CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-18 11:46:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

CUDA: fix race conditions FlashAttention kernels (#13438)

Browse Source
This commit is contained in:
Johannes Gäßler
2025-05-10 22:22:48 +02:00
committed by GitHub
parent d2a4ef05c6
commit 0208355f42
2 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
ggml/src/ggml-cuda/fattn-vec-f16.cuh
View File

@@ -168,6 +168,7 @@ static __global__ void flash_attn_vec_ext_f16(
for (int j = 0; j < ncols; ++j) {
KQ[j*D + tid] = -HALF_MAX_HALF;
}
__syncthreads();
half2 VKQ[ncols] = {{0.0f, 0.0f}};