/
synth_stereo_avx_accurate.S
286 lines (264 loc) · 6.78 KB
/
synth_stereo_avx_accurate.S
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
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
/*
synth_stereo_avx_accurate: AVX optimized synth for x86-64 (stereo specific, MPEG-compliant 16bit output version)
copyright 1995-2013 by the mpg123 project - free software under the terms of the LGPL 2.1
see COPYING and AUTHORS files in distribution or http://mpg123.org
initially written by Taihei Monma
*/
#include "mangle.h"
#ifdef IS_MSABI
/* real *window; */
#define WINDOW %r10
/* real *b0l; */
#define B0L %rdx
/* real *b0r; */
#define B0R %r8
/* real *samples; */
#define SAMPLES %r9
#else
/* real *window; */
#define WINDOW %rdi
/* real *b0l; */
#define B0L %rsi
/* real *b0r; */
#define B0R %rdx
/* real *samples; */
#define SAMPLES %r9
#endif
/*
int synth_1to1_s_avx_accurate_asm(real *window, real *b0l, real *b0r, real *samples, int bo1);
return value: number of clipped samples
*/
#ifndef __APPLE__
.section .rodata
#else
.data
#endif
ALIGN32
maxmin_avx:
.long 1191182335
.long 1191182335
.long 1191182335
.long 1191182335
.long 1191182335
.long 1191182335
.long 1191182335
.long 1191182335
.long -956301312
.long -956301312
.long -956301312
.long -956301312
.long -956301312
.long -956301312
.long -956301312
.long -956301312
.text
ALIGN16
.globl ASM_NAME(synth_1to1_s_avx_accurate_asm)
ASM_NAME(synth_1to1_s_avx_accurate_asm):
#ifdef IS_MSABI /* should save xmm6-15 */
push %rbp
mov %rsp, %rbp
sub $144, %rsp
movaps %xmm6, (%rsp)
movaps %xmm7, 16(%rsp)
movaps %xmm8, 32(%rsp)
movaps %xmm9, 48(%rsp)
movaps %xmm10, 64(%rsp)
movaps %xmm11, 80(%rsp)
movaps %xmm12, 96(%rsp)
movaps %xmm13, 112(%rsp)
movaps %xmm14, 128(%rsp)
movl 48(%rbp), %eax /* 5th argument; placed after 32-byte shadow space */
#endif
#ifdef IS_MSABI
shl $2, %eax
mov %rcx, WINDOW
#else
mov %r8d, %eax
shl $2, %eax
mov %rcx, SAMPLES
#endif
add $64, WINDOW
sub %rax, WINDOW
mov $128, %rax
mov $4, %ecx
vpxor %xmm14, %xmm14, %xmm14
ALIGN16
1:
vmovups (WINDOW), %ymm8
vmovups 32(WINDOW), %ymm9
vmovups (WINDOW,%rax), %ymm10
vmovups 32(WINDOW,%rax), %ymm11
vmulps (B0L), %ymm8, %ymm0
vmulps 32(B0L), %ymm9, %ymm1
vmulps (B0R), %ymm8, %ymm2
vmulps 32(B0R), %ymm9, %ymm3
vmulps 64(B0L), %ymm10, %ymm4
vmulps 96(B0L), %ymm11, %ymm5
vmulps 64(B0R), %ymm10, %ymm6
vmulps 96(B0R), %ymm11, %ymm7
vaddps %ymm1, %ymm0, %ymm8
vaddps %ymm3, %ymm2, %ymm0
vaddps %ymm5, %ymm4, %ymm9
vaddps %ymm7, %ymm6, %ymm1
lea (WINDOW,%rax,2), WINDOW
add %rax, B0L
add %rax, B0R
vmovups (WINDOW), %ymm10
vmovups 32(WINDOW), %ymm11
vmovups (WINDOW,%rax), %ymm12
vmovups 32(WINDOW,%rax), %ymm13
vmulps (B0L), %ymm10, %ymm2
vmulps 32(B0L), %ymm11, %ymm3
vmulps (B0R), %ymm10, %ymm4
vmulps 32(B0R), %ymm11, %ymm5
vmulps 64(B0L), %ymm12, %ymm6
vmulps 96(B0L), %ymm13, %ymm10
vmulps 64(B0R), %ymm12, %ymm7
vmulps 96(B0R), %ymm13, %ymm11
vaddps %ymm3, %ymm2, %ymm2
vaddps %ymm5, %ymm4, %ymm3
vaddps %ymm6, %ymm10, %ymm4
vaddps %ymm7, %ymm11, %ymm5
lea (WINDOW,%rax,2), WINDOW
add %rax, B0L
add %rax, B0R
vunpcklps %ymm0, %ymm8, %ymm6
vunpckhps %ymm0, %ymm8, %ymm0
vunpcklps %ymm1, %ymm9, %ymm7
vunpckhps %ymm1, %ymm9, %ymm1
vaddps %ymm6, %ymm0, %ymm0
vaddps %ymm7, %ymm1, %ymm1
vunpcklps %ymm3, %ymm2, %ymm6
vunpckhps %ymm3, %ymm2, %ymm2
vunpcklps %ymm5, %ymm4, %ymm7
vunpckhps %ymm5, %ymm4, %ymm3
vaddps %ymm6, %ymm2, %ymm2
vaddps %ymm7, %ymm3, %ymm3
vunpcklpd %ymm1, %ymm0, %ymm4
vunpckhpd %ymm1, %ymm0, %ymm0
vunpcklpd %ymm3, %ymm2, %ymm5
vunpckhpd %ymm3, %ymm2, %ymm1
vsubps %ymm0, %ymm4, %ymm0
vsubps %ymm1, %ymm5, %ymm1
vperm2f128 $0x20, %ymm1, %ymm0, %ymm2
vperm2f128 $0x31, %ymm1, %ymm0, %ymm3
vaddps %ymm3, %ymm2, %ymm0
vcmpnleps maxmin_avx(%rip), %ymm0, %ymm1
vcmpltps 32+maxmin_avx(%rip), %ymm0, %ymm2
vextractf128 $0x1, %ymm1, %xmm3
vextractf128 $0x1, %ymm2, %xmm4
vpackssdw %xmm2, %xmm1, %xmm1
vpackssdw %xmm4, %xmm3, %xmm3
vpaddw %xmm3, %xmm1, %xmm1
vpaddw %xmm1, %xmm14, %xmm14
vcvtps2dq %ymm0, %ymm0
vextractf128 $0x1, %ymm0, %xmm1
vpackssdw %xmm1, %xmm0, %xmm0
vmovups %xmm0, (SAMPLES)
add $16, SAMPLES
dec %ecx
jnz 1b
mov $4, %ecx
ALIGN16
1:
vmovups (WINDOW), %ymm8
vmovups 32(WINDOW), %ymm9
vmovups (WINDOW,%rax), %ymm10
vmovups 32(WINDOW,%rax), %ymm11
vmulps (B0L), %ymm8, %ymm0
vmulps 32(B0L), %ymm9, %ymm1
vmulps (B0R), %ymm8, %ymm2
vmulps 32(B0R), %ymm9, %ymm3
vmulps -64(B0L), %ymm10, %ymm4
vmulps -32(B0L), %ymm11, %ymm5
vmulps -64(B0R), %ymm10, %ymm6
vmulps -32(B0R), %ymm11, %ymm7
vaddps %ymm1, %ymm0, %ymm8
vaddps %ymm3, %ymm2, %ymm0
vaddps %ymm5, %ymm4, %ymm9
vaddps %ymm7, %ymm6, %ymm1
lea (WINDOW,%rax,2), WINDOW
sub %rax, B0L
sub %rax, B0R
vmovups (WINDOW), %ymm10
vmovups 32(WINDOW), %ymm11
vmovups (WINDOW,%rax), %ymm12
vmovups 32(WINDOW,%rax), %ymm13
vmulps (B0L), %ymm10, %ymm2
vmulps 32(B0L), %ymm11, %ymm3
vmulps (B0R), %ymm10, %ymm4
vmulps 32(B0R), %ymm11, %ymm5
vmulps -64(B0L), %ymm12, %ymm6
vmulps -32(B0L), %ymm13, %ymm10
vmulps -64(B0R), %ymm12, %ymm7
vmulps -32(B0R), %ymm13, %ymm11
vaddps %ymm3, %ymm2, %ymm2
vaddps %ymm5, %ymm4, %ymm3
vaddps %ymm6, %ymm10, %ymm4
vaddps %ymm7, %ymm11, %ymm5
lea (WINDOW,%rax,2), WINDOW
sub %rax, B0L
sub %rax, B0R
vunpcklps %ymm0, %ymm8, %ymm6
vunpckhps %ymm0, %ymm8, %ymm0
vunpcklps %ymm1, %ymm9, %ymm7
vunpckhps %ymm1, %ymm9, %ymm1
vaddps %ymm6, %ymm0, %ymm0
vaddps %ymm7, %ymm1, %ymm1
vunpcklps %ymm3, %ymm2, %ymm6
vunpckhps %ymm3, %ymm2, %ymm2
vunpcklps %ymm5, %ymm4, %ymm7
vunpckhps %ymm5, %ymm4, %ymm3
vaddps %ymm6, %ymm2, %ymm2
vaddps %ymm7, %ymm3, %ymm3
vunpcklpd %ymm1, %ymm0, %ymm4
vunpckhpd %ymm1, %ymm0, %ymm0
vunpcklpd %ymm3, %ymm2, %ymm5
vunpckhpd %ymm3, %ymm2, %ymm1
vaddps %ymm0, %ymm4, %ymm0
vaddps %ymm1, %ymm5, %ymm1
vperm2f128 $0x20, %ymm1, %ymm0, %ymm2
vperm2f128 $0x31, %ymm1, %ymm0, %ymm3
vaddps %ymm3, %ymm2, %ymm0
vcmpnleps maxmin_avx(%rip), %ymm0, %ymm1
vcmpltps 32+maxmin_avx(%rip), %ymm0, %ymm2
vextractf128 $0x1, %ymm1, %xmm3
vextractf128 $0x1, %ymm2, %xmm4
vpackssdw %xmm2, %xmm1, %xmm1
vpackssdw %xmm4, %xmm3, %xmm3
vpaddw %xmm3, %xmm1, %xmm1
vpaddw %xmm1, %xmm14, %xmm14
vcvtps2dq %ymm0, %ymm0
vextractf128 $0x1, %ymm0, %xmm1
vpackssdw %xmm1, %xmm0, %xmm0
vmovups %xmm0, (SAMPLES)
add $16, SAMPLES
dec %ecx
jnz 1b
vzeroupper
pxor %xmm1, %xmm1
psubw %xmm14, %xmm1
pshufd $0x4e, %xmm1, %xmm0
paddw %xmm1, %xmm0
pshuflw $0x4e, %xmm0, %xmm1
paddw %xmm1, %xmm0
pshuflw $0x11, %xmm0, %xmm1
paddw %xmm1, %xmm0
movd %xmm0, %eax
and $0x7f, %eax
#ifdef IS_MSABI
movaps (%rsp), %xmm6
movaps 16(%rsp), %xmm7
movaps 32(%rsp), %xmm8
movaps 48(%rsp), %xmm9
movaps 64(%rsp), %xmm10
movaps 80(%rsp), %xmm11
movaps 96(%rsp), %xmm12
movaps 112(%rsp), %xmm13
movaps 128(%rsp), %xmm14
mov %rbp, %rsp
pop %rbp
#endif
ret
NONEXEC_STACK