* Make CAS loops use the TTAS idiom.
* More efficient re-formulation of TTAS.
* Fix typo.
* The typo was not quite a typo
[ROCm/clr commit: 26914ec76e]
* [hip] add support for implicit kernel argument for multi-grid sync
* modified code for calculating the prev_sum
* change the impCoopArg type to size_t
* add memory clean up
* launch init_gws and main kernels into two separate loops
[ROCm/clr commit: 93c688a0c9]
[Algorithm]
[Release]
If CMAKE_INSTALL_PREFIX is set by the user:
If BIN_INSTALL_DIR is set by HIP, use it as CMAKE_INSTALL_PREFIX, otherwise CMAKE_INSTALL_PREFIX is used unchanged.
If the user does not set CMAKE_INSTALL_PREFIX (CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT):
If BIN_INSTALL_DIR is set by HIP, use it as CMAKE_INSTALL_PREFIX, otherwise use PROJECT_BINARY_DIR/bin for installation.
[Debug]
If CMAKE_INSTALL_PREFIX is set by the user:
CMAKE_INSTALL_PREFIX is used unchanged.
If the user does not set CMAKE_INSTALL_PREFIX (CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT):
use CMAKE_CURRENT_SOURCE_DIR/bin for installation.
Standalone build left unchanged: CMAKE_INSTALL_PREFIX is used if set.
[ROCm/clr commit: 2435567e70]
By implicit unconditional passing -fno-delayed-template-parsing option (which appeared in LLVM 3.8.0, thus doesn't need compatibility wrapping) to hipify-clang.
[Reason] To parse uncalled template functions otherwise they are not parsed without calling, thus not hipified.
Affects cub_03.cu test, which has uncalled global template function.
[ROCm/clr commit: 7ab06b3892]
[Reason] To support maximum CUDA features in offline tests
+ Add defined(__CUDA_ARCH__) && __CUDA_ARCH__ >= 600 restriction for atomicAdd on doubles in atomics.cu.
So if LLVM < 7 and --cuda-gpu-arch doesn't work, __CUDA_ARCH__ is unset too (350 by default in clang);
if LLVM >= 7 --cuda-gpu-arch is used and __CUDA_ARCH__ is set based on it.
[ROCm/clr commit: 3233a845f6]
[Reason] To support maximum CUDA features in offline tests
+ Add CUDA_VERSION >= 800 restriction for atomics.cu
[TODO] Find a way to use or exclude atomicAdd for doubles if LLVM < 7, because
LLVM 6.0.1 and older do not use --cuda-gpu-arch in clang's Driver code at all (option is only declared)
[ROCm/clr commit: 9fc7afa738]
SWDEV-207366 - [HIP] 'hipErrorInvalidValue' (1011) with hipMemcpy3D
We need to divide by sizeByte and not multiply the WidthInBytes to get pixel width
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#79 edit
[ROCm/clr commit: 5b6f922d59]
SWDEV-207100 - [HIP CQE][HIPonPAL][WIN][QR] 5 hiptests failed in 19H1 Windows on all ASICs
1. Reshuffle locations of the hipMemset functions to make them all next to each other.
2. Update the declarations of hipMemsetD8, hipMemsetD8Async, hipMemsetD16, hipMemsetD16Async. These functions are type aware and take in as their third argument the number of elements in the buffer, not the buffer size. Change the name of this argument from sizeBytes to count to align with the above description. Changes for the header are tracked here https://github.com/ROCm-Developer-Tools/HIP/pull/1544
3. Add the actual implementation of hipMemsetD8, hipMemsetD8Async, hipMemsetD16, hipMemsetD16Async.
4. Remove ihipMemset2D() as it is essentially a copy of ihipMemset(). Change hipMemset2D()/hipMemset2DAsync() to use ihipMemset().
5. Implement hipMemset3DAsync().
6. Update the test script to pick up the updated command line options for hipMemset and hipMemset3D.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#32 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#30 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#78 edit
... //depot/stg/opencl/drivers/opencl/make/hip.git/tests/scripts/hip_runtimeapi_tests.txt#13 edit
[ROCm/clr commit: 93ad67f8a6]
SWDEV-207449 - [HIP CQE][HIPonPAL][LNX][QR] 6 hiptests failed on all ASICs
hipTestHalf fails to build on Windows due to linker error "unresolved external symbol __gnu_h2f_ieee"
1. Expose __gnu_h2f_ieee() and __gnu_f2h_ieee() for Windows builds.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/18127/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#31 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#45 edit
[ROCm/clr commit: 6f43afddfb]
SWDEV-205925 - Update HIP texture APIs for issue in hipTexRefSetAddress in HIP/PAL on Windows
- Remove the nullptr possibility
http://ocltc.amd.com/reviews/r/18121/
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_texture.cpp#23 edit
[ROCm/clr commit: daf8977c12]
SWDEV-184710 - Support hipLaunchCooperativeKernelMultiDevice()
- Add support for multi grid launch in hip
- Detect the new hidden argument and pass the required information for the kernel launch
- Memory for synchronization is allocated as a single object and then the offset for each GPU is found
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#44 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#343 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devkernel.cpp#25 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devkernel.hpp#17 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palkernel.cpp#82 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdevice.cpp#136 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdevice.hpp#42 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocvirtual.cpp#90 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocvirtual.hpp#30 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/command.cpp#99 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/command.hpp#97 edit
[ROCm/clr commit: 07626f132b]
gandryey
Alex Voicu
satyanveshd
searlmc1
Aryan Salmanpour
amd-lthakur
Rahul Garg
Evgeny Mankov
foreman
Vladislav Sytchenko
Aaron Enye Shi