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
|
/*
* Copyright 2011 The LibYuv Project Authors. All rights reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include <stdlib.h>
#include <time.h>
#include "libyuv/cpu_id.h"
#include "libyuv/scale.h"
#include "../unit_test/unit_test.h"
namespace libyuv {
// Test scaling with C vs Opt and return maximum pixel difference. 0 = exact.
static int TestFilter(int src_width, int src_height,
int dst_width, int dst_height,
FilterMode f, int benchmark_iterations) {
int i, j;
const int b = 128;
int src_width_uv = (Abs(src_width) + 1) >> 1;
int src_height_uv = (Abs(src_height) + 1) >> 1;
int src_y_plane_size = (Abs(src_width) + b * 2) * (Abs(src_height) + b * 2);
int src_uv_plane_size = (src_width_uv + b * 2) * (src_height_uv + b * 2);
int src_stride_y = b * 2 + Abs(src_width);
int src_stride_uv = b * 2 + src_width_uv;
align_buffer_page_end(src_y, src_y_plane_size)
align_buffer_page_end(src_u, src_uv_plane_size)
align_buffer_page_end(src_v, src_uv_plane_size)
srandom(time(NULL));
MemRandomize(src_y, src_y_plane_size);
MemRandomize(src_u, src_uv_plane_size);
MemRandomize(src_v, src_uv_plane_size);
int dst_width_uv = (dst_width + 1) >> 1;
int dst_height_uv = (dst_height + 1) >> 1;
int dst_y_plane_size = (dst_width + b * 2) * (dst_height + b * 2);
int dst_uv_plane_size = (dst_width_uv + b * 2) * (dst_height_uv + b * 2);
int dst_stride_y = b * 2 + dst_width;
int dst_stride_uv = b * 2 + dst_width_uv;
align_buffer_page_end(dst_y_c, dst_y_plane_size)
align_buffer_page_end(dst_u_c, dst_uv_plane_size)
align_buffer_page_end(dst_v_c, dst_uv_plane_size)
align_buffer_page_end(dst_y_opt, dst_y_plane_size)
align_buffer_page_end(dst_u_opt, dst_uv_plane_size)
align_buffer_page_end(dst_v_opt, dst_uv_plane_size)
MaskCpuFlags(0); // Disable all CPU optimization.
double c_time = get_time();
I420Scale(src_y + (src_stride_y * b) + b, src_stride_y,
src_u + (src_stride_uv * b) + b, src_stride_uv,
src_v + (src_stride_uv * b) + b, src_stride_uv,
src_width, src_height,
dst_y_c + (dst_stride_y * b) + b, dst_stride_y,
dst_u_c + (dst_stride_uv * b) + b, dst_stride_uv,
dst_v_c + (dst_stride_uv * b) + b, dst_stride_uv,
dst_width, dst_height, f);
c_time = (get_time() - c_time);
MaskCpuFlags(-1); // Enable all CPU optimization.
double opt_time = get_time();
for (i = 0; i < benchmark_iterations; ++i) {
I420Scale(src_y + (src_stride_y * b) + b, src_stride_y,
src_u + (src_stride_uv * b) + b, src_stride_uv,
src_v + (src_stride_uv * b) + b, src_stride_uv,
src_width, src_height,
dst_y_opt + (dst_stride_y * b) + b, dst_stride_y,
dst_u_opt + (dst_stride_uv * b) + b, dst_stride_uv,
dst_v_opt + (dst_stride_uv * b) + b, dst_stride_uv,
dst_width, dst_height, f);
}
opt_time = (get_time() - opt_time) / benchmark_iterations;
// Report performance of C vs OPT
printf("filter %d - %8d us C - %8d us OPT\n",
f,
static_cast<int>(c_time * 1e6),
static_cast<int>(opt_time * 1e6));
// C version may be a little off from the optimized. Order of
// operations may introduce rounding somewhere. So do a difference
// of the buffers and look to see that the max difference isn't
// over 2.
int max_diff = 0;
for (i = b; i < (dst_height + b); ++i) {
for (j = b; j < (dst_width + b); ++j) {
int abs_diff = Abs(dst_y_c[(i * dst_stride_y) + j] -
dst_y_opt[(i * dst_stride_y) + j]);
if (abs_diff > max_diff) {
max_diff = abs_diff;
}
}
}
for (i = b; i < (dst_height_uv + b); ++i) {
for (j = b; j < (dst_width_uv + b); ++j) {
int abs_diff = Abs(dst_u_c[(i * dst_stride_uv) + j] -
dst_u_opt[(i * dst_stride_uv) + j]);
if (abs_diff > max_diff) {
max_diff = abs_diff;
}
abs_diff = Abs(dst_v_c[(i * dst_stride_uv) + j] -
dst_v_opt[(i * dst_stride_uv) + j]);
if (abs_diff > max_diff) {
max_diff = abs_diff;
}
}
}
free_aligned_buffer_page_end(dst_y_c)
free_aligned_buffer_page_end(dst_u_c)
free_aligned_buffer_page_end(dst_v_c)
free_aligned_buffer_page_end(dst_y_opt)
free_aligned_buffer_page_end(dst_u_opt)
free_aligned_buffer_page_end(dst_v_opt)
free_aligned_buffer_page_end(src_y)
free_aligned_buffer_page_end(src_u)
free_aligned_buffer_page_end(src_v)
return max_diff;
}
#define TEST_FACTOR1(name, filter, hfactor, vfactor, max_diff) \
TEST_F(libyuvTest, ScaleDownBy##name##_##filter) { \
int diff = TestFilter(benchmark_width_, benchmark_height_, \
Abs(benchmark_width_) * hfactor, \
Abs(benchmark_height_) * vfactor, \
kFilter##filter, benchmark_iterations_); \
EXPECT_LE(diff, max_diff); \
}
// Test a scale factor with all 4 filters. Expect unfiltered to be exact, but
// filtering is different fixed point implementations for SSSE3, Neon and C.
#define TEST_FACTOR(name, hfactor, vfactor) \
TEST_FACTOR1(name, None, hfactor, vfactor, 0) \
TEST_FACTOR1(name, Linear, hfactor, vfactor, 3) \
TEST_FACTOR1(name, Bilinear, hfactor, vfactor, 3) \
TEST_FACTOR1(name, Box, hfactor, vfactor, 3) \
// TODO(fbarchard): ScaleDownBy1 should be lossless, but Box has error of 2.
TEST_FACTOR(1, 1 / 1, 1 / 1)
TEST_FACTOR(2, 1 / 2, 1 / 2)
TEST_FACTOR(4, 1 / 4, 1 / 4)
TEST_FACTOR(8, 1 / 8, 1 / 8)
TEST_FACTOR(16, 1 / 16, 1 / 16)
TEST_FACTOR(2by3, 2 / 3, 2 / 3)
TEST_FACTOR(3by4, 3 / 4, 3 / 4)
TEST_FACTOR(3by8, 3 / 8, 3 / 8)
TEST_FACTOR(Vertical2by3, 1, 2 / 3)
#undef TEST_FACTOR1
#undef TEST_FACTOR
#define TEST_SCALETO1(name, width, height, filter, max_diff) \
TEST_F(libyuvTest, name##To##width##x##height##_##filter) { \
int diff = TestFilter(benchmark_width_, benchmark_height_, \
width, height, \
kFilter##filter, benchmark_iterations_); \
EXPECT_LE(diff, max_diff); \
} \
TEST_F(libyuvTest, name##From##width##x##height##_##filter) { \
int diff = TestFilter(width, height, \
Abs(benchmark_width_), Abs(benchmark_height_), \
kFilter##filter, benchmark_iterations_); \
EXPECT_LE(diff, max_diff); \
}
// Test scale to a specified size with all 4 filters.
#define TEST_SCALETO(name, width, height) \
TEST_SCALETO1(name, width, height, None, 0) \
TEST_SCALETO1(name, width, height, Linear, 3) \
TEST_SCALETO1(name, width, height, Bilinear, 3) \
TEST_SCALETO1(name, width, height, Box, 3)
TEST_SCALETO(Scale, 1, 1)
TEST_SCALETO(Scale, 320, 240)
TEST_SCALETO(Scale, 352, 288)
TEST_SCALETO(Scale, 640, 360)
TEST_SCALETO(Scale, 853, 480)
TEST_SCALETO(Scale, 1280, 720)
TEST_SCALETO(Scale, 1920, 1080)
#undef TEST_SCALETO1
#undef TEST_SCALETO
} // namespace libyuv
|
