aboutsummaryrefslogtreecommitdiffstats
path: root/libavcodec/x86/dsputil_init.c
blob: 2ac93516ca3ec4ac6cf76b2a189f32c57ffa5de9 (plain) (blame)
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
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
/*
 * This file is part of Libav.
 *
 * Libav is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * Libav is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with Libav; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "config.h"
#include "libavutil/attributes.h"
#include "libavutil/cpu.h"
#include "libavutil/internal.h"
#include "libavutil/x86/asm.h"
#include "libavutil/x86/cpu.h"
#include "libavcodec/dsputil.h"
#include "libavcodec/simple_idct.h"
#include "dsputil_x86.h"
#include "idct_xvid.h"

void ff_put_pixels8_l2_mmxext(uint8_t *dst, uint8_t *src1, uint8_t *src2,
                              int dstStride, int src1Stride, int h);
void ff_put_no_rnd_pixels8_l2_mmxext(uint8_t *dst, uint8_t *src1,
                                     uint8_t *src2, int dstStride,
                                     int src1Stride, int h);
void ff_avg_pixels8_l2_mmxext(uint8_t *dst, uint8_t *src1, uint8_t *src2,
                              int dstStride, int src1Stride, int h);
void ff_put_pixels16_l2_mmxext(uint8_t *dst, uint8_t *src1, uint8_t *src2,
                               int dstStride, int src1Stride, int h);
void ff_avg_pixels16_l2_mmxext(uint8_t *dst, uint8_t *src1, uint8_t *src2,
                               int dstStride, int src1Stride, int h);
void ff_put_no_rnd_pixels16_l2_mmxext(uint8_t *dst, uint8_t *src1, uint8_t *src2,
                                      int dstStride, int src1Stride, int h);
void ff_put_mpeg4_qpel16_h_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                         int dstStride, int srcStride, int h);
void ff_avg_mpeg4_qpel16_h_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                         int dstStride, int srcStride, int h);
void ff_put_no_rnd_mpeg4_qpel16_h_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                                 int dstStride, int srcStride,
                                                 int h);
void ff_put_mpeg4_qpel8_h_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                        int dstStride, int srcStride, int h);
void ff_avg_mpeg4_qpel8_h_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                        int dstStride, int srcStride, int h);
void ff_put_no_rnd_mpeg4_qpel8_h_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                                int dstStride, int srcStride,
                                                int h);
void ff_put_mpeg4_qpel16_v_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                         int dstStride, int srcStride);
void ff_avg_mpeg4_qpel16_v_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                         int dstStride, int srcStride);
void ff_put_no_rnd_mpeg4_qpel16_v_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                                 int dstStride, int srcStride);
void ff_put_mpeg4_qpel8_v_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                        int dstStride, int srcStride);
void ff_avg_mpeg4_qpel8_v_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                        int dstStride, int srcStride);
void ff_put_no_rnd_mpeg4_qpel8_v_lowpass_mmxext(uint8_t *dst, uint8_t *src,
                                                int dstStride, int srcStride);
#define ff_put_no_rnd_pixels16_mmxext ff_put_pixels16_mmxext
#define ff_put_no_rnd_pixels8_mmxext ff_put_pixels8_mmxext

int32_t ff_scalarproduct_int16_mmxext(const int16_t *v1, const int16_t *v2,
                                      int order);
int32_t ff_scalarproduct_int16_sse2(const int16_t *v1, const int16_t *v2,
                                    int order);
int32_t ff_scalarproduct_and_madd_int16_mmxext(int16_t *v1, const int16_t *v2,
                                               const int16_t *v3,
                                               int order, int mul);
int32_t ff_scalarproduct_and_madd_int16_sse2(int16_t *v1, const int16_t *v2,
                                             const int16_t *v3,
                                             int order, int mul);
int32_t ff_scalarproduct_and_madd_int16_ssse3(int16_t *v1, const int16_t *v2,
                                              const int16_t *v3,
                                              int order, int mul);

void ff_apply_window_int16_round_mmxext(int16_t *output, const int16_t *input,
                                        const int16_t *window, unsigned int len);
void ff_apply_window_int16_round_sse2(int16_t *output, const int16_t *input,
                                      const int16_t *window, unsigned int len);
void ff_apply_window_int16_mmxext(int16_t *output, const int16_t *input,
                                  const int16_t *window, unsigned int len);
void ff_apply_window_int16_sse2(int16_t *output, const int16_t *input,
                                const int16_t *window, unsigned int len);
void ff_apply_window_int16_ssse3(int16_t *output, const int16_t *input,
                                 const int16_t *window, unsigned int len);
void ff_apply_window_int16_ssse3_atom(int16_t *output, const int16_t *input,
                                      const int16_t *window, unsigned int len);

void ff_bswap32_buf_ssse3(uint32_t *dst, const uint32_t *src, int w);
void ff_bswap32_buf_sse2(uint32_t *dst, const uint32_t *src, int w);

void ff_add_hfyu_median_prediction_mmxext(uint8_t *dst, const uint8_t *top,
                                          const uint8_t *diff, int w,
                                          int *left, int *left_top);
int  ff_add_hfyu_left_prediction_ssse3(uint8_t *dst, const uint8_t *src,
                                       int w, int left);
int  ff_add_hfyu_left_prediction_sse4(uint8_t *dst, const uint8_t *src,
                                      int w, int left);

void ff_vector_clip_int32_mmx     (int32_t *dst, const int32_t *src,
                                   int32_t min, int32_t max, unsigned int len);
void ff_vector_clip_int32_sse2    (int32_t *dst, const int32_t *src,
                                   int32_t min, int32_t max, unsigned int len);
void ff_vector_clip_int32_int_sse2(int32_t *dst, const int32_t *src,
                                   int32_t min, int32_t max, unsigned int len);
void ff_vector_clip_int32_sse4    (int32_t *dst, const int32_t *src,
                                   int32_t min, int32_t max, unsigned int len);

#if HAVE_YASM

PIXELS16(static, ff_avg, , , _mmxext)
PIXELS16(static, ff_put, , , _mmxext)

#define QPEL_OP(OPNAME, RND, MMX)                                       \
static void OPNAME ## qpel8_mc00_ ## MMX (uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    ff_ ## OPNAME ## pixels8_ ## MMX(dst, src, stride, 8);              \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc10_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t temp[8];                                                   \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(half, src, 8,        \
                                                   stride, 8);          \
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, src, half,                 \
                                        stride, stride, 8);             \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc20_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    ff_ ## OPNAME ## mpeg4_qpel8_h_lowpass_ ## MMX(dst, src, stride,    \
                                                   stride, 8);          \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc30_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t temp[8];                                                   \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(half, src, 8,        \
                                                   stride, 8);          \
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, src + 1, half, stride,     \
                                        stride, 8);                     \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc01_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t temp[8];                                                   \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(half, src,           \
                                                   8, stride);          \
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, src, half,                 \
                                        stride, stride, 8);             \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc02_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    ff_ ## OPNAME ## mpeg4_qpel8_v_lowpass_ ## MMX(dst, src,            \
                                                   stride, stride);     \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc03_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t temp[8];                                                   \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(half, src,           \
                                                   8, stride);          \
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, src + stride, half, stride,\
                                        stride, 8);                     \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc11_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH  = ((uint8_t*)half) + 64;                     \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## pixels8_l2_ ## MMX(halfH, src, halfH, 8,           \
                                        stride, 9);                     \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(halfHV, halfH, 8, 8);\
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, halfH, halfHV,             \
                                        stride, 8, 8);                  \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc31_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH  = ((uint8_t*)half) + 64;                     \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## pixels8_l2_ ## MMX(halfH, src + 1, halfH, 8,       \
                                        stride, 9);                     \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(halfHV, halfH, 8, 8);\
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, halfH, halfHV,             \
                                        stride, 8, 8);                  \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc13_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH  = ((uint8_t*)half) + 64;                     \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## pixels8_l2_ ## MMX(halfH, src, halfH, 8,           \
                                        stride, 9);                     \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(halfHV, halfH, 8, 8);\
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, halfH + 8, halfHV,         \
                                        stride, 8, 8);                  \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc33_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH  = ((uint8_t*)half) + 64;                     \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## pixels8_l2_ ## MMX(halfH, src + 1, halfH, 8,       \
                                        stride, 9);                     \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(halfHV, halfH, 8, 8);\
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, halfH + 8, halfHV,         \
                                        stride, 8, 8);                  \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc21_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH  = ((uint8_t*)half) + 64;                     \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(halfHV, halfH, 8, 8);\
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, halfH, halfHV,             \
                                        stride, 8, 8);                  \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc23_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH  = ((uint8_t*)half) + 64;                     \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## mpeg4_qpel8_v_lowpass_ ## MMX(halfHV, halfH, 8, 8);\
    ff_ ## OPNAME ## pixels8_l2_ ## MMX(dst, halfH + 8, halfHV,         \
                                        stride, 8, 8);                  \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc12_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH = ((uint8_t*)half);                           \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## pixels8_l2_ ## MMX(halfH, src, halfH,              \
                                        8, stride, 9);                  \
    ff_ ## OPNAME ## mpeg4_qpel8_v_lowpass_ ## MMX(dst, halfH,          \
                                                   stride, 8);          \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc32_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[8 + 9];                                               \
    uint8_t * const halfH = ((uint8_t*)half);                           \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_put ## RND ## pixels8_l2_ ## MMX(halfH, src + 1, halfH, 8,       \
                                        stride, 9);                     \
    ff_ ## OPNAME ## mpeg4_qpel8_v_lowpass_ ## MMX(dst, halfH,          \
                                                   stride, 8);          \
}                                                                       \
                                                                        \
static void OPNAME ## qpel8_mc22_ ## MMX(uint8_t *dst, uint8_t *src,    \
                                         ptrdiff_t stride)              \
{                                                                       \
    uint64_t half[9];                                                   \
    uint8_t * const halfH = ((uint8_t*)half);                           \
    ff_put ## RND ## mpeg4_qpel8_h_lowpass_ ## MMX(halfH, src, 8,       \
                                                   stride, 9);          \
    ff_ ## OPNAME ## mpeg4_qpel8_v_lowpass_ ## MMX(dst, halfH,          \
                                                   stride, 8);          \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc00_ ## MMX (uint8_t *dst, uint8_t *src,  \
                                           ptrdiff_t stride)            \
{                                                                       \
    ff_ ## OPNAME ## pixels16_ ## MMX(dst, src, stride, 16);            \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc10_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t temp[32];                                                  \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(half, src, 16,      \
                                                    stride, 16);        \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, src, half, stride,        \
                                         stride, 16);                   \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc20_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    ff_ ## OPNAME ## mpeg4_qpel16_h_lowpass_ ## MMX(dst, src,           \
                                                    stride, stride, 16);\
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc30_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t temp[32];                                                  \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(half, src, 16,      \
                                                    stride, 16);        \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, src + 1, half,            \
                                         stride, stride, 16);           \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc01_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t temp[32];                                                  \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(half, src, 16,      \
                                                    stride);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, src, half, stride,        \
                                         stride, 16);                   \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc02_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    ff_ ## OPNAME ## mpeg4_qpel16_v_lowpass_ ## MMX(dst, src,           \
                                                    stride, stride);    \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc03_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t temp[32];                                                  \
    uint8_t * const half = (uint8_t*)temp;                              \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(half, src, 16,      \
                                                    stride);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, src+stride, half,         \
                                         stride, stride, 16);           \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc11_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[16 * 2 + 17 * 2];                                     \
    uint8_t * const halfH  = ((uint8_t*)half) + 256;                    \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## pixels16_l2_ ## MMX(halfH, src, halfH, 16,         \
                                         stride, 17);                   \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(halfHV, halfH,      \
                                                    16, 16);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, halfH, halfHV,            \
                                         stride, 16, 16);               \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc31_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[16 * 2 + 17 * 2];                                     \
    uint8_t * const halfH  = ((uint8_t*)half) + 256;                    \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## pixels16_l2_ ## MMX(halfH, src + 1, halfH, 16,     \
                                         stride, 17);                   \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(halfHV, halfH,      \
                                                    16, 16);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, halfH, halfHV,            \
                                         stride, 16, 16);               \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc13_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[16 * 2 + 17 * 2];                                     \
    uint8_t * const halfH  = ((uint8_t*)half) + 256;                    \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## pixels16_l2_ ## MMX(halfH, src, halfH, 16,         \
                                         stride, 17);                   \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(halfHV, halfH,      \
                                                    16, 16);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, halfH + 16, halfHV,       \
                                         stride, 16, 16);               \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc33_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[16 * 2 + 17 * 2];                                     \
    uint8_t * const halfH  = ((uint8_t*)half) + 256;                    \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## pixels16_l2_ ## MMX(halfH, src + 1, halfH, 16,     \
                                         stride, 17);                   \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(halfHV, halfH,      \
                                                    16, 16);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, halfH + 16, halfHV,       \
                                         stride, 16, 16);               \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc21_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[16 * 2 + 17 * 2];                                     \
    uint8_t * const halfH  = ((uint8_t*)half) + 256;                    \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(halfHV, halfH,      \
                                                    16, 16);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, halfH, halfHV,            \
                                         stride, 16, 16);               \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc23_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[16 * 2 + 17 * 2];                                     \
    uint8_t * const halfH  = ((uint8_t*)half) + 256;                    \
    uint8_t * const halfHV = ((uint8_t*)half);                          \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## mpeg4_qpel16_v_lowpass_ ## MMX(halfHV, halfH,      \
                                                    16, 16);            \
    ff_ ## OPNAME ## pixels16_l2_ ## MMX(dst, halfH + 16, halfHV,       \
                                         stride, 16, 16);               \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc12_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[17 * 2];                                              \
    uint8_t * const halfH = ((uint8_t*)half);                           \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## pixels16_l2_ ## MMX(halfH, src, halfH, 16,         \
                                         stride, 17);                   \
    ff_ ## OPNAME ## mpeg4_qpel16_v_lowpass_ ## MMX(dst, halfH,         \
                                                    stride, 16);        \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc32_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[17 * 2];                                              \
    uint8_t * const halfH = ((uint8_t*)half);                           \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_put ## RND ## pixels16_l2_ ## MMX(halfH, src + 1, halfH, 16,     \
                                         stride, 17);                   \
    ff_ ## OPNAME ## mpeg4_qpel16_v_lowpass_ ## MMX(dst, halfH,         \
                                                    stride, 16);        \
}                                                                       \
                                                                        \
static void OPNAME ## qpel16_mc22_ ## MMX(uint8_t *dst, uint8_t *src,   \
                                          ptrdiff_t stride)             \
{                                                                       \
    uint64_t half[17 * 2];                                              \
    uint8_t * const halfH = ((uint8_t*)half);                           \
    ff_put ## RND ## mpeg4_qpel16_h_lowpass_ ## MMX(halfH, src, 16,     \
                                                    stride, 17);        \
    ff_ ## OPNAME ## mpeg4_qpel16_v_lowpass_ ## MMX(dst, halfH,         \
                                                    stride, 16);        \
}

QPEL_OP(put_,        _,        mmxext)
QPEL_OP(avg_,        _,        mmxext)
QPEL_OP(put_no_rnd_, _no_rnd_, mmxext)
#endif /* HAVE_YASM */

#define SET_QPEL_FUNCS(PFX, IDX, SIZE, CPU, PREFIX)                          \
    do {                                                                     \
    c->PFX ## _pixels_tab[IDX][ 0] = PREFIX ## PFX ## SIZE ## _mc00_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 1] = PREFIX ## PFX ## SIZE ## _mc10_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 2] = PREFIX ## PFX ## SIZE ## _mc20_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 3] = PREFIX ## PFX ## SIZE ## _mc30_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 4] = PREFIX ## PFX ## SIZE ## _mc01_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 5] = PREFIX ## PFX ## SIZE ## _mc11_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 6] = PREFIX ## PFX ## SIZE ## _mc21_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 7] = PREFIX ## PFX ## SIZE ## _mc31_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 8] = PREFIX ## PFX ## SIZE ## _mc02_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][ 9] = PREFIX ## PFX ## SIZE ## _mc12_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][10] = PREFIX ## PFX ## SIZE ## _mc22_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][11] = PREFIX ## PFX ## SIZE ## _mc32_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][12] = PREFIX ## PFX ## SIZE ## _mc03_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][13] = PREFIX ## PFX ## SIZE ## _mc13_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][14] = PREFIX ## PFX ## SIZE ## _mc23_ ## CPU; \
    c->PFX ## _pixels_tab[IDX][15] = PREFIX ## PFX ## SIZE ## _mc33_ ## CPU; \
    } while (0)

static av_cold void dsputil_init_mmx(DSPContext *c, AVCodecContext *avctx,
                                     int cpu_flags)
{
#if HAVE_MMX_INLINE
    const int high_bit_depth = avctx->bits_per_raw_sample > 8;

    c->put_pixels_clamped        = ff_put_pixels_clamped_mmx;
    c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_mmx;
    c->add_pixels_clamped        = ff_add_pixels_clamped_mmx;

    if (!high_bit_depth) {
        c->clear_block  = ff_clear_block_mmx;
        c->clear_blocks = ff_clear_blocks_mmx;
        c->draw_edges   = ff_draw_edges_mmx;

        switch (avctx->idct_algo) {
        case FF_IDCT_AUTO:
        case FF_IDCT_SIMPLEMMX:
            c->idct_put              = ff_simple_idct_put_mmx;
            c->idct_add              = ff_simple_idct_add_mmx;
            c->idct                  = ff_simple_idct_mmx;
            c->idct_permutation_type = FF_SIMPLE_IDCT_PERM;
            break;
        case FF_IDCT_XVIDMMX:
            c->idct_put              = ff_idct_xvid_mmx_put;
            c->idct_add              = ff_idct_xvid_mmx_add;
            c->idct                  = ff_idct_xvid_mmx;
            break;
        }
    }

    c->gmc = ff_gmc_mmx;

    c->add_bytes = ff_add_bytes_mmx;
#endif /* HAVE_MMX_INLINE */

#if HAVE_MMX_EXTERNAL
    c->vector_clip_int32 = ff_vector_clip_int32_mmx;
#endif /* HAVE_MMX_EXTERNAL */
}

static av_cold void dsputil_init_mmxext(DSPContext *c, AVCodecContext *avctx,
                                        int cpu_flags)
{
#if HAVE_MMXEXT_INLINE
    const int high_bit_depth = avctx->bits_per_raw_sample > 8;

    if (!high_bit_depth && avctx->idct_algo == FF_IDCT_XVIDMMX) {
        c->idct_put = ff_idct_xvid_mmxext_put;
        c->idct_add = ff_idct_xvid_mmxext_add;
        c->idct     = ff_idct_xvid_mmxext;
    }
#endif /* HAVE_MMXEXT_INLINE */

#if HAVE_MMXEXT_EXTERNAL
    SET_QPEL_FUNCS(avg_qpel,        0, 16, mmxext, );
    SET_QPEL_FUNCS(avg_qpel,        1,  8, mmxext, );

    SET_QPEL_FUNCS(put_qpel,        0, 16, mmxext, );
    SET_QPEL_FUNCS(put_qpel,        1,  8, mmxext, );
    SET_QPEL_FUNCS(put_no_rnd_qpel, 0, 16, mmxext, );
    SET_QPEL_FUNCS(put_no_rnd_qpel, 1,  8, mmxext, );

    /* slower than cmov version on AMD */
    if (!(cpu_flags & AV_CPU_FLAG_3DNOW))
        c->add_hfyu_median_prediction = ff_add_hfyu_median_prediction_mmxext;

    c->scalarproduct_int16          = ff_scalarproduct_int16_mmxext;
    c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_mmxext;

    if (avctx->flags & CODEC_FLAG_BITEXACT) {
        c->apply_window_int16 = ff_apply_window_int16_mmxext;
    } else {
        c->apply_window_int16 = ff_apply_window_int16_round_mmxext;
    }
#endif /* HAVE_MMXEXT_EXTERNAL */
}

static av_cold void dsputil_init_sse(DSPContext *c, AVCodecContext *avctx,
                                     int cpu_flags)
{
#if HAVE_SSE_INLINE
    const int high_bit_depth = avctx->bits_per_raw_sample > 8;

    if (!high_bit_depth) {
#if FF_API_XVMC
FF_DISABLE_DEPRECATION_WARNINGS
        if (!(CONFIG_MPEG_XVMC_DECODER && avctx->xvmc_acceleration > 1)) {
            /* XvMCCreateBlocks() may not allocate 16-byte aligned blocks */
FF_ENABLE_DEPRECATION_WARNINGS
#endif /* FF_API_XVMC */
        c->clear_block  = ff_clear_block_sse;
        c->clear_blocks = ff_clear_blocks_sse;
#if FF_API_XVMC
        }
#endif /* FF_API_XVMC */
    }

    c->vector_clipf = ff_vector_clipf_sse;
#endif /* HAVE_SSE_INLINE */
}

static av_cold void dsputil_init_sse2(DSPContext *c, AVCodecContext *avctx,
                                      int cpu_flags)
{
#if HAVE_SSE2_INLINE
    const int high_bit_depth = avctx->bits_per_raw_sample > 8;

    if (!high_bit_depth && avctx->idct_algo == FF_IDCT_XVIDMMX) {
        c->idct_put              = ff_idct_xvid_sse2_put;
        c->idct_add              = ff_idct_xvid_sse2_add;
        c->idct                  = ff_idct_xvid_sse2;
        c->idct_permutation_type = FF_SSE2_IDCT_PERM;
    }
#endif /* HAVE_SSE2_INLINE */

#if HAVE_SSE2_EXTERNAL
    c->scalarproduct_int16          = ff_scalarproduct_int16_sse2;
    c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_sse2;
    if (cpu_flags & AV_CPU_FLAG_ATOM) {
        c->vector_clip_int32 = ff_vector_clip_int32_int_sse2;
    } else {
        c->vector_clip_int32 = ff_vector_clip_int32_sse2;
    }
    if (avctx->flags & CODEC_FLAG_BITEXACT) {
        c->apply_window_int16 = ff_apply_window_int16_sse2;
    } else if (!(cpu_flags & AV_CPU_FLAG_SSE2SLOW)) {
        c->apply_window_int16 = ff_apply_window_int16_round_sse2;
    }
    c->bswap_buf = ff_bswap32_buf_sse2;
#endif /* HAVE_SSE2_EXTERNAL */
}

static av_cold void dsputil_init_ssse3(DSPContext *c, AVCodecContext *avctx,
                                       int cpu_flags)
{
#if HAVE_SSSE3_EXTERNAL
    c->add_hfyu_left_prediction = ff_add_hfyu_left_prediction_ssse3;
    if (cpu_flags & AV_CPU_FLAG_SSE4) // not really SSE4, just slow on Conroe
        c->add_hfyu_left_prediction = ff_add_hfyu_left_prediction_sse4;

    if (cpu_flags & AV_CPU_FLAG_ATOM)
        c->apply_window_int16 = ff_apply_window_int16_ssse3_atom;
    else
        c->apply_window_int16 = ff_apply_window_int16_ssse3;
    if (!(cpu_flags & (AV_CPU_FLAG_SSE42 | AV_CPU_FLAG_3DNOW))) // cachesplit
        c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_ssse3;
    c->bswap_buf = ff_bswap32_buf_ssse3;
#endif /* HAVE_SSSE3_EXTERNAL */
}

static av_cold void dsputil_init_sse4(DSPContext *c, AVCodecContext *avctx,
                                      int cpu_flags)
{
#if HAVE_SSE4_EXTERNAL
    c->vector_clip_int32 = ff_vector_clip_int32_sse4;
#endif /* HAVE_SSE4_EXTERNAL */
}

av_cold void ff_dsputil_init_x86(DSPContext *c, AVCodecContext *avctx)
{
    int cpu_flags = av_get_cpu_flags();

#if HAVE_7REGS && HAVE_INLINE_ASM
    if (cpu_flags & AV_CPU_FLAG_CMOV)
        c->add_hfyu_median_prediction = ff_add_hfyu_median_prediction_cmov;
#endif

    if (X86_MMX(cpu_flags))
        dsputil_init_mmx(c, avctx, cpu_flags);

    if (X86_MMXEXT(cpu_flags))
        dsputil_init_mmxext(c, avctx, cpu_flags);

    if (X86_SSE(cpu_flags))
        dsputil_init_sse(c, avctx, cpu_flags);

    if (X86_SSE2(cpu_flags))
        dsputil_init_sse2(c, avctx, cpu_flags);

    if (EXTERNAL_SSSE3(cpu_flags))
        dsputil_init_ssse3(c, avctx, cpu_flags);

    if (EXTERNAL_SSE4(cpu_flags))
        dsputil_init_sse4(c, avctx, cpu_flags);

    if (CONFIG_ENCODERS)
        ff_dsputilenc_init_mmx(c, avctx);
}