Skip to content

opencl: refactor q8_0 set_tensor and mul_mat host side dispatch for Adreno#21938

Merged
lhez merged 3 commits intoggml-org:masterfrom
qualcomm:lh/q8_0-mm-host-refactor
Apr 17, 2026
Merged

opencl: refactor q8_0 set_tensor and mul_mat host side dispatch for Adreno#21938
lhez merged 3 commits intoggml-org:masterfrom
qualcomm:lh/q8_0-mm-host-refactor

Conversation

@lhez
Copy link
Copy Markdown
Contributor

@lhez lhez commented Apr 15, 2026

Overview

The q8_0 set_tensor and mul_mat host side dispatch code for Adreno is a bit messy. This PR does some refactoring to make it cleaner and follow the same pattern as more recently added quantizations, e.g., q4_1, etc.

Additional information

Requirements

@github-actions github-actions bot added ggml changes relating to the ggml tensor library for machine learning OpenCL Issues specific to the OpenCL backend labels Apr 15, 2026
@lhez lhez marked this pull request as ready for review April 15, 2026 22:20
@lhez lhez requested a review from a team as a code owner April 15, 2026 22:20
@lhez
Copy link
Copy Markdown
Contributor Author

lhez commented Apr 17, 2026

Failures are irrelevant - they are with webgpu.

@lhez
Copy link
Copy Markdown
Contributor Author

lhez commented Apr 17, 2026

@ggml-org/maintainers - can I get a second approval?

@lhez lhez merged commit 5e6c0e1 into ggml-org:master Apr 17, 2026
127 of 140 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 OpenCL Issues specific to the OpenCL backend

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants