Skip to content

[TRT RTX EP] Fixing the stream parameter in CopyTensors API and passing cudastream by value in createNotification API #25937

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

Merged
yuslepukhin merged 2 commits into from
Sep 4, 2025
Merged

[TRT RTX EP] Fixing the stream parameter in CopyTensors API and passing cudastream by value in createNotification API #25937

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f294fd7
Fixing the stream parameter in CopyTensors API and passing cudastream…
praneshgo Sep 3, 2025
f6380cc
Passing the cudastream as value and not pointer inside
praneshgo Sep 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion onnxruntime/core/framework/plugin_data_transfer.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,7 @@ Status DataTransfer::CopyTensors(const std::vector<SrcDstPair>& src_dst_pairs) c
for (size_t i = 0; i < src_dst_pairs.size(); ++i) {
src_values.push_back(&values[i * 2]);
dst_values.push_back(&values[i * 2 + 1]);
streams.push_back(nullptr); // static_cast<OrtSyncStream*>(src_dst_pairs[i].src_stream));
streams.push_back(reinterpret_cast<OrtSyncStream*>(src_dst_pairs[i].src_stream));
}

auto* status = impl_.CopyTensors(&impl_, src_values.data(), dst_values.data(), streams.data(),
Expand Down
6 changes: 3 additions & 3 deletions onnxruntime/core/providers/cuda/cuda_provider_factory.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -586,7 +586,7 @@ struct CudaSyncNotificationImpl : OrtSyncNotificationImpl {
Release = ReleaseImpl;
}

cudaStream_t& stream_;
cudaStream_t stream_;
cudaEvent_t event_;

const OrtApi& ort_api;
Expand Down Expand Up @@ -632,9 +632,9 @@ struct CudaSyncStreamImpl : OrtSyncStreamImpl {
*notification_impl = nullptr;

std::unique_ptr<CudaSyncNotificationImpl> notification;
cudaStream_t* cuda_stream = static_cast<cudaStream_t*>(impl.stream_.GetHandle());
cudaStream_t cuda_stream = static_cast<cudaStream_t>(impl.stream_.GetHandle());

RETURN_IF_ERROR(CudaSyncNotificationImpl::Create(*cuda_stream, impl.ort_api, notification));
RETURN_IF_ERROR(CudaSyncNotificationImpl::Create(cuda_stream, impl.ort_api, notification));
*notification_impl = notification.release();

return nullptr;
Expand Down
6 changes: 3 additions & 3 deletions onnxruntime/core/providers/nv_tensorrt_rtx/nv_provider_factory.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -431,7 +431,7 @@ struct NvTrtRtxSyncNotificationImpl : OrtSyncNotificationImpl {
Release = ReleaseImpl;
}

cudaStream_t& stream_;
cudaStream_t stream_;
cudaEvent_t event_;

const OrtApi& ort_api;
Expand Down Expand Up @@ -477,9 +477,9 @@ struct NvTrtRtxSyncStreamImpl : OrtSyncStreamImpl {
*notification_impl = nullptr;

std::unique_ptr<NvTrtRtxSyncNotificationImpl> notification;
cudaStream_t* cuda_stream = static_cast<cudaStream_t*>(impl.stream_.GetHandle());
cudaStream_t cuda_stream = static_cast<cudaStream_t>(impl.stream_.GetHandle());

RETURN_IF_ERROR(NvTrtRtxSyncNotificationImpl::Create(*cuda_stream, impl.ort_api, notification));
RETURN_IF_ERROR(NvTrtRtxSyncNotificationImpl::Create(cuda_stream, impl.ort_api, notification));
*notification_impl = notification.release();

return nullptr;
Expand Down
Loading