1fe7f1be91
Revert "HIP program state re-initialization logic"
...
This reverts commit 04640992dc
2018-08-03 17:03:04 -05:00
f6178c2159
Revert "Improve performance of re-initialization logic"
...
This reverts commit b883ea759d
2018-08-03 17:02:58 -05:00
329f3b7a5e
Revert "Keep the map which tracks GPU kernel symbols to grow monotonically"
...
This reverts commit 8f521edff1
2018-08-03 17:02:50 -05:00
f4add6058a
P4 to Git Change 1589138 by cpaquot@cpaquot-ocl-lc-lnx on 2018/08/03 01:04:05
...
SWDEV-145570 - [HIP] Added __gnu_h2f_ieee and __gnu_f2h_ieee
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#9 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#10 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#16 edit
2018-08-03 01:15:15 -04:00
ac18a3b701
Merge pull request #588 from Jorghi12/patch-10
...
Fixing a HIP bug that causes DEBUG symbols to be dropped.
2018-08-03 08:48:56 +05:30
b485957858
Merge pull request #603 from b-sumner/master
...
Drop use of NVCC_COMPAT since it is always set
2018-08-03 08:45:18 +05:30
ad44a77ed4
Merge pull request #601 from sarunyap/fix-shareWithAll-mem-dev-mapping
...
Remove device mapping from shareWithAll memory
2018-08-03 08:45:04 +05:30
678764d113
P4 to Git Change 1589087 by skudchad@skudchad_rocm on 2018/08/02 20:10:02
...
SWDEV-145570 - [HIP] - hipHccModuleLaunchKernel needs to be a c++ sym
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#9 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#14 edit
2018-08-02 20:20:23 -04:00
08a308dd59
P4 to Git Change 1589062 by cpaquot@cpaquot-ocl-lc-lnx on 2018/08/02 18:34:20
...
SWDEV-145570 - [HIP] Fix a segfault exposed by rocBlas.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#39 edit
2018-08-02 18:52:37 -04:00
388d597321
P4 to Git Change 1589030 by cpaquot@cpaquot-ocl-lc-lnx on 2018/08/02 17:40:15
...
SWDEV-145570 - [HIP] Fixed the underline
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#15 edit
2018-08-02 18:00:05 -04:00
2ff260c727
P4 to Git Change 1588779 by cpaquot@cpaquot-ocl-lc-lnx on 2018/08/02 12:25:51
...
SWDEV-145570 - [HIP] Refactored some g_* stuff
Refactored g_functions into a platform state.
Added a _vars for registered variables.
Added an execution stack similar to Hcc-clang.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#14 edit
2018-08-02 12:33:55 -04:00
e9d7584b02
Remove adipose extension from genco output
2018-08-02 15:16:36 +05:30
e638c1377f
P4 to Git Change 1588556 by skudchad@skudchad_rocm on 2018/08/02 02:04:15
...
SWDEV-145570 - [HIP] - Release eStop if it not a nullptr.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#13 edit
2018-08-02 02:18:57 -04:00
11a91fbe0b
P4 to Git Change 1588551 by skudchad@skudchad_rocm on 2018/08/02 01:55:45
...
SWDEV-145570 - [HIP] - Refactor ihipModuleLaunchKernel and add hipModuleLaunchKernelExt
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#8 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#8 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#12 edit
2018-08-02 02:14:04 -04:00
0354367a07
Merge pull request #610 from aaronenyeshi/enable-hcc-module-launch
...
Enable hipHccModuleLaunchKernel on HIP-Clang
2018-08-02 08:08:46 +05:30
4544e681ac
Merge pull request #600 from ROCm-Developer-Tools/enable-texture-functions
...
Enable HIP texture functions with HIP vector types
2018-08-02 08:01:00 +05:30
1692af19cc
Merge pull request #593 from mangupta/patch2-removerdynamic
...
[FindHIP] Remove rdynamic from CMAKE_SHARED_LIBRARY_LINK_HIP_FLAGS
2018-08-02 07:58:34 +05:30
2960739c43
Merge pull request #514 from luckynikki/GETandALGN
...
[GETandALGN]: Get channel and alignment details
2018-08-02 07:57:50 +05:30
2ba45a23b8
Let hipcc handle static library for hip-clang
2018-08-01 22:03:01 -04:00
7560956233
Enable HIP_COMPILER to choose HIP-Clang lib links
2018-08-01 20:37:46 +00:00
9c28758fa0
Add hc_get_workitem_absolute_id for hip-clang HCC compatibility mode
2018-08-01 15:14:38 -04:00
b91691966f
Add HCC compatibility mode
2018-08-01 15:14:38 -04:00
85777ea939
P4 to Git Change 1588185 by vsytchen@vsytchen-win10 on 2018/08/01 11:46:00
...
SWDEV-145570 - [HIP] Correctly handle calculating the offset when retrieving the amd::Memory object from a prepinned pointer
ReviewBoardURL = http://ocltc.amd.com/reviews/r/15539/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#38 edit
2018-08-01 12:08:35 -04:00
a244ddd659
Merge pull request #612 from emankov/docs
...
[HIPIFY][docs] Formatting
2018-08-01 17:30:35 +03:00
55ee734b22
[HIPIFY][docs] Formatting
2018-08-01 17:29:00 +03:00
58d9df900c
Merge pull request #609 from emankov/master
...
[HIPIFY][FFT] Add cuFFT support
2018-08-01 17:11:19 +03:00
b7a4b2c300
Added offset value verification.
...
Added offset value verification.
2018-08-01 10:41:30 +05:30
0ad36dfa58
Enable hipHccModuleLaunchKernel on HIP-Clang
2018-07-31 20:38:20 +00:00
c57bfd50e3
[HIPIFY][FFT] Add cuFFT support
...
+ Add API_FFT support (data types and functions)
+ Add CUFFT_API_supported_by_HIP.md
+ Add simple_cufft.cu test
+ Update README.md
2018-07-31 22:25:48 +03:00
b02040e934
P4 to Git Change 1587799 by vsytchen@vsytchen-win10 on 2018/07/31 14:47:31
...
SWDEV-145570 - [HIP] refactor hipMemcpy* functions to correctly handle copies using prepinned memory
The current implementation of hipMemcpy functions picks the copy type based on a flag that the user passes. However, one can use the hipMemcpyHostToDevice/hipMemcpyDeviceToHost flag in a combination with prepinned memory. By using the WriteMemoryCommand/ReadMemoyCommand in this case, we will pin the same host memory twice. This is fine on PAL/Linux, since pinning the same VA range is a noop, but this will start failing once we switch to using device memory with HIP/VDI/HSA.
The solution is to ignore the hipMemcpyKind flag and let the runtime decide what kind of copy is best to do. Except for the case when hipMemcpyHostToHost is passed, since both host pointers may be prepinned.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/15482/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#37 edit
2018-07-31 15:07:56 -04:00
5eb11b58f3
Added support for pinned 2D SDMA copy
2018-07-31 14:05:35 +05:30
e42af991d1
Protect macro TYPEDEF_VECTOR_VALUE_TYPE and prefix __hip_ for vector_value_types
2018-07-30 20:41:24 +00:00
ff1e49d191
Ensure correct vector value type used in texture
...
Implement the vector value types in terms of SCALAR_TYPE##SIZE_vector_value_type for float, int, uint and sizes 2, 3, 4, 8, and 16.
2018-07-30 19:43:12 +00:00
84aadb9274
Remove device mapping from shareWithAll memory
...
When shareWithAll memory (e.g., host memory) is allocated, set appId
in hc::AmPointerInfo to -1 to indicate that this memory is not mapped
to any device. Peer checking in ihipStream_t::canSeeMemory is not
necessary if memory is shared with all devices. Thus, it is skipped.
Note that earlier host memory is always mapped to device 0 and HIP
always performs peer checking for all kinds of hipMemcpy. Since the
peer checking process requires context locking, hipMemcpy from/to host
memory always grabs device 0's context lock. Therefore, if there is
another thread holding the context lock of device 0 (e.g.,
hipDeviceSynchronize on device 0), hipMemcpy will have to wait for the
lock until it can actually perform memcpy. This can significantly
deteriorate execution performance.
Signed-off-by: Sarunya Pumma <sarunya.pumma@amd.com >
2018-07-28 23:15:16 -07:00
f2e1b9d220
Merge pull request #607 from emankov/hipBLAS
...
[HIPIFY][BLAS] cuBLAS support revise
2018-07-28 21:11:03 +03:00
e00cfefd4f
[HIPIFY][BLAS] cuBLAS support revise
...
+ doc for CUBLAS API supported by HIP
+ 3 tests: 0 & 1 based indexing, sgemm matmul
2018-07-28 21:00:44 +03:00
fc1ed91ddb
Do not use std::complex in test hipDeviceMalloc
2018-07-27 17:07:00 -04:00
f08f26942c
Fix thread index calculation in __hip_malloc
2018-07-27 16:24:51 -04:00
d17dc21c9c
Support malloc/free for hip-clang
2018-07-27 16:24:51 -04:00
8d80abb30f
Drop use of NVCC_COMPAT since it is always set
2018-07-27 09:29:29 -07:00
17f612def1
Merge pull request #596 from mangupta/patch5-fixhiplaunchparm
...
[tests] Fixed hipLaunchParm test on nvcc path
2018-07-26 06:17:50 +05:30
86188f04c8
Merge pull request #595 from mangupta/patch4-fixhipStdComplex
...
[tests] Fixed build & disabled run of hipStdComplex on nvcc path
2018-07-26 06:17:04 +05:30
ec6995485c
Merge pull request #592 from mangupta/patch1-initialize-cudapath
...
[hipcc] Initialize CUDA_PATH before printing
2018-07-26 06:16:31 +05:30
ddcb85a77b
Merge pull request #590 from ROCm-Developer-Tools/fix-ballot-anyall-return
...
Fix return EXIT_FAILURE if ballot or anyall fails
2018-07-26 06:16:23 +05:30
bf6fdca1e8
Enable HIP texture functions with HIP vector types
...
Include texture_functions.h on HIP-Clang path. Use HIP vector types instead of hc::short_vectors for texture_functions, and remove HCC headers.
2018-07-25 19:52:19 +00:00
51307ac705
[tests] Fixed hipLaunchParm test on nvcc path
...
- Uses c++11 features. Added it to nvcc options
- Arguments for some kernels exceeded 4096 bytes which is the limit
imposed by nvcc. Reduced BLOCK_DIM_SIZE to 512 to handle this
- Fixed compilation issues on nvcc path
Change-Id: I14f6b28afcb7c6b24a085fd707b2104e2ed64627
2018-07-25 15:41:04 +05:30
843020bfe6
[tests] Fixed build & disabled run of hipStdComplex on nvcc path
...
Change-Id: I885e96902faa065f6559183f33b6a7aa1f37e393
2018-07-25 15:39:45 +05:30
7a5987e289
[FindHIP] Remove rdynamic from CMAKE_SHARED_LIBRARY_LINK_HIP_FLAGS
...
CMAKE_SHARED_LIBRARY_LINK_CXX_FLAGS contains -rdynamic which does not go
well with nvcc
Change-Id: I10e6d21553de0e43c6a7ae5fb3ec2418087959ff
2018-07-25 15:32:14 +05:30
2027675b3f
[hipcc] Initialize CUDA_PATH before printing
...
Change-Id: Ifd448b628fa4a7055f4a439ed2c817c5f11f9bef
2018-07-25 15:29:57 +05:30
8c195a3588
Merge pull request #589 from gargrahul/set_offset_gettexalignoffset
...
Set offset in hipGetTextureAlignmentOffset
2018-07-25 12:13:31 +05:30
Wen-Heng (Jack) Chung
foreman
Maneesh Gupta
Rahul Garg
Yaxun Sam Liu
Aaron Enye Shi
Evgeny Mankov
Evgeny Mankov
lthakur
Sarunya Pumma
Brian Sumner