fix ub overflow in move_cache_dynamic_last_kernel_h_block kernel#477
Open
Li-brua wants to merge 1 commit into
Open
fix ub overflow in move_cache_dynamic_last_kernel_h_block kernel#477Li-brua wants to merge 1 commit into
Li-brua wants to merge 1 commit into
Conversation
Contributor
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fix Triton compilation failure caused by UB overflow in
move_cache_dynamic_last_kernel_h_blockon Ascend NPUs. See issue: sgl-project/sglang#25330The previous implementation only tiled the
Hdimension while fully materializingV x Kblocks:For large V/K values, this could generate oversized temporary buffers during
BiShengHIR lowering and fail with:
This PR additionally tiles V and K dimensions to reduce per-tile UB usage and added upper bounds for Triton block sizes to avoide overflow.