This is an automated email from the ASF dual-hosted git repository.
apeforest pushed a change to branch benchmark
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from c0560fc add log message and TODO
add 77beeb6 add cutlass as 3rdparty dependency
add 6728de8 add cutlass to compilation flags
add 543098b remove all cutlass stuff
add 9b5ee9a add better error message and description and remove cutlass
from compilation flags
add 7ef32ee change credit for the approach since the code have changed
add 3115fb1 fix typos
add 67f5aa9 correct another typo
add 0485ed2 Add all the cuda/cublas helper functions
add 7f5194d remove tests using kAddTo
add d4ffa4e only use cublasStridedBatchedGemm if CUDA >= 9.1
add e88fa4b add equivalent mxnet code in description of mha ops
add e76c38a remove a wrong copy-paste
add 0b25025 add _contrib for namespace and add GPU only on description
add f8bd1cb add warning in bwd_ignore_zero_init description, also test
with fp32
add 68ded77 add error return if bwd_ignore_zero_init is used without
MXNET_EXEC_ENABLE_ADDTO
add a93ad94 remove std::move for clang
add 95cb2fd remove bwd_ignore_zero_init flag
add cec1ab2 remove bwd_ignore_zero_init in test_operator_gpu.py
add 344f3fd fix typo
add 579b9dd fix another typo
No new revisions were added by this update.
Summary of changes:
src/common/cuda_utils.h | 74 +++++
src/operator/contrib/transformer-inl.h | 9 +
src/operator/contrib/transformer.cc | 270 ++++++++++++++++
src/operator/contrib/transformer.cu | 554 +++++++++++++++++++++++++++++++++
tests/python/gpu/test_operator_gpu.py | 314 +++++++++++++++++++
5 files changed, 1221 insertions(+)
