llama.cpp
opencl: add optimized q8_0 mm kernel for adreno
#18871

Merged

opencl: add optimized q8_0 mm kernel for adreno #18871

lhez merged 5 commits into ggml-org:master from qualcomm:sq/q8_0_mm_opencl_kernel

shaofeiqi requested a review from

lhez 153 days ago

shaofeiqi requested a review from

max-krasnyansky 153 days ago

github-actions added ggml

github-actions added OpenCL

lhez marked this pull request as draft 153 days ago

lhez force pushed from b1b5284b to bcacdc32 144 days ago

lhez force pushed from bcacdc32 to 061bbff4 144 days ago

Add Q8_0 OpenCL kernel

5b320b83

opencl: fix build for non-adreno

8e968277

opencl: refactor q8_0

e15ba0e2

opencl: enforce subgroup size of 64 for adreno for q8_0

7ae603ed

lhez force pushed from 061bbff4 to 7ae603ed 141 days ago

lhez marked this pull request as ready for review 139 days ago

opencl: suppress warning when adreno kernels are disabled

791524c4

lhez approved these changes on 2026-01-30

lhez merged 971facc3 into master 138 days ago

Reviewers

lhez

max-krasnyansky

Assignees

No one assigned

Labels

ggml OpenCL

Milestone

No milestone