happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix case where DV % GGML_F32_EPR !=0

This commit is contained in:
Aman Gupta 2026-02-11 21:44:39 +05:30
parent a1e1420b46
commit 734f76fbc4
1 changed files with 4 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
ggml/src/ggml-cpu/ops.cpp
View File

@ -8420,9 +8420,6 @@ static void ggml_compute_forward_flash_attn_ext_tiled(
static constexpr int Q_TILE_SZ = ggml_fa_tile_config::Q;
static constexpr int KV_TILE_SZ = ggml_fa_tile_config::KV;
#ifdef GGML_SIMD
GGML_ASSERT(DV % GGML_F32_EPR == 0);
#endif
int ir = ir0;
while (ir < ir1) {
@ -8812,12 +8809,14 @@ static void ggml_compute_forward_flash_attn_ext_f16(
const int64_t dr = (nr + nchunk - 1) / nchunk;
static constexpr int64_t Q_TILE_SZ = ggml_fa_tile_config::Q;
const bool use_tiled = !use_ref &&
bool use_tiled = !use_ref &&
(q->type == GGML_TYPE_F32 &&
kv_is_f32_or_f16 &&
k->type == v->type &&
neq1 >= Q_TILE_SZ);
#ifdef GGML_SIMD
use_tiled &= (DV % GGML_F32_EPR == 0);
#endif
int current_chunk = ith;
while (current_chunk < nchunk) {