]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
opencl: improve get_rows, cpy, concat and q6_k flat gemv (#24160)
authorlhez <redacted>
Fri, 5 Jun 2026 20:45:25 +0000 (13:45 -0700)
committerGitHub <redacted>
Fri, 5 Jun 2026 20:45:25 +0000 (13:45 -0700)
commit308f61c31f083251ce8150f10b9ef97679b500b5
tree9c9ea110f2cdc3a4d88dd60fbcc5788108e08714
parentda87e9b612592ccd2a2e783eff631346459b9de0
opencl: improve get_rows, cpy, concat and q6_k flat gemv (#24160)

* opencl: allow multiple workgroups for large rows

* opencl: improve small cpy

* opencl: packed concat for small input

* opencl: tweak flat q6_K gemv, increase N_DST and remap threads
ggml/src/ggml-opencl/ggml-opencl.cpp
ggml/src/ggml-opencl/kernels/concat.cl
ggml/src/ggml-opencl/kernels/cpy.cl
ggml/src/ggml-opencl/kernels/get_rows.cl
ggml/src/ggml-opencl/kernels/mul_mv_q6_k_f32_flat.cl