PR #19054 opencl: add flattened q6_K mv

opencl: add flattened q6_K mv #19054

lhez merged 11 commits into ggml-org:master from qualcomm:lh/q6_k-flat

opencl: flatten `q6_K` and add `kernel_mul_mv_q6_K_f32_flat`

5ad71d4e

opencl: clean up

4cd7a356

opencl: refactor q6_K mv - put loop body in `block_q_6_K_dot_y_flat`

a1383d53

opencl: tweak the workgroup size a bit

635d7452

opencl: output 4 values per subgroup for `kernel_mul_mv_q6_K_f32_flat`

d8bb2f6e

opencl: proper alignment for q6_K

28ff8afb

opencl: boundary handling for flattened q6_K mv

10aa97e1

opencl: rename q6_K mv kernel file

cd7667e9

opencl: put flattened q6_K mv in its own file

46fd3fbc

opencl: use lower k in file name

9a7ec6b3

opencl: use K in variable names

6fa04dad

github-actions added ggml

github-actions added OpenCL

lhez marked this pull request as ready for review 182 days ago

lhez requested a review from

max-krasnyansky 182 days ago

max-krasnyansky approved these changes on 2026-01-25

lhez merged 94eeb596 into master 179 days ago

Reviewers

max-krasnyansky

Assignees

No one assigned

Labels

ggml OpenCL

Milestone

No milestone