MathFunctions.h 2.7 KB
Newer Older
1
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
Z
zhangjinchao01 已提交
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

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. */

#ifndef MATHFUNCTIONS_H_
#define MATHFUNCTIONS_H_

T
tensor-tang 已提交
18
#ifdef PADDLE_USE_MKLML
19 20 21 22 23
#include <mkl_cblas.h>
#include <mkl_lapacke.h>
#include <mkl_vml_functions.h>
#endif

Z
zhangjinchao01 已提交
24 25
#ifdef PADDLE_USE_MKL
#include <mkl.h>
L
lzhao4ever 已提交
26
#include <mkl_lapacke.h>
L
liaogang 已提交
27 28
#endif

29
#if defined(PADDLE_USE_ATLAS) || defined(PADDLE_USE_VECLIB)
L
lzhao4ever 已提交
30
extern "C" {
L
liaogang 已提交
31
#include <cblas.h>
L
lzhao4ever 已提交
32 33
#include <clapack.h>
}
L
liaogang 已提交
34 35 36 37
#endif

#ifdef PADDLE_USE_OPENBLAS
#include <cblas.h>
L
lzhao4ever 已提交
38 39
#include <lapacke.h>
#endif
L
liaogang 已提交
40 41 42

#ifndef LAPACK_FOUND
extern "C" {
L
liaogang 已提交
43 44 45 46 47 48 49 50 51
#include <cblas.h>
int LAPACKE_sgetrf(
    int matrix_layout, int m, int n, float* a, int lda, int* ipiv);
int LAPACKE_dgetrf(
    int matrix_layout, int m, int n, double* a, int lda, int* ipiv);
int LAPACKE_sgetri(
    int matrix_layout, int n, float* a, int lda, const int* ipiv);
int LAPACKE_dgetri(
    int matrix_layout, int n, double* a, int lda, const int* ipiv);
L
liaogang 已提交
52
}
Z
zhangjinchao01 已提交
53 54
#endif

L
liaogang 已提交
55 56
#include <cmath>

Z
zhangjinchao01 已提交
57 58
namespace paddle {

59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
template <class T>
void gemm(const CBLAS_TRANSPOSE transA,
          const CBLAS_TRANSPOSE transB,
          const int M,
          const int N,
          const int K,
          const T alpha,
          const T* A,
          const int lda,
          const T* B,
          const int ldb,
          const T beta,
          T* C,
          const int ldc);

template <class T>
int getrf(const CBLAS_ORDER Order,
          const int M,
          const int N,
          T* A,
          const int lda,
          int* ipiv);

template <class T>
int getri(
    const CBLAS_ORDER Order, const int N, T* A, const int lda, const int* ipiv);

template <class T>
Z
zhangjinchao01 已提交
87 88
void axpy(const int n, const T alpha, const T* x, T* y);

89
template <class T>
Z
zhangjinchao01 已提交
90 91
T dotProduct(const int n, const T* x, const T* y);

92
template <class T>
Z
zhangjinchao01 已提交
93 94
void vExp(const int n, const T* a, T* r);

95
template <class T>
Z
zhangjinchao01 已提交
96 97
void vPow(const int n, const T* a, const T b, T* r);

98
template <class T>
Z
zhangjinchao01 已提交
99 100
void vLog(const int n, const T* a, T* r);

101
template <class T>
Z
zhangjinchao01 已提交
102 103
void vAdd(const int n, const T* a, const T* b, T* r);

104
template <class T>
Z
zhangjinchao01 已提交
105 106
void vInvSqrt(const int n, const T* a, T* r);

107
template <class T>
Z
zhangjinchao01 已提交
108 109
void vLog1p(const int n, const T* a, T* r);

110
template <class T>
Z
zhangjinchao01 已提交
111 112 113 114 115
void vTanh(const int n, const T* a, T* r);

}  // namespace paddle

#endif  // MATHFUNCTIONS_H_