Commit Graph

5492 Commits

Author SHA1 Message Date
foreman 1d60371d01 P4 to Git Change 1780704 by yaxunl@yaxunl-lc10 on 2019/05/09 13:26:13
SWDEV-145570 - Fix device name mismatch for gfx906.

	For now hip-clang can only emits gfx906 ISA with conservative configurations, i.e. with ecc on and xnack on, therefore it is always gfx906. It is still under discussion how to encode the target id for xnack off or ecc off.

	Therefore, the reasonable solution for now is just allow code object marked as gfx906 to be loaded on any device name that starts with gfx906. We will have more detailed control once hip-clang is able to emit code object for gfx906 with ecc off or xnack off.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#30 edit
2019-05-09 14:19:54 -04:00
Maneesh Gupta 5b607e14a6 Merge pull request #1084 from mhbliao/hliao/master/api_ext
[hip] Add API `hipExtModuleLaunchKernel` in HIP runtime
2019-05-09 18:26:31 +05:30
Maneesh Gupta 88abfde2f8 [dtests] Fix hipModule test for nvcc path
Change-Id: If918b87b848a825242e06b0d552a7be188a1c4b6
2019-05-09 18:17:19 +05:30
Maneesh Gupta eb637766b9 [dtests] Add complex_loading_behavior test
Change-Id: Iadf135cb727a1a3761abef20336d652b159c7dcd
2019-05-09 18:03:42 +05:30
Maneesh Gupta 79843f3b12 [dtests] Add hipModule test to unit tests
Change-Id: I1dac38f8580265e2e9c82d88e4f070a2ff87f60b
2019-05-09 11:36:46 +05:30
Maneesh Gupta 49a2d785d0 [hit] Add support for BUILD_CMD 2019-05-09 11:36:26 +05:30
Maneesh Gupta 622ea32964 [hit] Remove CUSTOM_CMD
Change-Id: Ia156fe6aab9cfcc11284823ea5131e33eaf962bc
2019-05-09 09:59:18 +05:30
Maneesh Gupta 9f2d1453fb [hit] Rename RUN -> TEST & RUN_NAMED -> TEST_NAMED
Change-Id: I75e24f15129973cee15fc9dac65d678bd2172074
2019-05-09 09:59:18 +05:30
Evgeny Mankov bfb0524e13 Merge pull request #1090 from emankov/master
[HIPIFY][python] Initial support of hipify-python generation from hipify-clang
2019-05-08 19:12:08 +03:00
Evgeny Mankov 6b370e7743 [HIPIFY][python] Initial support of hipify-python generation from hipify-clang
+ Only a generation of transformation map of CUDA entities is implemented.
+ 2 hipify-clang options are added: -python, -o-python-map-dir.
+ Explicitly set -roc option for cuda_to_hip_mappings.py generation.
+ Generated file already might be used by pytorch team.
2019-05-08 19:08:55 +03:00
Evgeny Mankov 86fe1d3f1a Merge pull request #1089 from emankov/master
[HIPIFY][perl] Support of hipify-perl generation from hipify-clang: n…
2019-05-08 15:59:45 +03:00
Evgeny Mankov 9ddc316fa7 [HIPIFY][perl] Support of hipify-perl generation from hipify-clang: next steps
+ Generate transformation map sorted by entity type.
+ Add a generation of supported header files.
2019-05-08 15:25:06 +03:00
Maneesh Gupta 450c7ab295 Merge pull request #1088 from ROCm-Developer-Tools/mangupta-patch-1
[ci] Enable tests on ROCm 2.4
2019-05-08 12:44:02 +05:30
Maneesh Gupta 07bd3ecfad [ci] Enable tests on ROCm 2.4 2019-05-08 12:07:33 +05:30
Jenkins 2b8431f59b Merge 'master' into 'amd-master'
Change-Id: I0b8ccb710d733f41d62227750daf4adc38f401da
2019-05-07 12:15:02 -07:00
Evgeny Mankov 26d4677091 Merge pull request #1085 from emankov/master
[HIPIFY][perl] Initial support of hipify-perl generation from hipify-clang
2019-05-07 17:30:39 +03:00
Evgeny Mankov 5a3d33a338 [HIPIFY][perl] Initial support of hipify-perl generation from hipify-clang
+ Only a generation of transformation map of CUDA entities supported by HIP is implemented.
+ 3 hipify-clang options are added: -perl, -o-perl-map, -o-perl-map-dir.
+ OptionsParser mode is changed from OneOrMore to Optional to support hipify-perl generation without actual hipification.
+ Add explicit control of source files specification absence in case of no perl generation.
2019-05-07 17:27:34 +03:00
Maneesh Gupta 7264f6b64e Merge pull request #1082 from gargrahul/fix_hipmemcpy_symbol_nvcc
Fix symbol address issue on NVCC path
2019-05-07 16:17:01 +05:30
Maneesh Gupta d5abe65668 Merge pull request #1081 from mangupta/swdev-181624
Implement hipExtGetLinkTypeAndHopCount for ROCm devices
2019-05-07 16:15:41 +05:30
Maneesh Gupta f931152280 Merge pull request #1075 from mhbliao/hliao/master/test_fix2
[test] Add device variant of `std::declval`.
2019-05-07 16:15:01 +05:30
Maneesh Gupta 1d4941e487 Merge pull request #1074 from mhbliao/hliao/master/test_fix
[test] Use explicit cast for address space cast.
2019-05-07 16:09:15 +05:30
Maneesh Gupta 98ab402fcb Merge pull request #1073 from kpyzhov/multi-thread-device-test
hipMultiThreadDevice test: Reduced maximum number of created HIP stre…
2019-05-07 16:08:37 +05:30
Maneesh Gupta fea21dc6d5 Merge pull request #1072 from kpyzhov/master
Refined hipSetDevice test.
2019-05-07 16:07:36 +05:30
Maneesh Gupta 730763c817 Merge pull request #1069 from mhbliao/hliao/master/test_cleanup
[test] Remove unused common routines.
2019-05-07 16:02:57 +05:30
Maneesh Gupta d82d6b499e Merge pull request #1068 from mhbliao/hliao/master/dev_vec_func
[devfunc] Add necessary `__device__` and `__host__` attributes.
2019-05-07 16:01:48 +05:30
Yaxun (Sam) Liu 669d177079 Add documentation for supported clang options (#1065)
* Add documentation for supported clang options

* Fix typo
2019-05-07 15:59:40 +05:30
wkwchau 29b3b46b42 Return hipErrorInsufficientDriver status when CPU device not found (#1064)
* Return hipErrorInsufficientDriver status when CPU device not found - no exception thrown

* Return hipErrorInsufficientDriver status when CPU device not found
2019-05-07 15:58:25 +05:30
Maneesh Gupta 7eff09edad Merge pull request #1061 from mhbliao/hliao/master/hipcc
[hip] Repace `--rpath` with `--rpath-link`
2019-05-07 15:57:57 +05:30
Maneesh Gupta 927fd0a4bc Merge pull request #1054 from ssahasra/dry
minor cleanup: eliminate repetition
2019-05-07 15:57:46 +05:30
Michael LIAO 5150f1297a [hip] Add API hipExtModuleLaunchKernel in HIP runtime 2019-05-06 21:20:28 -04:00
Rahul Garg 620a07102d Maintain HIP_VISIBLE_DEVICES for kernel launch 2019-05-07 05:09:02 +05:30
Rahul Garg e906bd69ce Fix symbol address issue on NVCC path 2019-05-07 03:59:43 +05:30
foreman 1e856beba9 P4 to Git Change 1778985 by cpaquot@cpaquot-ocl-lc-lnx on 2019/05/06 17:21:57
SWDEV-188668 - [HIP] Implemented hipExtModuleLaunchKernel. Ignoring the flag for now.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#14 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#15 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#25 edit
2019-05-06 18:02:54 -04:00
foreman d41643c39f P4 to Git Change 1778971 by cpaquot@cpaquot-ocl-lc-lnx on 2019/05/06 17:03:12
SWDEV-187125 - [HIP] Protect hip::Event with lock and add threadId to logs.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_event.cpp#9 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_event.hpp#3 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#30 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#24 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_stream.cpp#18 edit
2019-05-06 17:43:06 -04:00
foreman 09920e6024 P4 to Git Change 1778949 by michliao@hliao-dev-00-hip.rocm-workspace on 2019/05/06 16:36:53
SWDEV-144570 - Fix build failure after switching to gcc-7

	- Hex representation of float needs gnu++11. We'd better not relying on
	  that. Change the float in hex format into alternative representation.

	RBT: http://ocltc.amd.com/reviews/r/17300/

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#29 edit
2019-05-06 17:20:00 -04:00
Maneesh Gupta 035ef04e1f Implement hipExtGetLinkTypeAndHopCount for ROCm devices
Change-Id: Ie5bb4f640ac6d189c7fceeab22627a7494fd10bd
2019-05-06 15:54:31 +05:30
Michael LIAO 4fbfb646f6 [test] Add device variant of std::declval.
- Current clang disallows any invocation of wrong-side functions even
  under context with type-inspection only. Work around that by adding a
  variant of `std::decl` with `__device__` attribute.
2019-05-03 15:58:31 -04:00
Michael LIAO 2746d95906 [test] Use explicit cast for address space cast. 2019-05-03 14:56:00 -04:00
Maneesh Gupta 117bdd8774 Merge pull request #1062 from mhbliao/hliao/master/icmp
[hip] Re-implement ballot using AMDGCN builtins
2019-05-03 17:48:19 +05:30
Maneesh Gupta 37d01a7da9 Merge pull request #1058 from mhbliao/hliao/master/devfunc
[Device Function] Fix implementation
2019-05-03 17:47:51 +05:30
Konstantin Pyzhov e29c9d50a6 hipMultiThreadDevice test: Reduced maximum number of created HIP streams on Windows. 2019-05-03 05:43:30 -04:00
Konstantin Pyzhov 069fb7e486 Refined hipSetDevice test. 2019-05-03 05:39:45 -04:00
Michael LIAO 16a27e06ee [test] Remove unused common routines.
- Remove used common routines, which variadic functions, not supported
  feature, are used.
2019-05-02 12:34:07 -04:00
Michael LIAO 44692978fe [devfunc] Add necessary __device__ and __host__ attributes.
- Minor clean up to keep consistent function declaration.
2019-05-01 22:26:35 -04:00
foreman 3ad9c3f984 P4 to Git Change 1777225 by cpaquot@cpaquot-ocl-lc-lnx on 2019/05/01 18:47:25
SWDEV-188233 - [HIP] Implemented hipGetChannelDesc.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_texture.cpp#11 edit
2019-05-01 19:53:29 -04:00
foreman ba93e1ae21 P4 to Git Change 1777158 by cpaquot@cpaquot-ocl-lc-lnx on 2019/05/01 17:12:01
SWDEV-188219 - [HIP] Add hipStreamCreateWithPriority. NOP right now.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_device_runtime.cpp#14 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#13 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#14 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_stream.cpp#17 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/commandqueue.hpp#21 edit
2019-05-01 18:43:47 -04:00
Evgeny Mankov 4e09081554 Merge pull request #1063 from emankov/master
[HIPIFY][tests] Add cuSPARSE CSR-BCSR-SPMV-conversions example
2019-04-30 17:40:05 +03:00
emankov e3082f5142 [HIPIFY][tests] Add cuSPARSE CSR-BCSR-SPMV-conversions example 2019-04-30 17:37:34 +03:00
Michael LIAO 9bd2d5746d [Device Function] Fix implementation of __bitinsert_u64
- It's a common mistake by assuming 1 << shamt would be promoted to
  64-bit, if shamt is a 64-bit integer. That's not the case. Replace
  that left shift to a 64-bit one to ensure it won't fall into undefined
  behavior.
- Fix the host-side implementation as well for device function testing.
2019-04-30 08:59:13 -04:00
Michael LIAO a64637da2c [devfunc] Re-implement ballot using AMDGCN builtins
- As the signature of `amdgcn.icmp` is changed for next-gen chip, using
  clang builtins is portable way to hide that details.
2019-04-29 17:21:25 -04:00