Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[ROCm] Fix for the broken --config=rocm build #31433

Merged
merged 1 commit into from
Aug 8, 2019
Merged

[ROCm] Fix for the broken --config=rocm build #31433

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
16 changes: 10 additions & 6 deletions tensorflow/core/kernels/conv_grad_input_ops.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1068,15 +1068,9 @@ void LaunchConv2DBackpropInputOp<GPUDevice, T>::operator()(
auto filter_ptr =
AsDeviceMemory(transformed_filter.template flat<T>().data(),
transformed_filter.template flat<T>().size());
se::TfAllocatorAdapter tf_allocator_adapter(stream->parent()->platform(),
ctx->device()->GetAllocator({}));
se::cuda::RedzoneAllocator rz_allocator(stream, &tf_allocator_adapter,
se::cuda::PtxCompilationOptions());
auto in_backprop_ptr =
AsDeviceMemory(pre_transformed_in_backprop.template flat<T>().data(),
pre_transformed_in_backprop.template flat<T>().size());
se::DeviceMemory<T> in_backprop_ptr_rz(
WrapRedzoneBestEffort(&rz_allocator, in_backprop_ptr));

static int64 ConvolveBackwardDataScratchSize = GetDnnWorkspaceLimit(
"TF_CUDNN_WORKSPACE_LIMIT_IN_MB", 1LL << 32 // 4GB by default
Expand Down Expand Up @@ -1107,6 +1101,16 @@ void LaunchConv2DBackpropInputOp<GPUDevice, T>::operator()(
if (cudnn_use_autotune && !AutoTuneConvBwdData::GetInstance()->Find(
conv_parameters, &algorithm_config)) {
#if GOOGLE_CUDA

se::TfAllocatorAdapter tf_allocator_adapter(
stream->parent()->platform(), ctx->device()->GetAllocator({}));

se::cuda::RedzoneAllocator rz_allocator(stream, &tf_allocator_adapter,
se::cuda::PtxCompilationOptions());

se::DeviceMemory<T> in_backprop_ptr_rz(
WrapRedzoneBestEffort(&rz_allocator, in_backprop_ptr));

std::vector<AlgorithmDesc> algorithms;
CHECK(stream->parent()->GetConvolveBackwardDataAlgorithms(
conv_parameters.ShouldIncludeWinogradNonfusedAlgo<T>(stream->parent()),
Expand Down