Skip to content

vulkan: fix data race and OOB access in argsort(large) - #28705

Merged
ggerganov merged 1 commit into
ggml-org:masterfrom
jeffbolznv:argsort_race_oob
Sep 11, 2026
Merged

ggerganov merged 1 commit into
ggml-org:masterfrom
jeffbolznv:argsort_race_oob

Conversation

@jeffbolznv

Copy link
Copy Markdown
Contributor

Overview

argsort had a data race in the inner loop, which VVL caught. But I don't think this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I couldn't reproduce it locally and I don't think it's a convincing explanation of the failures.

I think these fixes are correct and I want to see if they help with CI, I'm just not confident they will.

Requirements

argsort had a data race in the inner loop, which VVL caught. But I don't think
this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I
couldn't reproduce it locally and I don't think it's a convincing explanation
of the failures.
@jeffbolznv
jeffbolznv requested a review from a team as a code owner September 10, 2026 14:15
@github-actions github-actions Bot added Vulkan Issues specific to the Vulkan backend ggml changes relating to the ggml tensor library for machine learning labels Sep 10, 2026
@jeffbolznv

Copy link
Copy Markdown
Contributor Author

The CI passed. I guess let's go with this and then just monitor for any intermittent failures.

@0cc4m 0cc4m added the merge ready A maintainer can use this label to indicate that they consider the changes final and ready to merge. label Sep 11, 2026
@ggerganov
ggerganov merged commit 481c65f into ggml-org:master Sep 11, 2026
21 checks passed
Patt92 pushed a commit to Patt92/llama.cpp that referenced this pull request Sep 11, 2026
argsort had a data race in the inner loop, which VVL caught. But I don't think
this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I
couldn't reproduce it locally and I don't think it's a convincing explanation
of the failures.

(cherry picked from commit 481c65f)
pl752 pushed a commit to pl752/llama.cpp that referenced this pull request Sep 15, 2026
argsort had a data race in the inner loop, which VVL caught. But I don't think
this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I
couldn't reproduce it locally and I don't think it's a convincing explanation
of the failures.
quimmedes pushed a commit to quimmedes/cafe-llama.cpp that referenced this pull request Sep 16, 2026
argsort had a data race in the inner loop, which VVL caught. But I don't think
this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I
couldn't reproduce it locally and I don't think it's a convincing explanation
of the failures.
zsogitbe pushed a commit to zsogitbe/llama.cpp that referenced this pull request Sep 17, 2026
argsort had a data race in the inner loop, which VVL caught. But I don't think
this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I
couldn't reproduce it locally and I don't think it's a convincing explanation
of the failures.
x1250 pushed a commit to x1250/llama.cpp that referenced this pull request Sep 22, 2026
argsort had a data race in the inner loop, which VVL caught. But I don't think
this was causing failures in practice.

argsort_large has OOB accesses which might explain the failures in CI, but I
couldn't reproduce it locally and I don't think it's a convincing explanation
of the failures.

(cherry picked from commit 481c65f)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ggml changes relating to the ggml tensor library for machine learning merge ready A maintainer can use this label to indicate that they consider the changes final and ready to merge. Vulkan Issues specific to the Vulkan backend

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants