Repository navigation
[SYCL][E2E][Matrix] Re-enable SG32 bfloat16 row_major B case on PVC - #23567
Merged
Merged
Conversation
The 8x16x16 row_major B case was commented out because it produced wrong results with IGC_JointMatrixLoadStoreOpt=0/1/2. The root cause was the IGC scalar VNNI load for sub-group size 32: it indexed columns with the full sub-group local id (0..31) although B has only 16 columns, and wrote Rows instead of WiRows elements into the work-item slice. IGC fixed this in intel/intel-graphics-compiler@e262aa386 (v2.35.0), and CI now uses IGC v2.41.5. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
dkhaldi
marked this pull request as draft
October 20, 2026 18:16
dkhaldi
marked this pull request as ready for review
October 20, 2026 18:17
vmaksimo
approved these changes
Oct 22, 2026
Contributor
Pick up the fix for the flaky bindless_images external semaphore test.
Contributor
Author
|
@intel/llvm-gatekeepers, this is ready to be merged. |
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
4 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The 8x16x16
row_majorB case inSG32/joint_matrix_bfloat16.cppwas commented out because it produced wrong results withIGC_JointMatrixLoadStoreOpt=0/1/2(the default 2D block load path was fine).The root cause was the IGC scalar VNNI load for sub-group size 32: it indexed columns with the full sub-group local id (0..31) although B has only 16 columns, and wrote
Rowsinstead ofWiRowselements into the work-item slice. IGC fixed this in intel/intel-graphics-compiler@e262aa386 (v2.35.0); CI uses IGC v2.41.5.Tested on PVC with all
IGC_JointMatrixLoadStoreOptsettings: fails every run with IGC 2.11, passes every run with IGC master.🤖 Generated with Claude Code