Skip to content

CUDA: fix alignment check for FA #19023

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
JohannesGaessler merged 1 commit into from
Jan 22, 2026
Merged

CUDA: fix alignment check for FA #19023

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions ggml/src/ggml-cuda/fattn.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,7 +46,7 @@ static void ggml_cuda_flash_attn_ext_mma_f16_switch_ncols2(ggml_backend_cuda_con
// are put into the template specialization without GQA optimizations.
bool use_gqa_opt = mask && max_bias == 0.0f && K->ne[1] % FATTN_KQ_STRIDE == 0;
for (const ggml_tensor * t : {Q, K, V, mask}) {
if (t == nullptr) {
if (t == nullptr || ggml_is_quantized(t->type)) {
continue;
}
for (size_t i = 1; i < GGML_MAX_DIMS; ++i) {
Expand Down Expand Up @@ -236,7 +236,7 @@ static best_fattn_kernel ggml_cuda_get_best_fattn_kernel(const int device, const
// The kernel versions without this optimization are also used for ALiBi, if there is no mask, or if the KV cache is not padded,
bool gqa_opt_applies = gqa_ratio % 2 == 0 && mask && max_bias == 0.0f && K->ne[1] % FATTN_KQ_STRIDE == 0;
for (const ggml_tensor * t : {Q, K, V, mask}) {
if (t == nullptr) {
if (t == nullptr || ggml_is_quantized(t->type)) {
continue;
}
for (size_t i = 1; i < GGML_MAX_DIMS; ++i) {
Expand Down
Loading