Skip to content

Conversation

petercad
Copy link
Contributor

Addresses MFDNN-14012.

There was a bug in the code calculating the number of stream-k slabs that could lead to suboptimal L3 cache usage.

@petercad petercad requested a review from a team as a code owner October 14, 2025 23:47
@github-actions github-actions bot added the platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel label Oct 14, 2025
@petercad
Copy link
Contributor Author

make test perf-gpu
set primitive=gpu:gemm

@petercad petercad merged commit 7b3f2af into main Oct 15, 2025
10 of 11 checks passed
@petercad petercad deleted the petercad/kv_slab_fix branch October 15, 2025 17:05
@petercad petercad restored the petercad/kv_slab_fix branch October 15, 2025 23:02
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants