This commit is contained in:
Eric Wong
2025-11-09 15:44:33 +08:00
parent a01ab1aabc
commit 5f8a71a864
2 changed files with 11 additions and 3 deletions
@@ -195,7 +195,7 @@ static void sm90_m_grouped_fp8_gemm_contiguous_1d2d(const torch::Tensor& a, cons
MAYBE_LAUNCH(SM90FP8Gemm1D2DRuntime::launch(runtime, args));
}
static void sm90_m_grouped_fp8_gemm_masked_1d2d(const torch::Tensor& a, const torch::Tensor& sfa,
static std::optional<std::pair<int, int>> sm90_m_grouped_fp8_gemm_masked_1d2d(const torch::Tensor& a, const torch::Tensor& sfa,
const torch::Tensor& b, const torch::Tensor& sfb,
const torch::Tensor& d,
const torch::Tensor& masked_m,