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
|
/*******************************************************************************
* Copyright 2016-2018 Intel Corporation
*
* 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 TEST_CONVOLUTION_BACKWARD_DATA_COMMON_H
#define TEST_CONVOLUTION_BACKWARD_DATA_COMMON_H
#include "mkldnn_test_common.hpp"
#include "gtest/gtest.h"
#include "mkldnn.hpp"
namespace mkldnn {
template <typename data_t_diff_dst, typename data_t_wei,
typename data_t_acc, typename data_t_diff_src>
void compute_ref_conv_bwd_data(const test_convolution_sizes_t &c,
const memory &diff_src, const memory &weights, const memory &diff_dst)
{
data_t_diff_dst *diff_dst_data = (data_t_diff_dst *)diff_dst.get_data_handle();
data_t_wei *weights_data = (data_t_wei *)weights.get_data_handle();
data_t_diff_src *diff_src_data = (data_t_diff_src *)diff_src.get_data_handle();
const memory::desc diff_src_d = diff_src.get_primitive_desc().desc();
const memory::desc weights_d = weights.get_primitive_desc().desc();
const memory::desc diff_dst_d = diff_dst.get_primitive_desc().desc();
size_t padded_ic = diff_src_d.data.layout_desc.blocking.padding_dims[1];
size_t padded_oc = diff_dst_d.data.layout_desc.blocking.padding_dims[1];
mkldnn::impl::parallel_nd(c.mb, c.ng, c.ic / c.ng, c.ih, c.iw,
[&](int mb, int g, int ic, int ih, int iw) {
size_t sidx = mb * padded_ic * c.ih * c.iw
+ g * padded_ic / c.ng * c.ih * c.iw
+ ic * c.ih * c.iw + ih * c.iw + iw;
data_t_acc a = data_t_acc(0);
for (int oc = 0; oc < c.oc / c.ng; oc++) {
for (int kh = 0; kh < c.kh; kh++) {
for (int kw = 0; kw < c.kw; kw++) {
if (iw + c.padw < kw * (1 + c.dilw)
|| ih + c.padh < kh * (1 + c.dilh))
continue;
int ow = iw - kw * (1 + c.dilw) + c.padw;
int oh = ih - kh * (1 + c.dilh) + c.padh;
if (ow % c.strw != 0 || oh % c.strh != 0)
continue;
ow /= c.strw;
oh /= c.strh;
if (oh < c.oh && ow < c.ow) {
size_t didx = mb * padded_oc * c.oh * c.ow
+ g * padded_oc / c.ng * c.oh * c.ow
+ oc * c.oh * c.ow + oh * c.ow + ow;
size_t widx =
g * padded_oc / c.ng * padded_ic
/ c.ng * c.kh * c.kw
+ oc * padded_ic / c.ng * c.kh * c.kw
+ ic * c.kh * c.kw + kh * c.kw + kw;
a += (data_t_acc)(
diff_dst_data[map_index(diff_dst_d, didx)]
* weights_data[map_index(weights_d, widx)]);
}
}
}
}
diff_src_data[map_index(diff_src_d, sidx)] = (data_t_diff_src)a;
});
}
template <typename data_t_diff_dst, typename data_t_wei,
typename data_t_acc, typename data_t_diff_src>
class convolution_backward_data_test
: public ::testing::TestWithParam<test_convolution_params_t> {
protected:
virtual void SetUp() {
auto p = ::testing::TestWithParam<test_convolution_params_t>::GetParam();
catch_expected_failures([=](){Test();}, p.expect_to_fail,
p.expected_status);
}
void Test() {
auto p = ::testing::TestWithParam<test_convolution_params_t>::GetParam();
ASSERT_TRUE(p.engine_kind == engine::kind::cpu);
ASSERT_EQ(p.aalgorithm, convolution_direct);
auto eng = engine(p.engine_kind, 0);
auto data_type_diff_src = data_traits<data_t_diff_src>::data_type;
auto data_type_diff_dst = data_traits<data_t_diff_dst>::data_type;
auto data_type_wei = data_traits<data_t_wei>::data_type;
test_convolution_sizes_t cd = p.sizes;
auto c_src_desc = create_md({ cd.mb, cd.ic, cd.ih, cd.iw },
data_type_diff_src, p.formats.src_format);
auto c_weights_desc = cd.ng > 1
? create_md({ cd.ng, cd.oc / cd.ng, cd.ic / cd.ng, cd.kh, cd.kw },
data_type_wei, p.formats.weights_format)
: create_md({ cd.oc, cd.ic, cd.kh, cd.kw },
data_type_wei, p.formats.weights_format);
auto c_dst_desc = create_md({ cd.mb, cd.oc, cd.oh, cd.ow },
data_type_diff_dst, p.formats.dst_format);
auto c_src_desc_f = create_md({ cd.mb, cd.ic, cd.ih, cd.iw },
data_type_diff_dst, p.formats.src_format);
auto c_dst_desc_f = create_md({ cd.mb, cd.oc, cd.oh, cd.ow },
data_type_diff_src, p.formats.dst_format);
auto c_diff_src = test_memory(c_src_desc, eng);
auto c_weights = test_memory(c_weights_desc, eng);
auto c_diff_dst = test_memory(c_dst_desc, eng);
std::vector<int> padR = {
right_padding(cd.ih, cd.oh, cd.kh, cd.padh, cd.strh, cd.dilh),
right_padding(cd.iw, cd.ow, cd.kw, cd.padw, cd.strw, cd.dilw)
};
// Only true for dense format
fill_data<data_t_wei>(c_weights.get_size() / sizeof(data_t_wei),
(data_t_wei *)c_weights.get().get_data_handle());
fill_data<data_t_diff_dst>(
c_diff_dst.get_size() / sizeof(data_t_diff_dst),
(data_t_diff_dst *)c_diff_dst.get().get_data_handle());
fill_data<data_t_diff_src>(
c_diff_src.get_size() / sizeof(data_t_diff_src),
(data_t_diff_src *)c_diff_src.get().get_data_handle());
check_zero_tail<data_t_diff_dst>(1, c_diff_dst.get());
check_zero_tail<data_t_wei>(1, c_weights.get());
check_zero_tail<data_t_diff_src>(1, c_diff_src.get());
auto conv_desc = convolution_forward::desc(
prop_kind::forward_training, p.aalgorithm, c_src_desc_f,
c_weights_desc, c_dst_desc_f,
{ cd.strh, cd.strw }, { cd.dilh, cd.dilw },
{ cd.padh, cd.padw }, padR, padding_kind::zero);
auto conv_primitive_desc = convolution_forward::primitive_desc(
conv_desc, eng);
auto conv_bwd_data_desc = convolution_backward_data::desc(
p.aalgorithm, c_src_desc, c_weights_desc, c_dst_desc,
{ cd.strh, cd.strw }, { cd.dilh, cd.dilw },
{ cd.padh, cd.padw }, padR, padding_kind::zero);
auto conv_bwd_data_primitive_desc
= convolution_backward_data::primitive_desc(
conv_bwd_data_desc, eng, conv_primitive_desc);
auto conv_bwd_data = convolution_backward_data(
conv_bwd_data_primitive_desc,
c_diff_dst.get(), c_weights.get(), c_diff_src.get());
std::vector<primitive> pipeline;
pipeline.push_back(conv_bwd_data);
stream(stream::kind::lazy).submit(pipeline).wait();
auto ref_memory = memory(memory::primitive_desc(c_src_desc, eng));
compute_ref_conv_bwd_data
<data_t_diff_dst, data_t_wei, data_t_acc, data_t_diff_src>(
cd, ref_memory, c_weights.get(), c_diff_dst.get());
check_zero_tail<data_t_diff_src>(1, ref_memory);
compare_data<data_t_diff_src>(ref_memory, c_diff_src.get());
check_zero_tail<data_t_diff_src>(0, c_diff_src.get());
}
};
}
#endif
|
