[ROCm] No-fence global reduce #2584
Merged
Conversation
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
Collaborator
Author
|
! cherry-pick --onto release/2.8 rocm7.1_internal_testing |
dhonnappa-amd
pushed a commit
that referenced
this pull request
Aug 27, 2025
This change removes need for fences in global_reduce by converting the stores to reduce_buffer[] into atomics+return. This is crucial for perf in architectures with split caches (e.g. MI300), where fences are inherently costly. cherry-pick of pytorch#161180
dhonnappa-amd
pushed a commit
that referenced
this pull request
Aug 27, 2025
This change removes need for fences in global_reduce by converting the stores to reduce_buffer[] into atomics+return. This is crucial for perf in architectures with split caches (e.g. MI300), where fences are inherently costly. cherry-pick of pytorch#161180
|
Created branch autogenerated/release/2.8_cherry-pick_pr-2584 and #2585 Created branch autogenerated/rocm7.1_internal_testing_cherry-pick_pr-2584 and #2586 Comment processed by Build |
jerrymannil
added a commit
that referenced
this pull request
Aug 27, 2025
Cherry-pick of #2584 Co-authored-by: Jerry Mannil <65309407+jerrymannil@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This change removes need for fences in global_reduce by converting the stores to reduce_buffer[] into atomics+return. This is crucial for perf in architectures with split caches (e.g. MI300), where fences are inherently costly.
cherry-pick of pytorch#161180
Cherry-picked to release/2.8 branch via #2585
Cherry-picked to rocm7.1_internal_testing branch via #2586