forked from PaddlePaddle/Paddle
/
elementwise_kernel.cu
151 lines (142 loc) · 5.04 KB
/
elementwise_kernel.cu
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
// Copyright (c) 2022 PaddlePaddle 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.
#include "paddle/phi/backends/gpu/gpu_context.h"
#ifndef PADDLE_WITH_XPU_KP
#include "paddle/phi/common/complex.h"
#include "paddle/phi/common/float16.h"
#endif
#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/impl/elementwise_kernel_impl.h"
namespace phi {
// Create the definition of Maximum
DEFINE_CUDA_ELEMENTWISE_OP(Maximum)
template <typename T, typename Context>
void MaximumKernel(const Context& dev_ctx,
const DenseTensor& x,
const DenseTensor& y,
DenseTensor* out) {
int axis = -1;
MaximumRawKernel<T>(dev_ctx, x, y, axis, out);
}
// Create the definition of Minimum
DEFINE_CUDA_ELEMENTWISE_OP(Minimum)
template <typename T, typename Context>
void MinimumKernel(const Context& dev_ctx,
const DenseTensor& x,
const DenseTensor& y,
DenseTensor* out) {
int axis = -1;
MinimumRawKernel<T>(dev_ctx, x, y, axis, out);
}
// Create the definition of Modulo
DEFINE_CUDA_ELEMENTWISE_OP(Modulo)
// Create the definition of FloorDivide
DEFINE_CUDA_ELEMENTWISE_OP(FloorDivide)
template <typename T, typename Context>
void FloorDivideKernel(const Context& dev_ctx,
const DenseTensor& x,
const DenseTensor& y,
DenseTensor* out) {
int axis = -1;
FloorDivideRawKernel<T>(dev_ctx, x, y, axis, out);
}
// Create the definition of Heaviside
DEFINE_CUDA_ELEMENTWISE_OP(ElementwiseHeaviside)
// Create the definition of Pow
DEFINE_CUDA_ELEMENTWISE_OP(ElementwisePow)
template <typename T, typename Context>
void ElementwisePowKernel(const Context& dev_ctx,
const DenseTensor& x,
const DenseTensor& y,
DenseTensor* out) {
int axis = -1;
ElementwisePowRawKernel<T>(dev_ctx, x, y, axis, out);
}
} // namespace phi
#ifdef PADDLE_WITH_XPU_KP
PD_REGISTER_KERNEL(maximum, KPS, ALL_LAYOUT, phi::MaximumKernel, float) {}
PD_REGISTER_KERNEL(maximum_raw, KPS, ALL_LAYOUT, phi::MaximumRawKernel, float) {
}
PD_REGISTER_KERNEL(minimum, KPS, ALL_LAYOUT, phi::MinimumKernel, float) {}
PD_REGISTER_KERNEL(minimum_raw, KPS, ALL_LAYOUT, phi::MinimumRawKernel, float) {
}
PD_REGISTER_KERNEL(floor_divide, KPS, ALL_LAYOUT, phi::FloorDivideKernel, int) {
}
PD_REGISTER_KERNEL(
floor_divide_raw, KPS, ALL_LAYOUT, phi::FloorDivideRawKernel, int) {}
PD_REGISTER_KERNEL(
elementwise_pow, KPS, ALL_LAYOUT, phi::ElementwisePowKernel, float) {}
PD_REGISTER_KERNEL(
elementwise_pow_raw, KPS, ALL_LAYOUT, phi::ElementwisePowRawKernel, float) {
}
#else
using float16 = phi::dtype::float16;
using bfloat16 = phi::dtype::bfloat16;
using complex64 = ::phi::dtype::complex<float>;
using complex128 = ::phi::dtype::complex<double>;
PD_REGISTER_KERNEL(
fmax, KPS, ALL_LAYOUT, phi::FMaxKernel, float, double, int, int64_t) {}
PD_REGISTER_KERNEL(
fmin, KPS, ALL_LAYOUT, phi::FMinKernel, float, double, int, int64_t) {}
PD_REGISTER_KERNEL(maximum_raw,
KPS,
ALL_LAYOUT,
phi::MaximumRawKernel,
float,
double,
int,
int64_t,
float16,
bfloat16) {}
PD_REGISTER_KERNEL(minimum_raw,
KPS,
ALL_LAYOUT,
phi::MinimumRawKernel,
float,
double,
int,
int64_t,
float16,
bfloat16) {}
PD_REGISTER_KERNEL(modulo_raw,
KPS,
ALL_LAYOUT,
phi::ModuloRawKernel,
float,
double,
int,
int64_t) {}
PD_REGISTER_KERNEL(floor_divide_raw,
KPS,
ALL_LAYOUT,
phi::FloorDivideRawKernel,
int,
int64_t) {}
PD_REGISTER_KERNEL(elementwise_heaviside_raw,
KPS,
ALL_LAYOUT,
phi::ElementwiseHeavisideRawKernel,
float,
double,
int,
int64_t) {}
PD_REGISTER_KERNEL(elementwise_pow_raw,
KPS,
ALL_LAYOUT,
phi::ElementwisePowRawKernel,
float,
double,
int,
int64_t) {}
#endif