Optimize bf16 atomicCAS - #6294
Open
yvonne-lab wants to merge 1 commit into
Open
Conversation
Summary: Accumulate into FP32, then cast to BF16 Read BF16 gradients. Scatter-add into an FP32 temporary using hardware unsafeAtomicAdd. Convert the completed FP32 gradients back to BF16. This is essentially the strategy already used by the eligible path in D118739790. This is the safest option. It removes scalar BF16 CAS, improves numerical accuracy, and works on ROCm architectures without relying on packed BF16 atomic support. Its cost is an additional FP32 gradient buffer and final conversion. Differential Revision: D119529321
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary:
Accumulate into FP32, then cast to BF16
Read BF16 gradients. Scatter-add into an FP32 temporary using hardware unsafeAtomicAdd.
Convert the completed FP32 gradients back to BF16.
This is essentially the strategy already used by the eligible path in D118739790.
This is the safest option. It removes scalar BF16 CAS, improves numerical accuracy, and works on ROCm architectures without relying on packed BF16 atomic support. Its cost is an additional FP32 gradient buffer and final conversion.
Differential Revision: D119529321