[nccl-ep] Remove redundant cudaStreamSynchronize in ncclEpCreateHandle#2082
Open
kwen2501 wants to merge 1 commit intoNVIDIA:masterfrom
Open
[nccl-ep] Remove redundant cudaStreamSynchronize in ncclEpCreateHandle#2082kwen2501 wants to merge 1 commit intoNVIDIA:masterfrom
cudaStreamSynchronize in ncclEpCreateHandle#2082kwen2501 wants to merge 1 commit intoNVIDIA:masterfrom
Conversation
…ndle` The `ncclAllGather` and `call_metadata_preprocessing` are on the same CUDA stream, so intra-stream ordering already guarantees the allgather completes before the preprocessing kernel launches. Additionally, it seems nowhere else would use the allgather result `global_routing_map`. Co-authored-by: Claude <noreply@anthropic.com> Signed-off-by: Ke Wen <kwen@nvidia.com>
Collaborator
Author
|
/mirror |
Collaborator
|
/mirror-test |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
The
ncclAllGatherandcall_metadata_preprocessingare on the same CUDA stream, so intra-stream ordering already guarantees the allgather completes before the preprocessing kernel launches. Additionally, it seems nowhere else would use the allgather resultglobal_routing_map.Related Issues
Changes & Impact
Removed the
cudaStreamSynchronizePerformance Impact
Should save an unnecessary GPU-CPU sync.