llama.cpp
opencl: add several gemm variants for small batches and some fixes
#30850
Merged

opencl: add several gemm variants for small batches and some fixes #30850

lhez
lhez opencl: fix OOB read in `mul_mv_mxfp4_f32_flat`
1918a054
lhez opencl: remove large-N guard that forces CPU weight placement
79e5c003
lhez opencl: cap num of subgroups by workgroup size for q4_0 mc3
c51d44f1
lhez opencl: add small batch variant for q4_k dp4a GEMM
59d22dc1
lhez opencl: add a narrow variant of mul_mm_f16_f32_l4_lm
e1cc117d
lhez opencl: add a one subgroup variant for gemv_noshuffle_q1_0_f32 for big M
43e1f804
github-actions github-actions added ggml
github-actions github-actions added OpenCL
lhez lhez marked this pull request as ready for review 31 days ago
lhez lhez requested a review 31 days ago
lhez
lhez lhez added merge ready
max-krasnyansky
max-krasnyansky approved these changes on 2026-10-20
ggerganov ggerganov merged a01ce9ff into master 31 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone