+ '-D' might be set as hipify-clang option (before separator '--' or without specifying separator at all);
+ '-D' as a clang option might be specified as well (after options separator '--');
+ All defines specified as hipify-clang defines (before options separator '--') will be set after clang's ones and will redefine the same clang's ones.
[ROCm/clr commit: 4b8fdda651]
+ '-I' might be set as hipify-clang option (before separator '--' or without specifying separator at all);
+ '-I' as a clang option might be specified as well (after options separator '--').
[ROCm/clr commit: d5281a771a]
SWDEV-2 - Change OpenCL version number from 2813 to 2814.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2561 edit
[ROCm/clr commit: ae5ae500b0]
Introduce the Dot functions which are available in the device library. Forward those prototypes, and introduce HIP API to expose the usage of the dot functions.
[ROCm/clr commit: 84a028b9e1]
SWDEV-2 - Change OpenCL version number from 2812 to 2813.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2560 edit
[ROCm/clr commit: a8fd09d6d9]
+ Append '--' to the end of hipify-clang options in case of its absence.
[Reason]
+ Avoid the necessity for the user in JSON compilation database specification.
Otherwise:
Error while trying to load a compilation database:
Could not auto-detect compilation database for file "test.cu"
Though the hipification will be performed correctly without JSON database.
[ROCm/clr commit: 63b7fc5714]
Repeats clang's '--cuda-path' option.
[Reason]
In case of absence of any other clang's options setting '-cuda-path' allows not to specify separator '--' before clang's '--cuda-path'.
+ Tests and scripts are updated accordingly.
[ROCm/clr commit: c4268855e3]
SWDEV-2 - Change OpenCL version number from 2811 to 2812.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2559 edit
[ROCm/clr commit: ca14911693]
A hash calculated via FNV-1a algorithm is introduced in ihipModule_t, the
internal of hipModule_t. The hash is used by HIP module APIs such as
- read_agent_global_from_module
to determine whether the agent-scope globals for a module have been iterated.
This commit fixes one issue that applications which load / unload modules
frequently would occasionally fail. After deep investigation of the issue it
turns out the old implementation in read_agent_global_from_module uses
hipModule_t as the key, which is not robust enough, as hipModule_t instances
are allocated dynamically so there are cases that one memory address may be
used by multiple hipModule_t instances. The real solution is to introduce a
uniquely identifiable hash for the code object associated with the HIP module.
And that's the rationale behind this commit.
[ROCm/clr commit: 6e68d44220]
SWDEV-2 - Change OpenCL version number from 2810 to 2811.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2558 edit
[ROCm/clr commit: 0c6cf47532]
SWDEV-132899 - [OCL][GFX10] report number of WGP by default on gfx10 ASICs
Both HSAIL/SC and LC compilers use WGP mode by default on gfx10 ASICs (i.e., COMPUTE_PGM_RSRC1.WGP_MODE is set to 1 by both compilers) therefore runtime should report number of WGP (i.e., CU/2) on gfx10 ASICs by default.
The new environment variable (GPU_ENABLE_WGP_MODE = 0) can be used to force CU mode on LC (i.e., -mcumode option) if its needed (HSAIL/SC doesn't have any compiler option for forcing the CU mode)
Also, using the new environment variable (GPU_ENABLE_WAVE32_MODE) to control the wave32 mode on gfx10+.
ReviewRequestURL = http://ocltc.amd.com/reviews/r/16435/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#329 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devprogram.cpp#27 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldevice.cpp#121 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palsettings.cpp#65 edit
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#301 edit
[ROCm/clr commit: e631b3978f]
SWDEV-2 - Change OpenCL version number from 2809 to 2810.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2557 edit
[ROCm/clr commit: 612ea757c4]
SWDEV-2 - Change OpenCL version number from 2808 to 2809.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2556 edit
[ROCm/clr commit: e349f1fbb8]
hipify-clang could be built by Visual Studio 14 2015 or higher.
[Reason]
hipify-clang contains c++14 features which are not supported even by the latest VS 2013 and the Platform Toolset "Visual C++ Compiler Nov 2013 CTP".
[ROCm/clr commit: 207bbad778]
SWDEV-2 - Change OpenCL version number from 2807 to 2808.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2555 edit
[ROCm/clr commit: 3fc7bb0329]
- Improved install instructions for AMD-hcc and NVIDIA-nvcc.
- Removed redundant section related to HCC-Native-GCN-ISA.
Change-Id: I33a9e71d95df51dbdcade4a48834840146bdbe6f
[ROCm/clr commit: 92a556c20a]
SWDEV-79445 - OCL generic changes and code clean-up
- Add IS_LIGHTNING check for the rocr initialization, because currently for LC builds GPU_ENABLE_PAL is forced to 1.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#241 edit
[ROCm/clr commit: c65248d5a8]
SWDEV-79445 - OCL generic changes and code clean-up
- Allow ROCr and PAL support from a single runtime binary. Runtime will use ROCr path by default with GPU_ENABLE_PAL=1 forcing PAL.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/opencl/amdocl/build/Makefile.api#183 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#240 edit
[ROCm/clr commit: 0a5b58ddec]
Aaron Enye Shi
Evgeny Mankov
Maneesh Gupta
foreman
Evgeny Mankov
Lucas Campos
Laurent Morichetti
Alex Voicu
Wen-Heng (Jack) Chung
Sameer Sahasrabuddhe