vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 09:34:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'upstream' into concedo_experimental

Browse source 
# Conflicts:
#	.github/actions/windows-setup-curl/action.yml
#	.github/workflows/build-linux-cross.yml
#	README.md
#	common/CMakeLists.txt
#	examples/parallel/README.md
#	examples/parallel/parallel.cpp
#	ggml/src/ggml-sycl/element_wise.cpp
#	ggml/src/ggml-vulkan/CMakeLists.txt
#	tools/server/README.md
This commit is contained in:
Concedo 2025-05-18 23:27:53 +08:00
commit 59300dbdf5
25 changed files with 694 additions and 550 deletions
Download patch file Download diff file Expand all files Collapse all files

7
ggml/src/ggml-vulkan/ggml-vulkan.cpp
View file

@ -5896,10 +5896,17 @@ static void ggml_vk_flash_attn(ggml_backend_vk_context * ctx, vk_context& subctx
vk_pipeline *pipelines;
bool small_rows = N <= get_fa_num_small_rows(path);
// coopmat1 does not actually support "small rows" (it needs 16 rows).
// So use scalar instead.
if (small_rows && path == FA_COOPMAT1) {
path = FA_SCALAR;
}
// scalar is faster than coopmat2 when N==1
if (N == 1 && path == FA_COOPMAT2) {
path = FA_SCALAR;
}
bool f32acc = path == FA_SCALAR || dst->op_params[3] == GGML_PREC_F32;
switch (path) {