fix crash issue on triton paged_pa_mqa#1853
Merged
Conversation
Signed-off-by: ganyi <ygan@amd.com>
Contributor
There was a problem hiding this comment.
Pull request overview
This PR fixes an accuracy issue in the Triton paged_pa_mqa kernel by addressing an overflow problem in the deepgemm_fp8_paged_mqa_logits_stage1 function. The fix converts stride_out_heads to a 64-bit integer to prevent overflow in pointer arithmetic calculations.
Changes:
- Fixed overflow issue in pointer arithmetic by converting
stride_out_headstotl.int64 - Updated copyright year to 2026
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
ganyi1996ppo
changed the title
paged_pa_mqapaged_pa_mqa
5 tasks
valarLip
reviewed
Jan 16, 2026
Collaborator
There was a problem hiding this comment.
stride_out_heads : tl.int64,
Signed-off-by: ganyi <ygan@amd.com>
valarLip
approved these changes
Jan 16, 2026
yzhou103
pushed a commit
that referenced
this pull request
Jan 28, 2026
* fix accuracy issue on triton paged_pa_mqa Signed-off-by: ganyi <ygan@amd.com> * add int64 annotation for input stride Signed-off-by: ganyi <ygan@amd.com> --------- Signed-off-by: ganyi <ygan@amd.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Since vllm's nightly docker still remains in triton 3.4.0, We might still use this kernel until new triton version available. This PR fix the overflow issue of
deepgemm_fp8_paged_mqa_logits_stage1. And after the triton upgraded to 3.5.0, vllm will fully adopt gluon implementationTechnical Details
Test Plan
Test Result
Submission Checklist