from_image ) // Print system information for the log hip_build_image.inside( inside_args ) From 48292521ee9c2e8f6a99435f4b618a432f25e83f Mon Sep 17 00:00:00 2001 From: Kent Knox Date: Fri, 18 Aug 2017 13:21:15 -0500 Subject: [PATCH 103/108] Adding boolean parameter to job to push to docker-hub Remove the restriction to build only on gfx803 --- Jenkinsfile | 81 +++++++++++++++++++++++++++-------------------------- 1 file changed, 41 insertions(+), 40 deletions(-) diff --git a/Jenkinsfile b/Jenkinsfile index 12d76803b8..ce59476626 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -7,7 +7,7 @@ properties([buildDiscarder(logRotator( daysToKeepStr: '', numToKeepStr: '10')), disableConcurrentBuilds(), - // parameters([string(name: 'sample_string', defaultValue: '', description: 'description of a sample string')]), + parameters([booleanParam( name: 'push_image_to_docker_hub', defaultValue: false, description: 'Push hip & hcc image to rocm docker-hub' )]), [$class: 'CopyArtifactPermissionProperty', projectNames: '*'] ]) @@ -52,6 +52,21 @@ String build_directory_rel( String build_config ) } } +//////////////////////////////////////////////////////////////////////// +// Lots of images are created above; no apparent way to delete images:tags with docker global variable +def docker_clean_images( String org, String image_name ) +{ + // Check if any images exist first grepping for image names + int docker_images = sh( script: "docker images | grep \"${org}/${image_name}\"", returnStatus: true ) + + // The script returns a 0 for success (images were found ) + if( docker_images == 0 ) + { + // run bash script to clean images:tags after successful pushing + sh "docker images | grep \"${org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" + } +} + //////////////////////////////////////////////////////////////////////// // -- BUILD RELATED FUNCTIONS @@ -226,29 +241,25 @@ String docker_upload_artifactory( String hcc_ver, String artifactory_org, String //////////////////////////////////////////////////////////////////////// // Uploads the new docker image to the public docker-hub -def docker_upload_dockerhub( String artifactory_org, String image_name ) +def docker_upload_dockerhub( String local_org, String image_name, String remote_org ) { stage( 'docker-hub' ) { // Do not treat failures to push to docker-hub as a build fail try { - // Only push changes to the master branch to docker-hub - if( env.BRANCH_NAME.toLowerCase( ).startsWith( 'docker' ) ) + sh """#!/usr/bin/env bash + set -x + echo inside sh + docker tag ${local_org}/${image_name} ${remote_org}/${image_name} + """ + + docker_hub_image = docker.image( "${remote_org}/${image_name}" ) + + docker.withRegistry('https://registry.hub.docker.com', 'docker-hub-cred' ) { - sh """#!/usr/bin/env bash - set -x - echo inside sh - docker tag ${artifactory_org}/${image_name} rocm/${image_name} - """ - - docker_hub_image = docker.image( "rocm/${image_name}" ) - - // docker.withRegistry('https://registry.hub.docker.com', 'docker-hub-cred' ) - // { - // docker_hub_image.push( "${env.BUILD_NUMBER}" ) - // docker_hub_image.push( 'latest' ) - // } + docker_hub_image.push( "${env.BUILD_NUMBER}" ) + docker_hub_image.push( 'latest' ) } } catch( err ) @@ -258,19 +269,6 @@ def docker_upload_dockerhub( String artifactory_org, String image_name ) } } -// Lots of images with tags are created above; no apparent way to delete images:tags with docker global variable -def docker_clean_images( String org, String image_name ) -{ - // Check if any images exist first, the script returns a 0 for success, indicating grep found images - def docker_images = sh( script: "docker images | grep \"${org}/${image_name}\"", returnStatus: true ) - - if( docker_images == 0 ) - { - // run bash script to clean images:tags after successful pushing - sh "docker images | grep \"${org}/${image_name}\" | awk '{print \$1 \":\" \$2}' | xargs docker rmi" - } -} - //////////////////////////////////////////////////////////////////////// // hcc_integration_testing // This function is sets up compilation and testing of HiP on a compiler downloaded from an upstream build @@ -339,6 +337,7 @@ String job_name = env.JOB_NAME.toLowerCase( ) // Integration testing is a special path which implies testing of an upsteam build of hcc, // but does not need testing across older builds of hcc or cuda. This is more of a compiler // hcc unit test +// params.hcc_integration_test is set in HCC build if( params.hcc_integration_test ) { println "HCC integration testing" @@ -354,7 +353,7 @@ if( params.hcc_integration_test ) // The following launches 3 builds in parallel: hcc-ctu, hcc-1.6 and cuda parallel hcc_ctu: { - node('docker && rocm && gfx803') + node('docker && rocm') { String hcc_ver = 'hcc-ctu' String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/clang_tot_upgrade/hcc-lc-ubuntu-16.04:latest' @@ -383,15 +382,19 @@ parallel hcc_ctu: docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) // After a successful build, upload a docker image of the results - hip_install_image = docker_upload_artifactory( hcc_ver, job_name, from_image, source_hip_rel, build_hip_rel ) - docker_upload_dockerhub( job_name, hip_image_name ) + String hip_image_name = docker_upload_artifactory( hcc_ver, job_name, from_image, source_hip_rel, build_hip_rel ) + + if( params.push_image_to_docker_hub ) + { + docker_upload_dockerhub( job_name, hip_image_name, 'rocm' ) + docker_clean_images( 'rocm', hip_image_name ) + } docker_clean_images( job_name, hip_image_name ) - docker_clean_images( 'rocm', hip_image_name ) } }, hcc_1_6: { - node('docker && rocm && gfx803') + node('docker && rocm') { String hcc_ver = 'hcc-1.6' String from_image = 'compute-artifactory:5001/radeonopencompute/hcc/roc-1.6.x/hcc-lc-ubuntu-16.04:latest' @@ -419,9 +422,9 @@ hcc_1_6: // Build hip inside of the build environment docker_build_inside_image( hip_build_image, inside_args, hcc_ver, '', build_config, source_hip_rel, build_hip_rel ) - // Not pushing hip-hcc-1.6 builds at this time - hip_image_name = docker_upload_artifactory( hcc_ver, job_name, from_image, source_hip_rel, build_hip_rel ) - docker_clean_images( job_name, hip_image_name ) + // Not pushing hip-hcc-1.6 builds at this time; saves a minute and nobody needs? + // String hip_image_name = docker_upload_artifactory( hcc_ver, job_name, from_image, source_hip_rel, build_hip_rel ) + // docker_clean_images( job_name, hip_image_name ) } }, nvcc: @@ -459,7 +462,5 @@ nvcc: // Build hip inside of the build environment docker_build_inside_image( hip_build_image, inside_args, nvcc_ver, "-DHIP_NVCC_FLAGS=--Wno-deprecated-gpu-targets", build_config, source_hip_rel, build_hip_rel ) - - // Not pushing an Nvidia based HiP to artifactory at this time } } From c74d3fe2cb7efb69d93881af92b456646b4ef5a5 Mon Sep 17 00:00:00 2001 From: "Wen-Heng (Jack) Chung" Date: Fri, 15 Sep 2017 15:47:39 +0000 Subject: [PATCH 104/108] Bump device major version from 2 to 3 This would significantly improve performance for certain apps in kernel selection logic. --- src/hip_hcc.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hip_hcc.cpp b/src/hip_hcc.cpp index 445e74cbeb..2c98ac804f 100644 --- a/src/hip_hcc.cpp +++ b/src/hip_hcc.cpp @@ -824,7 +824,7 @@ hipError_t ihipDevice_t::initProperties(hipDeviceProp_t* prop) // Masquerade as a 3.0-level device. This will change as more HW functions are properly supported. // Application code should use the arch.has* to do detailed feature detection. - prop->major = 2; + prop->major = 3; prop->minor = 0; // Get number of Compute Unit From 2ec8de3b2719724020eff04efb9558a6a75c7ede Mon Sep 17 00:00:00 2001 From: wsttiger Date: Fri, 15 Sep 2017 15:05:05 -0500 Subject: [PATCH 105/108] Remove the conversion of warpSize to be consistent with hipDeviceProp --- bin/hipify-perl | 3 --- 1 file changed, 3 deletions(-) diff --git a/bin/hipify-perl b/bin/hipify-perl index a37f728bd1..ada017abf4 100755 --- a/bin/hipify-perl +++ b/bin/hipify-perl @@ -327,9 +327,6 @@ while (@ARGV) { $ft{'coord_func'} += s/\bgridDim\.y\b/hipGridDim_y/g; $ft{'coord_func'} += s/\bgridDim\.z\b/hipGridDim_z/g; - # hack to avoid replacing hipDeviceProp.warpSize call - $ft{'special_func'} += s/([^.])\bwarpSize\b/$1hipWarpSize/g; - #-------- # Events From 902d017e286c8ed7d437a1ad51fcdf2374f53db9 Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Mon, 18 Sep 2017 17:26:33 +0300 Subject: [PATCH 106/108] [HIPIFY] Fix typo in doc --- docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index cb8d359853..8c1ca16c47 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -125,7 +125,7 @@ | `cudaHostUnregister` | `hipHostUnregister` | | `cudaMalloc` | `hipMalloc` | | `cudaMalloc3D` | | -| `cudaMalloc3DArray` | `hipMalloc2DArray` | +| `cudaMalloc3DArray` | `hipMalloc3DArray` | | `cudaMallocArray` | `hipMallocArray` | | `cudaMallocHost` | `hipHostMalloc` | | `cudaMallocManaged` | | From 28a300aed261df100b8752aed3d30051dc420e27 Mon Sep 17 00:00:00 2001 From: Evgeny Mankov Date: Mon, 18 Sep 2017 17:52:50 +0300 Subject: [PATCH 107/108] [HIPIFY] CUDA RT Textures support update --- .../CUDA_Runtime_API_functions_supported_by_HIP.md | 8 ++++---- hipify-clang/src/Cuda2Hip.cpp | 8 ++++---- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md index 8c1ca16c47..73994026e3 100644 --- a/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md +++ b/docs/markdown/CUDA_Runtime_API_functions_supported_by_HIP.md @@ -307,15 +307,15 @@ | **CUDA** | **HIP** | |-----------------------------------------------------------|-------------------------------| -| `cudaBindTexture` | | +| `cudaBindTexture` | `hipBindTexture` | | `cudaBindTexture2D` | `hipBindTexture2D` | | `cudaBindTextureToArray` | `hipBindTextureToArray` | | `cudaBindTextureToMipmappedArray` | | -| `cudaCreateChannelDesc` | | -| `cudaGetChannelDesc` | | +| `cudaCreateChannelDesc` | `hipCreateChannelDesc` | +| `cudaGetChannelDesc` | `hipGetChannelDesc` | | `cudaGetTextureAlignmentOffset` | | | `cudaGetTextureReference` | | -| `cudaUnbindTexture` | | +| `cudaUnbindTexture` | `hipUnbindTexture` | ## **24. Surface Reference Management** diff --git a/hipify-clang/src/Cuda2Hip.cpp b/hipify-clang/src/Cuda2Hip.cpp index 79be71967c..553ea5d8af 100644 --- a/hipify-clang/src/Cuda2Hip.cpp +++ b/hipify-clang/src/Cuda2Hip.cpp @@ -1824,9 +1824,9 @@ struct cuda2hipMap { cuda2hipRename["cudaUnbindTexture"] = {"hipUnbindTexture", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaBindTexture2D"] = {"hipBindTexture2D", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaBindTextureToArray"] = {"hipBindTextureToArray", CONV_TEX, API_RUNTIME}; - cuda2hipRename["cudaBindTextureToMipmappedArray"] = {"hipBindTextureToMipmappedArray", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureAlignmentOffset"] = {"hipGetTextureAlignmentOffset", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; - cuda2hipRename["cudaGetTextureReference"] = {"hipGetTextureReference", CONV_TEX, API_RUNTIME, HIP_UNSUPPORTED}; + cuda2hipRename["cudaBindTextureToMipmappedArray"] = {"hipBindTextureToMipmappedArray", CONV_TEX, API_RUNTIME}; // Unsupported yet on NVCC path + cuda2hipRename["cudaGetTextureAlignmentOffset"] = {"hipGetTextureAlignmentOffset", CONV_TEX, API_RUNTIME}; // Unsupported yet on NVCC path + cuda2hipRename["cudaGetTextureReference"] = {"hipGetTextureReference", CONV_TEX, API_RUNTIME}; // Unsupported yet on NVCC path // Channel cuda2hipRename["cudaChannelFormatKind"] = {"hipChannelFormatKind", CONV_TEX, API_RUNTIME}; @@ -1845,7 +1845,7 @@ struct cuda2hipMap { cuda2hipRename["cudaResourceViewDesc"] = {"hipResourceViewDesc", CONV_TEX, API_RUNTIME}; cuda2hipRename["cudaTextureDesc"] = {"hipTextureDesc", CONV_TEX, API_RUNTIME}; cuda2hipRename["surfaceReference"] = {"hipSurfaceReference", CONV_SURFACE, API_RUNTIME, HIP_UNSUPPORTED}; - // Leaved unchanged + // Left unchanged // cuda2hipRename["textureReference"] = {"textureReference", CONV_TEX, API_RUNTIME}; // typedefs From 83d9e889a98984f6a1b9e710bfbf754a03e5d065 Mon Sep 17 00:00:00 2001 From: Siu Chi Chan Date: Wed, 20 Sep 2017 01:38:28 -0400 Subject: [PATCH 108/108] ensure the condition is evaluated before applying the negation operator --- include/hip/hcc_detail/hip_runtime.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/hip/hcc_detail/hip_runtime.h b/include/hip/hcc_detail/hip_runtime.h index 2386ea08cb..379fc05f5b 100644 --- a/include/hip/hcc_detail/hip_runtime.h +++ b/include/hip/hcc_detail/hip_runtime.h @@ -99,7 +99,7 @@ extern int HIP_TRACE_API; // TODO-HCC add a dummy implementation of assert, need to replace with a proper kernel exit call. #if __HIP_DEVICE_COMPILE__ == 1 #undef assert - #define assert(COND) { if (!COND) {abort();} } + #define assert(COND) { if (!(COND)) {abort();} } #endif