Skip to content

sycl: Hotfix for non dnnl codepath #14677

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 14, 2025

Conversation

ShanoToni
Copy link
Contributor

PR #14617 introduced new logic for dnnl codepath for batched mulmats to improve broadcast logic and support strided data. Due to this the use of get_to_fp16_nc_sycl which has overhead compared to get_to_fp16_sycl.

The non DNNL codepath however still does not support non continuous src1 without the nc_sycl function which caused failures when not running DNNL.

This PR ensures correct behaviour.

Copy link
Collaborator

@Alcpz Alcpz left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the quick fix. I don't mind the split codepath since we aim to remove MKL in the future.

@github-actions github-actions bot added ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language labels Jul 14, 2025
@ShanoToni ShanoToni force-pushed the sycl_batched_mul_mat_fix branch from 942a18d to b93e848 Compare July 14, 2025 14:18
@Alcpz Alcpz merged commit bdca383 into ggml-org:master Jul 14, 2025
45 of 48 checks passed
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 SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants