public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-06 14:31:29 +02:00
Code Issues Packages Projects Releases Wiki Activity

CUDA: fix FA tg at long context for CC >= 8.9 (llama/13852)

Browse Source
This commit is contained in:
Johannes Gäßler
2025-06-01 15:14:44 +03:00
committed by Georgi Gerganov
parent 0035b8527c
commit 9a500394ad
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/fattn-common.cuh
+2 -2
View File
@@ -623,8 +623,8 @@ static __global__ void flash_attn_combine_results(
__builtin_assume(tid < D);
extern __shared__ float2 meta[];
if (tid < 2*parallel_blocks) {
((float *) meta)[threadIdx.x] = ((const float *)VKQ_meta) [blockIdx.z*(2*parallel_blocks) + tid];
for (int i = tid; i < 2*parallel_blocks; i += D) {
((float *) meta)[i] = ((const float *)VKQ_meta) [blockIdx.z*(2*parallel_blocks) + i];
}
__syncthreads();