Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ggml-cuda: add flash-attn support for DKQ=320/DV=256 with ncols2=32 (… (#22286)

Browse Source 
* ggml-cuda: add flash-attn support for DKQ=320/DV=256 with ncols2=32 (GQA=32)

Adds MMA-f16 and tile kernel configs, dispatch logic, template instances,
and tile .cu file for Mistral Small 4 (head sizes 320/256), restricting to
ncols2=32 to support GQA ratio 32 only.

* Adding check to return BEST_FATTN_KERNEL_NONE in case GQA!=32

* Apply suggestions from code review

Address review comments

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* Address review comments and making kernel config default to DQK=512, DV=512 instead of DQK=256,DV=256

* Fixed bug with sinks=1, with ncols=32, there are two warp-groups created but sinks index is same(0,...,15) for both the groups hence with sinks=1, output is not matching with CPU output. Added sink_base which will be base index for each warp_group (threadIdx.y / np)

* Apply suggestions from code review

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* Update ggml/src/ggml-cuda/template-instances/generate_cu_files.py

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
This commit is contained in:
lnigam
2026-04-28 21:37:35 +02:00
committed by GitHub
co-authored by Johannes Gäßler
parent 5d56effdee
commit 7b8443ac78
8 changed files with 86 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/fattn-tile.cu
+4
View File
@@ -38,6 +38,10 @@ void ggml_cuda_flash_attn_ext_tile(ggml_backend_cuda_context & ctx, ggml_tensor
GGML_ASSERT(V->ne[0] == K->ne[0]);
ggml_cuda_flash_attn_ext_tile_case<256, 256>(ctx, dst);
} break;
case 320: {
GGML_ASSERT(V->ne[0] == 256);
ggml_cuda_flash_attn_ext_tile_case<320, 256>(ctx, dst);
} break;
case 512: {
GGML_ASSERT(V->ne[0] == K->ne[0]);
ggml_cuda_flash_attn_ext_tile_case<512, 512>(ctx, dst);