vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

try CI fix

Browse source
This commit is contained in:
Johannes Gäßler 2024-07-01 13:30:28 +02:00
parent df9e9c9fcf
commit 30f85eba85
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml/src/ggml-cuda/fattn-common.cuh
View file

@ -58,7 +58,7 @@ static __device__ __forceinline__ T vec_dot_fattn_vec_KQ_q4_0(
const block_q4_0 * K_q4_0 = (const block_q4_0 *) K_c; const block_q4_0 * K_q4_0 = (const block_q4_0 *) K_c;
GGML_UNUSED(Q_v); GGML_UNUSED(Q_v);
half sum = 0.0f; T sum = 0.0f;
#pragma unroll #pragma unroll
for (int k_KQ_0 = 0; k_KQ_0 < D/sizeof(int); k_KQ_0 += WARP_SIZE) { for (int k_KQ_0 = 0; k_KQ_0 < D/sizeof(int); k_KQ_0 += WARP_SIZE) {