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
|
;******************************************************************************
;*
;* Copyright (c) 2000-2001 Fabrice Bellard <fabrice@bellard.org>
;* Copyright (c) Nick Kurshev <nickols_k@mail.ru>
;* Copyright (c) 2002 Michael Niedermayer <michaelni@gmx.at>
;* Copyright (c) 2002 Zdenek Kabelac <kabi@informatics.muni.cz>
;* Copyright (c) 2013 Daniel Kang
;*
;* SIMD-optimized halfpel functions
;*
;* This file is part of FFmpeg.
;*
;* FFmpeg 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.
;*
;* FFmpeg 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 FFmpeg; if not, write to the Free Software
;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
;******************************************************************************
%include "libavutil/x86/x86util.asm"
SECTION_RODATA
cextern pb_1
cextern pw_2
pb_interleave16: db 0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15
pb_interleave8: db 0, 4, 1, 5, 2, 6, 3, 7
cextern pw_8192
SECTION .text
; void ff_put_pixels8_x2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_PIXELS8_X2 0
%if cpuflag(sse2)
cglobal put_pixels16_x2, 4,5,4
%else
cglobal put_pixels8_x2, 4,5
%endif
lea r4, [r2*2]
.loop:
movu m0, [r1+1]
movu m1, [r1+r2+1]
%if cpuflag(sse2)
movu m2, [r1]
movu m3, [r1+r2]
pavgb m0, m2
pavgb m1, m3
%else
PAVGB m0, [r1]
PAVGB m1, [r1+r2]
%endif
mova [r0], m0
mova [r0+r2], m1
add r1, r4
add r0, r4
movu m0, [r1+1]
movu m1, [r1+r2+1]
%if cpuflag(sse2)
movu m2, [r1]
movu m3, [r1+r2]
pavgb m0, m2
pavgb m1, m3
%else
PAVGB m0, [r1]
PAVGB m1, [r1+r2]
%endif
add r1, r4
mova [r0], m0
mova [r0+r2], m1
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_PIXELS8_X2
INIT_MMX 3dnow
PUT_PIXELS8_X2
; void ff_put_pixels16_x2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_PIXELS_16 0
cglobal put_pixels16_x2, 4,5
lea r4, [r2*2]
.loop:
mova m0, [r1]
mova m1, [r1+r2]
mova m2, [r1+8]
mova m3, [r1+r2+8]
PAVGB m0, [r1+1]
PAVGB m1, [r1+r2+1]
PAVGB m2, [r1+9]
PAVGB m3, [r1+r2+9]
mova [r0], m0
mova [r0+r2], m1
mova [r0+8], m2
mova [r0+r2+8], m3
add r1, r4
add r0, r4
mova m0, [r1]
mova m1, [r1+r2]
mova m2, [r1+8]
mova m3, [r1+r2+8]
PAVGB m0, [r1+1]
PAVGB m1, [r1+r2+1]
PAVGB m2, [r1+9]
PAVGB m3, [r1+r2+9]
add r1, r4
mova [r0], m0
mova [r0+r2], m1
mova [r0+8], m2
mova [r0+r2+8], m3
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_PIXELS_16
INIT_MMX 3dnow
PUT_PIXELS_16
; The 8_X2 macro can easily be used here
INIT_XMM sse2
PUT_PIXELS8_X2
; void ff_put_no_rnd_pixels8_x2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_NO_RND_PIXELS8_X2 0
cglobal put_no_rnd_pixels8_x2, 4,5
mova m6, [pb_1]
lea r4, [r2*2]
.loop:
mova m0, [r1]
mova m2, [r1+r2]
mova m1, [r1+1]
mova m3, [r1+r2+1]
add r1, r4
psubusb m0, m6
psubusb m2, m6
PAVGB m0, m1
PAVGB m2, m3
mova [r0], m0
mova [r0+r2], m2
mova m0, [r1]
mova m1, [r1+1]
mova m2, [r1+r2]
mova m3, [r1+r2+1]
add r0, r4
add r1, r4
psubusb m0, m6
psubusb m2, m6
PAVGB m0, m1
PAVGB m2, m3
mova [r0], m0
mova [r0+r2], m2
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_NO_RND_PIXELS8_X2
INIT_MMX 3dnow
PUT_NO_RND_PIXELS8_X2
; void ff_put_no_rnd_pixels8_x2_exact(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_NO_RND_PIXELS8_X2_EXACT 0
cglobal put_no_rnd_pixels8_x2_exact, 4,5
lea r4, [r2*3]
pcmpeqb m6, m6
.loop:
mova m0, [r1]
mova m2, [r1+r2]
mova m1, [r1+1]
mova m3, [r1+r2+1]
pxor m0, m6
pxor m2, m6
pxor m1, m6
pxor m3, m6
PAVGB m0, m1
PAVGB m2, m3
pxor m0, m6
pxor m2, m6
mova [r0], m0
mova [r0+r2], m2
mova m0, [r1+r2*2]
mova m1, [r1+r2*2+1]
mova m2, [r1+r4]
mova m3, [r1+r4+1]
pxor m0, m6
pxor m1, m6
pxor m2, m6
pxor m3, m6
PAVGB m0, m1
PAVGB m2, m3
pxor m0, m6
pxor m2, m6
mova [r0+r2*2], m0
mova [r0+r4], m2
lea r1, [r1+r2*4]
lea r0, [r0+r2*4]
sub r3d, 4
jg .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_NO_RND_PIXELS8_X2_EXACT
INIT_MMX 3dnow
PUT_NO_RND_PIXELS8_X2_EXACT
; void ff_put_pixels8_y2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_PIXELS8_Y2 0
%if cpuflag(sse2)
cglobal put_pixels16_y2, 4,5,3
%else
cglobal put_pixels8_y2, 4,5
%endif
lea r4, [r2*2]
movu m0, [r1]
sub r0, r2
.loop:
movu m1, [r1+r2]
movu m2, [r1+r4]
add r1, r4
PAVGB m0, m1
PAVGB m1, m2
mova [r0+r2], m0
mova [r0+r4], m1
movu m1, [r1+r2]
movu m0, [r1+r4]
add r0, r4
add r1, r4
PAVGB m2, m1
PAVGB m1, m0
mova [r0+r2], m2
mova [r0+r4], m1
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_PIXELS8_Y2
INIT_MMX 3dnow
PUT_PIXELS8_Y2
; actually, put_pixels16_y2_sse2
INIT_XMM sse2
PUT_PIXELS8_Y2
; void ff_put_no_rnd_pixels8_y2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_NO_RND_PIXELS8_Y2 0
cglobal put_no_rnd_pixels8_y2, 4,5
mova m6, [pb_1]
lea r4, [r2+r2]
mova m0, [r1]
sub r0, r2
.loop:
mova m1, [r1+r2]
mova m2, [r1+r4]
add r1, r4
psubusb m1, m6
PAVGB m0, m1
PAVGB m1, m2
mova [r0+r2], m0
mova [r0+r4], m1
mova m1, [r1+r2]
mova m0, [r1+r4]
add r0, r4
add r1, r4
psubusb m1, m6
PAVGB m2, m1
PAVGB m1, m0
mova [r0+r2], m2
mova [r0+r4], m1
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_NO_RND_PIXELS8_Y2
INIT_MMX 3dnow
PUT_NO_RND_PIXELS8_Y2
; void ff_put_no_rnd_pixels8_y2_exact(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro PUT_NO_RND_PIXELS8_Y2_EXACT 0
cglobal put_no_rnd_pixels8_y2_exact, 4,5
lea r4, [r2*3]
mova m0, [r1]
pcmpeqb m6, m6
add r1, r2
pxor m0, m6
.loop:
mova m1, [r1]
mova m2, [r1+r2]
pxor m1, m6
pxor m2, m6
PAVGB m0, m1
PAVGB m1, m2
pxor m0, m6
pxor m1, m6
mova [r0], m0
mova [r0+r2], m1
mova m1, [r1+r2*2]
mova m0, [r1+r4]
pxor m1, m6
pxor m0, m6
PAVGB m2, m1
PAVGB m1, m0
pxor m2, m6
pxor m1, m6
mova [r0+r2*2], m2
mova [r0+r4], m1
lea r1, [r1+r2*4]
lea r0, [r0+r2*4]
sub r3d, 4
jg .loop
REP_RET
%endmacro
INIT_MMX mmxext
PUT_NO_RND_PIXELS8_Y2_EXACT
INIT_MMX 3dnow
PUT_NO_RND_PIXELS8_Y2_EXACT
; void ff_avg_pixels8(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro AVG_PIXELS8 0
cglobal avg_pixels8, 4,5
lea r4, [r2*2]
.loop:
mova m0, [r0]
mova m1, [r0+r2]
PAVGB m0, [r1]
PAVGB m1, [r1+r2]
mova [r0], m0
mova [r0+r2], m1
add r1, r4
add r0, r4
mova m0, [r0]
mova m1, [r0+r2]
PAVGB m0, [r1]
PAVGB m1, [r1+r2]
add r1, r4
mova [r0], m0
mova [r0+r2], m1
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX 3dnow
AVG_PIXELS8
; void ff_avg_pixels8_x2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro AVG_PIXELS8_X2 0
%if cpuflag(sse2)
cglobal avg_pixels16_x2, 4,5,4
%else
cglobal avg_pixels8_x2, 4,5
%endif
lea r4, [r2*2]
%if notcpuflag(mmxext)
pcmpeqd m5, m5
paddb m5, m5
%endif
.loop:
movu m0, [r1]
movu m2, [r1+r2]
%if cpuflag(sse2)
movu m1, [r1+1]
movu m3, [r1+r2+1]
pavgb m0, m1
pavgb m2, m3
%else
PAVGB m0, [r1+1], m3, m5
PAVGB m2, [r1+r2+1], m4, m5
%endif
PAVGB m0, [r0], m3, m5
PAVGB m2, [r0+r2], m4, m5
add r1, r4
mova [r0], m0
mova [r0+r2], m2
movu m0, [r1]
movu m2, [r1+r2]
%if cpuflag(sse2)
movu m1, [r1+1]
movu m3, [r1+r2+1]
pavgb m0, m1
pavgb m2, m3
%else
PAVGB m0, [r1+1], m3, m5
PAVGB m2, [r1+r2+1], m4, m5
%endif
add r0, r4
add r1, r4
PAVGB m0, [r0], m3, m5
PAVGB m2, [r0+r2], m4, m5
mova [r0], m0
mova [r0+r2], m2
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmx
AVG_PIXELS8_X2
INIT_MMX mmxext
AVG_PIXELS8_X2
INIT_MMX 3dnow
AVG_PIXELS8_X2
; actually avg_pixels16_x2
INIT_XMM sse2
AVG_PIXELS8_X2
; void ff_avg_pixels8_y2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro AVG_PIXELS8_Y2 0
%if cpuflag(sse2)
cglobal avg_pixels16_y2, 4,5,3
%else
cglobal avg_pixels8_y2, 4,5
%endif
lea r4, [r2*2]
movu m0, [r1]
sub r0, r2
.loop:
movu m1, [r1+r2]
movu m2, [r1+r4]
add r1, r4
PAVGB m0, m1
PAVGB m1, m2
PAVGB m0, [r0+r2]
PAVGB m1, [r0+r4]
mova [r0+r2], m0
mova [r0+r4], m1
movu m1, [r1+r2]
movu m0, [r1+r4]
PAVGB m2, m1
PAVGB m1, m0
add r0, r4
add r1, r4
PAVGB m2, [r0+r2]
PAVGB m1, [r0+r4]
mova [r0+r2], m2
mova [r0+r4], m1
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
AVG_PIXELS8_Y2
INIT_MMX 3dnow
AVG_PIXELS8_Y2
; actually avg_pixels16_y2
INIT_XMM sse2
AVG_PIXELS8_Y2
; void ff_avg_pixels8_xy2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
; Note this is not correctly rounded, and is therefore used for
; not-bitexact output
%macro AVG_APPROX_PIXELS8_XY2 0
cglobal avg_approx_pixels8_xy2, 4,5
mova m6, [pb_1]
lea r4, [r2*2]
mova m0, [r1]
PAVGB m0, [r1+1]
.loop:
mova m2, [r1+r4]
mova m1, [r1+r2]
psubusb m2, m6
PAVGB m1, [r1+r2+1]
PAVGB m2, [r1+r4+1]
add r1, r4
PAVGB m0, m1
PAVGB m1, m2
PAVGB m0, [r0]
PAVGB m1, [r0+r2]
mova [r0], m0
mova [r0+r2], m1
mova m1, [r1+r2]
mova m0, [r1+r4]
PAVGB m1, [r1+r2+1]
PAVGB m0, [r1+r4+1]
add r0, r4
add r1, r4
PAVGB m2, m1
PAVGB m1, m0
PAVGB m2, [r0]
PAVGB m1, [r0+r2]
mova [r0], m2
mova [r0+r2], m1
add r0, r4
sub r3d, 4
jne .loop
REP_RET
%endmacro
INIT_MMX mmxext
AVG_APPROX_PIXELS8_XY2
INIT_MMX 3dnow
AVG_APPROX_PIXELS8_XY2
; void ff_avg_pixels16_xy2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
%macro SET_PIXELS_XY2 1
%if cpuflag(sse2)
cglobal %1_pixels16_xy2, 4,5,8
%else
cglobal %1_pixels8_xy2, 4,5
%endif
pxor m7, m7
mova m6, [pw_2]
movu m0, [r1]
movu m4, [r1+1]
mova m1, m0
mova m5, m4
punpcklbw m0, m7
punpcklbw m4, m7
punpckhbw m1, m7
punpckhbw m5, m7
paddusw m4, m0
paddusw m5, m1
xor r4, r4
add r1, r2
.loop:
movu m0, [r1+r4]
movu m2, [r1+r4+1]
mova m1, m0
mova m3, m2
punpcklbw m0, m7
punpcklbw m2, m7
punpckhbw m1, m7
punpckhbw m3, m7
paddusw m0, m2
paddusw m1, m3
paddusw m4, m6
paddusw m5, m6
paddusw m4, m0
paddusw m5, m1
psrlw m4, 2
psrlw m5, 2
%ifidn %1, avg
mova m3, [r0+r4]
packuswb m4, m5
PAVGB m4, m3
%else
packuswb m4, m5
%endif
mova [r0+r4], m4
add r4, r2
movu m2, [r1+r4]
movu m4, [r1+r4+1]
mova m3, m2
mova m5, m4
punpcklbw m2, m7
punpcklbw m4, m7
punpckhbw m3, m7
punpckhbw m5, m7
paddusw m4, m2
paddusw m5, m3
paddusw m0, m6
paddusw m1, m6
paddusw m0, m4
paddusw m1, m5
psrlw m0, 2
psrlw m1, 2
%ifidn %1, avg
mova m3, [r0+r4]
packuswb m0, m1
PAVGB m0, m3
%else
packuswb m0, m1
%endif
mova [r0+r4], m0
add r4, r2
sub r3d, 2
jnz .loop
REP_RET
%endmacro
INIT_MMX mmxext
SET_PIXELS_XY2 avg
INIT_MMX 3dnow
SET_PIXELS_XY2 avg
INIT_XMM sse2
SET_PIXELS_XY2 put
SET_PIXELS_XY2 avg
%macro SSSE3_PIXELS_XY2 1-2
%if %0 == 2 ; sse2
cglobal %1_pixels16_xy2, 4,5,%2
mova m4, [pb_interleave16]
%else
cglobal %1_pixels8_xy2, 4,5
mova m4, [pb_interleave8]
%endif
mova m5, [pb_1]
movu m0, [r1]
movu m1, [r1+1]
pmaddubsw m0, m5
pmaddubsw m1, m5
xor r4, r4
add r1, r2
.loop:
movu m2, [r1+r4]
movu m3, [r1+r4+1]
pmaddubsw m2, m5
pmaddubsw m3, m5
paddusw m0, m2
paddusw m1, m3
pmulhrsw m0, [pw_8192]
pmulhrsw m1, [pw_8192]
%ifidn %1, avg
mova m6, [r0+r4]
packuswb m0, m1
pshufb m0, m4
pavgb m0, m6
%else
packuswb m0, m1
pshufb m0, m4
%endif
mova [r0+r4], m0
add r4, r2
movu m0, [r1+r4]
movu m1, [r1+r4+1]
pmaddubsw m0, m5
pmaddubsw m1, m5
paddusw m2, m0
paddusw m3, m1
pmulhrsw m2, [pw_8192]
pmulhrsw m3, [pw_8192]
%ifidn %1, avg
mova m6, [r0+r4]
packuswb m2, m3
pshufb m2, m4
pavgb m2, m6
%else
packuswb m2, m3
pshufb m2, m4
%endif
mova [r0+r4], m2
add r4, r2
sub r3d, 2
jnz .loop
REP_RET
%endmacro
INIT_MMX ssse3
SSSE3_PIXELS_XY2 put
SSSE3_PIXELS_XY2 avg
INIT_XMM ssse3
SSSE3_PIXELS_XY2 put, 6
SSSE3_PIXELS_XY2 avg, 7
|
