-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathjidctfst.cpp
1214 lines (831 loc) · 37.1 KB
/
jidctfst.cpp
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
#include "jpegdecoder.h"
#pragma warning(push)
#pragma warning( disable : 4035 4799 )
/*
* jidctfst.c
*
* Copyright (C) 1994-1996, Thomas G. Lane.
* This file is part of the Independent JPEG Group's software.
* For conditions of distribution and use, see the accompanying README file.
*
* This file contains a fast, not so accurate integer implementation of the
* inverse DCT (Discrete Cosine Transform). In the IJG code, this routine
* must also perform dequantization of the input coefficients.
*
* A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT
* on each row (or vice versa, but it's more convenient to emit a row at
* a time). Direct algorithms are also available, but they are much more
* complex and seem not to be any faster when reduced to code.
*
* This implementation is based on Arai, Agui, and Nakajima's algorithm for
* scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in
* Japanese, but the algorithm is described in the Pennebaker & Mitchell
* JPEG textbook (see REFERENCES section in file README). The following code
* is based directly on figure 4-8 in P&M.
* While an 8-point DCT cannot be done in less than 11 multiplies, it is
* possible to arrange the computation so that many of the multiplies are
* simple scalings of the final outputs. These multiplies can then be
* folded into the multiplications or divisions by the JPEG quantization
* table entries. The AA&N method leaves only 5 multiplies and 29 adds
* to be done in the DCT itself.
* The primary disadvantage of this method is that with fixed-point math,
* accuracy is lost due to imprecise representation of the scaled
* quantization values. The smaller the quantization table entry, the less
* precise the scaled value, so this implementation does worse with high-
* quality-setting files than with low-quality ones.
*/
/***************************************************************************
*
* This program has been developed by Intel Corporation.
* You have Intel's permission to incorporate this code
* into your product, royalty free. Intel has various
* intellectual property rights which it may assert under
* certain circumstances, such as if another manufacturer's
* processor mis-identifies itself as being "GenuineIntel"
* when the CPUID instruction is executed.
*
* Intel specifically disclaims all warranties, express or
* implied, and all liability, including consequential and
* other indirect damages, for the use of this code,
* including liability for infringement of any proprietary
* rights, and including the warranties of merchantability
* and fitness for a particular purpose. Intel does not
* assume any responsibility for any errors which may
* appear in this code nor any responsibility to update it.
*
* * Other brands and names are the property of their respective
* owners.
*
* Copyright (c) 1997, Intel Corporation. All rights reserved.
***************************************************************************/
/* Scaling decisions are generally the same as in the LL&M algorithm;
* see jidctint.c for more details. However, we choose to descale
* (right shift) multiplication products as soon as they are formed,
* rather than carrying additional fractional bits into subsequent additions.
* This compromises accuracy slightly, but it lets us save a few shifts.
* More importantly, 16-bit arithmetic is then adequate (for 8-bit samples)
* everywhere except in the multiplications proper; this saves a good deal
* of work on 16-bit-int machines.
*
* The dequantized coefficients are not integers because the AA&N scaling
* factors have been incorporated. We represent them scaled up by PASS1_BITS,
* so that the first and second IDCT rounds have the same input scaling.
* For 8-bit JSAMPLEs, we choose IFAST_SCALE_BITS = PASS1_BITS so as to
* avoid a descaling shift; this compromises accuracy rather drastically
* for small quantization table entries, but it saves a lot of shifts.
* For 12-bit JSAMPLEs, there's no hope of using 16x16 multiplies anyway,
* so we use a much larger scaling factor to preserve accuracy.
*
* A final compromise is to represent the multiplicative constants to only
* 8 fractional bits, rather than 13. This saves some shifting work on some
* machines, and may also reduce the cost of multiplication (since there
* are fewer one-bits in the constants).
*/
#define FCONST_BITS 8
#define PASS1_BITS 2
#define SCALEDONE ((int32) 1)
#define CONST_SCALE (SCALEDONE << FCONST_BITS)
#define FIX(x) ((int32) ((x) * CONST_SCALE + 0.5))
/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus
* causing a lot of useless floating-point operations at run time.
* To get around this we use the following pre-calculated constants.
* If you change FCONST_BITS you may want to add appropriate values.
* (With a reasonable C compiler, you can just rely on the FIX() macro...)
*/
#if FCONST_BITS == 8
#define FIX_1_082392200 ((INT32) 277) /* FIX(1.082392200) */
#define FIX_1_414213562 ((INT32) 362) /* FIX(1.414213562) */
#define FIX_1_847759065 ((INT32) 473) /* FIX(1.847759065) */
#define FIX_2_613125930 ((INT32) 669) /* FIX(2.613125930) */
#else
#define FIX_1_082392200 FIX(1.082392200)
#define FIX_1_414213562 FIX(1.414213562)
#define FIX_1_847759065 FIX(1.847759065)
#define FIX_2_613125930 FIX(2.613125930)
#endif
/* We can gain a little more speed, with a further compromise in accuracy,
* by omitting the addition in a descaling shift. This yields an incorrectly
* rounded result half the time...
*/
#ifndef USE_ACCURATE_ROUNDING
#undef DESCALE
#define DESCALE(x,n) RIGHT_SHIFT(x, n)
#endif
#define DCTELEM BLOCK_TYPE
/* Multiply a DCTELEM variable by an INT32 constant, and immediately
* descale to yield a DCTELEM result.
*/
#define MULTIPLY(var,const) ((DCTELEM) DESCALE((var) * (const), FCONST_BITS))
/* Dequantize a coefficient by multiplying it by the multiplier-table
* entry; produce a DCTELEM result. For 8-bit data a 16x16->16
* multiplication will do. For 12-bit data, the multiplier table is
* declared INT32, so a 32-bit multiply will be used.
*/
#define DEQUANTIZE(coef,quantval) (((IFAST_MULT_TYPE) (coef)) * (quantval))
/* Like DESCALE, but applies to a DCTELEM and produces an int.
* We assume that int right shift is unsigned if INT32 right shift is.
*/
#define ISHIFT_TEMPS
#define IRIGHT_SHIFT(x,shft) ((x) >> (shft))
#ifdef USE_ACCURATE_ROUNDING
#define IDESCALE(x,n) ((int) IRIGHT_SHIFT((x) + (1 << ((n)-1)), n))
#else
#define IDESCALE(x,n) ((int) IRIGHT_SHIFT(x, n))
#endif
/*
* Perform dequantization and inverse DCT on one block of coefficients.
*/
static __int64 fix_141 = 0x5a825a825a825a82;
static __int64 fix_184n261 = 0xcf04cf04cf04cf04;
static __int64 fix_184 = 0x7641764176417641;
static __int64 fix_n184 = 0x896f896f896f896f;
static __int64 fix_108n184 = 0xcf04cf04cf04cf04;
static __int64 const_0x0080 = 0x0080008000800080;
void jpeg_idct_ifast (
BLOCK_TYPE *inptr,
short *quantptr,
uchar * *outptr,
int output_col)
{
short workspace[64 + 4]; /* buffers data between passes */
short *wsptr=workspace;
__asm{
mov edi, quantptr
mov ebx, inptr
mov esi, wsptr
add esi, 0x07 ;align wsptr to qword
and esi, 0xfffffff8 ;align wsptr to qword
mov eax, esi
/* Odd part */
movq mm1, [ebx + 8*10] ;load inptr[DCTSIZE*5]
pmullw mm1, [edi + 8*10] ;tmp6 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]);
movq mm0, [ebx + 8*6] ;load inptr[DCTSIZE*3]
pmullw mm0, [edi + 8*6] ;tmp5 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]);
movq mm3, [ebx + 8*2] ;load inptr[DCTSIZE*1]
movq mm2, mm1 ;copy tmp6 /* phase 6 */
pmullw mm3, [edi + 8*2] ;tmp4 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]);
movq mm4, [ebx + 8*14] ;load inptr[DCTSIZE*1]
paddw mm1, mm0 ;z13 = tmp6 + tmp5;
pmullw mm4, [edi + 8*14] ;tmp7 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]);
psubw mm2, mm0 ;z10 = tmp6 - tmp5
psllw mm2, 2 ;shift z10
movq mm0, mm2 ;copy z10
pmulhw mm2, fix_184n261 ;MULTIPLY( z12, FIX_1_847759065); /* 2*c2 */
movq mm5, mm3 ;copy tmp4
pmulhw mm0, fix_n184 ;MULTIPLY(z10, -FIX_1_847759065); /* 2*c2 */
paddw mm3, mm4 ;z11 = tmp4 + tmp7;
movq mm6, mm3 ;copy z11 /* phase 5 */
psubw mm5, mm4 ;z12 = tmp4 - tmp7;
psubw mm6, mm1 ;z11-z13
psllw mm5, 2 ;shift z12
movq mm4, [ebx + 8*12] ;load inptr[DCTSIZE*6], even part
movq mm7, mm5 ;copy z12
pmulhw mm5, fix_108n184 ;MULT(z12, (FIX_1_08-FIX_1_84)) //- z5; /* 2*(c2-c6) */ even part
paddw mm3, mm1 ;tmp7 = z11 + z13;
/* Even part */
pmulhw mm7, fix_184 ;MULTIPLY(z10,(FIX_1_847759065 - FIX_2_613125930)) //+ z5; /* -2*(c2+c6) */
psllw mm6, 2
movq mm1, [ebx + 8*4] ;load inptr[DCTSIZE*2]
pmullw mm1, [edi + 8*4] ;tmp1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]);
paddw mm0, mm5 ;tmp10
pmullw mm4, [edi + 8*12] ;tmp3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]);
paddw mm2, mm7 ;tmp12
pmulhw mm6, fix_141 ;tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562); /* 2*c4 */
psubw mm2, mm3 ;tmp6 = tmp12 - tmp7
movq mm5, mm1 ;copy tmp1
paddw mm1, mm4 ;tmp13= tmp1 + tmp3; /* phases 5-3 */
psubw mm5, mm4 ;tmp1-tmp3
psubw mm6, mm2 ;tmp5 = tmp11 - tmp6;
movq [esi+8*0], mm1 ;save tmp13 in workspace
psllw mm5, 2 ;shift tmp1-tmp3
movq mm7, [ebx + 8*0] ;load inptr[DCTSIZE*0]
pmulhw mm5, fix_141 ;MULTIPLY(tmp1 - tmp3, FIX_1_414213562)
paddw mm0, mm6 ;tmp4 = tmp10 + tmp5;
pmullw mm7, [edi + 8*0] ;tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]);
movq mm4, [ebx + 8*8] ;load inptr[DCTSIZE*4]
pmullw mm4, [edi + 8*8] ;tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]);
psubw mm5, mm1 ;tmp12 = MULTIPLY(tmp1 - tmp3, FIX_1_414213562) - tmp13; /* 2*c4 */
movq [esi+8*4], mm0 ;save tmp4 in workspace
movq mm1, mm7 ;copy tmp0 /* phase 3 */
movq [esi+8*2], mm5 ;save tmp12 in workspace
psubw mm1, mm4 ;tmp11 = tmp0 - tmp2;
paddw mm7, mm4 ;tmp10 = tmp0 + tmp2;
movq mm5, mm1 ;copy tmp11
paddw mm1, [esi+8*2] ;tmp1 = tmp11 + tmp12;
movq mm4, mm7 ;copy tmp10 /* phase 2 */
paddw mm7, [esi+8*0] ;tmp0 = tmp10 + tmp13;
psubw mm4, [esi+8*0] ;tmp3 = tmp10 - tmp13;
movq mm0, mm7 ;copy tmp0
psubw mm5, [esi+8*2] ;tmp2 = tmp11 - tmp12;
paddw mm7, mm3 ;wsptr[DCTSIZE*0] = (int) (tmp0 + tmp7);
psubw mm0, mm3 ;wsptr[DCTSIZE*7] = (int) (tmp0 - tmp7);
movq [esi + 8*0], mm7 ;wsptr[DCTSIZE*0]
movq mm3, mm1 ;copy tmp1
movq [esi + 8*14], mm0 ;wsptr[DCTSIZE*7]
paddw mm1, mm2 ;wsptr[DCTSIZE*1] = (int) (tmp1 + tmp6);
psubw mm3, mm2 ;wsptr[DCTSIZE*6] = (int) (tmp1 - tmp6);
movq [esi + 8*2], mm1 ;wsptr[DCTSIZE*1]
movq mm1, mm4 ;copy tmp3
movq [esi + 8*12], mm3 ;wsptr[DCTSIZE*6]
paddw mm4, [esi+8*4] ;wsptr[DCTSIZE*4] = (int) (tmp3 + tmp4);
psubw mm1, [esi+8*4] ;wsptr[DCTSIZE*3] = (int) (tmp3 - tmp4);
movq [esi + 8*8], mm4
movq mm7, mm5 ;copy tmp2
paddw mm5, mm6 ;wsptr[DCTSIZE*2] = (int) (tmp2 + tmp5)
movq [esi+8*6], mm1 ;
psubw mm7, mm6 ;wsptr[DCTSIZE*5] = (int) (tmp2 - tmp5);
movq [esi + 8*4], mm5
movq [esi + 8*10], mm7
/*****************************************************************/
add edi, 8
add ebx, 8
add esi, 8
/*****************************************************************/
movq mm1, [ebx + 8*10] ;load inptr[DCTSIZE*5]
pmullw mm1, [edi + 8*10] ;tmp6 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]);
movq mm0, [ebx + 8*6] ;load inptr[DCTSIZE*3]
pmullw mm0, [edi + 8*6] ;tmp5 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]);
movq mm3, [ebx + 8*2] ;load inptr[DCTSIZE*1]
movq mm2, mm1 ;copy tmp6 /* phase 6 */
pmullw mm3, [edi + 8*2] ;tmp4 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]);
movq mm4, [ebx + 8*14] ;load inptr[DCTSIZE*1]
paddw mm1, mm0 ;z13 = tmp6 + tmp5;
pmullw mm4, [edi + 8*14] ;tmp7 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]);
psubw mm2, mm0 ;z10 = tmp6 - tmp5
psllw mm2, 2 ;shift z10
movq mm0, mm2 ;copy z10
pmulhw mm2, fix_184n261 ;MULTIPLY( z12, FIX_1_847759065); /* 2*c2 */
movq mm5, mm3 ;copy tmp4
pmulhw mm0, fix_n184 ;MULTIPLY(z10, -FIX_1_847759065); /* 2*c2 */
paddw mm3, mm4 ;z11 = tmp4 + tmp7;
movq mm6, mm3 ;copy z11 /* phase 5 */
psubw mm5, mm4 ;z12 = tmp4 - tmp7;
psubw mm6, mm1 ;z11-z13
psllw mm5, 2 ;shift z12
movq mm4, [ebx + 8*12] ;load inptr[DCTSIZE*6], even part
movq mm7, mm5 ;copy z12
pmulhw mm5, fix_108n184 ;MULT(z12, (FIX_1_08-FIX_1_84)) //- z5; /* 2*(c2-c6) */ even part
paddw mm3, mm1 ;tmp7 = z11 + z13;
/* Even part */
pmulhw mm7, fix_184 ;MULTIPLY(z10,(FIX_1_847759065 - FIX_2_613125930)) //+ z5; /* -2*(c2+c6) */
psllw mm6, 2
movq mm1, [ebx + 8*4] ;load inptr[DCTSIZE*2]
pmullw mm1, [edi + 8*4] ;tmp1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]);
paddw mm0, mm5 ;tmp10
pmullw mm4, [edi + 8*12] ;tmp3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]);
paddw mm2, mm7 ;tmp12
pmulhw mm6, fix_141 ;tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562); /* 2*c4 */
psubw mm2, mm3 ;tmp6 = tmp12 - tmp7
movq mm5, mm1 ;copy tmp1
paddw mm1, mm4 ;tmp13= tmp1 + tmp3; /* phases 5-3 */
psubw mm5, mm4 ;tmp1-tmp3
psubw mm6, mm2 ;tmp5 = tmp11 - tmp6;
movq [esi+8*0], mm1 ;save tmp13 in workspace
psllw mm5, 2 ;shift tmp1-tmp3
movq mm7, [ebx + 8*0] ;load inptr[DCTSIZE*0]
paddw mm0, mm6 ;tmp4 = tmp10 + tmp5;
pmulhw mm5, fix_141 ;MULTIPLY(tmp1 - tmp3, FIX_1_414213562)
pmullw mm7, [edi + 8*0] ;tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]);
movq mm4, [ebx + 8*8] ;load inptr[DCTSIZE*4]
pmullw mm4, [edi + 8*8] ;tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]);
psubw mm5, mm1 ;tmp12 = MULTIPLY(tmp1 - tmp3, FIX_1_414213562) - tmp13; /* 2*c4 */
movq [esi+8*4], mm0 ;save tmp4 in workspace
movq mm1, mm7 ;copy tmp0 /* phase 3 */
movq [esi+8*2], mm5 ;save tmp12 in workspace
psubw mm1, mm4 ;tmp11 = tmp0 - tmp2;
paddw mm7, mm4 ;tmp10 = tmp0 + tmp2;
movq mm5, mm1 ;copy tmp11
paddw mm1, [esi+8*2] ;tmp1 = tmp11 + tmp12;
movq mm4, mm7 ;copy tmp10 /* phase 2 */
paddw mm7, [esi+8*0] ;tmp0 = tmp10 + tmp13;
psubw mm4, [esi+8*0] ;tmp3 = tmp10 - tmp13;
movq mm0, mm7 ;copy tmp0
psubw mm5, [esi+8*2] ;tmp2 = tmp11 - tmp12;
paddw mm7, mm3 ;wsptr[DCTSIZE*0] = (int) (tmp0 + tmp7);
psubw mm0, mm3 ;wsptr[DCTSIZE*7] = (int) (tmp0 - tmp7);
movq [esi + 8*0], mm7 ;wsptr[DCTSIZE*0]
movq mm3, mm1 ;copy tmp1
movq [esi + 8*14], mm0 ;wsptr[DCTSIZE*7]
paddw mm1, mm2 ;wsptr[DCTSIZE*1] = (int) (tmp1 + tmp6);
psubw mm3, mm2 ;wsptr[DCTSIZE*6] = (int) (tmp1 - tmp6);
movq [esi + 8*2], mm1 ;wsptr[DCTSIZE*1]
movq mm1, mm4 ;copy tmp3
movq [esi + 8*12], mm3 ;wsptr[DCTSIZE*6]
paddw mm4, [esi+8*4] ;wsptr[DCTSIZE*4] = (int) (tmp3 + tmp4);
psubw mm1, [esi+8*4] ;wsptr[DCTSIZE*3] = (int) (tmp3 - tmp4);
movq [esi + 8*8], mm4
movq mm7, mm5 ;copy tmp2
paddw mm5, mm6 ;wsptr[DCTSIZE*2] = (int) (tmp2 + tmp5)
movq [esi+8*6], mm1 ;
psubw mm7, mm6 ;wsptr[DCTSIZE*5] = (int) (tmp2 - tmp5);
movq [esi + 8*4], mm5
movq [esi + 8*10], mm7
/*****************************************************************/
/* Pass 2: process rows from work array, store into output array. */
/* Note that we must descale the results by a factor of 8 == 2**3, */
/* and also undo the PASS1_BITS scaling. */
/*****************************************************************/
/* Even part */
mov esi, eax
mov eax, outptr
// tmp10 = ((DCTELEM) wsptr[0] + (DCTELEM) wsptr[4]);
// tmp13 = ((DCTELEM) wsptr[2] + (DCTELEM) wsptr[6]);
// tmp11 = ((DCTELEM) wsptr[0] - (DCTELEM) wsptr[4]);
// tmp14 = ((DCTELEM) wsptr[2] - (DCTELEM) wsptr[6]);
movq mm0, [esi+8*0] ;wsptr[0,0],[0,1],[0,2],[0,3]
movq mm1, [esi+8*1] ;wsptr[0,4],[0,5],[0,6],[0,7]
movq mm2, mm0
movq mm3, [esi+8*2] ;wsptr[1,0],[1,1],[1,2],[1,3]
paddw mm0, mm1 ;wsptr[0,tmp10],[xxx],[0,tmp13],[xxx]
movq mm4, [esi+8*3] ;wsptr[1,4],[1,5],[1,6],[1,7]
psubw mm2, mm1 ;wsptr[0,tmp11],[xxx],[0,tmp14],[xxx]
movq mm6, mm0
movq mm5, mm3
paddw mm3, mm4 ;wsptr[1,tmp10],[xxx],[1,tmp13],[xxx]
movq mm1, mm2
psubw mm5, mm4 ;wsptr[1,tmp11],[xxx],[1,tmp14],[xxx]
punpcklwd mm0, mm3 ;wsptr[0,tmp10],[1,tmp10],[xxx],[xxx]
movq mm7, [esi+8*7] ;wsptr[3,4],[3,5],[3,6],[3,7]
punpckhwd mm6, mm3 ;wsptr[0,tmp13],[1,tmp13],[xxx],[xxx]
movq mm3, [esi+8*4] ;wsptr[2,0],[2,1],[2,2],[2,3]
punpckldq mm0, mm6 ;wsptr[0,tmp10],[1,tmp10],[0,tmp13],[1,tmp13]
punpcklwd mm1, mm5 ;wsptr[0,tmp11],[1,tmp11],[xxx],[xxx]
movq mm4, mm3
movq mm6, [esi+8*6] ;wsptr[3,0],[3,1],[3,2],[3,3]
punpckhwd mm2, mm5 ;wsptr[0,tmp14],[1,tmp14],[xxx],[xxx]
movq mm5, [esi+8*5] ;wsptr[2,4],[2,5],[2,6],[2,7]
punpckldq mm1, mm2 ;wsptr[0,tmp11],[1,tmp11],[0,tmp14],[1,tmp14]
paddw mm3, mm5 ;wsptr[2,tmp10],[xxx],[2,tmp13],[xxx]
movq mm2, mm6
psubw mm4, mm5 ;wsptr[2,tmp11],[xxx],[2,tmp14],[xxx]
paddw mm6, mm7 ;wsptr[3,tmp10],[xxx],[3,tmp13],[xxx]
movq mm5, mm3
punpcklwd mm3, mm6 ;wsptr[2,tmp10],[3,tmp10],[xxx],[xxx]
psubw mm2, mm7 ;wsptr[3,tmp11],[xxx],[3,tmp14],[xxx]
punpckhwd mm5, mm6 ;wsptr[2,tmp13],[3,tmp13],[xxx],[xxx]
movq mm7, mm4
punpckldq mm3, mm5 ;wsptr[2,tmp10],[3,tmp10],[2,tmp13],[3,tmp13]
punpcklwd mm4, mm2 ;wsptr[2,tmp11],[3,tmp11],[xxx],[xxx]
punpckhwd mm7, mm2 ;wsptr[2,tmp14],[3,tmp14],[xxx],[xxx]
punpckldq mm4, mm7 ;wsptr[2,tmp11],[3,tmp11],[2,tmp14],[3,tmp14]
movq mm6, mm1
// mm0 = ;wsptr[0,tmp10],[1,tmp10],[0,tmp13],[1,tmp13]
// mm1 = ;wsptr[0,tmp11],[1,tmp11],[0,tmp14],[1,tmp14]
movq mm2, mm0
punpckhdq mm6, mm4 ;wsptr[0,tmp14],[1,tmp14],[2,tmp14],[3,tmp14]
punpckldq mm1, mm4 ;wsptr[0,tmp11],[1,tmp11],[2,tmp11],[3,tmp11]
psllw mm6, 2
pmulhw mm6, fix_141
punpckldq mm0, mm3 ;wsptr[0,tmp10],[1,tmp10],[2,tmp10],[3,tmp10]
punpckhdq mm2, mm3 ;wsptr[0,tmp13],[1,tmp13],[2,tmp13],[3,tmp13]
movq mm7, mm0
// tmp0 = tmp10 + tmp13;
// tmp3 = tmp10 - tmp13;
paddw mm0, mm2 ;[0,tmp0],[1,tmp0],[2,tmp0],[3,tmp0]
psubw mm7, mm2 ;[0,tmp3],[1,tmp3],[2,tmp3],[3,tmp3]
// tmp12 = MULTIPLY(tmp14, FIX_1_414213562) - tmp13;
psubw mm6, mm2 ;wsptr[0,tmp12],[1,tmp12],[2,tmp12],[3,tmp12]
// tmp1 = tmp11 + tmp12;
// tmp2 = tmp11 - tmp12;
movq mm5, mm1
/* Odd part */
// z13 = (DCTELEM) wsptr[5] + (DCTELEM) wsptr[3];
// z10 = (DCTELEM) wsptr[5] - (DCTELEM) wsptr[3];
// z11 = (DCTELEM) wsptr[1] + (DCTELEM) wsptr[7];
// z12 = (DCTELEM) wsptr[1] - (DCTELEM) wsptr[7];
movq mm3, [esi+8*0] ;wsptr[0,0],[0,1],[0,2],[0,3]
paddw mm1, mm6 ;[0,tmp1],[1,tmp1],[2,tmp1],[3,tmp1]
movq mm4, [esi+8*1] ;wsptr[0,4],[0,5],[0,6],[0,7]
psubw mm5, mm6 ;[0,tmp2],[1,tmp2],[2,tmp2],[3,tmp2]
movq mm6, mm3
punpckldq mm3, mm4 ;wsptr[0,0],[0,1],[0,4],[0,5]
punpckhdq mm4, mm6 ;wsptr[0,6],[0,7],[0,2],[0,3]
movq mm2, mm3
//Save tmp0 and tmp1 in wsptr
movq [esi+8*0], mm0 ;save tmp0
paddw mm2, mm4 ;wsptr[xxx],[0,z11],[xxx],[0,z13]
//Continue with z10 --- z13
movq mm6, [esi+8*2] ;wsptr[1,0],[1,1],[1,2],[1,3]
psubw mm3, mm4 ;wsptr[xxx],[0,z12],[xxx],[0,z10]
movq mm0, [esi+8*3] ;wsptr[1,4],[1,5],[1,6],[1,7]
movq mm4, mm6
movq [esi+8*1], mm1 ;save tmp1
punpckldq mm6, mm0 ;wsptr[1,0],[1,1],[1,4],[1,5]
punpckhdq mm0, mm4 ;wsptr[1,6],[1,7],[1,2],[1,3]
movq mm1, mm6
//Save tmp2 and tmp3 in wsptr
paddw mm6, mm0 ;wsptr[xxx],[1,z11],[xxx],[1,z13]
movq mm4, mm2
//Continue with z10 --- z13
movq [esi+8*2], mm5 ;save tmp2
punpcklwd mm2, mm6 ;wsptr[xxx],[xxx],[0,z11],[1,z11]
psubw mm1, mm0 ;wsptr[xxx],[1,z12],[xxx],[1,z10]
punpckhwd mm4, mm6 ;wsptr[xxx],[xxx],[0,z13],[1,z13]
movq mm0, mm3
punpcklwd mm3, mm1 ;wsptr[xxx],[xxx],[0,z12],[1,z12]
movq [esi+8*3], mm7 ;save tmp3
punpckhwd mm0, mm1 ;wsptr[xxx],[xxx],[0,z10],[1,z10]
movq mm6, [esi+8*4] ;wsptr[2,0],[2,1],[2,2],[2,3]
punpckhdq mm0, mm2 ;wsptr[0,z10],[1,z10],[0,z11],[1,z11]
movq mm7, [esi+8*5] ;wsptr[2,4],[2,5],[2,6],[2,7]
punpckhdq mm3, mm4 ;wsptr[0,z12],[1,z12],[0,z13],[1,z13]
movq mm1, [esi+8*6] ;wsptr[3,0],[3,1],[3,2],[3,3]
movq mm4, mm6
punpckldq mm6, mm7 ;wsptr[2,0],[2,1],[2,4],[2,5]
movq mm5, mm1
punpckhdq mm7, mm4 ;wsptr[2,6],[2,7],[2,2],[2,3]
movq mm2, mm6
movq mm4, [esi+8*7] ;wsptr[3,4],[3,5],[3,6],[3,7]
paddw mm6, mm7 ;wsptr[xxx],[2,z11],[xxx],[2,z13]
psubw mm2, mm7 ;wsptr[xxx],[2,z12],[xxx],[2,z10]
punpckldq mm1, mm4 ;wsptr[3,0],[3,1],[3,4],[3,5]
punpckhdq mm4, mm5 ;wsptr[3,6],[3,7],[3,2],[3,3]
movq mm7, mm1
paddw mm1, mm4 ;wsptr[xxx],[3,z11],[xxx],[3,z13]
psubw mm7, mm4 ;wsptr[xxx],[3,z12],[xxx],[3,z10]
movq mm5, mm6
punpcklwd mm6, mm1 ;wsptr[xxx],[xxx],[2,z11],[3,z11]
punpckhwd mm5, mm1 ;wsptr[xxx],[xxx],[2,z13],[3,z13]
movq mm4, mm2
punpcklwd mm2, mm7 ;wsptr[xxx],[xxx],[2,z12],[3,z12]
punpckhwd mm4, mm7 ;wsptr[xxx],[xxx],[2,z10],[3,z10]
punpckhdq mm4, mm6 ;wsptr[2,z10],[3,z10],[2,z11],[3,z11]
punpckhdq mm2, mm5 ;wsptr[2,z12],[3,z12],[2,z13],[3,z13]
movq mm5, mm0
punpckldq mm0, mm4 ;wsptr[0,z10],[1,z10],[2,z10],[3,z10]
punpckhdq mm5, mm4 ;wsptr[0,z11],[1,z11],[2,z11],[3,z11]
movq mm4, mm3
punpckhdq mm4, mm2 ;wsptr[0,z13],[1,z13],[2,z13],[3,z13]
movq mm1, mm5
punpckldq mm3, mm2 ;wsptr[0,z12],[1,z12],[2,z12],[3,z12]
// tmp7 = z11 + z13; /* phase 5 */
// tmp8 = z11 - z13; /* phase 5 */
psubw mm1, mm4 ;tmp8
paddw mm5, mm4 ;tmp7
// tmp21 = MULTIPLY(tmp8, FIX_1_414213562); /* 2*c4 */
psllw mm1, 2
psllw mm0, 2
pmulhw mm1, fix_141 ;tmp21
// tmp20 = MULTIPLY(z12, (FIX_1_082392200- FIX_1_847759065)) /* 2*(c2-c6) */
// + MULTIPLY(z10, - FIX_1_847759065); /* 2*c2 */
psllw mm3, 2
movq mm7, mm0
pmulhw mm7, fix_n184
movq mm6, mm3
movq mm2, [esi+8*0] ;tmp0,final1
pmulhw mm6, fix_108n184
// tmp22 = MULTIPLY(z10,(FIX_1_847759065 - FIX_2_613125930)) /* -2*(c2+c6) */
// + MULTIPLY(z12, FIX_1_847759065); /* 2*c2 */
movq mm4, mm2 ;final1
pmulhw mm0, fix_184n261
paddw mm2, mm5 ;tmp0+tmp7,final1
pmulhw mm3, fix_184
psubw mm4, mm5 ;tmp0-tmp7,final1
// tmp6 = tmp22 - tmp7; /* phase 2 */
psraw mm2, 5 ;outptr[0,0],[1,0],[2,0],[3,0],final1
paddsw mm2, const_0x0080 ;final1
paddw mm7, mm6 ;tmp20
psraw mm4, 5 ;outptr[0,7],[1,7],[2,7],[3,7],final1
paddsw mm4, const_0x0080 ;final1
paddw mm3, mm0 ;tmp22
// tmp5 = tmp21 - tmp6;
psubw mm3, mm5 ;tmp6
// tmp4 = tmp20 + tmp5;
movq mm0, [esi+8*1] ;tmp1,final2
psubw mm1, mm3 ;tmp5
movq mm6, mm0 ;final2
paddw mm0, mm3 ;tmp1+tmp6,final2
/* Final output stage: scale down by a factor of 8 and range-limit */
// outptr[0] = range_limit[IDESCALE(tmp0 + tmp7, PASS1_BITS+3)
// & RANGE_MASK];
// outptr[7] = range_limit[IDESCALE(tmp0 - tmp7, PASS1_BITS+3)
// & RANGE_MASK]; final1
// outptr[1] = range_limit[IDESCALE(tmp1 + tmp6, PASS1_BITS+3)
// & RANGE_MASK];
// outptr[6] = range_limit[IDESCALE(tmp1 - tmp6, PASS1_BITS+3)
// & RANGE_MASK]; final2
psubw mm6, mm3 ;tmp1-tmp6,final2
psraw mm0, 5 ;outptr[0,1],[1,1],[2,1],[3,1]
paddsw mm0, const_0x0080
psraw mm6, 5 ;outptr[0,6],[1,6],[2,6],[3,6]
paddsw mm6, const_0x0080 ;need to check this value
packuswb mm0, mm4 ;out[0,1],[1,1],[2,1],[3,1],[0,7],[1,7],[2,7],[3,7]
movq mm5, [esi+8*2] ;tmp2,final3
packuswb mm2, mm6 ;out[0,0],[1,0],[2,0],[3,0],[0,6],[1,6],[2,6],[3,6]
// outptr[2] = range_limit[IDESCALE(tmp2 + tmp5, PASS1_BITS+3)
// & RANGE_MASK];
// outptr[5] = range_limit[IDESCALE(tmp2 - tmp5, PASS1_BITS+3)
// & RANGE_MASK]; final3
paddw mm7, mm1 ;tmp4
movq mm3, mm5
paddw mm5, mm1 ;tmp2+tmp5
psubw mm3, mm1 ;tmp2-tmp5
psraw mm5, 5 ;outptr[0,2],[1,2],[2,2],[3,2]
paddsw mm5, const_0x0080
movq mm4, [esi+8*3] ;tmp3,final4
psraw mm3, 5 ;outptr[0,5],[1,5],[2,5],[3,5]
paddsw mm3, const_0x0080
// outptr[4] = range_limit[IDESCALE(tmp3 + tmp4, PASS1_BITS+3)
// & RANGE_MASK];
// outptr[3] = range_limit[IDESCALE(tmp3 - tmp4, PASS1_BITS+3)
// & RANGE_MASK]; final4
movq mm6, mm4
paddw mm4, mm7 ;tmp3+tmp4
psubw mm6, mm7 ;tmp3-tmp4
psraw mm4, 5 ;outptr[0,4],[1,4],[2,4],[3,4]
mov ecx, [eax]
paddsw mm4, const_0x0080
psraw mm6, 5 ;outptr[0,3],[1,3],[2,3],[3,3]
paddsw mm6, const_0x0080
packuswb mm5, mm4 ;out[0,2],[1,2],[2,2],[3,2],[0,4],[1,4],[2,4],[3,4]
packuswb mm6, mm3 ;out[0,3],[1,3],[2,3],[3,3],[0,5],[1,5],[2,5],[3,5]
movq mm4, mm2
movq mm7, mm5
punpcklbw mm2, mm0 ;out[0,0],[0,1],[1,0],[1,1],[2,0],[2,1],[3,0],[3,1]
punpckhbw mm4, mm0 ;out[0,6],[0,7],[1,6],[1,7],[2,6],[2,7],[3,6],[3,7]
movq mm1, mm2
punpcklbw mm5, mm6 ;out[0,2],[0,3],[1,2],[1,3],[2,2],[2,3],[3,2],[3,3]
add eax, 4
punpckhbw mm7, mm6 ;out[0,4],[0,5],[1,4],[1,5],[2,4],[2,5],[3,4],[3,5]
punpcklwd mm2, mm5 ;out[0,0],[0,1],[0,2],[0,3],[1,0],[1,1],[1,2],[1,3]
add ecx, output_col
movq mm6, mm7
punpckhwd mm1, mm5 ;out[2,0],[2,1],[2,2],[2,3],[3,0],[3,1],[3,2],[3,3]
movq mm0, mm2
punpcklwd mm6, mm4 ;out[0,4],[0,5],[0,6],[0,7],[1,4],[1,5],[1,6],[1,7]
mov ebx, [eax]
punpckldq mm2, mm6 ;out[0,0],[0,1],[0,2],[0,3],[0,4],[0,5],[0,6],[0,7]
add eax, 4
movq mm3, mm1
add ebx, output_col
punpckhwd mm7, mm4 ;out[2,4],[2,5],[2,6],[2,7],[3,4],[3,5],[3,6],[3,7]
movq [ecx], mm2
punpckhdq mm0, mm6 ;out[1,0],[1,1],[1,2],[1,3],[1,4],[1,5],[1,6],[1,7]
mov ecx, [eax]
add eax, 4
add ecx, output_col
movq [ebx], mm0
punpckldq mm1, mm7 ;out[2,0],[2,1],[2,2],[2,3],[2,4],[2,5],[2,6],[2,7]
mov ebx, [eax]
add ebx, output_col
punpckhdq mm3, mm7 ;out[3,0],[3,1],[3,2],[3,3],[3,4],[3,5],[3,6],[3,7]
movq [ecx], mm1
movq [ebx], mm3
/*******************************************************************/
add esi, 64
add eax, 4
/*******************************************************************/
// tmp10 = ((DCTELEM) wsptr[0] + (DCTELEM) wsptr[4]);
// tmp13 = ((DCTELEM) wsptr[2] + (DCTELEM) wsptr[6]);
// tmp11 = ((DCTELEM) wsptr[0] - (DCTELEM) wsptr[4]);
// tmp14 = ((DCTELEM) wsptr[2] - (DCTELEM) wsptr[6]);
movq mm0, [esi+8*0] ;wsptr[0,0],[0,1],[0,2],[0,3]
movq mm1, [esi+8*1] ;wsptr[0,4],[0,5],[0,6],[0,7]
movq mm2, mm0
movq mm3, [esi+8*2] ;wsptr[1,0],[1,1],[1,2],[1,3]
paddw mm0, mm1 ;wsptr[0,tmp10],[xxx],[0,tmp13],[xxx]
movq mm4, [esi+8*3] ;wsptr[1,4],[1,5],[1,6],[1,7]
psubw mm2, mm1 ;wsptr[0,tmp11],[xxx],[0,tmp14],[xxx]
movq mm6, mm0
movq mm5, mm3
paddw mm3, mm4 ;wsptr[1,tmp10],[xxx],[1,tmp13],[xxx]
movq mm1, mm2
psubw mm5, mm4 ;wsptr[1,tmp11],[xxx],[1,tmp14],[xxx]
punpcklwd mm0, mm3 ;wsptr[0,tmp10],[1,tmp10],[xxx],[xxx]
movq mm7, [esi+8*7] ;wsptr[3,4],[3,5],[3,6],[3,7]
punpckhwd mm6, mm3 ;wsptr[0,tmp13],[1,tmp13],[xxx],[xxx]
movq mm3, [esi+8*4] ;wsptr[2,0],[2,1],[2,2],[2,3]
punpckldq mm0, mm6 ;wsptr[0,tmp10],[1,tmp10],[0,tmp13],[1,tmp13]
punpcklwd mm1, mm5 ;wsptr[0,tmp11],[1,tmp11],[xxx],[xxx]
movq mm4, mm3
movq mm6, [esi+8*6] ;wsptr[3,0],[3,1],[3,2],[3,3]
punpckhwd mm2, mm5 ;wsptr[0,tmp14],[1,tmp14],[xxx],[xxx]
movq mm5, [esi+8*5] ;wsptr[2,4],[2,5],[2,6],[2,7]
punpckldq mm1, mm2 ;wsptr[0,tmp11],[1,tmp11],[0,tmp14],[1,tmp14]
paddw mm3, mm5 ;wsptr[2,tmp10],[xxx],[2,tmp13],[xxx]
movq mm2, mm6
psubw mm4, mm5 ;wsptr[2,tmp11],[xxx],[2,tmp14],[xxx]
paddw mm6, mm7 ;wsptr[3,tmp10],[xxx],[3,tmp13],[xxx]
movq mm5, mm3
punpcklwd mm3, mm6 ;wsptr[2,tmp10],[3,tmp10],[xxx],[xxx]
psubw mm2, mm7 ;wsptr[3,tmp11],[xxx],[3,tmp14],[xxx]
punpckhwd mm5, mm6 ;wsptr[2,tmp13],[3,tmp13],[xxx],[xxx]
movq mm7, mm4
punpckldq mm3, mm5 ;wsptr[2,tmp10],[3,tmp10],[2,tmp13],[3,tmp13]
punpcklwd mm4, mm2 ;wsptr[2,tmp11],[3,tmp11],[xxx],[xxx]
punpckhwd mm7, mm2 ;wsptr[2,tmp14],[3,tmp14],[xxx],[xxx]
punpckldq mm4, mm7 ;wsptr[2,tmp11],[3,tmp11],[2,tmp14],[3,tmp14]
movq mm6, mm1
// mm0 = ;wsptr[0,tmp10],[1,tmp10],[0,tmp13],[1,tmp13]
// mm1 = ;wsptr[0,tmp11],[1,tmp11],[0,tmp14],[1,tmp14]
movq mm2, mm0
punpckhdq mm6, mm4 ;wsptr[0,tmp14],[1,tmp14],[2,tmp14],[3,tmp14]
punpckldq mm1, mm4 ;wsptr[0,tmp11],[1,tmp11],[2,tmp11],[3,tmp11]
psllw mm6, 2
pmulhw mm6, fix_141
punpckldq mm0, mm3 ;wsptr[0,tmp10],[1,tmp10],[2,tmp10],[3,tmp10]
punpckhdq mm2, mm3 ;wsptr[0,tmp13],[1,tmp13],[2,tmp13],[3,tmp13]
movq mm7, mm0
// tmp0 = tmp10 + tmp13;
// tmp3 = tmp10 - tmp13;
paddw mm0, mm2 ;[0,tmp0],[1,tmp0],[2,tmp0],[3,tmp0]
psubw mm7, mm2 ;[0,tmp3],[1,tmp3],[2,tmp3],[3,tmp3]
// tmp12 = MULTIPLY(tmp14, FIX_1_414213562) - tmp13;
psubw mm6, mm2 ;wsptr[0,tmp12],[1,tmp12],[2,tmp12],[3,tmp12]
// tmp1 = tmp11 + tmp12;
// tmp2 = tmp11 - tmp12;
movq mm5, mm1
/* Odd part */
// z13 = (DCTELEM) wsptr[5] + (DCTELEM) wsptr[3];
// z10 = (DCTELEM) wsptr[5] - (DCTELEM) wsptr[3];
// z11 = (DCTELEM) wsptr[1] + (DCTELEM) wsptr[7];
// z12 = (DCTELEM) wsptr[1] - (DCTELEM) wsptr[7];
movq mm3, [esi+8*0] ;wsptr[0,0],[0,1],[0,2],[0,3]
paddw mm1, mm6 ;[0,tmp1],[1,tmp1],[2,tmp1],[3,tmp1]
movq mm4, [esi+8*1] ;wsptr[0,4],[0,5],[0,6],[0,7]
psubw mm5, mm6 ;[0,tmp2],[1,tmp2],[2,tmp2],[3,tmp2]
movq mm6, mm3
punpckldq mm3, mm4 ;wsptr[0,0],[0,1],[0,4],[0,5]
punpckhdq mm4, mm6 ;wsptr[0,6],[0,7],[0,2],[0,3]
movq mm2, mm3
//Save tmp0 and tmp1 in wsptr
movq [esi+8*0], mm0 ;save tmp0
paddw mm2, mm4 ;wsptr[xxx],[0,z11],[xxx],[0,z13]
//Continue with z10 --- z13
movq mm6, [esi+8*2] ;wsptr[1,0],[1,1],[1,2],[1,3]
psubw mm3, mm4 ;wsptr[xxx],[0,z12],[xxx],[0,z10]
movq mm0, [esi+8*3] ;wsptr[1,4],[1,5],[1,6],[1,7]
movq mm4, mm6
movq [esi+8*1], mm1 ;save tmp1
punpckldq mm6, mm0 ;wsptr[1,0],[1,1],[1,4],[1,5]
punpckhdq mm0, mm4 ;wsptr[1,6],[1,7],[1,2],[1,3]
movq mm1, mm6
//Save tmp2 and tmp3 in wsptr
paddw mm6, mm0 ;wsptr[xxx],[1,z11],[xxx],[1,z13]
movq mm4, mm2
//Continue with z10 --- z13
movq [esi+8*2], mm5 ;save tmp2
punpcklwd mm2, mm6 ;wsptr[xxx],[xxx],[0,z11],[1,z11]
psubw mm1, mm0 ;wsptr[xxx],[1,z12],[xxx],[1,z10]
punpckhwd mm4, mm6 ;wsptr[xxx],[xxx],[0,z13],[1,z13]
movq mm0, mm3
punpcklwd mm3, mm1 ;wsptr[xxx],[xxx],[0,z12],[1,z12]
movq [esi+8*3], mm7 ;save tmp3
punpckhwd mm0, mm1 ;wsptr[xxx],[xxx],[0,z10],[1,z10]
movq mm6, [esi+8*4] ;wsptr[2,0],[2,1],[2,2],[2,3]
punpckhdq mm0, mm2 ;wsptr[0,z10],[1,z10],[0,z11],[1,z11]
movq mm7, [esi+8*5] ;wsptr[2,4],[2,5],[2,6],[2,7]
punpckhdq mm3, mm4 ;wsptr[0,z12],[1,z12],[0,z13],[1,z13]
movq mm1, [esi+8*6] ;wsptr[3,0],[3,1],[3,2],[3,3]
movq mm4, mm6
punpckldq mm6, mm7 ;wsptr[2,0],[2,1],[2,4],[2,5]
movq mm5, mm1
punpckhdq mm7, mm4 ;wsptr[2,6],[2,7],[2,2],[2,3]