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
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
|
;
; Copyright (c) 2010 The VP8 project authors. All Rights Reserved.
;
; Use of this source code is governed by a BSD-style license and patent
; grant that can be found in the LICENSE file in the root of the source
; tree. All contributing project authors may be found in the AUTHORS
; file in the root of the source tree.
;
%include "vpx_ports/x86_abi_support.asm"
section .text
global sym(vp8_short_fdct4x4_mmx)
global sym(vp8_fast_fdct4x4_mmx)
global sym(vp8_fast_fdct8x4_wmt)
%define DCTCONSTANTSBITS (16)
%define DCTROUNDINGVALUE (1<< (DCTCONSTANTSBITS-1))
%define x_c1 (60547) ; cos(pi /8) * (1<<15)
%define x_c2 (46341) ; cos(pi*2/8) * (1<<15)
%define x_c3 (25080) ; cos(pi*3/8) * (1<<15)
%define _1STSTAGESHIFT 14
%define _2NDSTAGESHIFT 16
; using matrix multiply with source and destbuffer has a pitch
;void vp8_short_fdct4x4_mmx(short *input, short *output, int pitch)
sym(vp8_short_fdct4x4_mmx):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 3
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rsi, arg(0) ;input
mov rdi, arg(1) ;output
movsxd rax, dword ptr arg(2) ;pitch
lea rdx, [dct_matrix GLOBAL]
movq mm0, [rsi ]
movq mm1, [rsi + rax]
movq mm2, [rsi + rax*2]
lea rsi, [rsi + rax*2]
movq mm3, [rsi + rax]
; first column
movq mm4, mm0
movq mm7, [rdx]
pmaddwd mm4, mm7
movq mm5, mm1
pmaddwd mm5, mm7
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, mm7
movq mm6, mm3
pmaddwd mm6, mm7
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct1st_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _1STSTAGESHIFT
psrad mm5, _1STSTAGESHIFT
packssdw mm4, mm5
movq [rdi], mm4
;second column
movq mm4, mm0
pmaddwd mm4, [rdx+8]
movq mm5, mm1
pmaddwd mm5, [rdx+8]
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, [rdx+8]
movq mm6, mm3
pmaddwd mm6, [rdx+8]
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct1st_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _1STSTAGESHIFT
psrad mm5, _1STSTAGESHIFT
packssdw mm4, mm5
movq [rdi+8], mm4
;third column
movq mm4, mm0
pmaddwd mm4, [rdx+16]
movq mm5, mm1
pmaddwd mm5, [rdx+16]
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, [rdx+16]
movq mm6, mm3
pmaddwd mm6, [rdx+16]
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct1st_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _1STSTAGESHIFT
psrad mm5, _1STSTAGESHIFT
packssdw mm4, mm5
movq [rdi+16], mm4
;fourth column (this is the last column, so we do not have save the source any more)
pmaddwd mm0, [rdx+24]
pmaddwd mm1, [rdx+24]
movq mm6, mm0
punpckldq mm0, mm1
punpckhdq mm6, mm1
paddd mm0, mm6
pmaddwd mm2, [rdx+24]
pmaddwd mm3, [rdx+24]
movq mm7, mm2
punpckldq mm2, mm3
punpckhdq mm7, mm3
paddd mm2, mm7
movq mm6, [dct1st_stage_rounding_mmx GLOBAL]
paddd mm0, mm6
paddd mm2, mm6
psrad mm0, _1STSTAGESHIFT
psrad mm2, _1STSTAGESHIFT
packssdw mm0, mm2
movq mm3, mm0
; done with one pass
; now start second pass
movq mm0, [rdi ]
movq mm1, [rdi+ 8]
movq mm2, [rdi+ 16]
movq mm4, mm0
pmaddwd mm4, [rdx]
movq mm5, mm1
pmaddwd mm5, [rdx]
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, [rdx]
movq mm6, mm3
pmaddwd mm6, [rdx]
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct2nd_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _2NDSTAGESHIFT
psrad mm5, _2NDSTAGESHIFT
packssdw mm4, mm5
movq [rdi], mm4
;second column
movq mm4, mm0
pmaddwd mm4, [rdx+8]
movq mm5, mm1
pmaddwd mm5, [rdx+8]
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, [rdx+8]
movq mm6, mm3
pmaddwd mm6, [rdx+8]
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct2nd_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _2NDSTAGESHIFT
psrad mm5, _2NDSTAGESHIFT
packssdw mm4, mm5
movq [rdi+8], mm4
;third column
movq mm4, mm0
pmaddwd mm4, [rdx+16]
movq mm5, mm1
pmaddwd mm5, [rdx+16]
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, [rdx+16]
movq mm6, mm3
pmaddwd mm6, [rdx+16]
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct2nd_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _2NDSTAGESHIFT
psrad mm5, _2NDSTAGESHIFT
packssdw mm4, mm5
movq [rdi+16], mm4
;fourth column
movq mm4, mm0
pmaddwd mm4, [rdx+24]
movq mm5, mm1
pmaddwd mm5, [rdx+24]
movq mm6, mm4
punpckldq mm4, mm5
punpckhdq mm6, mm5
paddd mm4, mm6
movq mm5, mm2
pmaddwd mm5, [rdx+24]
movq mm6, mm3
pmaddwd mm6, [rdx+24]
movq mm7, mm5
punpckldq mm5, mm6
punpckhdq mm7, mm6
paddd mm5, mm7
movq mm6, [dct2nd_stage_rounding_mmx GLOBAL]
paddd mm4, mm6
paddd mm5, mm6
psrad mm4, _2NDSTAGESHIFT
psrad mm5, _2NDSTAGESHIFT
packssdw mm4, mm5
movq [rdi+24], mm4
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_fast_fdct4x4_mmx(short *input, short *output, int pitch)
sym(vp8_fast_fdct4x4_mmx):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 3
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rsi, arg(0) ;input
mov rdi, arg(1) ;output
lea rdx, [dct_const_mmx GLOBAL]
movsxd rax, dword ptr arg(2) ;pitch
lea rcx, [rsi + rax*2]
; read the input data
movq mm0, [rsi]
movq mm1, [rsi + rax ]
movq mm2, [rcx]
movq mm3, [rcx + rax]
; get the constants
;shift to left by 1 for prescision
paddw mm0, mm0
paddw mm1, mm1
psllw mm2, 1
psllw mm3, 1
; transpose for the second stage
movq mm4, mm0 ; 00 01 02 03
movq mm5, mm2 ; 10 11 12 03
punpcklwd mm0, mm1 ; 00 10 01 11
punpckhwd mm4, mm1 ; 02 12 03 13
punpcklwd mm2, mm3 ; 20 30 21 31
punpckhwd mm5, mm3 ; 22 32 23 33
movq mm1, mm0 ; 00 10 01 11
punpckldq mm0, mm2 ; 00 10 20 30
punpckhdq mm1, mm2 ; 01 11 21 31
movq mm2, mm4 ; 02 12 03 13
punpckldq mm2, mm5 ; 02 12 22 32
punpckhdq mm4, mm5 ; 03 13 23 33
movq mm3, mm4
; first stage
movq mm5, mm0
movq mm4, mm1
paddw mm0, mm3 ; a = 0 + 3
paddw mm1, mm2 ; b = 1 + 2
psubw mm4, mm2 ; c = 1 - 2
psubw mm5, mm3 ; d = 0 - 3
; output 0 and 2
movq mm6, [rdx + 16] ; c2
movq mm2, mm0 ; a
paddw mm0, mm1 ; a + b
psubw mm2, mm1 ; a - b
movq mm1, mm0 ; a + b
pmulhw mm0, mm6 ; 00 01 02 03
paddw mm0, mm1 ; output 00 01 02 03
pmulhw mm6, mm2 ; 20 21 22 23
paddw mm2, mm6 ; output 20 21 22 23
; output 1 and 3
movq mm6, [rdx + 8] ; c1
movq mm7, [rdx + 24] ; c3
movq mm1, mm4 ; c
movq mm3, mm5 ; d
pmulhw mm1, mm7 ; c * c3
pmulhw mm3, mm6 ; d * c1
paddw mm3, mm5 ; d * c1 rounded
paddw mm1, mm3 ; output 10 11 12 13
movq mm3, mm4 ; c
pmulhw mm5, mm7 ; d * c3
pmulhw mm4, mm6 ; c * c1
paddw mm3, mm4 ; round c* c1
psubw mm5, mm3 ; output 30 31 32 33
movq mm3, mm5
; done with vertical
; transpose for the second stage
movq mm4, mm0 ; 00 01 02 03
movq mm5, mm2 ; 10 11 12 03
punpcklwd mm0, mm1 ; 00 10 01 11
punpckhwd mm4, mm1 ; 02 12 03 13
punpcklwd mm2, mm3 ; 20 30 21 31
punpckhwd mm5, mm3 ; 22 32 23 33
movq mm1, mm0 ; 00 10 01 11
punpckldq mm0, mm2 ; 00 10 20 30
punpckhdq mm1, mm2 ; 01 11 21 31
movq mm2, mm4 ; 02 12 03 13
punpckldq mm2, mm5 ; 02 12 22 32
punpckhdq mm4, mm5 ; 03 13 23 33
movq mm3, mm4
; first stage
movq mm5, mm0
movq mm4, mm1
paddw mm0, mm3 ; a = 0 + 3
paddw mm1, mm2 ; b = 1 + 2
psubw mm4, mm2 ; c = 1 - 2
psubw mm5, mm3 ; d = 0 - 3
; output 0 and 2
movq mm6, [rdx + 16] ; c2
movq mm2, mm0 ; a
paddw mm0, mm1 ; a + b
psubw mm2, mm1 ; a - b
movq mm1, mm0 ; a + b
pmulhw mm0, mm6 ; 00 01 02 03
paddw mm0, mm1 ; output 00 01 02 03
pmulhw mm6, mm2 ; 20 21 22 23
paddw mm2, mm6 ; output 20 21 22 23
; output 1 and 3
movq mm6, [rdx + 8] ; c1
movq mm7, [rdx + 24] ; c3
movq mm1, mm4 ; c
movq mm3, mm5 ; d
pmulhw mm1, mm7 ; c * c3
pmulhw mm3, mm6 ; d * c1
paddw mm3, mm5 ; d * c1 rounded
paddw mm1, mm3 ; output 10 11 12 13
movq mm3, mm4 ; c
pmulhw mm5, mm7 ; d * c3
pmulhw mm4, mm6 ; c * c1
paddw mm3, mm4 ; round c* c1
psubw mm5, mm3 ; output 30 31 32 33
movq mm3, mm5
; done with vertical
pcmpeqw mm4, mm4
pcmpeqw mm5, mm5
psrlw mm4, 15
psrlw mm5, 15
paddw mm0, mm4
paddw mm1, mm5
paddw mm2, mm4
paddw mm3, mm5
psraw mm0, 1
psraw mm1, 1
psraw mm2, 1
psraw mm3, 1
movq [rdi ], mm0
movq [rdi+ 8], mm1
movq [rdi+16], mm2
movq [rdi+24], mm3
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_fast_fdct8x4_wmt(short *input, short *output, int pitch)
sym(vp8_fast_fdct8x4_wmt):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 3
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rsi, arg(0) ;input
mov rdi, arg(1) ;output
lea rdx, [dct_const_xmm GLOBAL]
movsxd rax, dword ptr arg(2) ;pitch
lea rcx, [rsi + rax*2]
; read the input data
movdqa xmm0, [rsi]
movdqa xmm2, [rsi + rax]
movdqa xmm4, [rcx]
movdqa xmm3, [rcx + rax]
; get the constants
;shift to left by 1 for prescision
psllw xmm0, 1
psllw xmm2, 1
psllw xmm4, 1
psllw xmm3, 1
; transpose for the second stage
movdqa xmm1, xmm0 ; 00 01 02 03 04 05 06 07
movdqa xmm5, xmm4 ; 20 21 22 23 24 25 26 27
punpcklwd xmm0, xmm2 ; 00 10 01 11 02 12 03 13
punpckhwd xmm1, xmm2 ; 04 14 05 15 06 16 07 17
punpcklwd xmm4, xmm3 ; 20 30 21 31 22 32 23 33
punpckhwd xmm5, xmm3 ; 24 34 25 35 26 36 27 37
movdqa xmm2, xmm0 ; 00 10 01 11 02 12 03 13
punpckldq xmm0, xmm4 ; 00 10 20 30 01 11 21 31
punpckhdq xmm2, xmm4 ; 02 12 22 32 03 13 23 33
movdqa xmm4, xmm1 ; 04 14 05 15 06 16 07 17
punpckldq xmm4, xmm5 ; 04 14 24 34 05 15 25 35
punpckhdq xmm1, xmm5 ; 06 16 26 36 07 17 27 37
movdqa xmm3, xmm2 ; 02 12 22 32 03 13 23 33
punpckhqdq xmm3, xmm1 ; 03 13 23 33 07 17 27 37
punpcklqdq xmm2, xmm1 ; 02 12 22 32 06 16 26 36
movdqa xmm1, xmm0 ; 00 10 20 30 01 11 21 31
punpcklqdq xmm0, xmm4 ; 00 10 20 30 04 14 24 34
punpckhqdq xmm1, xmm4 ; 01 11 21 32 05 15 25 35
; xmm0 0
; xmm1 1
; xmm2 2
; xmm3 3
; first stage
movdqa xmm5, xmm0
movdqa xmm4, xmm1
paddw xmm0, xmm3 ; a = 0 + 3
paddw xmm1, xmm2 ; b = 1 + 2
psubw xmm4, xmm2 ; c = 1 - 2
psubw xmm5, xmm3 ; d = 0 - 3
; output 0 and 2
movdqa xmm6, [rdx + 32] ; c2
movdqa xmm2, xmm0 ; a
paddw xmm0, xmm1 ; a + b
psubw xmm2, xmm1 ; a - b
movdqa xmm1, xmm0 ; a + b
pmulhw xmm0, xmm6 ; 00 01 02 03
paddw xmm0, xmm1 ; output 00 01 02 03
pmulhw xmm6, xmm2 ; 20 21 22 23
paddw xmm2, xmm6 ; output 20 21 22 23
; output 1 and 3
movdqa xmm6, [rdx + 16] ; c1
movdqa xmm7, [rdx + 48] ; c3
movdqa xmm1, xmm4 ; c
movdqa xmm3, xmm5 ; d
pmulhw xmm1, xmm7 ; c * c3
pmulhw xmm3, xmm6 ; d * c1
paddw xmm3, xmm5 ; d * c1 rounded
paddw xmm1, xmm3 ; output 10 11 12 13
movdqa xmm3, xmm4 ; c
pmulhw xmm5, xmm7 ; d * c3
pmulhw xmm4, xmm6 ; c * c1
paddw xmm3, xmm4 ; round c* c1
psubw xmm5, xmm3 ; output 30 31 32 33
movdqa xmm3, xmm5
; done with vertical
; transpose for the second stage
movdqa xmm4, xmm2 ; 02 12 22 32 06 16 26 36
movdqa xmm2, xmm1 ; 01 11 21 31 05 15 25 35
movdqa xmm1, xmm0 ; 00 10 20 30 04 14 24 34
movdqa xmm5, xmm4 ; 02 12 22 32 06 16 26 36
punpcklwd xmm0, xmm2 ; 00 01 10 11 20 21 30 31
punpckhwd xmm1, xmm2 ; 04 05 14 15 24 25 34 35
punpcklwd xmm4, xmm3 ; 02 03 12 13 22 23 32 33
punpckhwd xmm5, xmm3 ; 06 07 16 17 26 27 36 37
movdqa xmm2, xmm0 ; 00 01 10 11 20 21 30 31
punpckldq xmm0, xmm4 ; 00 01 02 03 10 11 12 13
punpckhdq xmm2, xmm4 ; 20 21 22 23 30 31 32 33
movdqa xmm4, xmm1 ; 04 05 14 15 24 25 34 35
punpckldq xmm4, xmm5 ; 04 05 06 07 14 15 16 17
punpckhdq xmm1, xmm5 ; 24 25 26 27 34 35 36 37
movdqa xmm3, xmm2 ; 20 21 22 23 30 31 32 33
punpckhqdq xmm3, xmm1 ; 30 31 32 33 34 35 36 37
punpcklqdq xmm2, xmm1 ; 20 21 22 23 24 25 26 27
movdqa xmm1, xmm0 ; 00 01 02 03 10 11 12 13
punpcklqdq xmm0, xmm4 ; 00 01 02 03 04 05 06 07
punpckhqdq xmm1, xmm4 ; 10 11 12 13 14 15 16 17
; first stage
movdqa xmm5, xmm0
movdqa xmm4, xmm1
paddw xmm0, xmm3 ; a = 0 + 3
paddw xmm1, xmm2 ; b = 1 + 2
psubw xmm4, xmm2 ; c = 1 - 2
psubw xmm5, xmm3 ; d = 0 - 3
; output 0 and 2
movdqa xmm6, [rdx + 32] ; c2
movdqa xmm2, xmm0 ; a
paddw xmm0, xmm1 ; a + b
psubw xmm2, xmm1 ; a - b
movdqa xmm1, xmm0 ; a + b
pmulhw xmm0, xmm6 ; 00 01 02 03
paddw xmm0, xmm1 ; output 00 01 02 03
pmulhw xmm6, xmm2 ; 20 21 22 23
paddw xmm2, xmm6 ; output 20 21 22 23
; output 1 and 3
movdqa xmm6, [rdx + 16] ; c1
movdqa xmm7, [rdx + 48] ; c3
movdqa xmm1, xmm4 ; c
movdqa xmm3, xmm5 ; d
pmulhw xmm1, xmm7 ; c * c3
pmulhw xmm3, xmm6 ; d * c1
paddw xmm3, xmm5 ; d * c1 rounded
paddw xmm1, xmm3 ; output 10 11 12 13
movdqa xmm3, xmm4 ; c
pmulhw xmm5, xmm7 ; d * c3
pmulhw xmm4, xmm6 ; c * c1
paddw xmm3, xmm4 ; round c* c1
psubw xmm5, xmm3 ; output 30 31 32 33
movdqa xmm3, xmm5
; done with vertical
pcmpeqw xmm4, xmm4
pcmpeqw xmm5, xmm5;
psrlw xmm4, 15
psrlw xmm5, 15
paddw xmm0, xmm4
paddw xmm1, xmm5
paddw xmm2, xmm4
paddw xmm3, xmm5
psraw xmm0, 1
psraw xmm1, 1
psraw xmm2, 1
psraw xmm3, 1
movq QWORD PTR[rdi ], xmm0
movq QWORD PTR[rdi+ 8], xmm1
movq QWORD PTR[rdi+16], xmm2
movq QWORD PTR[rdi+24], xmm3
psrldq xmm0, 8
psrldq xmm1, 8
psrldq xmm2, 8
psrldq xmm3, 8
movq QWORD PTR[rdi+32], xmm0
movq QWORD PTR[rdi+40], xmm1
movq QWORD PTR[rdi+48], xmm2
movq QWORD PTR[rdi+56], xmm3
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
SECTION_RODATA
;static const unsigned int dct1st_stage_rounding_mmx[2] =
align 16
dct1st_stage_rounding_mmx:
times 2 dd 8192
;static const unsigned int dct2nd_stage_rounding_mmx[2] =
align 16
dct2nd_stage_rounding_mmx:
times 2 dd 32768
;static const short dct_matrix[4][4]=
align 16
dct_matrix:
times 4 dw 23170
dw 30274
dw 12540
dw -12540
dw -30274
dw 23170
times 2 dw -23170
dw 23170
dw 12540
dw -30274
dw 30274
dw -12540
;static const unsigned short dct_const_mmx[4 * 4]=
align 16
dct_const_mmx:
times 4 dw 0
times 4 dw 60547
times 4 dw 46341
times 4 dw 25080
;static const unsigned short dct_const_xmm[8 * 4]=
align 16
dct_const_xmm:
times 8 dw 0
times 8 dw 60547
times 8 dw 46341
times 8 dw 25080
|
