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

Bc 4 for scalar FA is not a valid configuration

This commit is contained in:
Ruben Ortlam 2026-02-12 16:38:57 +01:00
parent cd54ba2b86
commit 16cb912442
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ggml/src/ggml-vulkan/ggml-vulkan.cpp
View File

@ -2765,11 +2765,9 @@ static constexpr uint32_t flash_attention_num_small_rows = 32;
static uint32_t get_fa_scalar_num_rows(uint32_t hsk, uint32_t hsv, FaRows rows, bool small_cache) {
if (rows == FA_ROWS_1) {
return 1;
} else if (rows == FA_ROWS_SMALL) {
return 4;
}
if (hsv >= 192 || (hsv | hsk) & 8 || small_cache) {
if (rows == FA_ROWS_SMALL || hsv >= 192 || (hsv | hsk) & 8 || small_cache) {
return 8;
}