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
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
|
/* Copyright 2015 The TensorFlow Authors. 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.
==============================================================================*/
#ifndef TENSORFLOW_CORE_KERNELS_AGGREGATE_OPS_CPU_H_
#define TENSORFLOW_CORE_KERNELS_AGGREGATE_OPS_CPU_H_
#include "third_party/eigen3/unsupported/Eigen/CXX11/Tensor"
#include "tensorflow/core/framework/tensor_types.h"
#include "tensorflow/core/kernels/aggregate_ops.h"
typedef Eigen::ThreadPoolDevice CPUDevice;
#ifdef TENSORFLOW_USE_SYCL
typedef Eigen::SyclDevice SYCLDevice;
#endif // TENSORFLOW_USE_SYCL
namespace tensorflow {
// Partial specializations for a CPUDevice, that uses the Eigen implementation
// from AddNEigenImpl.
namespace functor {
template <typename T>
struct Add2Functor<CPUDevice, T> {
void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2) {
Add2EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2);
}
};
template <typename T>
struct Add3Functor<CPUDevice, T> {
void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3) {
Add3EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3);
}
};
template <typename T>
struct Add4Functor<CPUDevice, T> {
void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4) {
Add4EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4);
}
};
template <typename T>
struct Add5Functor<CPUDevice, T> {
void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5) {
Add5EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4, in5);
}
};
template <typename T>
struct Add6Functor<CPUDevice, T> {
void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5,
typename TTypes<T>::ConstFlat in6) {
Add6EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6);
}
};
template <typename T>
struct Add7Functor<CPUDevice, T> {
void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5,
typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7) {
Add7EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7);
}
};
template <typename T>
struct Add8Functor<CPUDevice, T> {
void operator()(
const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1, typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3, typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5, typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7, typename TTypes<T>::ConstFlat in8) {
Add8EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7, in8);
}
};
template <typename T>
struct Add8pFunctor<CPUDevice, T> {
void operator()(
const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1, typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3, typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5, typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7, typename TTypes<T>::ConstFlat in8) {
Add8pEigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7, in8);
}
};
template <typename T>
struct Add9Functor<CPUDevice, T> {
void operator()(
const CPUDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1, typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3, typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5, typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7, typename TTypes<T>::ConstFlat in8,
typename TTypes<T>::ConstFlat in9) {
Add9EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7, in8, in9);
}
};
#ifdef TENSORFLOW_USE_SYCL
// Partial specializations for a SYCLDevice, that uses the Eigen implementation
// from AddNEigenImpl.
template <typename T>
struct Add2Functor<SYCLDevice, T> {
void operator()(const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2) {
Add2EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2);
}
};
template <typename T>
struct Add3Functor<SYCLDevice, T> {
void operator()(const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3) {
Add3EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3);
}
};
template <typename T>
struct Add4Functor<SYCLDevice, T> {
void operator()(const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4) {
Add4EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4);
}
};
template <typename T>
struct Add5Functor<SYCLDevice, T> {
void operator()(const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5) {
Add5EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4, in5);
}
};
template <typename T>
struct Add6Functor<SYCLDevice, T> {
void operator()(const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5,
typename TTypes<T>::ConstFlat in6) {
Add6EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6);
}
};
template <typename T>
struct Add7Functor<SYCLDevice, T> {
void operator()(const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1,
typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3,
typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5,
typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7) {
Add7EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7);
}
};
template <typename T>
struct Add8Functor<SYCLDevice, T> {
void operator()(
const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1, typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3, typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5, typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7, typename TTypes<T>::ConstFlat in8) {
Add8EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7, in8);
}
};
template <typename T>
struct Add8pFunctor<SYCLDevice, T> {
void operator()(
const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1, typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3, typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5, typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7, typename TTypes<T>::ConstFlat in8) {
Add8pEigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7, in8);
}
};
template <typename T>
struct Add9Functor<SYCLDevice, T> {
void operator()(
const SYCLDevice& d, typename TTypes<T>::Flat out,
typename TTypes<T>::ConstFlat in1, typename TTypes<T>::ConstFlat in2,
typename TTypes<T>::ConstFlat in3, typename TTypes<T>::ConstFlat in4,
typename TTypes<T>::ConstFlat in5, typename TTypes<T>::ConstFlat in6,
typename TTypes<T>::ConstFlat in7, typename TTypes<T>::ConstFlat in8,
typename TTypes<T>::ConstFlat in9) {
Add9EigenImpl<SYCLDevice, T>::Compute(d, out, in1, in2, in3, in4, in5, in6,
in7, in8, in9);
}
};
#endif // TENSORFLOW_USE_SYCL
} // namespace functor
} // namespace tensorflow
#endif // TENSORFLOW_CORE_KERNELS_AGGREGATE_OPS_CPU_H_
|
