Skip to content

[Kernel] Add enable_sm120_or_later for SM121 (DGX Spark) CUTLASS support #33517

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
vllm-bot merged 4 commits into from
Feb 7, 2026
+13 −1
Merged

[Kernel] Add enable_sm120_or_later for SM121 (DGX Spark) CUTLASS support #33517

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
9179e0a
[Kernel] Add enable_sm120_or_later for SM121 (DGX Spark) CUTLASS support
Code4me2 Feb 1, 2026
b21d37d
Rename enable_sm120_or_later to enable_sm120_family
Code4me2 Feb 2, 2026
1212876
Merge branch 'main' into fix/sm121-cpp-kernel-support
Code4me2 Feb 6, 2026
b9d5735
Merge branch 'main' into fix/sm121-cpp-kernel-support
Code4me2 Feb 6, 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
11 changes: 11 additions & 0 deletions csrc/cutlass_extensions/common.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,3 +152,14 @@ struct enable_sm120_only : Kernel {
#endif
}
};

// SM12x family includes SM120 (RTX 5090) and SM121 (DGX Spark GB10)
template <typename Kernel>
struct enable_sm120_family : Kernel {
template <typename... Args>
CUTLASS_DEVICE void operator()(Args&&... args) {
#if defined __CUDA_ARCH__ && (__CUDA_ARCH__ >= 1200 && __CUDA_ARCH__ < 1300)
Kernel::operator()(std::forward<Args>(args)...);
#endif
}
};
3 changes: 2 additions & 1 deletion csrc/quantization/w8a8/cutlass/c3x/scaled_mm_blockwise_sm120_fp8_dispatch.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,7 +103,8 @@ struct cutlass_3x_gemm_fp8_blockwise {
MainloopScheduler
>::CollectiveOp;

using KernelType = enable_sm120_only<cutlass::gemm::kernel::GemmUniversal<
// SM12x family to support both SM120 (RTX 5090) and SM121 (DGX Spark)
using KernelType = enable_sm120_family<cutlass::gemm::kernel::GemmUniversal<
Shape<int, int, int, int>, CollectiveMainloop, CollectiveEpilogue>>;

struct GemmKernel : public KernelType {};
Expand Down