Граф коммитов

13474 Коммитов

Автор SHA1 Сообщение Дата
Evgeny Mankov c1ac414f14 [HIPIFY][LLVMCompat] Support of upcoming LLVM 9.0.0 2019-05-24 15:14:14 +03:00
Jenkins 56e06de2fe Merge 'master' into 'amd-master'
Change-Id: I2cda8cde846bc575585e988573ea70b03123ee4b
2019-05-24 07:27:13 -04:00
Maneesh Gupta 68d30cd7e4 Merge pull request #1136 from yxsamliu/fix-rdc
Fix device lib path for hip-clang.
2019-05-24 14:16:30 +05:30
Maneesh Gupta 98d93eef6b Merge pull request #1129 from b-sumner/master
Update kernel language documentation
2019-05-24 14:16:21 +05:30
Maneesh Gupta f03a8cc1b0 Merge pull request #1128 from aaronenyeshi/fix-smid-func
Fix bug in __smid not setting correct size
2019-05-24 14:16:12 +05:30
Maneesh Gupta f30446c4bc Merge pull request #1124 from kpyzhov/hipModule-test
Updated kernel binary file for hipModule test.
2019-05-24 14:16:03 +05:30
Maneesh Gupta c289973030 Merge pull request #1121 from kpyzhov/hipStreamSync2-test
Disable hipStreamSync2 test for hip-clang.
2019-05-24 14:15:54 +05:30
Maneesh Gupta 30ccb66949 Merge pull request #1117 from aaronenyeshi/fix-exit-code-hipcc
Fix hipcc exit code when failing
2019-05-24 14:15:44 +05:30
foreman 3ffa7150da P4 to Git Change 1787115 by chui@ocl-promo-incrementor on 2019/05/24 03:00:02
SWDEV-2 - Change OpenCL version number from 2907 to 2908.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2655 edit
2019-05-24 03:08:12 -04:00
Laurent Morichetti de89102528 Add support for code object v3
Use the code object manager library to parse the code object metadata. Both
code object v2 and v3 formats are now supported for HCC generated binaries.
2019-05-23 18:03:32 -07:00
Yaxun Sam Liu 6360570532 Fix device lib path for hip-clang.
We now always need device lib path when compiling and not need it at linking.
2019-05-23 15:11:02 -04:00
Evgeny Mankov acaea5d7b2 Merge pull request #1132 from emankov/master
[HIPIFY] Add device texture fetch functions support
2019-05-23 12:48:53 +03:00
Evgeny Mankov 7db992cf44 [HIPIFY] Add device texture fetch functions support
+ Add a corresponding reverse engineered sample tex2dKernel with texture template
2019-05-23 12:47:08 +03:00
Evgeny Mankov c3606891a9 Merge pull request #1127 from emankov/master
[HIP][HIPIFY] Make hipMemcpyParam2D coherent with cuMemcpy2D
2019-05-23 10:45:14 +03:00
foreman d6882ece45 P4 to Git Change 1786576 by chui@ocl-promo-incrementor on 2019/05/23 03:00:03
SWDEV-2 - Change OpenCL version number from 2906 to 2907.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2654 edit
2019-05-23 03:14:32 -04:00
foreman bee817ba6c P4 to Git Change 1786426 by gandryey@gera-win10 on 2019/05/22 16:54:24
SWDEV-79445 - OCL generic changes and code clean-up
	- Fix a crash when SDMA is disabled

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldevice.cpp#137 edit
2019-05-22 17:36:57 -04:00
Brian Sumner e481012f43 Update kernel language documentation 2019-05-22 12:30:00 -07:00
Aaron Enye Shi 2b11a8bf0c Fix bug in __smid not setting correct size
The SZ field should minus by 1 since SIZE range is 1..32. Also add comments that results may vary.
2019-05-22 19:20:09 +00:00
Evgeny Mankov 49b9df7a9e [HIP] fix typo in #1127 2019-05-22 20:48:18 +03:00
Evgeny Mankov 6806ab6745 [HIP] fix nvcc path break in #1127 2019-05-22 20:04:45 +03:00
Evgeny Mankov 9cb3e9aa5e [HIP][HIPIFY] Make hipMemcpyParam2D coherent with cuMemcpy2D
+ Makes hip_Memcpy2D struct compatible with CUDA_MEMCPY2D struct
+ Add hipMemcpyParam2D support in nvcc fallback path
+ Update hipify-clang, tests and docs accordingly
2019-05-22 18:31:39 +03:00
Jenkins 1e526ae7c5 Merge 'master' into 'amd-master'
Change-Id: I2c8409732f0c2397d6c65f1c39c0fb847bd1bcd8
2019-05-22 07:30:16 -04:00
foreman 064f5d2c01 P4 to Git Change 1785966 by chui@ocl-promo-incrementor on 2019/05/22 03:00:06
SWDEV-2 - Change OpenCL version number from 2905 to 2906.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2653 edit
2019-05-22 03:13:17 -04:00
foreman 163046cd23 P4 to Git Change 1785925 by skeely@skeely_LC_linux on 2019/05/21 23:23:32
SWDEV-190142 - Switch os::TimeNanos to use CLOCK_MONOTONIC_RAW.

	Avoids NTP adjustment and re-aligns CPU clock usage with ROCm.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/os/os_posix.cpp#45 edit
2019-05-21 23:48:30 -04:00
foreman fde741c2b4 P4 to Git Change 1785707 by chesik@chesik_bdc-dtw-chesikw10_driver on 2019/05/21 15:03:23
SWDEV-189990, SWDEV-190337 - Update OpenCL/PAL gfx10 counter blocks again to match GPUPerfAPI.

	SWDEV-190337:
	  - Update the GFX10 block list (gfx10BlockIsPal) to match GPUPerfAPI, which expects the OpenCL block lists to match OpenGL (UGL). Note: The list now matches the expectations of GPUPerfAPI. For TA/TD/TCP, nearly all GFX10 ASICs only require 10 or 12 instances (Arden would require 14, Mero would require 8, but not sure if those are supported by OCL), but we are using 16 instances to match UGL.

	  - Make sure the blockIdToIndexSelect array contains all the blocks supported by PAL (add a static_assert to ensure this)

	  - Refactor the PCIndexSelect enum. This enum is used to determine how to sum up counters across multiple block instances. The following types are now supported:
	    Instance -- no autosumming; instances have a one-to-one correlation with PAL
	    ShaderEngine -- the block is instanced per shader engine, and OpenCL will autosum counters across all PAL instances, providing a single value for all of PAL's instances
	    ShaderArray -- the block is instanced per shader array, and OpenCL will autosum counters across shader arrays, providing a single value for each instance within a shader array. For example, if a block has four instances per shader array, PAL would expose 16 instances total on Navi10 (2 SEs, 2 SAs per SE), but OpenCL will expose four instances
	    ComputeUnit -- the block is instanced per compute unit, and OpenCL will autosum counters across shader arrays, providing a single value for each compute-unit-per-shader-array. For example, if a block is instanced per compute unit, then PAL would expose 40 instances on a 40CU Navi10. OpenCL would support 10 instances (2 CUs-per-WGP, 5 WGPs-per-SA), autosummed across shader arrays.

	SWDEV-189990:
	 - Revert GFX9 and GFX10 tests back to using the MCVML2 counter it was using previously (prior to CL 1766829).  This is counter index 2, which the test calls "BigK bank 0 hits".  In the aforementioned change list, I updated the counter index from 2 to 14, since index 14 is the actual counter that represents "BigK bank 0 hits".  Counter index 2 is the number of hits, not "bigK" hits. This previous change caused a test regression reported in SWDEV-189990. By reverting the code to use counter 2, the expected value in the test should be correct.  Perhaps a better update would be to change the description in the source from "BigK bank 0 hits" to "bank 0 hits", but for now, I'm just going to go back to what the test was doing before.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palcounters.cpp#21 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palcounters.hpp#11 edit
... //depot/stg/opencl/drivers/opencl/tests/ocltst/module/runtime/OCLPerfCounters.cpp#47 edit
2019-05-21 16:07:56 -04:00
Konstantin Pyzhov d2125f0325 Updated kernel binary file for hipModule test. 2019-05-21 13:21:48 -04:00
Maneesh Gupta 7068859ff4 Merge pull request #1123 from emankov/master
[HIP][tests] Copyright notice update
2019-05-21 09:35:13 +05:30
Evgeny Mankov 7664560f14 [HIP][tests] Copyright notice update 2019-05-20 18:46:23 +03:00
Evgeny Mankov be09662009 Merge pull request #1122 from emankov/master
[HIPIFY][tests] Add reverse engineered HIP sample "peer2peer"
2019-05-20 17:17:56 +03:00
Evgeny Mankov 5c205b75f5 [HIPIFY][tests] Add reverse engineered HIP sample "peer2peer"
+ Fix typo in sample "stream"
2019-05-20 17:16:11 +03:00
Evgeny Mankov 44f1be5a28 Merge pull request #1120 from emankov/master
[HIPIFY][doc] Update README.md
2019-05-20 13:09:50 +03:00
Evgeny Mankov c3c42834ab [HIPIFY][doc] Update README.md
+ Populate supported CUDA and LLVM releases with hyperlinks
2019-05-20 13:03:26 +03:00
foreman e5446b960b P4 to Git Change 1785103 by chui@ocl-promo-incrementor on 2019/05/20 03:00:04
SWDEV-2 - Change OpenCL version number from 2904 to 2905.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2652 edit
2019-05-20 03:21:53 -04:00
Konstantin Pyzhov adf3b8774e Disable hipStreamSync2 test for hip-clang. 2019-05-20 02:13:30 -04:00
foreman db183b7033 P4 to Git Change 1784979 by chui@ocl-promo-incrementor on 2019/05/19 03:00:03
SWDEV-2 - Change OpenCL version number from 2903 to 2904.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2651 edit
2019-05-19 03:16:23 -04:00
Jenkins 96f4b022ab Merge 'master' into 'amd-master'
Change-Id: Ic1757967f5f1d7982fa304c84f8a4c923dd606c4
2019-05-18 19:59:41 -07:00
foreman 55fdfd9228 P4 to Git Change 1784883 by chui@ocl-promo-incrementor on 2019/05/18 03:00:03
SWDEV-2 - Change OpenCL version number from 2902 to 2903.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2650 edit
2019-05-18 03:07:54 -04:00
foreman 22bd466daf P4 to Git Change 1784838 by gandryey@gera-win10 on 2019/05/17 17:36:58
SWDEV-189453 - [Navi10][OpenCl][x32][Converter] Process hang
	- Use the argument size from the caller. With LC path and 32 bit the both sizes are different and runtime has to use the caller's size, which matches the host bitness, because the optimized path updates 32 bit values only.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palblit.cpp#30 edit
2019-05-17 18:08:20 -04:00
Maneesh Gupta 71143b426e Merge pull request #1118 from yxsamliu/fix-build-hiprtc
Fix build failure due to hiprtc for HIP-Clang
2019-05-17 21:38:22 +05:30
Jenkins 0d805b92be Merge 'master' into 'amd-master'
Change-Id: I4189feaa83580761c6fb2d0b636040a40964cbeb
2019-05-17 10:50:29 -04:00
Yaxun Sam Liu d088a8c23b Fix build failure due to hiprtc for HIP-Clang 2019-05-17 09:52:43 -04:00
foreman ab218e9710 P4 to Git Change 1784378 by chui@ocl-promo-incrementor on 2019/05/17 03:00:03
SWDEV-2 - Change OpenCL version number from 2901 to 2902.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2649 edit
2019-05-17 03:11:59 -04:00
Aaron Enye Shi 303ba600d6 Fix hipcc exit code when failing 2019-05-16 20:36:58 +00:00
Evgeny Mankov 633323c981 Update README.md 2019-05-16 18:16:17 +03:00
Evgeny Mankov ab865dbf0e Merge pull request #1116 from emankov/master
[HIPIFY][doc] Update README.md
2019-05-16 18:10:09 +03:00
Evgeny Mankov df2a64cf0d [HIPIFY][doc] Update README.md
+ Update logs
+ Formatting
2019-05-16 18:07:11 +03:00
Evgeny Mankov 21518887bc Merge pull request #1115 from emankov/master
[HIPIFY][cuDNN] Add cudnnGetFilter4dDescriptor support
2019-05-16 16:39:34 +03:00
Evgeny Mankov 02f8dfe7d8 [HIPIFY][cuDNN] Add cudnnGetFilter4dDescriptor support
+ Update cudnn_convolution_forward test accordingly
2019-05-16 16:36:23 +03:00
Alex Voicu a538eb705a Add HIPRTC, glorious ersatz for NVRTC (#1097)
* Add ersatz for NVRTC.

* Fix extraneous paren and use correct namespace.

* Use lowerCamelCase (yuck, yuck) consistently.

* Link against FS when building hiprtc lib.

* Correctly mark Manipulators. Fix dual compile.

* Add unit tests. Extend HIT to accept linker options.

* Make sure the HIPRTC library is installed.

* Better logging. Try to auto-detect the target.

* Stop specifying the target explicitly.

* Add missing flavour of `hipModuleLaunchKernel`.

* Program was already destroyed.

* Don't use `--genco`. Fix mangled name trimming.

* Fix HIPRTC breakage due to upstream noise.

* [dtests] Replace RUN -> TEST in hiprtc tests

Change-Id: Ie499e92dfe4e5c94634b1c2b76cf52d241bcfea3

* [hit] Set HIP_PATH to HIP_ROOT_DIR for all tests

Change-Id: Ib0ad1f99bc71c03e363e055dd508a7a4a210680a
2019-05-16 18:28:54 +05:30
Maneesh Gupta 3178ae2f92 Merge pull request #1113 from wenkaidu/hop_count
Use NUMA distance for hop count calculation
2019-05-16 14:16:29 +05:30