llama.cpp
opencl: Q6_K GEMM/GEMV fix for ne01 of weights that are not multiples of 128.
#25464

Merged

opencl: Q6_K GEMM/GEMV fix for ne01 of weights that are not multiples of 128. #25464

max-krasnyansky merged 3 commits into ggml-org:master from qualcomm:hq/adreno-q6k-smallk-fix

opencl: fix garbled output for Q6_K weights with ne01 % 128 != 0 on A…

3ff664d2

opencl: reserve alignment slack for the SOA subbuffer carve in alloc …

f7eda48a

opencl: use lm based q6_k mm when ne1 is not multiple of 128

0f2bce23

wanghqc requested a review 55 days ago

github-actions added ggml

github-actions added OpenCL

lhez approved these changes on 2026-07-08

lhez requested a review from

max-krasnyansky 55 days ago

max-krasnyansky approved these changes on 2026-07-08

max-krasnyansky merged 92366df3 into master 54 days ago

Reviewers

max-krasnyansky

lhez

Assignees

No one assigned

Labels

ggml OpenCL

Milestone

No milestone