fix: add row boundary checks to Triton normalization kernels#103
Open
Coco58323 wants to merge 1 commit intothu-ml:mainfrom
Open
fix: add row boundary checks to Triton normalization kernels#103Coco58323 wants to merge 1 commit intothu-ml:mainfrom
Coco58323 wants to merge 1 commit intothu-ml:mainfrom
Conversation
Fixed illegal memory access in _rms_norm_fwd_fused, _layer_norm_param_fwd_fused, and _layer_norm_noparam_fwd_fused kernels. The kernels were missing row boundary checks (rows < M) which caused out-of-bounds memory access when the number of rows M is not divisible by BLOCK_M (32). Changes: - Added M parameter to all three kernels - Added row_mask = rows < M - Changed mask from 1D (cols only) to 2D (rows & cols) - Applied proper masking to all tl.load and tl.store operations
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixed illegal memory access in _rms_norm_fwd_fused, _layer_norm_param_fwd_fused, and _layer_norm_noparam_fwd_fused kernels.
The kernels were missing row boundary checks (rows < M) which caused out-of-bounds memory access when the number of rows M is not divisible by BLOCK_M (32).
Changes: