nvfp4_scaled_mm_entry.cu

/* Copyright 2025 SGLang Team. All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/

#include <torch/all.h>

#if defined ENABLE_NVFP4 && ENABLE_NVFP4
void cutlass_scaled_fp4_mm_sm100a_sm120a(
    torch::Tensor& D,
    torch::Tensor const& A,
    torch::Tensor const& B,
    torch::Tensor const& A_sf,
    torch::Tensor const& B_sf,
    torch::Tensor const& alpha);

// SM120 specific dispatch functions
void cutlass_fp4_bf16_gemm_dispatch_sm120(
    torch::Tensor& D,
    torch::Tensor const& A,
    torch::Tensor const& B,
    torch::Tensor const& A_sf,
    torch::Tensor const& B_sf,
    torch::Tensor const& alpha,
    int m,
    int n,
    int k,
    cudaStream_t stream);

void cutlass_fp4_f16_gemm_dispatch_sm120(
    torch::Tensor& D,
    torch::Tensor const& A,
    torch::Tensor const& B,
    torch::Tensor const& A_sf,
    torch::Tensor const& B_sf,
    torch::Tensor const& alpha,
    int m,
    int n,
    int k,
    cudaStream_t stream);
#endif

void cutlass_scaled_fp4_mm(
    torch::Tensor& D,
    torch::Tensor const& A,
    torch::Tensor const& B,
    torch::Tensor const& A_sf,
    torch::Tensor const& B_sf,
    torch::Tensor const& alpha) {
#if defined ENABLE_NVFP4 && ENABLE_NVFP4
  return cutlass_scaled_fp4_mm_sm100a_sm120a(D, A, B, A_sf, B_sf, alpha);
#endif
  TORCH_CHECK_NOT_IMPLEMENTED(false, "No compiled nvfp4 mm kernel.");
}