File: fused_kernel.h

package info (click to toggle)
pytorch 1.7.1-7
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 80,340 kB
  • sloc: cpp: 670,830; python: 343,991; ansic: 67,845; asm: 5,503; sh: 2,924; java: 2,888; xml: 266; makefile: 244; ruby: 148; yacc: 144; objc: 51; lex: 44
file content (45 lines) | stat: -rw-r--r-- 1,120 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
#pragma once

#include <ATen/ATen.h>
#include <torch/csrc/WindowsTorchApiMacro.h>
#include <torch/csrc/jit/codegen/fuser/fused_kernel.h>
#include <torch/csrc/utils/disallow_copy.h>

#include <cstdint>
#include <memory>
#include <string>

namespace torch {
namespace jit {
namespace fuser {
namespace cpu {

// Represents a compiled CPU kernel and the metadata necessary to run it
struct TORCH_API FusedKernelCPU : public ::torch::jit::fuser::FusedKernel {
  FusedKernelCPU(
      std::string name,
      std::string code,
      std::vector<TensorDesc> input_desc,
      std::vector<TensorDesc> output_desc,
      std::vector<PartitionDesc> chunk_desc,
      std::vector<PartitionDesc> concat_desc,
      bool has_random);

  at::Backend backend() const override {
    return at::Backend::CPU;
  }

  void launch_raw(const uint32_t numel, std::vector<void*>& arguments)
      const override {
    kernel(numel, arguments.data());
  }

 private:
  std::unique_ptr<at::DynamicLibrary> so_lib;
  void (*kernel)(uint32_t, void**) = nullptr;
};

} // namespace cpu
} // namespace fuser
} // namespace jit
} // namespace torch