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
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
|
;
; 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"
%define BLOCK_HEIGHT_WIDTH 4
%define VP8_FILTER_WEIGHT 128
%define VP8_FILTER_SHIFT 7
;/************************************************************************************
; Notes: filter_block1d_h6 applies a 6 tap filter horizontally to the input pixels. The
; input pixel array has output_height rows. This routine assumes that output_height is an
; even number. This function handles 8 pixels in horizontal direction, calculating ONE
; rows each iteration to take advantage of the 128 bits operations.
;*************************************************************************************/
;void vp8_filter_block1d8_h6_sse2
;(
; unsigned char *src_ptr,
; unsigned short *output_ptr,
; unsigned int src_pixels_per_line,
; unsigned int pixel_step,
; unsigned int output_height,
; unsigned int output_width,
; short *vp8_filter
;)
global sym(vp8_filter_block1d8_h6_sse2)
sym(vp8_filter_block1d8_h6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 7
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rdx, arg(6) ;vp8_filter
mov rsi, arg(0) ;src_ptr
mov rdi, arg(1) ;output_ptr
movsxd rcx, dword ptr arg(4) ;output_height
movsxd rax, dword ptr arg(2) ;src_pixels_per_line ; Pitch for Source
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(5) ;output_width
%endif
pxor xmm0, xmm0 ; clear xmm0 for unpack
filter_block1d8_h6_rowloop:
movq xmm3, MMWORD PTR [rsi - 2]
movq xmm1, MMWORD PTR [rsi + 6]
prefetcht2 [rsi+rax-2]
pslldq xmm1, 8
por xmm1, xmm3
movdqa xmm4, xmm1
movdqa xmm5, xmm1
movdqa xmm6, xmm1
movdqa xmm7, xmm1
punpcklbw xmm3, xmm0 ; xx05 xx04 xx03 xx02 xx01 xx01 xx-1 xx-2
psrldq xmm4, 1 ; xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00 -1
pmullw xmm3, XMMWORD PTR [rdx] ; x[-2] * H[-2]; Tap 1
punpcklbw xmm4, xmm0 ; xx06 xx05 xx04 xx03 xx02 xx01 xx00 xx-1
psrldq xmm5, 2 ; xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00
pmullw xmm4, XMMWORD PTR [rdx+16] ; x[-1] * H[-1]; Tap 2
punpcklbw xmm5, xmm0 ; xx07 xx06 xx05 xx04 xx03 xx02 xx01 xx00
psrldq xmm6, 3 ; xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01
pmullw xmm5, [rdx+32] ; x[ 0] * H[ 0]; Tap 3
punpcklbw xmm6, xmm0 ; xx08 xx07 xx06 xx05 xx04 xx03 xx02 xx01
psrldq xmm7, 4 ; xx xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02
pmullw xmm6, [rdx+48] ; x[ 1] * h[ 1] ; Tap 4
punpcklbw xmm7, xmm0 ; xx09 xx08 xx07 xx06 xx05 xx04 xx03 xx02
psrldq xmm1, 5 ; xx xx xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03
pmullw xmm7, [rdx+64] ; x[ 2] * h[ 2] ; Tap 5
punpcklbw xmm1, xmm0 ; xx0a xx09 xx08 xx07 xx06 xx05 xx04 xx03
pmullw xmm1, [rdx+80] ; x[ 3] * h[ 3] ; Tap 6
paddsw xmm4, xmm7
paddsw xmm4, xmm5
paddsw xmm4, xmm3
paddsw xmm4, xmm6
paddsw xmm4, xmm1
paddsw xmm4, [rd GLOBAL]
psraw xmm4, 7
packuswb xmm4, xmm0
punpcklbw xmm4, xmm0
movdqa XMMWORD Ptr [rdi], xmm4
lea rsi, [rsi + rax]
%if ABI_IS_32BIT
add rdi, DWORD Ptr arg(5) ;[output_width]
%else
add rdi, r8
%endif
dec rcx
jnz filter_block1d8_h6_rowloop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_filter_block1d16_h6_sse2
;(
; unsigned char *src_ptr,
; unsigned short *output_ptr,
; unsigned int src_pixels_per_line,
; unsigned int pixel_step,
; unsigned int output_height,
; unsigned int output_width,
; short *vp8_filter
;)
;/************************************************************************************
; Notes: filter_block1d_h6 applies a 6 tap filter horizontally to the input pixels. The
; input pixel array has output_height rows. This routine assumes that output_height is an
; even number. This function handles 8 pixels in horizontal direction, calculating ONE
; rows each iteration to take advantage of the 128 bits operations.
;*************************************************************************************/
global sym(vp8_filter_block1d16_h6_sse2)
sym(vp8_filter_block1d16_h6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 7
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rdx, arg(6) ;vp8_filter
mov rsi, arg(0) ;src_ptr
mov rdi, arg(1) ;output_ptr
movsxd rcx, dword ptr arg(4) ;output_height
movsxd rax, dword ptr arg(2) ;src_pixels_per_line ; Pitch for Source
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(5) ;output_width
%endif
pxor xmm0, xmm0 ; clear xmm0 for unpack
filter_block1d16_h6_sse2_rowloop:
movq xmm3, MMWORD PTR [rsi - 2]
movq xmm1, MMWORD PTR [rsi + 6]
movq xmm2, MMWORD PTR [rsi +14]
pslldq xmm2, 8
por xmm2, xmm1
prefetcht2 [rsi+rax-2]
pslldq xmm1, 8
por xmm1, xmm3
movdqa xmm4, xmm1
movdqa xmm5, xmm1
movdqa xmm6, xmm1
movdqa xmm7, xmm1
punpcklbw xmm3, xmm0 ; xx05 xx04 xx03 xx02 xx01 xx01 xx-1 xx-2
psrldq xmm4, 1 ; xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00 -1
pmullw xmm3, XMMWORD PTR [rdx] ; x[-2] * H[-2]; Tap 1
punpcklbw xmm4, xmm0 ; xx06 xx05 xx04 xx03 xx02 xx01 xx00 xx-1
psrldq xmm5, 2 ; xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00
pmullw xmm4, XMMWORD PTR [rdx+16] ; x[-1] * H[-1]; Tap 2
punpcklbw xmm5, xmm0 ; xx07 xx06 xx05 xx04 xx03 xx02 xx01 xx00
psrldq xmm6, 3 ; xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01
pmullw xmm5, [rdx+32] ; x[ 0] * H[ 0]; Tap 3
punpcklbw xmm6, xmm0 ; xx08 xx07 xx06 xx05 xx04 xx03 xx02 xx01
psrldq xmm7, 4 ; xx xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02
pmullw xmm6, [rdx+48] ; x[ 1] * h[ 1] ; Tap 4
punpcklbw xmm7, xmm0 ; xx09 xx08 xx07 xx06 xx05 xx04 xx03 xx02
psrldq xmm1, 5 ; xx xx xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03
pmullw xmm7, [rdx+64] ; x[ 2] * h[ 2] ; Tap 5
punpcklbw xmm1, xmm0 ; xx0a xx09 xx08 xx07 xx06 xx05 xx04 xx03
pmullw xmm1, [rdx+80] ; x[ 3] * h[ 3] ; Tap 6
paddsw xmm4, xmm7
paddsw xmm4, xmm5
paddsw xmm4, xmm3
paddsw xmm4, xmm6
paddsw xmm4, xmm1
paddsw xmm4, [rd GLOBAL]
psraw xmm4, 7
packuswb xmm4, xmm0
punpcklbw xmm4, xmm0
movdqa XMMWORD Ptr [rdi], xmm4
movdqa xmm3, xmm2
movdqa xmm4, xmm2
movdqa xmm5, xmm2
movdqa xmm6, xmm2
movdqa xmm7, xmm2
punpcklbw xmm3, xmm0 ; xx05 xx04 xx03 xx02 xx01 xx01 xx-1 xx-2
psrldq xmm4, 1 ; xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00 -1
pmullw xmm3, XMMWORD PTR [rdx] ; x[-2] * H[-2]; Tap 1
punpcklbw xmm4, xmm0 ; xx06 xx05 xx04 xx03 xx02 xx01 xx00 xx-1
psrldq xmm5, 2 ; xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00
pmullw xmm4, XMMWORD PTR [rdx+16] ; x[-1] * H[-1]; Tap 2
punpcklbw xmm5, xmm0 ; xx07 xx06 xx05 xx04 xx03 xx02 xx01 xx00
psrldq xmm6, 3 ; xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01
pmullw xmm5, [rdx+32] ; x[ 0] * H[ 0]; Tap 3
punpcklbw xmm6, xmm0 ; xx08 xx07 xx06 xx05 xx04 xx03 xx02 xx01
psrldq xmm7, 4 ; xx xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03 02
pmullw xmm6, [rdx+48] ; x[ 1] * h[ 1] ; Tap 4
punpcklbw xmm7, xmm0 ; xx09 xx08 xx07 xx06 xx05 xx04 xx03 xx02
psrldq xmm2, 5 ; xx xx xx xx xx 0d 0c 0b 0a 09 08 07 06 05 04 03
pmullw xmm7, [rdx+64] ; x[ 2] * h[ 2] ; Tap 5
punpcklbw xmm2, xmm0 ; xx0a xx09 xx08 xx07 xx06 xx05 xx04 xx03
pmullw xmm2, [rdx+80] ; x[ 3] * h[ 3] ; Tap 6
paddsw xmm4, xmm7
paddsw xmm4, xmm5
paddsw xmm4, xmm3
paddsw xmm4, xmm6
paddsw xmm4, xmm2
paddsw xmm4, [rd GLOBAL]
psraw xmm4, 7
packuswb xmm4, xmm0
punpcklbw xmm4, xmm0
movdqa XMMWORD Ptr [rdi+16], xmm4
lea rsi, [rsi + rax]
%if ABI_IS_32BIT
add rdi, DWORD Ptr arg(5) ;[output_width]
%else
add rdi, r8
%endif
dec rcx
jnz filter_block1d16_h6_sse2_rowloop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_filter_block1d8_v6_sse2
;(
; short *src_ptr,
; unsigned char *output_ptr,
; int dst_ptich,
; unsigned int pixels_per_line,
; unsigned int pixel_step,
; unsigned int output_height,
; unsigned int output_width,
; short * vp8_filter
;)
;/************************************************************************************
; Notes: filter_block1d8_v6 applies a 6 tap filter vertically to the input pixels. The
; input pixel array has output_height rows.
;*************************************************************************************/
global sym(vp8_filter_block1d8_v6_sse2)
sym(vp8_filter_block1d8_v6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 8
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rax, arg(7) ;vp8_filter
movsxd rdx, dword ptr arg(3) ;pixels_per_line
mov rdi, arg(1) ;output_ptr
mov rsi, arg(0) ;src_ptr
sub rsi, rdx
sub rsi, rdx
movsxd rcx, DWORD PTR arg(5) ;[output_height]
pxor xmm0, xmm0 ; clear xmm0
movdqa xmm7, XMMWORD PTR [rd GLOBAL]
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(2) ; dst_ptich
%endif
vp8_filter_block1d8_v6_sse2_loop:
movdqa xmm1, XMMWORD PTR [rsi]
pmullw xmm1, [rax]
movdqa xmm2, XMMWORD PTR [rsi + rdx]
pmullw xmm2, [rax + 16]
movdqa xmm3, XMMWORD PTR [rsi + rdx * 2]
pmullw xmm3, [rax + 32]
movdqa xmm5, XMMWORD PTR [rsi + rdx * 4]
pmullw xmm5, [rax + 64]
add rsi, rdx
movdqa xmm4, XMMWORD PTR [rsi + rdx * 2]
pmullw xmm4, [rax + 48]
movdqa xmm6, XMMWORD PTR [rsi + rdx * 4]
pmullw xmm6, [rax + 80]
paddsw xmm2, xmm5
paddsw xmm2, xmm3
paddsw xmm2, xmm1
paddsw xmm2, xmm4
paddsw xmm2, xmm6
paddsw xmm2, xmm7
psraw xmm2, 7
packuswb xmm2, xmm0 ; pack and saturate
movq QWORD PTR [rdi], xmm2 ; store the results in the destination
%if ABI_IS_32BIT
add rdi, DWORD PTR arg(2) ;[dst_ptich]
%else
add rdi, r8
%endif
dec rcx ; decrement count
jnz vp8_filter_block1d8_v6_sse2_loop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_unpack_block1d16_h6_sse2
;(
; unsigned char *src_ptr,
; unsigned short *output_ptr,
; unsigned int src_pixels_per_line,
; unsigned int output_height,
; unsigned int output_width
;)
global sym(vp8_unpack_block1d16_h6_sse2)
sym(vp8_unpack_block1d16_h6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 5
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rsi, arg(0) ;src_ptr
mov rdi, arg(1) ;output_ptr
movsxd rcx, dword ptr arg(3) ;output_height
movsxd rax, dword ptr arg(2) ;src_pixels_per_line ; Pitch for Source
pxor xmm0, xmm0 ; clear xmm0 for unpack
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(4) ;output_width ; Pitch for Source
%endif
unpack_block1d16_h6_sse2_rowloop:
movq xmm1, MMWORD PTR [rsi] ; 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00 -1 -2
movq xmm3, MMWORD PTR [rsi+8] ; make copy of xmm1
punpcklbw xmm3, xmm0 ; xx05 xx04 xx03 xx02 xx01 xx01 xx-1 xx-2
punpcklbw xmm1, xmm0
movdqa XMMWORD Ptr [rdi], xmm1
movdqa XMMWORD Ptr [rdi + 16], xmm3
lea rsi, [rsi + rax]
%if ABI_IS_32BIT
add rdi, DWORD Ptr arg(4) ;[output_width]
%else
add rdi, r8
%endif
dec rcx
jnz unpack_block1d16_h6_sse2_rowloop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_unpack_block1d8_h6_sse2
;(
; unsigned char *src_ptr,
; unsigned short *output_ptr,
; unsigned int src_pixels_per_line,
; unsigned int output_height,
; unsigned int output_width
;)
global sym(vp8_unpack_block1d8_h6_sse2)
sym(vp8_unpack_block1d8_h6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 5
GET_GOT rbx
push rsi
push rdi
; end prolog
mov rsi, arg(0) ;src_ptr
mov rdi, arg(1) ;output_ptr
movsxd rcx, dword ptr arg(3) ;output_height
movsxd rax, dword ptr arg(2) ;src_pixels_per_line ; Pitch for Source
pxor xmm0, xmm0 ; clear xmm0 for unpack
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(4) ;output_width ; Pitch for Source
%endif
unpack_block1d8_h6_sse2_rowloop:
movq xmm1, MMWORD PTR [rsi] ; 0d 0c 0b 0a 09 08 07 06 05 04 03 02 01 00 -1 -2
lea rsi, [rsi + rax]
punpcklbw xmm1, xmm0
movdqa XMMWORD Ptr [rdi], xmm1
%if ABI_IS_32BIT
add rdi, DWORD Ptr arg(4) ;[output_width]
%else
add rdi, r8
%endif
dec rcx
jnz unpack_block1d8_h6_sse2_rowloop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_pack_block1d8_v6_sse2
;(
; short *src_ptr,
; unsigned char *output_ptr,
; int dst_ptich,
; unsigned int pixels_per_line,
; unsigned int output_height,
; unsigned int output_width
;)
global sym(vp8_pack_block1d8_v6_sse2)
sym(vp8_pack_block1d8_v6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
GET_GOT rbx
push rsi
push rdi
; end prolog
movsxd rdx, dword ptr arg(3) ;pixels_per_line
mov rdi, arg(1) ;output_ptr
mov rsi, arg(0) ;src_ptr
movsxd rcx, DWORD PTR arg(4) ;[output_height]
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(5) ;output_width ; Pitch for Source
%endif
pack_block1d8_v6_sse2_loop:
movdqa xmm0, XMMWORD PTR [rsi]
packuswb xmm0, xmm0
movq QWORD PTR [rdi], xmm0 ; store the results in the destination
lea rsi, [rsi+rdx]
%if ABI_IS_32BIT
add rdi, DWORD Ptr arg(5) ;[output_width]
%else
add rdi, r8
%endif
dec rcx ; decrement count
jnz pack_block1d8_v6_sse2_loop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_pack_block1d16_v6_sse2
;(
; short *src_ptr,
; unsigned char *output_ptr,
; int dst_ptich,
; unsigned int pixels_per_line,
; unsigned int output_height,
; unsigned int output_width
;)
global sym(vp8_pack_block1d16_v6_sse2)
sym(vp8_pack_block1d16_v6_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
GET_GOT rbx
push rsi
push rdi
; end prolog
movsxd rdx, dword ptr arg(3) ;pixels_per_line
mov rdi, arg(1) ;output_ptr
mov rsi, arg(0) ;src_ptr
movsxd rcx, DWORD PTR arg(4) ;[output_height]
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(2) ;dst_pitch
%endif
pack_block1d16_v6_sse2_loop:
movdqa xmm0, XMMWORD PTR [rsi]
movdqa xmm1, XMMWORD PTR [rsi+16]
packuswb xmm0, xmm1
movdqa XMMWORD PTR [rdi], xmm0 ; store the results in the destination
add rsi, rdx
%if ABI_IS_32BIT
add rdi, DWORD Ptr arg(2) ;dst_pitch
%else
add rdi, r8
%endif
dec rcx ; decrement count
jnz pack_block1d16_v6_sse2_loop ; next row
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_bilinear_predict16x16_sse2
;(
; unsigned char *src_ptr,
; int src_pixels_per_line,
; int xoffset,
; int yoffset,
; unsigned char *dst_ptr,
; int dst_pitch
;)
extern sym(vp8_bilinear_filters_mmx)
global sym(vp8_bilinear_predict16x16_sse2)
sym(vp8_bilinear_predict16x16_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
GET_GOT rbx
push rsi
push rdi
; end prolog
;const short *HFilter = bilinear_filters_mmx[xoffset]
;const short *VFilter = bilinear_filters_mmx[yoffset]
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
movsxd rax, dword ptr arg(2) ;xoffset
cmp rax, 0 ;skip first_pass filter if xoffset=0
je b16x16_sp_only
shl rax, 5
add rax, rcx ;HFilter
mov rdi, arg(4) ;dst_ptr
mov rsi, arg(0) ;src_ptr
movsxd rdx, dword ptr arg(5) ;dst_pitch
movdqa xmm1, [rax]
movdqa xmm2, [rax+16]
movsxd rax, dword ptr arg(3) ;yoffset
cmp rax, 0 ;skip second_pass filter if yoffset=0
je b16x16_fp_only
shl rax, 5
add rax, rcx ;VFilter
lea rcx, [rdi+rdx*8]
lea rcx, [rcx+rdx*8]
movsxd rdx, dword ptr arg(1) ;src_pixels_per_line
pxor xmm0, xmm0
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(5) ;dst_pitch
%endif
; get the first horizontal line done
movdqu xmm3, [rsi] ; xx 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14
movdqa xmm4, xmm3 ; make a copy of current line
punpcklbw xmm3, xmm0 ; xx 00 01 02 03 04 05 06
punpckhbw xmm4, xmm0
pmullw xmm3, xmm1
pmullw xmm4, xmm1
movdqu xmm5, [rsi+1]
movdqa xmm6, xmm5
punpcklbw xmm5, xmm0
punpckhbw xmm6, xmm0
pmullw xmm5, xmm2
pmullw xmm6, xmm2
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
psraw xmm4, VP8_FILTER_SHIFT
movdqa xmm7, xmm3
packuswb xmm7, xmm4
add rsi, rdx ; next line
next_row:
movdqu xmm3, [rsi] ; xx 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14
movdqa xmm4, xmm3 ; make a copy of current line
punpcklbw xmm3, xmm0 ; xx 00 01 02 03 04 05 06
punpckhbw xmm4, xmm0
pmullw xmm3, xmm1
pmullw xmm4, xmm1
movdqu xmm5, [rsi+1]
movdqa xmm6, xmm5
punpcklbw xmm5, xmm0
punpckhbw xmm6, xmm0
pmullw xmm5, xmm2
pmullw xmm6, xmm2
paddw xmm3, xmm5
paddw xmm4, xmm6
movdqa xmm5, xmm7
movdqa xmm6, xmm7
punpcklbw xmm5, xmm0
punpckhbw xmm6, xmm0
pmullw xmm5, [rax]
pmullw xmm6, [rax]
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
psraw xmm4, VP8_FILTER_SHIFT
movdqa xmm7, xmm3
packuswb xmm7, xmm4
pmullw xmm3, [rax+16]
pmullw xmm4, [rax+16]
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
psraw xmm4, VP8_FILTER_SHIFT
packuswb xmm3, xmm4
movdqa [rdi], xmm3 ; store the results in the destination
add rsi, rdx ; next line
%if ABI_IS_32BIT
add rdi, DWORD PTR arg(5) ;dst_pitch
%else
add rdi, r8
%endif
cmp rdi, rcx
jne next_row
jmp done
b16x16_sp_only:
movsxd rax, dword ptr arg(3) ;yoffset
shl rax, 5
add rax, rcx ;VFilter
mov rdi, arg(4) ;dst_ptr
mov rsi, arg(0) ;src_ptr
movsxd rdx, dword ptr arg(5) ;dst_pitch
movdqa xmm1, [rax]
movdqa xmm2, [rax+16]
lea rcx, [rdi+rdx*8]
lea rcx, [rcx+rdx*8]
movsxd rax, dword ptr arg(1) ;src_pixels_per_line
pxor xmm0, xmm0
; get the first horizontal line done
movdqu xmm7, [rsi] ; xx 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14
add rsi, rax ; next line
next_row_spo:
movdqu xmm3, [rsi] ; xx 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14
movdqa xmm5, xmm7
movdqa xmm6, xmm7
movdqa xmm4, xmm3 ; make a copy of current line
movdqa xmm7, xmm3
punpcklbw xmm5, xmm0
punpckhbw xmm6, xmm0
punpcklbw xmm3, xmm0 ; xx 00 01 02 03 04 05 06
punpckhbw xmm4, xmm0
pmullw xmm5, xmm1
pmullw xmm6, xmm1
pmullw xmm3, xmm2
pmullw xmm4, xmm2
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
psraw xmm4, VP8_FILTER_SHIFT
packuswb xmm3, xmm4
movdqa [rdi], xmm3 ; store the results in the destination
add rsi, rax ; next line
add rdi, rdx ;dst_pitch
cmp rdi, rcx
jne next_row_spo
jmp done
b16x16_fp_only:
lea rcx, [rdi+rdx*8]
lea rcx, [rcx+rdx*8]
movsxd rax, dword ptr arg(1) ;src_pixels_per_line
pxor xmm0, xmm0
next_row_fpo:
movdqu xmm3, [rsi] ; xx 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14
movdqa xmm4, xmm3 ; make a copy of current line
punpcklbw xmm3, xmm0 ; xx 00 01 02 03 04 05 06
punpckhbw xmm4, xmm0
pmullw xmm3, xmm1
pmullw xmm4, xmm1
movdqu xmm5, [rsi+1]
movdqa xmm6, xmm5
punpcklbw xmm5, xmm0
punpckhbw xmm6, xmm0
pmullw xmm5, xmm2
pmullw xmm6, xmm2
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
psraw xmm4, VP8_FILTER_SHIFT
packuswb xmm3, xmm4
movdqa [rdi], xmm3 ; store the results in the destination
add rsi, rax ; next line
add rdi, rdx ; dst_pitch
cmp rdi, rcx
jne next_row_fpo
done:
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
;void vp8_bilinear_predict8x8_sse2
;(
; unsigned char *src_ptr,
; int src_pixels_per_line,
; int xoffset,
; int yoffset,
; unsigned char *dst_ptr,
; int dst_pitch
;)
extern sym(vp8_bilinear_filters_mmx)
global sym(vp8_bilinear_predict8x8_sse2)
sym(vp8_bilinear_predict8x8_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
GET_GOT rbx
push rsi
push rdi
; end prolog
ALIGN_STACK 16, rax
sub rsp, 144 ; reserve 144 bytes
;const short *HFilter = bilinear_filters_mmx[xoffset]
;const short *VFilter = bilinear_filters_mmx[yoffset]
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
mov rsi, arg(0) ;src_ptr
movsxd rdx, dword ptr arg(1) ;src_pixels_per_line
;Read 9-line unaligned data in and put them on stack. This gives a big
;performance boost.
movdqu xmm0, [rsi]
lea rax, [rdx + rdx*2]
movdqu xmm1, [rsi+rdx]
movdqu xmm2, [rsi+rdx*2]
add rsi, rax
movdqu xmm3, [rsi]
movdqu xmm4, [rsi+rdx]
movdqu xmm5, [rsi+rdx*2]
add rsi, rax
movdqu xmm6, [rsi]
movdqu xmm7, [rsi+rdx]
movdqa XMMWORD PTR [rsp], xmm0
movdqu xmm0, [rsi+rdx*2]
movdqa XMMWORD PTR [rsp+16], xmm1
movdqa XMMWORD PTR [rsp+32], xmm2
movdqa XMMWORD PTR [rsp+48], xmm3
movdqa XMMWORD PTR [rsp+64], xmm4
movdqa XMMWORD PTR [rsp+80], xmm5
movdqa XMMWORD PTR [rsp+96], xmm6
movdqa XMMWORD PTR [rsp+112], xmm7
movdqa XMMWORD PTR [rsp+128], xmm0
movsxd rax, dword ptr arg(2) ;xoffset
shl rax, 5
add rax, rcx ;HFilter
mov rdi, arg(4) ;dst_ptr
movsxd rdx, dword ptr arg(5) ;dst_pitch
movdqa xmm1, [rax]
movdqa xmm2, [rax+16]
movsxd rax, dword ptr arg(3) ;yoffset
shl rax, 5
add rax, rcx ;VFilter
lea rcx, [rdi+rdx*8]
movdqa xmm5, [rax]
movdqa xmm6, [rax+16]
pxor xmm0, xmm0
; get the first horizontal line done
movdqa xmm3, XMMWORD PTR [rsp]
movdqa xmm4, xmm3 ; make a copy of current line
psrldq xmm4, 1
punpcklbw xmm3, xmm0 ; 00 01 02 03 04 05 06 07
punpcklbw xmm4, xmm0 ; 01 02 03 04 05 06 07 08
pmullw xmm3, xmm1
pmullw xmm4, xmm2
paddw xmm3, xmm4
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
movdqa xmm7, xmm3
add rsp, 16 ; next line
next_row8x8:
movdqa xmm3, XMMWORD PTR [rsp] ; 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15
movdqa xmm4, xmm3 ; make a copy of current line
psrldq xmm4, 1
punpcklbw xmm3, xmm0 ; 00 01 02 03 04 05 06 07
punpcklbw xmm4, xmm0 ; 01 02 03 04 05 06 07 08
pmullw xmm3, xmm1
pmullw xmm4, xmm2
paddw xmm3, xmm4
pmullw xmm7, xmm5
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
movdqa xmm4, xmm3
pmullw xmm3, xmm6
paddw xmm3, xmm7
movdqa xmm7, xmm4
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
packuswb xmm3, xmm0
movq [rdi], xmm3 ; store the results in the destination
add rsp, 16 ; next line
add rdi, rdx
cmp rdi, rcx
jne next_row8x8
;add rsp, 144
pop rsp
; begin epilog
pop rdi
pop rsi
RESTORE_GOT
UNSHADOW_ARGS
pop rbp
ret
SECTION_RODATA
align 16
rd:
times 8 dw 0x40
|
