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
|
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <xmmintrin.h>
#include <emmintrin.h>
#include "macros.h"
#include "celt_lpc.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "pitch.h"
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
#include <smmintrin.h>
#include "x86cpu.h"
opus_val32 celt_inner_prod_sse4_1(const opus_val16 *x, const opus_val16 *y,
int N)
{
opus_int i, dataSize16;
opus_int32 sum;
__m128i inVec1_76543210, inVec1_FEDCBA98, acc1;
__m128i inVec2_76543210, inVec2_FEDCBA98, acc2;
__m128i inVec1_3210, inVec2_3210;
sum = 0;
dataSize16 = N & ~15;
acc1 = _mm_setzero_si128();
acc2 = _mm_setzero_si128();
for (i=0;i<dataSize16;i+=16) {
inVec1_76543210 = _mm_loadu_si128((__m128i *)(&x[i + 0]));
inVec2_76543210 = _mm_loadu_si128((__m128i *)(&y[i + 0]));
inVec1_FEDCBA98 = _mm_loadu_si128((__m128i *)(&x[i + 8]));
inVec2_FEDCBA98 = _mm_loadu_si128((__m128i *)(&y[i + 8]));
inVec1_76543210 = _mm_madd_epi16(inVec1_76543210, inVec2_76543210);
inVec1_FEDCBA98 = _mm_madd_epi16(inVec1_FEDCBA98, inVec2_FEDCBA98);
acc1 = _mm_add_epi32(acc1, inVec1_76543210);
acc2 = _mm_add_epi32(acc2, inVec1_FEDCBA98);
}
acc1 = _mm_add_epi32(acc1, acc2);
if (N - i >= 8)
{
inVec1_76543210 = _mm_loadu_si128((__m128i *)(&x[i + 0]));
inVec2_76543210 = _mm_loadu_si128((__m128i *)(&y[i + 0]));
inVec1_76543210 = _mm_madd_epi16(inVec1_76543210, inVec2_76543210);
acc1 = _mm_add_epi32(acc1, inVec1_76543210);
i += 8;
}
if (N - i >= 4)
{
inVec1_3210 = OP_CVTEPI16_EPI32_M64(&x[i + 0]);
inVec2_3210 = OP_CVTEPI16_EPI32_M64(&y[i + 0]);
inVec1_3210 = _mm_mullo_epi32(inVec1_3210, inVec2_3210);
acc1 = _mm_add_epi32(acc1, inVec1_3210);
i += 4;
}
acc1 = _mm_add_epi32(acc1, _mm_unpackhi_epi64(acc1, acc1));
acc1 = _mm_add_epi32(acc1, _mm_shufflelo_epi16(acc1, 0x0E));
sum += _mm_cvtsi128_si32(acc1);
for (;i<N;i++)
{
sum = silk_SMLABB(sum, x[i], y[i]);
}
return sum;
}
void xcorr_kernel_sse4_1(const opus_val16 * x, const opus_val16 * y, opus_val32 sum[ 4 ], int len)
{
int j;
__m128i vecX, vecX0, vecX1, vecX2, vecX3;
__m128i vecY0, vecY1, vecY2, vecY3;
__m128i sum0, sum1, sum2, sum3, vecSum;
__m128i initSum;
#ifdef OPUS_CHECK_ASM
opus_val32 sum_c[4];
for (j=0;j<4;j++) {
sum_c[j] = sum[j];
}
xcorr_kernel_c(x, y, sum_c, len);
#endif
celt_assert(len >= 3);
sum0 = _mm_setzero_si128();
sum1 = _mm_setzero_si128();
sum2 = _mm_setzero_si128();
sum3 = _mm_setzero_si128();
for (j=0;j<(len-7);j+=8)
{
vecX = _mm_loadu_si128((__m128i *)(&x[j + 0]));
vecY0 = _mm_loadu_si128((__m128i *)(&y[j + 0]));
vecY1 = _mm_loadu_si128((__m128i *)(&y[j + 1]));
vecY2 = _mm_loadu_si128((__m128i *)(&y[j + 2]));
vecY3 = _mm_loadu_si128((__m128i *)(&y[j + 3]));
sum0 = _mm_add_epi32(sum0, _mm_madd_epi16(vecX, vecY0));
sum1 = _mm_add_epi32(sum1, _mm_madd_epi16(vecX, vecY1));
sum2 = _mm_add_epi32(sum2, _mm_madd_epi16(vecX, vecY2));
sum3 = _mm_add_epi32(sum3, _mm_madd_epi16(vecX, vecY3));
}
sum0 = _mm_add_epi32(sum0, _mm_unpackhi_epi64( sum0, sum0));
sum0 = _mm_add_epi32(sum0, _mm_shufflelo_epi16( sum0, 0x0E));
sum1 = _mm_add_epi32(sum1, _mm_unpackhi_epi64( sum1, sum1));
sum1 = _mm_add_epi32(sum1, _mm_shufflelo_epi16( sum1, 0x0E));
sum2 = _mm_add_epi32(sum2, _mm_unpackhi_epi64( sum2, sum2));
sum2 = _mm_add_epi32(sum2, _mm_shufflelo_epi16( sum2, 0x0E));
sum3 = _mm_add_epi32(sum3, _mm_unpackhi_epi64( sum3, sum3));
sum3 = _mm_add_epi32(sum3, _mm_shufflelo_epi16( sum3, 0x0E));
vecSum = _mm_unpacklo_epi64(_mm_unpacklo_epi32(sum0, sum1),
_mm_unpacklo_epi32(sum2, sum3));
for (;j<(len-3);j+=4)
{
vecX = OP_CVTEPI16_EPI32_M64(&x[j + 0]);
vecX0 = _mm_shuffle_epi32(vecX, 0x00);
vecX1 = _mm_shuffle_epi32(vecX, 0x55);
vecX2 = _mm_shuffle_epi32(vecX, 0xaa);
vecX3 = _mm_shuffle_epi32(vecX, 0xff);
vecY0 = OP_CVTEPI16_EPI32_M64(&y[j + 0]);
vecY1 = OP_CVTEPI16_EPI32_M64(&y[j + 1]);
vecY2 = OP_CVTEPI16_EPI32_M64(&y[j + 2]);
vecY3 = OP_CVTEPI16_EPI32_M64(&y[j + 3]);
sum0 = _mm_mullo_epi32(vecX0, vecY0);
sum1 = _mm_mullo_epi32(vecX1, vecY1);
sum2 = _mm_mullo_epi32(vecX2, vecY2);
sum3 = _mm_mullo_epi32(vecX3, vecY3);
sum0 = _mm_add_epi32(sum0, sum1);
sum2 = _mm_add_epi32(sum2, sum3);
vecSum = _mm_add_epi32(vecSum, sum0);
vecSum = _mm_add_epi32(vecSum, sum2);
}
vecX = OP_CVTEPI16_EPI32_M64(&x[len - 4]);
if (len - j == 3)
{
vecX0 = _mm_shuffle_epi32(vecX, 0x55);
vecX1 = _mm_shuffle_epi32(vecX, 0xaa);
vecX2 = _mm_shuffle_epi32(vecX, 0xff);
vecY0 = OP_CVTEPI16_EPI32_M64(&y[j + 0]);
vecY1 = OP_CVTEPI16_EPI32_M64(&y[j + 1]);
vecY2 = OP_CVTEPI16_EPI32_M64(&y[j + 2]);
sum0 = _mm_mullo_epi32(vecX0, vecY0);
sum1 = _mm_mullo_epi32(vecX1, vecY1);
sum2 = _mm_mullo_epi32(vecX2, vecY2);
vecSum = _mm_add_epi32(vecSum, sum0);
vecSum = _mm_add_epi32(vecSum, sum1);
vecSum = _mm_add_epi32(vecSum, sum2);
}
else if (len - j == 2)
{
vecX0 = _mm_shuffle_epi32(vecX, 0xaa);
vecX1 = _mm_shuffle_epi32(vecX, 0xff);
vecY0 = OP_CVTEPI16_EPI32_M64(&y[j + 0]);
vecY1 = OP_CVTEPI16_EPI32_M64(&y[j + 1]);
sum0 = _mm_mullo_epi32(vecX0, vecY0);
sum1 = _mm_mullo_epi32(vecX1, vecY1);
vecSum = _mm_add_epi32(vecSum, sum0);
vecSum = _mm_add_epi32(vecSum, sum1);
}
else if (len - j == 1)
{
vecX0 = _mm_shuffle_epi32(vecX, 0xff);
vecY0 = OP_CVTEPI16_EPI32_M64(&y[j + 0]);
sum0 = _mm_mullo_epi32(vecX0, vecY0);
vecSum = _mm_add_epi32(vecSum, sum0);
}
initSum = _mm_loadu_si128((__m128i *)(&sum[0]));
initSum = _mm_add_epi32(initSum, vecSum);
_mm_storeu_si128((__m128i *)sum, initSum);
#ifdef OPUS_CHECK_ASM
celt_assert(!memcmp(sum_c, sum, sizeof(sum_c)));
#endif
}
#endif
|
