* First attempt for half2 vectorized memory access in SkipLayerNorm
* Add some functions for debugging
* Clean up the code
* Clean up the code
* Generalize the vectorized kernels with aligned_vector and remove cudaDeviceProp
* Add a unit test for a larger input size
* Fix some Lint C++ warnings
* Use ILP = 4 for the vectorized kernels
* Rewrite the vectorized kernel and templatize ComputeSkipLayerNorm
* Use conditional operator for input_v
* Refactor LaunchSkipLayerNormKernel and replace the original SkipLayerNormKernelSmall with the vectorized kernel
* Clean some comments and rename the layernorm function
* Use ComputeSkipLayerNorm to replace LaunchSkipLayerNormKernel
* Resolve a Lint C++ warning
* Fix SkipLayerNormBatch1_Float16_vec output data
* Add hipified code of bert SkipLayerNorm for ROCmEP
* Resolve some Lint C++ warnings
* Resolve some Lint C++ warnings
* Resolve some Lint C++ warnings
* Resolve Python formatting issue
* Using vectorized loads (float2) for fp16 to improve performance
* Fix a few warnings from cpplint
* Fix a few warnings from cpplint
* Use __float2half2_rn and fix some cpplint warnings
* Move some computaions to LaunchFastGeluKernel
* Fix some Lint C++ warning
* Using vectorized loads (float4) for fp16 to improve performance
* Switch whether to optimize FastGelu with float4 vectorization
* Switch to float4 memory access based on input_length in FastGelu
* Comment how to set the threshold of float2 and float4 vectorized kernels
* Add FastGelu fp16 unit tests for bias_length = 2 and 8
* Make vectorized kernels generic with aligned_vector
* Unify the vectorized kernels with/without bias
* Refactor the code to suppress cpplint warnings
* Solve formatting issues
* Remove cudaDeviceProp from FastGeluKernel and LaunchFastGeluKernel
* Move fast_gelu_impl.h to rocm/bert
* Fix some Lint C++ warnings and code alignment
Description: Format all python files under onnxruntime with black and isort.
After checking in, we can use .git-blame-ignore-revs to ignore the formatting PR in git blame.
#11315, #11316
* update layernorm to reflect the fix in ROCm 4.3.1
* fix UT
Co-authored-by: Weixing Zhang <wezhan@microsoft.com>
Co-authored-by: Ethan Tao <ettao@microsoft.com@orttrainingdev7.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net>
* Enable Attention op for ROCM EP.
As a note, potential hipify improvements: (1) handle math
contants (attention_softmax.h), (2) correctly generate transpose
options for the GEMM helpers, consider counterpart/dummy API for
CublasMathModeSetter (attention_impl.cu, attention_impl.cu). After
these improvements, we don't need to manually keep copies of the
above mentioned files any more.
* Clean up debugging code.
* [ROCm] update hipify-perl location
Depending on the ROCm version installed, hipify-perl might not always
live in the hard-coded path of /opt/rocm/bin. Use python 3.3's
shutil.which to locate the script.
* provide alternative locations for hipify-perl if not in PATH
* implement hipify-perl search as a function
This avoids running the logic during module import since all builds
import the amd_hipify module.
* fix flake8 errors
Support for device function pointers is not yet available for ROCm.
Instead, the device function pointers were converted to device functors.
Case statements, lambdas, and macros are used for dispatch; as a result,
all combinations of kernels are compiled with inlined functors. The
basis of this approach can be found in PyTorch.
Lastly, hipify and register Resize and Upsample for ROCm EP.
* re-hipify all rocm EP sources
* fix all other files affected by re-hipify
* add cuda_provider_factory.h to amd_hipify.py
* do not use cudnn_conv_algo_search in ROCm EP, missing reduce min registration
* Fix ReduceConsts template specialization introduced in #9101.
Fixes the error when building for ROCm 4.3.1:
error: too many template headers for onnxruntime::rocm::ReduceConsts<__half>::One (should be 0)
* fix flake8 error in amd_hipify.py
* speed up hipify with concurrent.futures
* flake8 fix in amd_hipify.py
* make work for both rocm 4.2 and rocm 4.3.1
* fix rocm 4.3.1 docker image reference
* fix CUDA_VERSION to ROCM_VERSION
* fix ReduceConsts conflict def
* add ifdef to miopen_common.h as well
* trailing ws
* correct batchnorm replacement output order;
remove bn replacement in grad graph builder
* update op defs and kernel class
* implement batch norm internal and grad.
* change saved_var into saved_inv_std
* cuda test case: bn internal
* remove redundant include
* fix comment; add support and UT for 1d input.
* exclude batch_norm_internal in amd_hipify
* run BNInternal UT for CUDA only
* fix CI error
* fix comment errors
* fix error
* add comment for inconsistency with cudnnBN doc
* additional comments for cudnnBN inconsistency
* Pass cuda stream to thrust function to not use default stream.
In the commit 299ace0, ORT has been changed to not use cuda default stream.
* update amd_hipify.py
* remove un-necessary stream sync
Co-authored-by: Weixing Zhang <wezhan@microsoft.com>
* ConvGrad CUDA impl
* Set up the test case for Deberta Conv1D
* Add fp16 test
Co-authored-by: Sherlock Huang <bahuang@OrtTrainingDev3.af05slrtruoetgaxwwjv5nsq5e.px.internal.cloudapp.net>
* Permit dense seq optimization on BERT-L pytorch export by enabling ReduceSumTraining, Equal, and NonZero on AMD
* enable Equal tests
* enable fast_matrix_reduction test case
* optimize cuda conv by fused activation
* remove needless print out
* exclude test from cpu
* handle status error from cudnn 8.x
* add reference to base class
* add hipify
Move CudaKernel from cuda_common.h to a new separate header, cuda_kernel.h. Update include sites to use cuda_kernel.h instead if they need CudaKernel. Inclusions of cuda_common.h are now more lightweight.
Make corresponding changes for ROCM execution provider code.
Other minor cleanup.
