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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
|
// Copyright 2004-present Facebook. All Rights Reserved.
#pragma once
#include "caffe2/core/context.h"
#include "caffe2/utils/math.h"
namespace caffe2 {
void custom_fp16_gemm(
const int m,
const int k,
const int n,
const float* A_fp16,
const float* B_fp16,
const float beta,
float* C,
const bool use_acc_fp16,
const bool use_temp_accumulator);
void custom_fp16_gemm_with_trans(
const CBLAS_TRANSPOSE trans_A,
const CBLAS_TRANSPOSE trans_B,
const int m,
const int k,
const int n,
const float* A_fp16,
const float* B_fp16,
const float beta,
float* C,
const bool use_acc_fp16,
const bool use_temp_accumulator);
void transpose(const float* A, float* A_trans, int M, int N);
void custom_fp16_gemv(
const bool use_acc_fp16,
const bool use_custom_acc32,
const bool use_temp_accumulator,
const CBLAS_TRANSPOSE trans_A,
const int M,
const int N,
const float alpha,
const float* A,
const float* x,
const float beta,
float* y,
CPUContext* context);
void custom_fp16_gemm_batched(
const bool use_acc_fp16,
const bool use_custom_acc32,
const bool use_temp_accumulator,
const CBLAS_TRANSPOSE trans_A,
const CBLAS_TRANSPOSE trans_B,
const int batch_size,
const int M,
const int N,
const int K,
const float alpha,
const float** A,
const float** B,
const float beta,
float** C,
CPUContext* context);
void custom_fp16_gemm_strided_batched(
const bool use_acc_fp16,
const bool use_custom_acc32,
const bool use_temp_accumulator,
const CBLAS_TRANSPOSE trans_A,
const CBLAS_TRANSPOSE trans_B,
const int batch_size,
const int M,
const int N,
const int K,
const float alpha /* unused */,
const float* A,
const int A_stride,
const float* B,
const int B_stride,
const float beta,
float* C,
const int C_stride,
CPUContext* context);
} // namespace caffe2
|
