summaryrefslogtreecommitdiffstats
path: root/media/libvpx/vp8/common/x86/subpixel_ssse3.asm
blob: c06f24556e687c2f6653bdbb0dfb2840a4e377f5 (plain)
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
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
;
;  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
;
;  Use of this source code is governed by a BSD-style license
;  that can be found in the LICENSE file in the root of the source
;  tree. An additional intellectual property rights grant can be found
;  in the file PATENTS.  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.
;
; This is an implementation of some of the SSE optimizations first seen in ffvp8
;
;*************************************************************************************/
;void vp8_filter_block1d8_h6_ssse3
;(
;    unsigned char  *src_ptr,
;    unsigned int    src_pixels_per_line,
;    unsigned char *output_ptr,
;    unsigned int    output_pitch,
;    unsigned int    output_height,
;    unsigned int    vp8_filter_index
;)
global sym(vp8_filter_block1d8_h6_ssse3) PRIVATE
sym(vp8_filter_block1d8_h6_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    movsxd      rdx, DWORD PTR arg(5)   ;table index
    xor         rsi, rsi
    shl         rdx, 4

    movdqa      xmm7, [GLOBAL(rd)]

    lea         rax, [GLOBAL(k0_k5)]
    add         rax, rdx
    mov         rdi, arg(2)             ;output_ptr

    cmp         esi, DWORD PTR [rax]
    je          vp8_filter_block1d8_h4_ssse3

    movdqa      xmm4, XMMWORD PTR [rax]         ;k0_k5
    movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3

    mov         rsi, arg(0)             ;src_ptr
    movsxd      rax, dword ptr arg(1)   ;src_pixels_per_line
    movsxd      rcx, dword ptr arg(4)   ;output_height

    movsxd      rdx, dword ptr arg(3)   ;output_pitch

    sub         rdi, rdx
;xmm3 free
.filter_block1d8_h6_rowloop_ssse3:
    movq        xmm0,   MMWORD PTR [rsi - 2]    ; -2 -1  0  1  2  3  4  5

    movq        xmm2,   MMWORD PTR [rsi + 3]    ;  3  4  5  6  7  8  9 10

    punpcklbw   xmm0,   xmm2                    ; -2  3 -1  4  0  5  1  6  2  7  3  8  4  9  5 10

    movdqa      xmm1,   xmm0
    pmaddubsw   xmm0,   xmm4

    movdqa      xmm2,   xmm1
    pshufb      xmm1,   [GLOBAL(shuf2bfrom1)]

    pshufb      xmm2,   [GLOBAL(shuf3bfrom1)]
    pmaddubsw   xmm1,   xmm5

    lea         rdi,    [rdi + rdx]
    pmaddubsw   xmm2,   xmm6

    lea         rsi,    [rsi + rax]
    dec         rcx

    paddsw      xmm0,   xmm1
    paddsw      xmm2,   xmm7

    paddsw      xmm0,   xmm2

    psraw       xmm0,   7

    packuswb    xmm0,   xmm0

    movq        MMWORD Ptr [rdi], xmm0
    jnz         .filter_block1d8_h6_rowloop_ssse3

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

vp8_filter_block1d8_h4_ssse3:
    movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3

    movdqa      xmm3, XMMWORD PTR [GLOBAL(shuf2bfrom1)]
    movdqa      xmm4, XMMWORD PTR [GLOBAL(shuf3bfrom1)]

    mov         rsi, arg(0)             ;src_ptr

    movsxd      rax, dword ptr arg(1)   ;src_pixels_per_line
    movsxd      rcx, dword ptr arg(4)   ;output_height

    movsxd      rdx, dword ptr arg(3)   ;output_pitch

    sub         rdi, rdx

.filter_block1d8_h4_rowloop_ssse3:
    movq        xmm0,   MMWORD PTR [rsi - 2]    ; -2 -1  0  1  2  3  4  5

    movq        xmm1,   MMWORD PTR [rsi + 3]    ;  3  4  5  6  7  8  9 10

    punpcklbw   xmm0,   xmm1                    ; -2  3 -1  4  0  5  1  6  2  7  3  8  4  9  5 10

    movdqa      xmm2,   xmm0
    pshufb      xmm0,   xmm3

    pshufb      xmm2,   xmm4
    pmaddubsw   xmm0,   xmm5

    lea         rdi,    [rdi + rdx]
    pmaddubsw   xmm2,   xmm6

    lea         rsi,    [rsi + rax]
    dec         rcx

    paddsw      xmm0,   xmm7

    paddsw      xmm0,   xmm2

    psraw       xmm0,   7

    packuswb    xmm0,   xmm0

    movq        MMWORD Ptr [rdi], xmm0

    jnz         .filter_block1d8_h4_rowloop_ssse3

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret
;void vp8_filter_block1d16_h6_ssse3
;(
;    unsigned char  *src_ptr,
;    unsigned int    src_pixels_per_line,
;    unsigned char  *output_ptr,
;    unsigned int    output_pitch,
;    unsigned int    output_height,
;    unsigned int    vp8_filter_index
;)
global sym(vp8_filter_block1d16_h6_ssse3) PRIVATE
sym(vp8_filter_block1d16_h6_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    movsxd      rdx, DWORD PTR arg(5)           ;table index
    xor         rsi, rsi
    shl         rdx, 4      ;

    lea         rax, [GLOBAL(k0_k5)]
    add         rax, rdx

    mov         rdi, arg(2)                     ;output_ptr

    mov         rsi, arg(0)                     ;src_ptr

    movdqa      xmm4, XMMWORD PTR [rax]         ;k0_k5
    movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3

    movsxd      rax, dword ptr arg(1)           ;src_pixels_per_line
    movsxd      rcx, dword ptr arg(4)           ;output_height
    movsxd      rdx, dword ptr arg(3)           ;output_pitch

.filter_block1d16_h6_rowloop_ssse3:
    movq        xmm0,   MMWORD PTR [rsi - 2]    ; -2 -1  0  1  2  3  4  5

    movq        xmm3,   MMWORD PTR [rsi + 3]    ;  3  4  5  6  7  8  9 10

    punpcklbw   xmm0,   xmm3                    ; -2  3 -1  4  0  5  1  6  2  7  3  8  4  9  5 10

    movdqa      xmm1,   xmm0
    pmaddubsw   xmm0,   xmm4

    movdqa      xmm2,   xmm1
    pshufb      xmm1,   [GLOBAL(shuf2bfrom1)]

    pshufb      xmm2,   [GLOBAL(shuf3bfrom1)]
    movq        xmm3,   MMWORD PTR [rsi +  6]

    pmaddubsw   xmm1,   xmm5
    movq        xmm7,   MMWORD PTR [rsi + 11]

    pmaddubsw   xmm2,   xmm6
    punpcklbw   xmm3,   xmm7

    paddsw      xmm0,   xmm1
    movdqa      xmm1,   xmm3

    pmaddubsw   xmm3,   xmm4
    paddsw      xmm0,   xmm2

    movdqa      xmm2,   xmm1
    paddsw      xmm0,   [GLOBAL(rd)]

    pshufb      xmm1,   [GLOBAL(shuf2bfrom1)]
    pshufb      xmm2,   [GLOBAL(shuf3bfrom1)]

    psraw       xmm0,   7
    pmaddubsw   xmm1,   xmm5

    pmaddubsw   xmm2,   xmm6
    packuswb    xmm0,   xmm0

    lea         rsi,    [rsi + rax]
    paddsw      xmm3,   xmm1

    paddsw      xmm3,   xmm2

    paddsw      xmm3,   [GLOBAL(rd)]

    psraw       xmm3,   7

    packuswb    xmm3,   xmm3

    punpcklqdq  xmm0,   xmm3

    movdqa      XMMWORD Ptr [rdi], xmm0

    lea         rdi,    [rdi + rdx]
    dec         rcx
    jnz         .filter_block1d16_h6_rowloop_ssse3

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

;void vp8_filter_block1d4_h6_ssse3
;(
;    unsigned char  *src_ptr,
;    unsigned int    src_pixels_per_line,
;    unsigned char  *output_ptr,
;    unsigned int    output_pitch,
;    unsigned int    output_height,
;    unsigned int    vp8_filter_index
;)
global sym(vp8_filter_block1d4_h6_ssse3) PRIVATE
sym(vp8_filter_block1d4_h6_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    movsxd      rdx, DWORD PTR arg(5)   ;table index
    xor         rsi, rsi
    shl         rdx, 4      ;

    lea         rax, [GLOBAL(k0_k5)]
    add         rax, rdx
    movdqa      xmm7, [GLOBAL(rd)]

    cmp         esi, DWORD PTR [rax]
    je          .vp8_filter_block1d4_h4_ssse3

    movdqa      xmm4, XMMWORD PTR [rax]         ;k0_k5
    movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3

    mov         rsi, arg(0)             ;src_ptr
    mov         rdi, arg(2)             ;output_ptr
    movsxd      rax, dword ptr arg(1)   ;src_pixels_per_line
    movsxd      rcx, dword ptr arg(4)   ;output_height

    movsxd      rdx, dword ptr arg(3)   ;output_pitch

;xmm3 free
.filter_block1d4_h6_rowloop_ssse3:
    movdqu      xmm0,   XMMWORD PTR [rsi - 2]

    movdqa      xmm1, xmm0
    pshufb      xmm0, [GLOBAL(shuf1b)]

    movdqa      xmm2, xmm1
    pshufb      xmm1, [GLOBAL(shuf2b)]
    pmaddubsw   xmm0, xmm4
    pshufb      xmm2, [GLOBAL(shuf3b)]
    pmaddubsw   xmm1, xmm5

;--
    pmaddubsw   xmm2, xmm6

    lea         rsi,    [rsi + rax]
;--
    paddsw      xmm0, xmm1
    paddsw      xmm0, xmm7
    pxor        xmm1, xmm1
    paddsw      xmm0, xmm2
    psraw       xmm0, 7
    packuswb    xmm0, xmm0

    movd        DWORD PTR [rdi], xmm0

    add         rdi, rdx
    dec         rcx
    jnz         .filter_block1d4_h6_rowloop_ssse3

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

.vp8_filter_block1d4_h4_ssse3:
    movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3
    movdqa      xmm0, XMMWORD PTR [GLOBAL(shuf2b)]
    movdqa      xmm3, XMMWORD PTR [GLOBAL(shuf3b)]

    mov         rsi, arg(0)             ;src_ptr
    mov         rdi, arg(2)             ;output_ptr
    movsxd      rax, dword ptr arg(1)   ;src_pixels_per_line
    movsxd      rcx, dword ptr arg(4)   ;output_height

    movsxd      rdx, dword ptr arg(3)   ;output_pitch

.filter_block1d4_h4_rowloop_ssse3:
    movdqu      xmm1,   XMMWORD PTR [rsi - 2]

    movdqa      xmm2, xmm1
    pshufb      xmm1, xmm0 ;;[GLOBAL(shuf2b)]
    pshufb      xmm2, xmm3 ;;[GLOBAL(shuf3b)]
    pmaddubsw   xmm1, xmm5

;--
    pmaddubsw   xmm2, xmm6

    lea         rsi,    [rsi + rax]
;--
    paddsw      xmm1, xmm7
    paddsw      xmm1, xmm2
    psraw       xmm1, 7
    packuswb    xmm1, xmm1

    movd        DWORD PTR [rdi], xmm1

    add         rdi, rdx
    dec         rcx
    jnz         .filter_block1d4_h4_rowloop_ssse3

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret



;void vp8_filter_block1d16_v6_ssse3
;(
;    unsigned char *src_ptr,
;    unsigned int   src_pitch,
;    unsigned char *output_ptr,
;    unsigned int   out_pitch,
;    unsigned int   output_height,
;    unsigned int   vp8_filter_index
;)
global sym(vp8_filter_block1d16_v6_ssse3) PRIVATE
sym(vp8_filter_block1d16_v6_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    movsxd      rdx, DWORD PTR arg(5)   ;table index
    xor         rsi, rsi
    shl         rdx, 4      ;

    lea         rax, [GLOBAL(k0_k5)]
    add         rax, rdx

    cmp         esi, DWORD PTR [rax]
    je          .vp8_filter_block1d16_v4_ssse3

    movdqa      xmm5, XMMWORD PTR [rax]         ;k0_k5
    movdqa      xmm6, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm7, XMMWORD PTR [rax+128]     ;k1_k3

    mov         rsi, arg(0)             ;src_ptr
    movsxd      rdx, DWORD PTR arg(1)   ;pixels_per_line
    mov         rdi, arg(2)             ;output_ptr

%if ABI_IS_32BIT=0
    movsxd      r8, DWORD PTR arg(3)    ;out_pitch
%endif
    mov         rax, rsi
    movsxd      rcx, DWORD PTR arg(4)   ;output_height
    add         rax, rdx


.vp8_filter_block1d16_v6_ssse3_loop:
    movq        xmm1, MMWORD PTR [rsi]                  ;A
    movq        xmm2, MMWORD PTR [rsi + rdx]            ;B
    movq        xmm3, MMWORD PTR [rsi + rdx * 2]        ;C
    movq        xmm4, MMWORD PTR [rax + rdx * 2]        ;D
    movq        xmm0, MMWORD PTR [rsi + rdx * 4]        ;E

    punpcklbw   xmm2, xmm4                  ;B D
    punpcklbw   xmm3, xmm0                  ;C E

    movq        xmm0, MMWORD PTR [rax + rdx * 4]        ;F

    pmaddubsw   xmm3, xmm6
    punpcklbw   xmm1, xmm0                  ;A F
    pmaddubsw   xmm2, xmm7
    pmaddubsw   xmm1, xmm5

    paddsw      xmm2, xmm3
    paddsw      xmm2, xmm1
    paddsw      xmm2, [GLOBAL(rd)]
    psraw       xmm2, 7
    packuswb    xmm2, xmm2

    movq        MMWORD PTR [rdi], xmm2          ;store the results

    movq        xmm1, MMWORD PTR [rsi + 8]                  ;A
    movq        xmm2, MMWORD PTR [rsi + rdx + 8]            ;B
    movq        xmm3, MMWORD PTR [rsi + rdx * 2 + 8]        ;C
    movq        xmm4, MMWORD PTR [rax + rdx * 2 + 8]        ;D
    movq        xmm0, MMWORD PTR [rsi + rdx * 4 + 8]        ;E

    punpcklbw   xmm2, xmm4                  ;B D
    punpcklbw   xmm3, xmm0                  ;C E

    movq        xmm0, MMWORD PTR [rax + rdx * 4 + 8]        ;F
    pmaddubsw   xmm3, xmm6
    punpcklbw   xmm1, xmm0                  ;A F
    pmaddubsw   xmm2, xmm7
    pmaddubsw   xmm1, xmm5

    add         rsi,  rdx
    add         rax,  rdx
;--
;--
    paddsw      xmm2, xmm3
    paddsw      xmm2, xmm1
    paddsw      xmm2, [GLOBAL(rd)]
    psraw       xmm2, 7
    packuswb    xmm2, xmm2

    movq        MMWORD PTR [rdi+8], xmm2

%if ABI_IS_32BIT
    add         rdi,        DWORD PTR arg(3) ;out_pitch
%else
    add         rdi,        r8
%endif
    dec         rcx
    jnz         .vp8_filter_block1d16_v6_ssse3_loop

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

.vp8_filter_block1d16_v4_ssse3:
    movdqa      xmm6, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm7, XMMWORD PTR [rax+128]     ;k1_k3

    mov         rsi, arg(0)             ;src_ptr
    movsxd      rdx, DWORD PTR arg(1)   ;pixels_per_line
    mov         rdi, arg(2)             ;output_ptr

%if ABI_IS_32BIT=0
    movsxd      r8, DWORD PTR arg(3)    ;out_pitch
%endif
    mov         rax, rsi
    movsxd      rcx, DWORD PTR arg(4)   ;output_height
    add         rax, rdx

.vp8_filter_block1d16_v4_ssse3_loop:
    movq        xmm2, MMWORD PTR [rsi + rdx]            ;B
    movq        xmm3, MMWORD PTR [rsi + rdx * 2]        ;C
    movq        xmm4, MMWORD PTR [rax + rdx * 2]        ;D
    movq        xmm0, MMWORD PTR [rsi + rdx * 4]        ;E

    punpcklbw   xmm2, xmm4                  ;B D
    punpcklbw   xmm3, xmm0                  ;C E

    pmaddubsw   xmm3, xmm6
    pmaddubsw   xmm2, xmm7
    movq        xmm5, MMWORD PTR [rsi + rdx + 8]            ;B
    movq        xmm1, MMWORD PTR [rsi + rdx * 2 + 8]        ;C
    movq        xmm4, MMWORD PTR [rax + rdx * 2 + 8]        ;D
    movq        xmm0, MMWORD PTR [rsi + rdx * 4 + 8]        ;E

    paddsw      xmm2, [GLOBAL(rd)]
    paddsw      xmm2, xmm3
    psraw       xmm2, 7
    packuswb    xmm2, xmm2

    punpcklbw   xmm5, xmm4                  ;B D
    punpcklbw   xmm1, xmm0                  ;C E

    pmaddubsw   xmm1, xmm6
    pmaddubsw   xmm5, xmm7

    movdqa      xmm4, [GLOBAL(rd)]
    add         rsi,  rdx
    add         rax,  rdx
;--
;--
    paddsw      xmm5, xmm1
    paddsw      xmm5, xmm4
    psraw       xmm5, 7
    packuswb    xmm5, xmm5

    punpcklqdq  xmm2, xmm5

    movdqa       XMMWORD PTR [rdi], xmm2

%if ABI_IS_32BIT
    add         rdi,        DWORD PTR arg(3) ;out_pitch
%else
    add         rdi,        r8
%endif
    dec         rcx
    jnz         .vp8_filter_block1d16_v4_ssse3_loop

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

;void vp8_filter_block1d8_v6_ssse3
;(
;    unsigned char *src_ptr,
;    unsigned int   src_pitch,
;    unsigned char *output_ptr,
;    unsigned int   out_pitch,
;    unsigned int   output_height,
;    unsigned int   vp8_filter_index
;)
global sym(vp8_filter_block1d8_v6_ssse3) PRIVATE
sym(vp8_filter_block1d8_v6_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    movsxd      rdx, DWORD PTR arg(5)   ;table index
    xor         rsi, rsi
    shl         rdx, 4      ;

    lea         rax, [GLOBAL(k0_k5)]
    add         rax, rdx

    movsxd      rdx, DWORD PTR arg(1)   ;pixels_per_line
    mov         rdi, arg(2)             ;output_ptr
%if ABI_IS_32BIT=0
    movsxd      r8, DWORD PTR arg(3)    ; out_pitch
%endif
    movsxd      rcx, DWORD PTR arg(4)   ;[output_height]

    cmp         esi, DWORD PTR [rax]
    je          .vp8_filter_block1d8_v4_ssse3

    movdqa      xmm5, XMMWORD PTR [rax]         ;k0_k5
    movdqa      xmm6, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm7, XMMWORD PTR [rax+128]     ;k1_k3

    mov         rsi, arg(0)             ;src_ptr

    mov         rax, rsi
    add         rax, rdx

.vp8_filter_block1d8_v6_ssse3_loop:
    movq        xmm1, MMWORD PTR [rsi]                  ;A
    movq        xmm2, MMWORD PTR [rsi + rdx]            ;B
    movq        xmm3, MMWORD PTR [rsi + rdx * 2]        ;C
    movq        xmm4, MMWORD PTR [rax + rdx * 2]        ;D
    movq        xmm0, MMWORD PTR [rsi + rdx * 4]        ;E

    punpcklbw   xmm2, xmm4                  ;B D
    punpcklbw   xmm3, xmm0                  ;C E

    movq        xmm0, MMWORD PTR [rax + rdx * 4]        ;F
    movdqa      xmm4, [GLOBAL(rd)]

    pmaddubsw   xmm3, xmm6
    punpcklbw   xmm1, xmm0                  ;A F
    pmaddubsw   xmm2, xmm7
    pmaddubsw   xmm1, xmm5
    add         rsi,  rdx
    add         rax,  rdx
;--
;--
    paddsw      xmm2, xmm3
    paddsw      xmm2, xmm1
    paddsw      xmm2, xmm4
    psraw       xmm2, 7
    packuswb    xmm2, xmm2

    movq        MMWORD PTR [rdi], xmm2

%if ABI_IS_32BIT
    add         rdi,        DWORD PTR arg(3) ;[out_pitch]
%else
    add         rdi,        r8
%endif
    dec         rcx
    jnz         .vp8_filter_block1d8_v6_ssse3_loop

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

.vp8_filter_block1d8_v4_ssse3:
    movdqa      xmm6, XMMWORD PTR [rax+256]     ;k2_k4
    movdqa      xmm7, XMMWORD PTR [rax+128]     ;k1_k3
    movdqa      xmm5, [GLOBAL(rd)]

    mov         rsi, arg(0)             ;src_ptr

    mov         rax, rsi
    add         rax, rdx

.vp8_filter_block1d8_v4_ssse3_loop:
    movq        xmm2, MMWORD PTR [rsi + rdx]            ;B
    movq        xmm3, MMWORD PTR [rsi + rdx * 2]        ;C
    movq        xmm4, MMWORD PTR [rax + rdx * 2]        ;D
    movq        xmm0, MMWORD PTR [rsi + rdx * 4]        ;E

    punpcklbw   xmm2, xmm4                  ;B D
    punpcklbw   xmm3, xmm0                  ;C E

    pmaddubsw   xmm3, xmm6
    pmaddubsw   xmm2, xmm7
    add         rsi,  rdx
    add         rax,  rdx
;--
;--
    paddsw      xmm2, xmm3
    paddsw      xmm2, xmm5
    psraw       xmm2, 7
    packuswb    xmm2, xmm2

    movq        MMWORD PTR [rdi], xmm2

%if ABI_IS_32BIT
    add         rdi,        DWORD PTR arg(3) ;[out_pitch]
%else
    add         rdi,        r8
%endif
    dec         rcx
    jnz         .vp8_filter_block1d8_v4_ssse3_loop

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret
;void vp8_filter_block1d4_v6_ssse3
;(
;    unsigned char *src_ptr,
;    unsigned int   src_pitch,
;    unsigned char *output_ptr,
;    unsigned int   out_pitch,
;    unsigned int   output_height,
;    unsigned int   vp8_filter_index
;)
global sym(vp8_filter_block1d4_v6_ssse3) PRIVATE
sym(vp8_filter_block1d4_v6_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    movsxd      rdx, DWORD PTR arg(5)   ;table index
    xor         rsi, rsi
    shl         rdx, 4      ;

    lea         rax, [GLOBAL(k0_k5)]
    add         rax, rdx

    movsxd      rdx, DWORD PTR arg(1)   ;pixels_per_line
    mov         rdi, arg(2)             ;output_ptr
%if ABI_IS_32BIT=0
    movsxd      r8, DWORD PTR arg(3)    ; out_pitch
%endif
    movsxd      rcx, DWORD PTR arg(4)   ;[output_height]

    cmp         esi, DWORD PTR [rax]
    je          .vp8_filter_block1d4_v4_ssse3

    movq        mm5, MMWORD PTR [rax]         ;k0_k5
    movq        mm6, MMWORD PTR [rax+256]     ;k2_k4
    movq        mm7, MMWORD PTR [rax+128]     ;k1_k3

    mov         rsi, arg(0)             ;src_ptr

    mov         rax, rsi
    add         rax, rdx

.vp8_filter_block1d4_v6_ssse3_loop:
    movd        mm1, DWORD PTR [rsi]                  ;A
    movd        mm2, DWORD PTR [rsi + rdx]            ;B
    movd        mm3, DWORD PTR [rsi + rdx * 2]        ;C
    movd        mm4, DWORD PTR [rax + rdx * 2]        ;D
    movd        mm0, DWORD PTR [rsi + rdx * 4]        ;E

    punpcklbw   mm2, mm4                  ;B D
    punpcklbw   mm3, mm0                  ;C E

    movd        mm0, DWORD PTR [rax + rdx * 4]        ;F

    movq        mm4, [GLOBAL(rd)]

    pmaddubsw   mm3, mm6
    punpcklbw   mm1, mm0                  ;A F
    pmaddubsw   mm2, mm7
    pmaddubsw   mm1, mm5
    add         rsi,  rdx
    add         rax,  rdx
;--
;--
    paddsw      mm2, mm3
    paddsw      mm2, mm1
    paddsw      mm2, mm4
    psraw       mm2, 7
    packuswb    mm2, mm2

    movd        DWORD PTR [rdi], mm2

%if ABI_IS_32BIT
    add         rdi,        DWORD PTR arg(3) ;[out_pitch]
%else
    add         rdi,        r8
%endif
    dec         rcx
    jnz         .vp8_filter_block1d4_v6_ssse3_loop

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    UNSHADOW_ARGS
    pop         rbp
    ret

.vp8_filter_block1d4_v4_ssse3:
    movq        mm6, MMWORD PTR [rax+256]     ;k2_k4
    movq        mm7, MMWORD PTR [rax+128]     ;k1_k3
    movq        mm5, MMWORD PTR [GLOBAL(rd)]

    mov         rsi, arg(0)             ;src_ptr

    mov         rax, rsi
    add         rax, rdx

.vp8_filter_block1d4_v4_ssse3_loop:
    movd        mm2, DWORD PTR [rsi + rdx]            ;B
    movd        mm3, DWORD PTR [rsi + rdx * 2]        ;C
    movd        mm4, DWORD PTR [rax + rdx * 2]        ;D
    movd        mm0, DWORD PTR [rsi + rdx * 4]        ;E

    punpcklbw   mm2, mm4                  ;B D
    punpcklbw   mm3, mm0                  ;C E

    pmaddubsw   mm3, mm6
    pmaddubsw   mm2, mm7
    add         rsi,  rdx
    add         rax,  rdx
;--
;--
    paddsw      mm2, mm3
    paddsw      mm2, mm5
    psraw       mm2, 7
    packuswb    mm2, mm2

    movd        DWORD PTR [rdi], mm2

%if ABI_IS_32BIT
    add         rdi,        DWORD PTR arg(3) ;[out_pitch]
%else
    add         rdi,        r8
%endif
    dec         rcx
    jnz         .vp8_filter_block1d4_v4_ssse3_loop

    ; begin epilog
    pop rdi
    pop rsi
    RESTORE_GOT
    UNSHADOW_ARGS
    pop         rbp
    ret

;void vp8_bilinear_predict16x16_ssse3
;(
;    unsigned char  *src_ptr,
;    int   src_pixels_per_line,
;    int  xoffset,
;    int  yoffset,
;    unsigned char *dst_ptr,
;    int dst_pitch
;)
global sym(vp8_bilinear_predict16x16_ssse3) PRIVATE
sym(vp8_bilinear_predict16x16_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

        lea         rcx,        [GLOBAL(vp8_bilinear_filters_ssse3)]
        movsxd      rax,        dword ptr arg(2)    ; xoffset

        cmp         rax,        0                   ; skip first_pass filter if xoffset=0
        je          .b16x16_sp_only

        shl         rax,        4
        lea         rax,        [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]

        movsxd      rax,        dword ptr arg(3)    ; yoffset

        cmp         rax,        0                   ; skip second_pass filter if yoffset=0
        je          .b16x16_fp_only

        shl         rax,        4
        lea         rax,        [rax + rcx]         ; VFilter

        lea         rcx,        [rdi+rdx*8]
        lea         rcx,        [rcx+rdx*8]
        movsxd      rdx,        dword ptr arg(1)    ; src_pixels_per_line

        movdqa      xmm2,       [rax]

%if ABI_IS_32BIT=0
        movsxd      r8,         dword ptr arg(5)    ; dst_pitch
%endif
        movq        xmm3,       [rsi]               ; 00 01 02 03 04 05 06 07
        movq        xmm5,       [rsi+1]             ; 01 02 03 04 05 06 07 08

        punpcklbw   xmm3,       xmm5                ; 00 01 01 02 02 03 03 04 04 05 05 06 06 07 07 08
        movq        xmm4,       [rsi+8]             ; 08 09 10 11 12 13 14 15

        movq        xmm5,       [rsi+9]             ; 09 10 11 12 13 14 15 16

        lea         rsi,        [rsi + rdx]         ; next line

        pmaddubsw   xmm3,       xmm1                ; 00 02 04 06 08 10 12 14

        punpcklbw   xmm4,       xmm5                ; 08 09 09 10 10 11 11 12 12 13 13 14 14 15 15 16
        pmaddubsw   xmm4,       xmm1                ; 01 03 05 07 09 11 13 15

        paddw       xmm3,       [GLOBAL(rd)]        ; xmm3 += round value
        psraw       xmm3,       VP8_FILTER_SHIFT    ; xmm3 /= 128

        paddw       xmm4,       [GLOBAL(rd)]        ; xmm4 += round value
        psraw       xmm4,       VP8_FILTER_SHIFT    ; xmm4 /= 128

        movdqa      xmm7,       xmm3
        packuswb    xmm7,       xmm4                ; 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

.next_row:
        movq        xmm6,       [rsi]               ; 00 01 02 03 04 05 06 07
        movq        xmm5,       [rsi+1]             ; 01 02 03 04 05 06 07 08

        punpcklbw   xmm6,       xmm5
        movq        xmm4,       [rsi+8]             ; 08 09 10 11 12 13 14 15

        movq        xmm5,       [rsi+9]             ; 09 10 11 12 13 14 15 16
        lea         rsi,        [rsi + rdx]         ; next line

        pmaddubsw   xmm6,       xmm1

        punpcklbw   xmm4,       xmm5
        pmaddubsw   xmm4,       xmm1

        paddw       xmm6,       [GLOBAL(rd)]        ; xmm6 += round value
        psraw       xmm6,       VP8_FILTER_SHIFT    ; xmm6 /= 128

        paddw       xmm4,       [GLOBAL(rd)]        ; xmm4 += round value
        psraw       xmm4,       VP8_FILTER_SHIFT    ; xmm4 /= 128

        packuswb    xmm6,       xmm4
        movdqa      xmm5,       xmm7

        punpcklbw   xmm5,       xmm6
        pmaddubsw   xmm5,       xmm2

        punpckhbw   xmm7,       xmm6
        pmaddubsw   xmm7,       xmm2

        paddw       xmm5,       [GLOBAL(rd)]        ; xmm5 += round value
        psraw       xmm5,       VP8_FILTER_SHIFT    ; xmm5 /= 128

        paddw       xmm7,       [GLOBAL(rd)]        ; xmm7 += round value
        psraw       xmm7,       VP8_FILTER_SHIFT    ; xmm7 /= 128

        packuswb    xmm5,       xmm7
        movdqa      xmm7,       xmm6

        movdqa      [rdi],      xmm5                ; store the results in the destination
%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,        4
        lea         rax,        [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]               ; VFilter

        lea         rcx,        [rdi+rdx*8]
        lea         rcx,        [rcx+rdx*8]
        movsxd      rax,        dword ptr arg(1)    ; src_pixels_per_line

        ; get the first horizontal line done
        movq        xmm4,       [rsi]               ; load row 0
        movq        xmm2,       [rsi + 8]           ; load row 0

        lea         rsi,        [rsi + rax]         ; next line
.next_row_sp:
        movq        xmm3,       [rsi]               ; load row + 1
        movq        xmm5,       [rsi + 8]           ; load row + 1

        punpcklbw   xmm4,       xmm3
        punpcklbw   xmm2,       xmm5

        pmaddubsw   xmm4,       xmm1
        movq        xmm7,       [rsi + rax]         ; load row + 2

        pmaddubsw   xmm2,       xmm1
        movq        xmm6,       [rsi + rax + 8]     ; load row + 2

        punpcklbw   xmm3,       xmm7
        punpcklbw   xmm5,       xmm6

        pmaddubsw   xmm3,       xmm1
        paddw       xmm4,       [GLOBAL(rd)]

        pmaddubsw   xmm5,       xmm1
        paddw       xmm2,       [GLOBAL(rd)]

        psraw       xmm4,       VP8_FILTER_SHIFT
        psraw       xmm2,       VP8_FILTER_SHIFT

        packuswb    xmm4,       xmm2
        paddw       xmm3,       [GLOBAL(rd)]

        movdqa      [rdi],      xmm4                ; store row 0
        paddw       xmm5,       [GLOBAL(rd)]

        psraw       xmm3,       VP8_FILTER_SHIFT
        psraw       xmm5,       VP8_FILTER_SHIFT

        packuswb    xmm3,       xmm5
        movdqa      xmm4,       xmm7

        movdqa      [rdi + rdx],xmm3                ; store row 1
        lea         rsi,        [rsi + 2*rax]

        movdqa      xmm2,       xmm6
        lea         rdi,        [rdi + 2*rdx]

        cmp         rdi,        rcx
        jne         .next_row_sp

        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

.next_row_fp:
        movq        xmm2,       [rsi]               ; 00 01 02 03 04 05 06 07
        movq        xmm4,       [rsi+1]             ; 01 02 03 04 05 06 07 08

        punpcklbw   xmm2,       xmm4
        movq        xmm3,       [rsi+8]             ; 08 09 10 11 12 13 14 15

        pmaddubsw   xmm2,       xmm1
        movq        xmm4,       [rsi+9]             ; 09 10 11 12 13 14 15 16

        lea         rsi,        [rsi + rax]         ; next line
        punpcklbw   xmm3,       xmm4

        pmaddubsw   xmm3,       xmm1
        movq        xmm5,       [rsi]

        paddw       xmm2,       [GLOBAL(rd)]
        movq        xmm7,       [rsi+1]

        movq        xmm6,       [rsi+8]
        psraw       xmm2,       VP8_FILTER_SHIFT

        punpcklbw   xmm5,       xmm7
        movq        xmm7,       [rsi+9]

        paddw       xmm3,       [GLOBAL(rd)]
        pmaddubsw   xmm5,       xmm1

        psraw       xmm3,       VP8_FILTER_SHIFT
        punpcklbw   xmm6,       xmm7

        packuswb    xmm2,       xmm3
        pmaddubsw   xmm6,       xmm1

        movdqa      [rdi],      xmm2                ; store the results in the destination
        paddw       xmm5,       [GLOBAL(rd)]

        lea         rdi,        [rdi + rdx]         ; dst_pitch
        psraw       xmm5,       VP8_FILTER_SHIFT

        paddw       xmm6,       [GLOBAL(rd)]
        psraw       xmm6,       VP8_FILTER_SHIFT

        packuswb    xmm5,       xmm6
        lea         rsi,        [rsi + rax]         ; next line

        movdqa      [rdi],      xmm5                ; store the results in the destination
        lea         rdi,        [rdi + rdx]         ; dst_pitch

        cmp         rdi,        rcx

        jne         .next_row_fp

.done:
    ; begin epilog
    pop         rdi
    pop         rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

;void vp8_bilinear_predict8x8_ssse3
;(
;    unsigned char  *src_ptr,
;    int   src_pixels_per_line,
;    int  xoffset,
;    int  yoffset,
;    unsigned char *dst_ptr,
;    int dst_pitch
;)
global sym(vp8_bilinear_predict8x8_ssse3) PRIVATE
sym(vp8_bilinear_predict8x8_ssse3):
    push        rbp
    mov         rbp, rsp
    SHADOW_ARGS_TO_STACK 6
    SAVE_XMM 7
    GET_GOT     rbx
    push        rsi
    push        rdi
    ; end prolog

    ALIGN_STACK 16, rax
    sub         rsp, 144                         ; reserve 144 bytes

        lea         rcx,        [GLOBAL(vp8_bilinear_filters_ssse3)]

        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
        cmp         rax,        0                   ; skip first_pass filter if xoffset=0
        je          .b8x8_sp_only

        shl         rax,        4
        add         rax,        rcx                 ; HFilter

        mov         rdi,        arg(4)              ; dst_ptr
        movsxd      rdx,        dword ptr arg(5)    ; dst_pitch

        movdqa      xmm0,       [rax]

        movsxd      rax,        dword ptr arg(3)    ; yoffset
        cmp         rax,        0                   ; skip second_pass filter if yoffset=0
        je          .b8x8_fp_only

        shl         rax,        4
        lea         rax,        [rax + rcx]         ; VFilter

        lea         rcx,        [rdi+rdx*8]

        movdqa      xmm1,       [rax]

        ; get the first horizontal line done
        movdqa      xmm3,       [rsp]               ; 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15
        movdqa      xmm5,       xmm3                ; 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 xx

        psrldq      xmm5,       1
        lea         rsp,        [rsp + 16]          ; next line

        punpcklbw   xmm3,       xmm5                ; 00 01 01 02 02 03 03 04 04 05 05 06 06 07 07 08
        pmaddubsw   xmm3,       xmm0                ; 00 02 04 06 08 10 12 14

        paddw       xmm3,       [GLOBAL(rd)]        ; xmm3 += round value
        psraw       xmm3,       VP8_FILTER_SHIFT    ; xmm3 /= 128

        movdqa      xmm7,       xmm3
        packuswb    xmm7,       xmm7                ; 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

.next_row:
        movdqa      xmm6,       [rsp]               ; 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15
        lea         rsp,        [rsp + 16]          ; next line

        movdqa      xmm5,       xmm6

        psrldq      xmm5,       1

        punpcklbw   xmm6,       xmm5
        pmaddubsw   xmm6,       xmm0

        paddw       xmm6,       [GLOBAL(rd)]        ; xmm6 += round value
        psraw       xmm6,       VP8_FILTER_SHIFT    ; xmm6 /= 128

        packuswb    xmm6,       xmm6

        punpcklbw   xmm7,       xmm6
        pmaddubsw   xmm7,       xmm1

        paddw       xmm7,       [GLOBAL(rd)]        ; xmm7 += round value
        psraw       xmm7,       VP8_FILTER_SHIFT    ; xmm7 /= 128

        packuswb    xmm7,       xmm7

        movq        [rdi],      xmm7                ; store the results in the destination
        lea         rdi,        [rdi + rdx]

        movdqa      xmm7,       xmm6

        cmp         rdi,        rcx
        jne         .next_row

        jmp         .done8x8

.b8x8_sp_only:
        movsxd      rax,        dword ptr arg(3)    ; yoffset
        shl         rax,        4
        lea         rax,        [rax + rcx]         ; VFilter

        mov         rdi,        arg(4) ;dst_ptr
        movsxd      rdx,        dword ptr arg(5)    ; dst_pitch

        movdqa      xmm0,       [rax]               ; VFilter

        movq        xmm1,       XMMWORD PTR [rsp]
        movq        xmm2,       XMMWORD PTR [rsp+16]

        movq        xmm3,       XMMWORD PTR [rsp+32]
        punpcklbw   xmm1,       xmm2

        movq        xmm4,       XMMWORD PTR [rsp+48]
        punpcklbw   xmm2,       xmm3

        movq        xmm5,       XMMWORD PTR [rsp+64]
        punpcklbw   xmm3,       xmm4

        movq        xmm6,       XMMWORD PTR [rsp+80]
        punpcklbw   xmm4,       xmm5

        movq        xmm7,       XMMWORD PTR [rsp+96]
        punpcklbw   xmm5,       xmm6

        pmaddubsw   xmm1,       xmm0
        pmaddubsw   xmm2,       xmm0

        pmaddubsw   xmm3,       xmm0
        pmaddubsw   xmm4,       xmm0

        pmaddubsw   xmm5,       xmm0
        punpcklbw   xmm6,       xmm7

        pmaddubsw   xmm6,       xmm0
        paddw       xmm1,       [GLOBAL(rd)]

        paddw       xmm2,       [GLOBAL(rd)]
        psraw       xmm1,       VP8_FILTER_SHIFT

        paddw       xmm3,       [GLOBAL(rd)]
        psraw       xmm2,       VP8_FILTER_SHIFT

        paddw       xmm4,       [GLOBAL(rd)]
        psraw       xmm3,       VP8_FILTER_SHIFT

        paddw       xmm5,       [GLOBAL(rd)]
        psraw       xmm4,       VP8_FILTER_SHIFT

        paddw       xmm6,       [GLOBAL(rd)]
        psraw       xmm5,       VP8_FILTER_SHIFT

        psraw       xmm6,       VP8_FILTER_SHIFT
        packuswb    xmm1,       xmm1

        packuswb    xmm2,       xmm2
        movq        [rdi],      xmm1

        packuswb    xmm3,       xmm3
        movq        [rdi+rdx],  xmm2

        packuswb    xmm4,       xmm4
        movq        xmm1,       XMMWORD PTR [rsp+112]

        lea         rdi,        [rdi + 2*rdx]
        movq        xmm2,       XMMWORD PTR [rsp+128]

        packuswb    xmm5,       xmm5
        movq        [rdi],      xmm3

        packuswb    xmm6,       xmm6
        movq        [rdi+rdx],  xmm4

        lea         rdi,        [rdi + 2*rdx]
        punpcklbw   xmm7,       xmm1

        movq        [rdi],      xmm5
        pmaddubsw   xmm7,       xmm0

        movq        [rdi+rdx],  xmm6
        punpcklbw   xmm1,       xmm2

        pmaddubsw   xmm1,       xmm0
        paddw       xmm7,       [GLOBAL(rd)]

        psraw       xmm7,       VP8_FILTER_SHIFT
        paddw       xmm1,       [GLOBAL(rd)]

        psraw       xmm1,       VP8_FILTER_SHIFT
        packuswb    xmm7,       xmm7

        packuswb    xmm1,       xmm1
        lea         rdi,        [rdi + 2*rdx]

        movq        [rdi],      xmm7

        movq        [rdi+rdx],  xmm1
        lea         rsp,        [rsp + 144]

        jmp         .done8x8

.b8x8_fp_only:
        lea         rcx,        [rdi+rdx*8]

.next_row_fp:
        movdqa      xmm1,       XMMWORD PTR [rsp]
        movdqa      xmm3,       XMMWORD PTR [rsp+16]

        movdqa      xmm2,       xmm1
        movdqa      xmm5,       XMMWORD PTR [rsp+32]

        psrldq      xmm2,       1
        movdqa      xmm7,       XMMWORD PTR [rsp+48]

        movdqa      xmm4,       xmm3
        psrldq      xmm4,       1

        movdqa      xmm6,       xmm5
        psrldq      xmm6,       1

        punpcklbw   xmm1,       xmm2
        pmaddubsw   xmm1,       xmm0

        punpcklbw   xmm3,       xmm4
        pmaddubsw   xmm3,       xmm0

        punpcklbw   xmm5,       xmm6
        pmaddubsw   xmm5,       xmm0

        movdqa      xmm2,       xmm7
        psrldq      xmm2,       1

        punpcklbw   xmm7,       xmm2
        pmaddubsw   xmm7,       xmm0

        paddw       xmm1,       [GLOBAL(rd)]
        psraw       xmm1,       VP8_FILTER_SHIFT

        paddw       xmm3,       [GLOBAL(rd)]
        psraw       xmm3,       VP8_FILTER_SHIFT

        paddw       xmm5,       [GLOBAL(rd)]
        psraw       xmm5,       VP8_FILTER_SHIFT

        paddw       xmm7,       [GLOBAL(rd)]
        psraw       xmm7,       VP8_FILTER_SHIFT

        packuswb    xmm1,       xmm1
        packuswb    xmm3,       xmm3

        packuswb    xmm5,       xmm5
        movq        [rdi],      xmm1

        packuswb    xmm7,       xmm7
        movq        [rdi+rdx],  xmm3

        lea         rdi,        [rdi + 2*rdx]
        movq        [rdi],      xmm5

        lea         rsp,        [rsp + 4*16]
        movq        [rdi+rdx],  xmm7

        lea         rdi,        [rdi + 2*rdx]
        cmp         rdi,        rcx

        jne         .next_row_fp

        lea         rsp,        [rsp + 16]

.done8x8:
    ;add rsp, 144
    pop         rsp
    ; begin epilog
    pop         rdi
    pop         rsi
    RESTORE_GOT
    RESTORE_XMM
    UNSHADOW_ARGS
    pop         rbp
    ret

SECTION_RODATA
align 16
shuf1b:
    db 0, 5, 1, 6, 2, 7, 3, 8, 4, 9, 5, 10, 6, 11, 7, 12
shuf2b:
    db 2, 4, 3, 5, 4, 6, 5, 7, 6, 8, 7, 9, 8, 10, 9, 11
shuf3b:
    db 1, 3, 2, 4, 3, 5, 4, 6, 5, 7, 6, 8, 7, 9, 8, 10

align 16
shuf2bfrom1:
    db  4, 8, 6, 1, 8, 3, 1, 5, 3, 7, 5, 9, 7,11, 9,13
align 16
shuf3bfrom1:
    db  2, 6, 4, 8, 6, 1, 8, 3, 1, 5, 3, 7, 5, 9, 7,11

align 16
rd:
    times 8 dw 0x40

align 16
k0_k5:
    times 8 db 0, 0             ;placeholder
    times 8 db 0, 0
    times 8 db 2, 1
    times 8 db 0, 0
    times 8 db 3, 3
    times 8 db 0, 0
    times 8 db 1, 2
    times 8 db 0, 0
k1_k3:
    times 8 db  0,    0         ;placeholder
    times 8 db  -6,  12
    times 8 db -11,  36
    times 8 db  -9,  50
    times 8 db -16,  77
    times 8 db  -6,  93
    times 8 db  -8, 108
    times 8 db  -1, 123
k2_k4:
    times 8 db 128,    0        ;placeholder
    times 8 db 123,   -1
    times 8 db 108,   -8
    times 8 db  93,   -6
    times 8 db  77,  -16
    times 8 db  50,   -9
    times 8 db  36,  -11
    times 8 db  12,   -6
align 16
vp8_bilinear_filters_ssse3:
    times 8 db 128, 0
    times 8 db 112, 16
    times 8 db 96,  32
    times 8 db 80,  48
    times 8 db 64,  64
    times 8 db 48,  80
    times 8 db 32,  96
    times 8 db 16,  112