quant.c
20.7 KB
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
/*****************************************************************************
* quant.c: ppc quantization
*****************************************************************************
* Copyright (C) 2007-2024 x264 project
*
* Authors: Guillaume Poirier <gpoirier@mplayerhq.hu>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
*
* This program is also available under a commercial proprietary license.
* For more information, contact us at licensing@x264.com.
*****************************************************************************/
#include "common/common.h"
#include "ppccommon.h"
#include "quant.h"
#if !HIGH_BIT_DEPTH
// quant of a whole 4x4 block, unrolled 2x and "pre-scheduled"
#define QUANT_16_U( idx0, idx1 ) \
{ \
temp1v = vec_ld((idx0), dct); \
temp2v = vec_ld((idx1), dct); \
mfvA = vec_ld((idx0), mf); \
mfvB = vec_ld((idx1), mf); \
biasvA = vec_ld((idx0), bias); \
biasvB = vec_ld((idx1), bias); \
mskA = vec_cmplt(temp1v, zero_s16v); \
mskB = vec_cmplt(temp2v, zero_s16v); \
coefvA = (vec_u16_t)vec_abs( temp1v ); \
coefvB = (vec_u16_t)vec_abs( temp2v ); \
coefvA = vec_adds(coefvA, biasvA); \
coefvB = vec_adds(coefvB, biasvB); \
multEvenvA = vec_mule(coefvA, mfvA); \
multOddvA = vec_mulo(coefvA, mfvA); \
multEvenvB = vec_mule(coefvB, mfvB); \
multOddvB = vec_mulo(coefvB, mfvB); \
multEvenvA = vec_sr(multEvenvA, i_qbitsv); \
multOddvA = vec_sr(multOddvA, i_qbitsv); \
multEvenvB = vec_sr(multEvenvB, i_qbitsv); \
multOddvB = vec_sr(multOddvB, i_qbitsv); \
temp1v = (vec_s16_t) vec_packs( multEvenvA, multOddvA ); \
tmpv = xxpermdi( temp1v, temp1v, 2 ); \
temp1v = vec_mergeh( temp1v, tmpv ); \
temp2v = (vec_s16_t) vec_packs( multEvenvB, multOddvB ); \
tmpv = xxpermdi( temp2v, temp2v, 2 ); \
temp2v = vec_mergeh( temp2v, tmpv ); \
temp1v = vec_xor(temp1v, mskA); \
temp2v = vec_xor(temp2v, mskB); \
temp1v = vec_adds(temp1v, vec_and(mskA, one)); \
vec_st(temp1v, (idx0), dct); \
temp2v = vec_adds(temp2v, vec_and(mskB, one)); \
nz = vec_or(nz, vec_or(temp1v, temp2v)); \
vec_st(temp2v, (idx1), dct); \
}
int x264_quant_4x4_altivec( int16_t dct[16], uint16_t mf[16], uint16_t bias[16] )
{
LOAD_ZERO;
vector bool short mskA;
vec_u32_t i_qbitsv = vec_splats( (uint32_t)16 );
vec_u16_t coefvA;
vec_u32_t multEvenvA, multOddvA;
vec_u16_t mfvA;
vec_u16_t biasvA;
vec_s16_t one = vec_splat_s16(1);
vec_s16_t nz = zero_s16v;
vector bool short mskB;
vec_u16_t coefvB;
vec_u32_t multEvenvB, multOddvB;
vec_u16_t mfvB;
vec_u16_t biasvB;
vec_s16_t temp1v, temp2v, tmpv;
QUANT_16_U( 0, 16 );
return vec_any_ne(nz, zero_s16v);
}
int x264_quant_4x4x4_altivec( dctcoef dcta[4][16], udctcoef mf[16], udctcoef bias[16] )
{
LOAD_ZERO;
vec_u32_t i_qbitsv = vec_splats( (uint32_t)16 );
vec_s16_t one = vec_splat_s16( 1 );
vec_s16_t nz0, nz1, nz2, nz3;
vector bool short mskA0;
vec_u16_t coefvA0;
vec_u32_t multEvenvA0, multOddvA0;
vec_u16_t mfvA0;
vec_u16_t biasvA0;
vector bool short mskB0;
vec_u16_t coefvB0;
vec_u32_t multEvenvB0, multOddvB0;
vec_u16_t mfvB0;
vec_u16_t biasvB0;
vector bool short mskA1;
vec_u16_t coefvA1;
vec_u32_t multEvenvA1, multOddvA1;
vec_u16_t mfvA1;
vec_u16_t biasvA1;
vector bool short mskB1;
vec_u16_t coefvB1;
vec_u32_t multEvenvB1, multOddvB1;
vec_u16_t mfvB1;
vec_u16_t biasvB1;
vector bool short mskA2;
vec_u16_t coefvA2;
vec_u32_t multEvenvA2, multOddvA2;
vec_u16_t mfvA2;
vec_u16_t biasvA2;
vector bool short mskB2;
vec_u16_t coefvB2;
vec_u32_t multEvenvB2, multOddvB2;
vec_u16_t mfvB2;
vec_u16_t biasvB2;
vector bool short mskA3;
vec_u16_t coefvA3;
vec_u32_t multEvenvA3, multOddvA3;
vec_u16_t mfvA3;
vec_u16_t biasvA3;
vector bool short mskB3;
vec_u16_t coefvB3;
vec_u32_t multEvenvB3, multOddvB3;
vec_u16_t mfvB3;
vec_u16_t biasvB3;
vec_s16_t temp1v, temp2v;
vec_s16_t tmpv0;
vec_s16_t tmpv1;
dctcoef *dct0 = dcta[0];
dctcoef *dct1 = dcta[1];
dctcoef *dct2 = dcta[2];
dctcoef *dct3 = dcta[3];
temp1v = vec_ld( 0, dct0 );
temp2v = vec_ld( 16, dct0 );
mfvA0 = vec_ld( 0, mf );
mfvB0 = vec_ld( 16, mf );
biasvA0 = vec_ld( 0, bias );
biasvB0 = vec_ld( 16, bias );
mskA0 = vec_cmplt( temp1v, zero_s16v );
mskB0 = vec_cmplt( temp2v, zero_s16v );
coefvA0 = (vec_u16_t)vec_abs( temp1v );
coefvB0 = (vec_u16_t)vec_abs( temp2v );
temp1v = vec_ld( 0, dct1 );
temp2v = vec_ld( 16, dct1 );
mfvA1 = vec_ld( 0, mf );
mfvB1 = vec_ld( 16, mf );
biasvA1 = vec_ld( 0, bias );
biasvB1 = vec_ld( 16, bias );
mskA1 = vec_cmplt( temp1v, zero_s16v );
mskB1 = vec_cmplt( temp2v, zero_s16v );
coefvA1 = (vec_u16_t)vec_abs( temp1v );
coefvB1 = (vec_u16_t)vec_abs( temp2v );
temp1v = vec_ld( 0, dct2 );
temp2v = vec_ld( 16, dct2 );
mfvA2 = vec_ld( 0, mf );
mfvB2 = vec_ld( 16, mf );
biasvA2 = vec_ld( 0, bias );
biasvB2 = vec_ld( 16, bias );
mskA2 = vec_cmplt( temp1v, zero_s16v );
mskB2 = vec_cmplt( temp2v, zero_s16v );
coefvA2 = (vec_u16_t)vec_abs( temp1v );
coefvB2 = (vec_u16_t)vec_abs( temp2v );
temp1v = vec_ld( 0, dct3 );
temp2v = vec_ld( 16, dct3 );
mfvA3 = vec_ld( 0, mf );
mfvB3 = vec_ld( 16, mf );
biasvA3 = vec_ld( 0, bias );
biasvB3 = vec_ld( 16, bias );
mskA3 = vec_cmplt( temp1v, zero_s16v );
mskB3 = vec_cmplt( temp2v, zero_s16v );
coefvA3 = (vec_u16_t)vec_abs( temp1v );
coefvB3 = (vec_u16_t)vec_abs( temp2v );
coefvA0 = vec_adds( coefvA0, biasvA0 );
coefvB0 = vec_adds( coefvB0, biasvB0 );
coefvA1 = vec_adds( coefvA1, biasvA1 );
coefvB1 = vec_adds( coefvB1, biasvB1 );
coefvA2 = vec_adds( coefvA2, biasvA2 );
coefvB2 = vec_adds( coefvB2, biasvB2 );
coefvA3 = vec_adds( coefvA3, biasvA3 );
coefvB3 = vec_adds( coefvB3, biasvB3 );
multEvenvA0 = vec_mule( coefvA0, mfvA0 );
multOddvA0 = vec_mulo( coefvA0, mfvA0 );
multEvenvB0 = vec_mule( coefvB0, mfvB0 );
multOddvB0 = vec_mulo( coefvB0, mfvB0 );
multEvenvA0 = vec_sr( multEvenvA0, i_qbitsv );
multOddvA0 = vec_sr( multOddvA0, i_qbitsv );
multEvenvB0 = vec_sr( multEvenvB0, i_qbitsv );
multOddvB0 = vec_sr( multOddvB0, i_qbitsv );
temp1v = (vec_s16_t)vec_packs( multEvenvA0, multOddvA0 );
temp2v = (vec_s16_t)vec_packs( multEvenvB0, multOddvB0 );
tmpv0 = xxpermdi( temp1v, temp1v, 2 );
tmpv1 = xxpermdi( temp2v, temp2v, 2 );
temp1v = vec_mergeh( temp1v, tmpv0 );
temp2v = vec_mergeh( temp2v, tmpv1 );
temp1v = vec_xor( temp1v, mskA0 );
temp2v = vec_xor( temp2v, mskB0 );
temp1v = vec_adds( temp1v, vec_and( mskA0, one ) );
temp2v = vec_adds( temp2v, vec_and( mskB0, one ) );
vec_st( temp1v, 0, dct0 );
vec_st( temp2v, 16, dct0 );
nz0 = vec_or( temp1v, temp2v );
multEvenvA1 = vec_mule( coefvA1, mfvA1 );
multOddvA1 = vec_mulo( coefvA1, mfvA1 );
multEvenvB1 = vec_mule( coefvB1, mfvB1 );
multOddvB1 = vec_mulo( coefvB1, mfvB1 );
multEvenvA1 = vec_sr( multEvenvA1, i_qbitsv );
multOddvA1 = vec_sr( multOddvA1, i_qbitsv );
multEvenvB1 = vec_sr( multEvenvB1, i_qbitsv );
multOddvB1 = vec_sr( multOddvB1, i_qbitsv );
temp1v = (vec_s16_t)vec_packs( multEvenvA1, multOddvA1 );
temp2v = (vec_s16_t)vec_packs( multEvenvB1, multOddvB1 );
tmpv0 = xxpermdi( temp1v, temp1v, 2 );
tmpv1 = xxpermdi( temp2v, temp2v, 2 );
temp1v = vec_mergeh( temp1v, tmpv0 );
temp2v = vec_mergeh( temp2v, tmpv1 );
temp1v = vec_xor( temp1v, mskA1 );
temp2v = vec_xor( temp2v, mskB1 );
temp1v = vec_adds( temp1v, vec_and( mskA1, one ) );
temp2v = vec_adds( temp2v, vec_and( mskB1, one ) );
vec_st( temp1v, 0, dct1 );
vec_st( temp2v, 16, dct1 );
nz1 = vec_or( temp1v, temp2v );
multEvenvA2 = vec_mule( coefvA2, mfvA2 );
multOddvA2 = vec_mulo( coefvA2, mfvA2 );
multEvenvB2 = vec_mule( coefvB2, mfvB2 );
multOddvB2 = vec_mulo( coefvB2, mfvB2 );
multEvenvA2 = vec_sr( multEvenvA2, i_qbitsv );
multOddvA2 = vec_sr( multOddvA2, i_qbitsv );
multEvenvB2 = vec_sr( multEvenvB2, i_qbitsv );
multOddvB2 = vec_sr( multOddvB2, i_qbitsv );
temp1v = (vec_s16_t)vec_packs( multEvenvA2, multOddvA2 );
temp2v = (vec_s16_t)vec_packs( multEvenvB2, multOddvB2 );
tmpv0 = xxpermdi( temp1v, temp1v, 2 );
tmpv1 = xxpermdi( temp2v, temp2v, 2 );
temp1v = vec_mergeh( temp1v, tmpv0 );
temp2v = vec_mergeh( temp2v, tmpv1 );
temp1v = vec_xor( temp1v, mskA2 );
temp2v = vec_xor( temp2v, mskB2 );
temp1v = vec_adds( temp1v, vec_and( mskA2, one ) );
temp2v = vec_adds( temp2v, vec_and( mskB2, one ) );
vec_st( temp1v, 0, dct2 );
vec_st( temp2v, 16, dct2 );
nz2 = vec_or( temp1v, temp2v );
multEvenvA3 = vec_mule( coefvA3, mfvA3 );
multOddvA3 = vec_mulo( coefvA3, mfvA3 );
multEvenvB3 = vec_mule( coefvB3, mfvB3 );
multOddvB3 = vec_mulo( coefvB3, mfvB3 );
multEvenvA3 = vec_sr( multEvenvA3, i_qbitsv );
multOddvA3 = vec_sr( multOddvA3, i_qbitsv );
multEvenvB3 = vec_sr( multEvenvB3, i_qbitsv );
multOddvB3 = vec_sr( multOddvB3, i_qbitsv );
temp1v = (vec_s16_t)vec_packs( multEvenvA3, multOddvA3 );
temp2v = (vec_s16_t)vec_packs( multEvenvB3, multOddvB3 );
tmpv0 = xxpermdi( temp1v, temp1v, 2 );
tmpv1 = xxpermdi( temp2v, temp2v, 2 );
temp1v = vec_mergeh( temp1v, tmpv0 );
temp2v = vec_mergeh( temp2v, tmpv1 );
temp1v = vec_xor( temp1v, mskA3 );
temp2v = vec_xor( temp2v, mskB3 );
temp1v = vec_adds( temp1v, vec_and( mskA3, one ) );
temp2v = vec_adds( temp2v, vec_and( mskB3, one ) );
vec_st( temp1v, 0, dct3 );
vec_st( temp2v, 16, dct3 );
nz3 = vec_or( temp1v, temp2v );
return (vec_any_ne( nz0, zero_s16v ) << 0) | (vec_any_ne( nz1, zero_s16v ) << 1) |
(vec_any_ne( nz2, zero_s16v ) << 2) | (vec_any_ne( nz3, zero_s16v ) << 3);
}
// DC quant of a whole 4x4 block, unrolled 2x and "pre-scheduled"
#define QUANT_16_U_DC( idx0, idx1 ) \
{ \
temp1v = vec_ld((idx0), dct); \
temp2v = vec_ld((idx1), dct); \
mskA = vec_cmplt(temp1v, zero_s16v); \
mskB = vec_cmplt(temp2v, zero_s16v); \
coefvA = (vec_u16_t)vec_max(vec_sub(zero_s16v, temp1v), temp1v);\
coefvB = (vec_u16_t)vec_max(vec_sub(zero_s16v, temp2v), temp2v);\
coefvA = vec_add(coefvA, biasv); \
coefvB = vec_add(coefvB, biasv); \
multEvenvA = vec_mule(coefvA, mfv); \
multOddvA = vec_mulo(coefvA, mfv); \
multEvenvB = vec_mule(coefvB, mfv); \
multOddvB = vec_mulo(coefvB, mfv); \
multEvenvA = vec_sr(multEvenvA, i_qbitsv); \
multOddvA = vec_sr(multOddvA, i_qbitsv); \
multEvenvB = vec_sr(multEvenvB, i_qbitsv); \
multOddvB = vec_sr(multOddvB, i_qbitsv); \
temp1v = (vec_s16_t) vec_packs(vec_mergeh(multEvenvA, multOddvA), vec_mergel(multEvenvA, multOddvA)); \
temp2v = (vec_s16_t) vec_packs(vec_mergeh(multEvenvB, multOddvB), vec_mergel(multEvenvB, multOddvB)); \
temp1v = vec_xor(temp1v, mskA); \
temp2v = vec_xor(temp2v, mskB); \
temp1v = vec_add(temp1v, vec_and(mskA, one)); \
vec_st(temp1v, (idx0), dct); \
temp2v = vec_add(temp2v, vec_and(mskB, one)); \
nz = vec_or(nz, vec_or(temp1v, temp2v)); \
vec_st(temp2v, (idx1), dct); \
}
int x264_quant_4x4_dc_altivec( int16_t dct[16], int mf, int bias )
{
LOAD_ZERO;
vector bool short mskA;
vec_u32_t i_qbitsv;
vec_u16_t coefvA;
vec_u32_t multEvenvA, multOddvA;
vec_s16_t one = vec_splat_s16(1);
vec_s16_t nz = zero_s16v;
vector bool short mskB;
vec_u16_t coefvB;
vec_u32_t multEvenvB, multOddvB;
vec_s16_t temp1v, temp2v;
vec_u16_t mfv;
vec_u16_t biasv;
mfv = vec_splats( (uint16_t)mf );
i_qbitsv = vec_splats( (uint32_t) 16 );
biasv = vec_splats( (uint16_t)bias );
QUANT_16_U_DC( 0, 16 );
return vec_any_ne(nz, zero_s16v);
}
// DC quant of a whole 2x2 block
#define QUANT_4_U_DC( idx0 ) \
{ \
const vec_u16_t sel = (vec_u16_t) CV(-1,-1,-1,-1,0,0,0,0); \
temp1v = vec_ld((idx0), dct); \
mskA = vec_cmplt(temp1v, zero_s16v); \
coefvA = (vec_u16_t)vec_max(vec_sub(zero_s16v, temp1v), temp1v);\
coefvA = vec_add(coefvA, biasv); \
multEvenvA = vec_mule(coefvA, mfv); \
multOddvA = vec_mulo(coefvA, mfv); \
multEvenvA = vec_sr(multEvenvA, i_qbitsv); \
multOddvA = vec_sr(multOddvA, i_qbitsv); \
temp2v = (vec_s16_t) vec_packs(vec_mergeh(multEvenvA, multOddvA), vec_mergel(multEvenvA, multOddvA)); \
temp2v = vec_xor(temp2v, mskA); \
temp2v = vec_add(temp2v, vec_and(mskA, one)); \
temp1v = vec_sel(temp1v, temp2v, sel); \
nz = vec_or(nz, temp1v); \
vec_st(temp1v, (idx0), dct); \
}
int x264_quant_2x2_dc_altivec( int16_t dct[4], int mf, int bias )
{
LOAD_ZERO;
vector bool short mskA;
vec_u32_t i_qbitsv;
vec_u16_t coefvA;
vec_u32_t multEvenvA, multOddvA;
vec_s16_t one = vec_splat_s16(1);
vec_s16_t nz = zero_s16v;
static const vec_s16_t mask2 = CV(-1, -1, -1, -1, 0, 0, 0, 0);
vec_s16_t temp1v, temp2v;
vec_u16_t mfv;
vec_u16_t biasv;
mfv = vec_splats( (uint16_t)mf );
i_qbitsv = vec_splats( (uint32_t) 16 );
biasv = vec_splats( (uint16_t)bias );
QUANT_4_U_DC(0);
return vec_any_ne(vec_and(nz, mask2), zero_s16v);
}
int x264_quant_8x8_altivec( int16_t dct[64], uint16_t mf[64], uint16_t bias[64] )
{
LOAD_ZERO;
vector bool short mskA;
vec_u32_t i_qbitsv;
vec_u16_t coefvA;
vec_u32_t multEvenvA, multOddvA;
vec_u16_t mfvA;
vec_u16_t biasvA;
vec_s16_t one = vec_splat_s16(1);
vec_s16_t nz = zero_s16v;
vector bool short mskB;
vec_u16_t coefvB;
vec_u32_t multEvenvB, multOddvB;
vec_u16_t mfvB;
vec_u16_t biasvB;
vec_s16_t temp1v, temp2v, tmpv;
i_qbitsv = vec_splats( (uint32_t)16 );
for( int i = 0; i < 4; i++ )
QUANT_16_U( i*2*16, i*2*16+16 );
return vec_any_ne(nz, zero_s16v);
}
#define DEQUANT_SHL() \
{ \
dctv = vec_ld(8*y, dct); \
mf1v = vec_ld(16*y, dequant_mf[i_mf]); \
mf2v = vec_ld(16+16*y, dequant_mf[i_mf]); \
mfv = vec_packs(mf1v, mf2v); \
\
multEvenvA = vec_mule(dctv, mfv); \
multOddvA = vec_mulo(dctv, mfv); \
dctv = (vec_s16_t) vec_packs( multEvenvA, multOddvA ); \
tmpv = xxpermdi( dctv, dctv, 2 ); \
dctv = vec_mergeh( dctv, tmpv ); \
dctv = vec_sl(dctv, i_qbitsv); \
vec_st(dctv, 8*y, dct); \
}
#ifdef WORDS_BIGENDIAN
#define VEC_MULE vec_mule
#define VEC_MULO vec_mulo
#else
#define VEC_MULE vec_mulo
#define VEC_MULO vec_mule
#endif
#define DEQUANT_SHR() \
{ \
dctv = vec_ld(8*y, dct); \
dct1v = vec_mergeh(dctv, dctv); \
dct2v = vec_mergel(dctv, dctv); \
mf1v = vec_ld(16*y, dequant_mf[i_mf]); \
mf2v = vec_ld(16+16*y, dequant_mf[i_mf]); \
\
multEvenvA = VEC_MULE(dct1v, (vec_s16_t)mf1v); \
multOddvA = VEC_MULO(dct1v, (vec_s16_t)mf1v); \
temp1v = vec_add(vec_sl(multEvenvA, sixteenv), multOddvA); \
temp1v = vec_add(temp1v, fv); \
temp1v = vec_sra(temp1v, i_qbitsv); \
\
multEvenvA = VEC_MULE(dct2v, (vec_s16_t)mf2v); \
multOddvA = VEC_MULO(dct2v, (vec_s16_t)mf2v); \
temp2v = vec_add(vec_sl(multEvenvA, sixteenv), multOddvA); \
temp2v = vec_add(temp2v, fv); \
temp2v = vec_sra(temp2v, i_qbitsv); \
\
dctv = (vec_s16_t)vec_packs(temp1v, temp2v); \
vec_st(dctv, y*8, dct); \
}
void x264_dequant_4x4_altivec( int16_t dct[16], int dequant_mf[6][16], int i_qp )
{
int i_mf = i_qp%6;
int i_qbits = i_qp/6 - 4;
vec_s16_t dctv, tmpv;
vec_s16_t dct1v, dct2v;
vec_s32_t mf1v, mf2v;
vec_s16_t mfv;
vec_s32_t multEvenvA, multOddvA;
vec_s32_t temp1v, temp2v;
if( i_qbits >= 0 )
{
vec_u16_t i_qbitsv;
i_qbitsv = vec_splats( (uint16_t) i_qbits );
for( int y = 0; y < 4; y+=2 )
DEQUANT_SHL();
}
else
{
const int f = 1 << (-i_qbits-1);
vec_s32_t fv;
fv = vec_splats( f );
vec_u32_t i_qbitsv;
i_qbitsv = vec_splats( (uint32_t)-i_qbits );
vec_u32_t sixteenv;
sixteenv = vec_splats( (uint32_t)16 );
for( int y = 0; y < 4; y+=2 )
DEQUANT_SHR();
}
}
void x264_dequant_8x8_altivec( int16_t dct[64], int dequant_mf[6][64], int i_qp )
{
int i_mf = i_qp%6;
int i_qbits = i_qp/6 - 6;
vec_s16_t dctv, tmpv;
vec_s16_t dct1v, dct2v;
vec_s32_t mf1v, mf2v;
vec_s16_t mfv;
vec_s32_t multEvenvA, multOddvA;
vec_s32_t temp1v, temp2v;
if( i_qbits >= 0 )
{
vec_u16_t i_qbitsv;
i_qbitsv = vec_splats((uint16_t)i_qbits );
for( int y = 0; y < 16; y+=2 )
DEQUANT_SHL();
}
else
{
const int f = 1 << (-i_qbits-1);
vec_s32_t fv;
fv = vec_splats( f );
vec_u32_t i_qbitsv;
i_qbitsv = vec_splats( (uint32_t)-i_qbits );
vec_u32_t sixteenv;
sixteenv = vec_splats( (uint32_t)16 );
for( int y = 0; y < 16; y+=2 )
DEQUANT_SHR();
}
}
#endif // !HIGH_BIT_DEPTH