RYDE-WORK/llama.cpp
1
0
Fork 0
mirror of https://github.com/RYDE-WORK/llama.cpp.git synced 2026-02-01 21:23:14 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

CUDA: fix shared memory access condition for mmv (#10740)

Browse Source
This commit is contained in:
Johannes Gäßler 2024-12-09 20:07:12 +01:00 committed by GitHub
parent c37fb4cf62
commit 26a8406ba9
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-cuda/mmv.cu
View File

@ -57,7 +57,7 @@ static __global__ void mul_mat_vec(
if (block_size > WARP_SIZE) { if (block_size > WARP_SIZE) {
buf_iw[tid/WARP_SIZE] = sumf; buf_iw[tid/WARP_SIZE] = sumf;
__syncthreads(); __syncthreads();
if (tid > WARP_SIZE) { if (tid >= WARP_SIZE) {
return; return;
} }
sumf = buf_iw[tid]; sumf = buf_iw[tid];