Public/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-19 17:24:57 +02:00
Code Issues Packages Projects Releases Wiki Activity

sycl: parallelize the non-contiguous concat kernel

Browse Source 
Launch geometry only: the non-contiguous concat kernel launched a single-lane
work-group (1, 1, 1), now it will launch a (1, 1, SYCL_CONCAT_BLOCK_SIZE) one.

SYCL_CONCAT_BLOCK_SIZE is defined in `ggml/src/ggml-sycl/presets.hpp`.

llama-bench (Arc Pro B70, Qwen3.6-27B-UD-Q4_K_XL, -fa on, q8_0 KV),
on top of upstream master: pp2048 920 -> 1006 t/s (+9.4%)
This commit is contained in:
Simon Gardling
2026-07-17 18:20:24 -04:00
parent 6bdd77f13c
commit dcc5f72053
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-sycl/concat.cpp
+2 -1
View File
@@ -127,7 +127,8 @@ static void concat_T_sycl_non_cont(
int64_t ne2, int64_t ne3, uint64_t nb0, uint64_t nb1, uint64_t nb2,
uint64_t nb3, int32_t dim) {
sycl::range<3> gridDim(ne3, ne2, ne1);
stream->parallel_for(sycl::nd_range<3>(gridDim, sycl::range<3>(1, 1, 1)), [=](sycl::nd_item<3> item_ct1) {
sycl::range<3> blockDim(1, 1, SYCL_CONCAT_BLOCK_SIZE);
stream->parallel_for(sycl::nd_range<3>(gridDim * blockDim, blockDim), [=](sycl::nd_item<3> item_ct1) {
int64_t i3 = item_ct1.get_group(0);
int64_t i2 = item_ct1.get_group(1);
int64_t i1 = item_ct1.get_group(2);