nvfp4_quant_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 scaled_fp4_quant_sm100a(
    torch::Tensor& output, torch::Tensor const& input, torch::Tensor& output_sf, torch::Tensor const& input_sf);

void scaled_fp4_experts_quant_sm100a(
    torch::Tensor& output,
    torch::Tensor& output_scale,
    torch::Tensor const& input,
    torch::Tensor const& input_global_scale,
    torch::Tensor const& input_offset_by_experts,
    torch::Tensor const& output_scale_offset_by_experts);

void silu_and_mul_scaled_fp4_experts_quant_sm100a(
    torch::Tensor& output,
    torch::Tensor& output_scale,
    torch::Tensor const& input,
    torch::Tensor const& input_global_scale,
    torch::Tensor const& mask,
    bool use_silu_and_mul);

#endif

void scaled_fp4_quant(
    torch::Tensor& output, torch::Tensor const& input, torch::Tensor& output_sf, torch::Tensor const& input_sf) {
#if defined ENABLE_NVFP4 && ENABLE_NVFP4
  return scaled_fp4_quant_sm100a(output, input, output_sf, input_sf);
#endif
  TORCH_CHECK_NOT_IMPLEMENTED(false, "No compiled nvfp4 quantization");
}

void scaled_fp4_experts_quant(
    torch::Tensor& output,
    torch::Tensor& output_scale,
    torch::Tensor const& input,
    torch::Tensor const& input_global_scale,
    torch::Tensor const& input_offset_by_experts,
    torch::Tensor const& output_scale_offset_by_experts) {
#if defined ENABLE_NVFP4 && ENABLE_NVFP4
  return scaled_fp4_experts_quant_sm100a(
      output, output_scale, input, input_global_scale, input_offset_by_experts, output_scale_offset_by_experts);
#endif
  TORCH_CHECK_NOT_IMPLEMENTED(false, "No compiled nvfp4 experts quantization kernel");
}

void silu_and_mul_scaled_fp4_experts_quant(
    torch::Tensor& output,
    torch::Tensor& output_scale,
    torch::Tensor const& input,
    torch::Tensor const& input_global_scale,
    torch::Tensor const& mask,
    bool use_silu_and_mul) {
#if defined ENABLE_NVFP4 && ENABLE_NVFP4
  return silu_and_mul_scaled_fp4_experts_quant_sm100a(
      output, output_scale, input, input_global_scale, mask, use_silu_and_mul);
#endif
  TORCH_CHECK_NOT_IMPLEMENTED(false, "No compiled nvfp4 experts quantization kernel");
}