diff options
| author | iotamudelta <dieterich@ogolem.org> | 2018-08-06 14:48:45 -0700 |
|---|---|---|
| committer | Facebook Github Bot <facebook-github-bot@users.noreply.github.com> | 2018-08-06 14:54:01 -0700 |
| commit | a38b572de3de15b6c59df97d721603ca85dd322e (patch) | |
| tree | e551b364598b097f1023bf66fc3868fbd7184961 /cmake | |
| parent | e0d43572c13108bb08887b4ed2e73eb8c8a32e56 (diff) | |
| download | pytorch-a38b572de3de15b6c59df97d721603ca85dd322e.tar.gz pytorch-a38b572de3de15b6c59df97d721603ca85dd322e.tar.bz2 pytorch-a38b572de3de15b6c59df97d721603ca85dd322e.zip | |
enable unit tests and other changes (#10266)
Summary:
This PR for the ROCm target does the following:
* enable some unit tests on ROCm
* fix a missing static_cast that breaks BatchNorm call on ROCm
* fix BatchNorm to work on ROCm w/ ROCm warp sizes etc
* improve the pyhipify script by introducing kernel scope to some transpilations and other improvements
* fix a linking issue on ROCm
* for more unit test sets: mark currently broken tests broken (to be fixed)
* enable THINLTO (phase one) to parallelize linking
* address the first failing of the elementwise kernel by removing non-working ROCm specialization
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10266
Differential Revision: D9184178
Pulled By: ezyang
fbshipit-source-id: 03bcd1fe4ca4dd3241f09634dbd42b6a4c350297
Diffstat (limited to 'cmake')
| -rw-r--r-- | cmake/public/LoadHIP.cmake | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/cmake/public/LoadHIP.cmake b/cmake/public/LoadHIP.cmake index b3ee2ae9e7..481f812852 100644 --- a/cmake/public/LoadHIP.cmake +++ b/cmake/public/LoadHIP.cmake @@ -111,6 +111,8 @@ IF(HIP_FOUND) set(CMAKE_HIP_ARCHIVE_CREATE ${CMAKE_CXX_ARCHIVE_CREATE}) set(CMAKE_HIP_ARCHIVE_APPEND ${CMAKE_CXX_ARCHIVE_APPEND}) set(CMAKE_HIP_ARCHIVE_FINISH ${CMAKE_CXX_ARCHIVE_FINISH}) + SET(CMAKE_HCC_FLAGS_DEBUG ${CMAKE_CXX_FLAGS_DEBUG}) + SET(CMAKE_HCC_FLAGS_RELEASE ${CMAKE_CXX_FLAGS_RELEASE}) ### Remove setting of Flags when FindHIP.CMake PR #558 is accepted.### set(rocrand_DIR ${ROCRAND_PATH}/lib/cmake/rocrand) |