[Kernel] CUTLASS grouped gemm fp8 MoE kernel (#13972)

Signed-off-by: ElizaWszola <eliza@neuralmagic.com>
Signed-off-by: ElizaWszola <ewszola@redhat.com>
Co-authored-by: Lucas Wilkinson <wilkinson.lucas@gmail.com>
This commit is contained in:
ElizaWszola
2025-03-27 00:54:44 +00:00
committed by GitHub
co-authored by Lucas Wilkinson
parent 7a6d45bc8a
commit 9239bf718e
22 changed files with 2317 additions and 15 deletions
+11 -1
View File
@@ -48,4 +48,14 @@ struct enable_sm90_or_later : Kernel {
Kernel::operator()(std::forward<Args>(args)...);
#endif
}
};
};
template <typename Kernel>
struct enable_sm90_only : Kernel {
template <typename... Args>
CUTLASS_DEVICE void operator()(Args&&... args) {
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ == 900
Kernel::operator()(std::forward<Args>(args)...);
#endif
}
};