Skip to content

cuda: use the transposing concat kernel on all GPUs, not only on GB10 #309

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

Open
sb32445 wants to merge 2 commits into
base: prism
Choose a base branch
from
Open

cuda: use the transposing concat kernel on all GPUs, not only on GB10 #309

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
73f09bd
cuda: use the transposing concat kernel on all GPUs, not only on GB10
sb32445 Oct 2, 2026
97aec1b
cuda: remove the GGML_CUDA_CONCAT_TRANSPOSE switch
sb32445 Oct 3, 2026
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
3 changes: 1 addition & 2 deletions ggml/src/ggml-cuda/concat.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -200,8 +200,7 @@ static void concat_cuda(const ggml_tensor * src0, const ggml_tensor * src1, ggml

dim3 grid_dim(dst->ne[1], dst->ne[2], dst->ne[3]);
if constexpr (sizeof(T) == sizeof(uint32_t)) {
const bool transpose_dim0 = ggml_cuda_info().devices[ggml_cuda_get_device()].cc == GGML_CUDA_CC_DGX_SPARK &&
dim == 0 && src0->ne[2] == 1 && src0->ne[3] == 1 && src1->ne[2] == 1 && src1->ne[3] == 1 &&
const bool transpose_dim0 = dim == 0 && src0->ne[2] == 1 && src0->ne[3] == 1 && src1->ne[2] == 1 && src1->ne[3] == 1 &&
dst->ne[2] == 1 && dst->ne[3] == 1 && src0->ne[0] <= 8 &&
src0->nb[0] == sizeof(uint32_t) && src0->nb[1] == (uint64_t) src0->ne[0]*sizeof(uint32_t) &&
src1->nb[1] == sizeof(uint32_t) && dst->nb[0] == sizeof(uint32_t) &&
Expand Down