summaryrefslogtreecommitdiffstats
path: root/media/libvpx/libvpx/vpx_dsp/ppc/fdct32x32_vsx.c
blob: 328b0e313012274efbfa7a696723700bd6015862 (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
/*
 *  Copyright (c) 2018 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_config.h"
#include "./vpx_dsp_rtcd.h"

#include "vpx_dsp/ppc/transpose_vsx.h"
#include "vpx_dsp/ppc/txfm_common_vsx.h"
#include "vpx_dsp/ppc/types_vsx.h"

// Returns ((a +/- b) * cospi16 + (2 << 13)) >> 14.
static INLINE void single_butterfly(int16x8_t a, int16x8_t b, int16x8_t *add,
                                    int16x8_t *sub) {
  // Since a + b can overflow 16 bits, the multiplication is distributed
  // (a * c +/- b * c).
  const int32x4_t ac_e = vec_mule(a, cospi16_v);
  const int32x4_t ac_o = vec_mulo(a, cospi16_v);
  const int32x4_t bc_e = vec_mule(b, cospi16_v);
  const int32x4_t bc_o = vec_mulo(b, cospi16_v);

  // Reuse the same multiplies for sum and difference.
  const int32x4_t sum_e = vec_add(ac_e, bc_e);
  const int32x4_t sum_o = vec_add(ac_o, bc_o);
  const int32x4_t diff_e = vec_sub(ac_e, bc_e);
  const int32x4_t diff_o = vec_sub(ac_o, bc_o);

  // Add rounding offset
  const int32x4_t rsum_o = vec_add(sum_o, vec_dct_const_rounding);
  const int32x4_t rsum_e = vec_add(sum_e, vec_dct_const_rounding);
  const int32x4_t rdiff_o = vec_add(diff_o, vec_dct_const_rounding);
  const int32x4_t rdiff_e = vec_add(diff_e, vec_dct_const_rounding);

  const int32x4_t ssum_o = vec_sra(rsum_o, vec_dct_const_bits);
  const int32x4_t ssum_e = vec_sra(rsum_e, vec_dct_const_bits);
  const int32x4_t sdiff_o = vec_sra(rdiff_o, vec_dct_const_bits);
  const int32x4_t sdiff_e = vec_sra(rdiff_e, vec_dct_const_bits);

  // There's no pack operation for even and odd, so we need to permute.
  *add = (int16x8_t)vec_perm(ssum_e, ssum_o, vec_perm_odd_even_pack);
  *sub = (int16x8_t)vec_perm(sdiff_e, sdiff_o, vec_perm_odd_even_pack);
}

// Returns (a * c1 +/- b * c2 + (2 << 13)) >> 14
static INLINE void double_butterfly(int16x8_t a, int16x8_t c1, int16x8_t b,
                                    int16x8_t c2, int16x8_t *add,
                                    int16x8_t *sub) {
  const int32x4_t ac1_o = vec_mulo(a, c1);
  const int32x4_t ac1_e = vec_mule(a, c1);
  const int32x4_t ac2_o = vec_mulo(a, c2);
  const int32x4_t ac2_e = vec_mule(a, c2);

  const int32x4_t bc1_o = vec_mulo(b, c1);
  const int32x4_t bc1_e = vec_mule(b, c1);
  const int32x4_t bc2_o = vec_mulo(b, c2);
  const int32x4_t bc2_e = vec_mule(b, c2);

  const int32x4_t sum_o = vec_add(ac1_o, bc2_o);
  const int32x4_t sum_e = vec_add(ac1_e, bc2_e);
  const int32x4_t diff_o = vec_sub(ac2_o, bc1_o);
  const int32x4_t diff_e = vec_sub(ac2_e, bc1_e);

  // Add rounding offset
  const int32x4_t rsum_o = vec_add(sum_o, vec_dct_const_rounding);
  const int32x4_t rsum_e = vec_add(sum_e, vec_dct_const_rounding);
  const int32x4_t rdiff_o = vec_add(diff_o, vec_dct_const_rounding);
  const int32x4_t rdiff_e = vec_add(diff_e, vec_dct_const_rounding);

  const int32x4_t ssum_o = vec_sra(rsum_o, vec_dct_const_bits);
  const int32x4_t ssum_e = vec_sra(rsum_e, vec_dct_const_bits);
  const int32x4_t sdiff_o = vec_sra(rdiff_o, vec_dct_const_bits);
  const int32x4_t sdiff_e = vec_sra(rdiff_e, vec_dct_const_bits);

  // There's no pack operation for even and odd, so we need to permute.
  *add = (int16x8_t)vec_perm(ssum_e, ssum_o, vec_perm_odd_even_pack);
  *sub = (int16x8_t)vec_perm(sdiff_e, sdiff_o, vec_perm_odd_even_pack);
}

// While other architecture combine the load and the stage 1 operations, Power9
// benchmarking show no benefit in such an approach.
static INLINE void load(const int16_t *a, int stride, int16x8_t *b) {
  // Tried out different combinations of load and shift instructions, this is
  // the fastest one.
  {
    const int16x8_t l0 = vec_vsx_ld(0, a);
    const int16x8_t l1 = vec_vsx_ld(0, a + stride);
    const int16x8_t l2 = vec_vsx_ld(0, a + 2 * stride);
    const int16x8_t l3 = vec_vsx_ld(0, a + 3 * stride);
    const int16x8_t l4 = vec_vsx_ld(0, a + 4 * stride);
    const int16x8_t l5 = vec_vsx_ld(0, a + 5 * stride);
    const int16x8_t l6 = vec_vsx_ld(0, a + 6 * stride);
    const int16x8_t l7 = vec_vsx_ld(0, a + 7 * stride);

    const int16x8_t l8 = vec_vsx_ld(0, a + 8 * stride);
    const int16x8_t l9 = vec_vsx_ld(0, a + 9 * stride);
    const int16x8_t l10 = vec_vsx_ld(0, a + 10 * stride);
    const int16x8_t l11 = vec_vsx_ld(0, a + 11 * stride);
    const int16x8_t l12 = vec_vsx_ld(0, a + 12 * stride);
    const int16x8_t l13 = vec_vsx_ld(0, a + 13 * stride);
    const int16x8_t l14 = vec_vsx_ld(0, a + 14 * stride);
    const int16x8_t l15 = vec_vsx_ld(0, a + 15 * stride);

    b[0] = vec_sl(l0, vec_dct_scale_log2);
    b[1] = vec_sl(l1, vec_dct_scale_log2);
    b[2] = vec_sl(l2, vec_dct_scale_log2);
    b[3] = vec_sl(l3, vec_dct_scale_log2);
    b[4] = vec_sl(l4, vec_dct_scale_log2);
    b[5] = vec_sl(l5, vec_dct_scale_log2);
    b[6] = vec_sl(l6, vec_dct_scale_log2);
    b[7] = vec_sl(l7, vec_dct_scale_log2);

    b[8] = vec_sl(l8, vec_dct_scale_log2);
    b[9] = vec_sl(l9, vec_dct_scale_log2);
    b[10] = vec_sl(l10, vec_dct_scale_log2);
    b[11] = vec_sl(l11, vec_dct_scale_log2);
    b[12] = vec_sl(l12, vec_dct_scale_log2);
    b[13] = vec_sl(l13, vec_dct_scale_log2);
    b[14] = vec_sl(l14, vec_dct_scale_log2);
    b[15] = vec_sl(l15, vec_dct_scale_log2);
  }
  {
    const int16x8_t l16 = vec_vsx_ld(0, a + 16 * stride);
    const int16x8_t l17 = vec_vsx_ld(0, a + 17 * stride);
    const int16x8_t l18 = vec_vsx_ld(0, a + 18 * stride);
    const int16x8_t l19 = vec_vsx_ld(0, a + 19 * stride);
    const int16x8_t l20 = vec_vsx_ld(0, a + 20 * stride);
    const int16x8_t l21 = vec_vsx_ld(0, a + 21 * stride);
    const int16x8_t l22 = vec_vsx_ld(0, a + 22 * stride);
    const int16x8_t l23 = vec_vsx_ld(0, a + 23 * stride);

    const int16x8_t l24 = vec_vsx_ld(0, a + 24 * stride);
    const int16x8_t l25 = vec_vsx_ld(0, a + 25 * stride);
    const int16x8_t l26 = vec_vsx_ld(0, a + 26 * stride);
    const int16x8_t l27 = vec_vsx_ld(0, a + 27 * stride);
    const int16x8_t l28 = vec_vsx_ld(0, a + 28 * stride);
    const int16x8_t l29 = vec_vsx_ld(0, a + 29 * stride);
    const int16x8_t l30 = vec_vsx_ld(0, a + 30 * stride);
    const int16x8_t l31 = vec_vsx_ld(0, a + 31 * stride);

    b[16] = vec_sl(l16, vec_dct_scale_log2);
    b[17] = vec_sl(l17, vec_dct_scale_log2);
    b[18] = vec_sl(l18, vec_dct_scale_log2);
    b[19] = vec_sl(l19, vec_dct_scale_log2);
    b[20] = vec_sl(l20, vec_dct_scale_log2);
    b[21] = vec_sl(l21, vec_dct_scale_log2);
    b[22] = vec_sl(l22, vec_dct_scale_log2);
    b[23] = vec_sl(l23, vec_dct_scale_log2);

    b[24] = vec_sl(l24, vec_dct_scale_log2);
    b[25] = vec_sl(l25, vec_dct_scale_log2);
    b[26] = vec_sl(l26, vec_dct_scale_log2);
    b[27] = vec_sl(l27, vec_dct_scale_log2);
    b[28] = vec_sl(l28, vec_dct_scale_log2);
    b[29] = vec_sl(l29, vec_dct_scale_log2);
    b[30] = vec_sl(l30, vec_dct_scale_log2);
    b[31] = vec_sl(l31, vec_dct_scale_log2);
  }
}

static INLINE void store(tran_low_t *a, const int16x8_t *b) {
  vec_vsx_st(b[0], 0, a);
  vec_vsx_st(b[8], 0, a + 8);
  vec_vsx_st(b[16], 0, a + 16);
  vec_vsx_st(b[24], 0, a + 24);

  vec_vsx_st(b[1], 0, a + 32);
  vec_vsx_st(b[9], 0, a + 40);
  vec_vsx_st(b[17], 0, a + 48);
  vec_vsx_st(b[25], 0, a + 56);

  vec_vsx_st(b[2], 0, a + 64);
  vec_vsx_st(b[10], 0, a + 72);
  vec_vsx_st(b[18], 0, a + 80);
  vec_vsx_st(b[26], 0, a + 88);

  vec_vsx_st(b[3], 0, a + 96);
  vec_vsx_st(b[11], 0, a + 104);
  vec_vsx_st(b[19], 0, a + 112);
  vec_vsx_st(b[27], 0, a + 120);

  vec_vsx_st(b[4], 0, a + 128);
  vec_vsx_st(b[12], 0, a + 136);
  vec_vsx_st(b[20], 0, a + 144);
  vec_vsx_st(b[28], 0, a + 152);

  vec_vsx_st(b[5], 0, a + 160);
  vec_vsx_st(b[13], 0, a + 168);
  vec_vsx_st(b[21], 0, a + 176);
  vec_vsx_st(b[29], 0, a + 184);

  vec_vsx_st(b[6], 0, a + 192);
  vec_vsx_st(b[14], 0, a + 200);
  vec_vsx_st(b[22], 0, a + 208);
  vec_vsx_st(b[30], 0, a + 216);

  vec_vsx_st(b[7], 0, a + 224);
  vec_vsx_st(b[15], 0, a + 232);
  vec_vsx_st(b[23], 0, a + 240);
  vec_vsx_st(b[31], 0, a + 248);
}

// Returns 1 if negative 0 if positive
static INLINE int16x8_t vec_sign_s16(int16x8_t a) {
  return vec_sr(a, vec_shift_sign_s16);
}

// Add 2 if positive, 1 if negative, and shift by 2.
static INLINE int16x8_t sub_round_shift(const int16x8_t a) {
  const int16x8_t sign = vec_sign_s16(a);
  return vec_sra(vec_sub(vec_add(a, vec_twos_s16), sign), vec_dct_scale_log2);
}

// Add 1 if positive, 2 if negative, and shift by 2.
// In practice, add 1, then add the sign bit, then shift without rounding.
static INLINE int16x8_t add_round_shift_s16(const int16x8_t a) {
  const int16x8_t sign = vec_sign_s16(a);
  return vec_sra(vec_add(vec_add(a, vec_ones_s16), sign), vec_dct_scale_log2);
}

static void fdct32_vsx(const int16x8_t *in, int16x8_t *out, int pass) {
  int16x8_t temp0[32];  // Hold stages: 1, 4, 7
  int16x8_t temp1[32];  // Hold stages: 2, 5
  int16x8_t temp2[32];  // Hold stages: 3, 6
  int i;

  // Stage 1
  // Unrolling this loops actually slows down Power9 benchmarks
  for (i = 0; i < 16; i++) {
    temp0[i] = vec_add(in[i], in[31 - i]);
    // pass through to stage 3.
    temp1[i + 16] = vec_sub(in[15 - i], in[i + 16]);
  }

  // Stage 2
  // Unrolling this loops actually slows down Power9 benchmarks
  for (i = 0; i < 8; i++) {
    temp1[i] = vec_add(temp0[i], temp0[15 - i]);
    temp1[i + 8] = vec_sub(temp0[7 - i], temp0[i + 8]);
  }

  // Apply butterflies (in place) on pass through to stage 3.
  single_butterfly(temp1[27], temp1[20], &temp1[27], &temp1[20]);
  single_butterfly(temp1[26], temp1[21], &temp1[26], &temp1[21]);
  single_butterfly(temp1[25], temp1[22], &temp1[25], &temp1[22]);
  single_butterfly(temp1[24], temp1[23], &temp1[24], &temp1[23]);

  // dump the magnitude by 4, hence the intermediate values are within
  // the range of 16 bits.
  if (pass) {
    temp1[0] = add_round_shift_s16(temp1[0]);
    temp1[1] = add_round_shift_s16(temp1[1]);
    temp1[2] = add_round_shift_s16(temp1[2]);
    temp1[3] = add_round_shift_s16(temp1[3]);
    temp1[4] = add_round_shift_s16(temp1[4]);
    temp1[5] = add_round_shift_s16(temp1[5]);
    temp1[6] = add_round_shift_s16(temp1[6]);
    temp1[7] = add_round_shift_s16(temp1[7]);
    temp1[8] = add_round_shift_s16(temp1[8]);
    temp1[9] = add_round_shift_s16(temp1[9]);
    temp1[10] = add_round_shift_s16(temp1[10]);
    temp1[11] = add_round_shift_s16(temp1[11]);
    temp1[12] = add_round_shift_s16(temp1[12]);
    temp1[13] = add_round_shift_s16(temp1[13]);
    temp1[14] = add_round_shift_s16(temp1[14]);
    temp1[15] = add_round_shift_s16(temp1[15]);

    temp1[16] = add_round_shift_s16(temp1[16]);
    temp1[17] = add_round_shift_s16(temp1[17]);
    temp1[18] = add_round_shift_s16(temp1[18]);
    temp1[19] = add_round_shift_s16(temp1[19]);
    temp1[20] = add_round_shift_s16(temp1[20]);
    temp1[21] = add_round_shift_s16(temp1[21]);
    temp1[22] = add_round_shift_s16(temp1[22]);
    temp1[23] = add_round_shift_s16(temp1[23]);
    temp1[24] = add_round_shift_s16(temp1[24]);
    temp1[25] = add_round_shift_s16(temp1[25]);
    temp1[26] = add_round_shift_s16(temp1[26]);
    temp1[27] = add_round_shift_s16(temp1[27]);
    temp1[28] = add_round_shift_s16(temp1[28]);
    temp1[29] = add_round_shift_s16(temp1[29]);
    temp1[30] = add_round_shift_s16(temp1[30]);
    temp1[31] = add_round_shift_s16(temp1[31]);
  }

  // Stage 3
  temp2[0] = vec_add(temp1[0], temp1[7]);
  temp2[1] = vec_add(temp1[1], temp1[6]);
  temp2[2] = vec_add(temp1[2], temp1[5]);
  temp2[3] = vec_add(temp1[3], temp1[4]);
  temp2[5] = vec_sub(temp1[2], temp1[5]);
  temp2[6] = vec_sub(temp1[1], temp1[6]);
  temp2[8] = temp1[8];
  temp2[9] = temp1[9];

  single_butterfly(temp1[13], temp1[10], &temp2[13], &temp2[10]);
  single_butterfly(temp1[12], temp1[11], &temp2[12], &temp2[11]);
  temp2[14] = temp1[14];
  temp2[15] = temp1[15];

  temp2[18] = vec_add(temp1[18], temp1[21]);
  temp2[19] = vec_add(temp1[19], temp1[20]);

  temp2[20] = vec_sub(temp1[19], temp1[20]);
  temp2[21] = vec_sub(temp1[18], temp1[21]);

  temp2[26] = vec_sub(temp1[29], temp1[26]);
  temp2[27] = vec_sub(temp1[28], temp1[27]);

  temp2[28] = vec_add(temp1[28], temp1[27]);
  temp2[29] = vec_add(temp1[29], temp1[26]);

  // Pass through Stage 4
  temp0[7] = vec_sub(temp1[0], temp1[7]);
  temp0[4] = vec_sub(temp1[3], temp1[4]);
  temp0[16] = vec_add(temp1[16], temp1[23]);
  temp0[17] = vec_add(temp1[17], temp1[22]);
  temp0[22] = vec_sub(temp1[17], temp1[22]);
  temp0[23] = vec_sub(temp1[16], temp1[23]);
  temp0[24] = vec_sub(temp1[31], temp1[24]);
  temp0[25] = vec_sub(temp1[30], temp1[25]);
  temp0[30] = vec_add(temp1[30], temp1[25]);
  temp0[31] = vec_add(temp1[31], temp1[24]);

  // Stage 4
  temp0[0] = vec_add(temp2[0], temp2[3]);
  temp0[1] = vec_add(temp2[1], temp2[2]);
  temp0[2] = vec_sub(temp2[1], temp2[2]);
  temp0[3] = vec_sub(temp2[0], temp2[3]);
  single_butterfly(temp2[6], temp2[5], &temp0[6], &temp0[5]);

  temp0[9] = vec_add(temp2[9], temp2[10]);
  temp0[10] = vec_sub(temp2[9], temp2[10]);
  temp0[13] = vec_sub(temp2[14], temp2[13]);
  temp0[14] = vec_add(temp2[14], temp2[13]);

  double_butterfly(temp2[29], cospi8_v, temp2[18], cospi24_v, &temp0[29],
                   &temp0[18]);
  double_butterfly(temp2[28], cospi8_v, temp2[19], cospi24_v, &temp0[28],
                   &temp0[19]);
  double_butterfly(temp2[27], cospi24_v, temp2[20], cospi8m_v, &temp0[27],
                   &temp0[20]);
  double_butterfly(temp2[26], cospi24_v, temp2[21], cospi8m_v, &temp0[26],
                   &temp0[21]);

  // Pass through Stage 5
  temp1[8] = vec_add(temp2[8], temp2[11]);
  temp1[11] = vec_sub(temp2[8], temp2[11]);
  temp1[12] = vec_sub(temp2[15], temp2[12]);
  temp1[15] = vec_add(temp2[15], temp2[12]);

  // Stage 5
  // 0 and 1 pass through to 0 and 16 at the end
  single_butterfly(temp0[0], temp0[1], &out[0], &out[16]);

  // 2 and 3 pass through to 8 and 24 at the end
  double_butterfly(temp0[3], cospi8_v, temp0[2], cospi24_v, &out[8], &out[24]);

  temp1[4] = vec_add(temp0[4], temp0[5]);
  temp1[5] = vec_sub(temp0[4], temp0[5]);
  temp1[6] = vec_sub(temp0[7], temp0[6]);
  temp1[7] = vec_add(temp0[7], temp0[6]);

  double_butterfly(temp0[14], cospi8_v, temp0[9], cospi24_v, &temp1[14],
                   &temp1[9]);
  double_butterfly(temp0[13], cospi24_v, temp0[10], cospi8m_v, &temp1[13],
                   &temp1[10]);

  temp1[17] = vec_add(temp0[17], temp0[18]);
  temp1[18] = vec_sub(temp0[17], temp0[18]);

  temp1[21] = vec_sub(temp0[22], temp0[21]);
  temp1[22] = vec_add(temp0[22], temp0[21]);

  temp1[25] = vec_add(temp0[25], temp0[26]);
  temp1[26] = vec_sub(temp0[25], temp0[26]);

  temp1[29] = vec_sub(temp0[30], temp0[29]);
  temp1[30] = vec_add(temp0[30], temp0[29]);

  // Pass through Stage 6
  temp2[16] = vec_add(temp0[16], temp0[19]);
  temp2[19] = vec_sub(temp0[16], temp0[19]);
  temp2[20] = vec_sub(temp0[23], temp0[20]);
  temp2[23] = vec_add(temp0[23], temp0[20]);
  temp2[24] = vec_add(temp0[24], temp0[27]);
  temp2[27] = vec_sub(temp0[24], temp0[27]);
  temp2[28] = vec_sub(temp0[31], temp0[28]);
  temp2[31] = vec_add(temp0[31], temp0[28]);

  // Stage 6
  // 4 and 7 pass through to 4 and 28 at the end
  double_butterfly(temp1[7], cospi4_v, temp1[4], cospi28_v, &out[4], &out[28]);
  // 5 and 6 pass through to 20 and 12 at the end
  double_butterfly(temp1[6], cospi20_v, temp1[5], cospi12_v, &out[20],
                   &out[12]);
  temp2[8] = vec_add(temp1[8], temp1[9]);
  temp2[9] = vec_sub(temp1[8], temp1[9]);
  temp2[10] = vec_sub(temp1[11], temp1[10]);
  temp2[11] = vec_add(temp1[11], temp1[10]);
  temp2[12] = vec_add(temp1[12], temp1[13]);
  temp2[13] = vec_sub(temp1[12], temp1[13]);
  temp2[14] = vec_sub(temp1[15], temp1[14]);
  temp2[15] = vec_add(temp1[15], temp1[14]);

  double_butterfly(temp1[30], cospi4_v, temp1[17], cospi28_v, &temp2[30],
                   &temp2[17]);
  double_butterfly(temp1[29], cospi28_v, temp1[18], cospi4m_v, &temp2[29],
                   &temp2[18]);
  double_butterfly(temp1[26], cospi20_v, temp1[21], cospi12_v, &temp2[26],
                   &temp2[21]);
  double_butterfly(temp1[25], cospi12_v, temp1[22], cospi20m_v, &temp2[25],
                   &temp2[22]);

  // Stage 7
  double_butterfly(temp2[15], cospi2_v, temp2[8], cospi30_v, &out[2], &out[30]);
  double_butterfly(temp2[14], cospi18_v, temp2[9], cospi14_v, &out[18],
                   &out[14]);
  double_butterfly(temp2[13], cospi10_v, temp2[10], cospi22_v, &out[10],
                   &out[22]);
  double_butterfly(temp2[12], cospi26_v, temp2[11], cospi6_v, &out[26],
                   &out[6]);

  temp0[16] = vec_add(temp2[16], temp2[17]);
  temp0[17] = vec_sub(temp2[16], temp2[17]);
  temp0[18] = vec_sub(temp2[19], temp2[18]);
  temp0[19] = vec_add(temp2[19], temp2[18]);
  temp0[20] = vec_add(temp2[20], temp2[21]);
  temp0[21] = vec_sub(temp2[20], temp2[21]);
  temp0[22] = vec_sub(temp2[23], temp2[22]);
  temp0[23] = vec_add(temp2[23], temp2[22]);
  temp0[24] = vec_add(temp2[24], temp2[25]);
  temp0[25] = vec_sub(temp2[24], temp2[25]);
  temp0[26] = vec_sub(temp2[27], temp2[26]);
  temp0[27] = vec_add(temp2[27], temp2[26]);
  temp0[28] = vec_add(temp2[28], temp2[29]);
  temp0[29] = vec_sub(temp2[28], temp2[29]);
  temp0[30] = vec_sub(temp2[31], temp2[30]);
  temp0[31] = vec_add(temp2[31], temp2[30]);

  // Final stage --- outputs indices are bit-reversed.
  double_butterfly(temp0[31], cospi1_v, temp0[16], cospi31_v, &out[1],
                   &out[31]);
  double_butterfly(temp0[30], cospi17_v, temp0[17], cospi15_v, &out[17],
                   &out[15]);
  double_butterfly(temp0[29], cospi9_v, temp0[18], cospi23_v, &out[9],
                   &out[23]);
  double_butterfly(temp0[28], cospi25_v, temp0[19], cospi7_v, &out[25],
                   &out[7]);
  double_butterfly(temp0[27], cospi5_v, temp0[20], cospi27_v, &out[5],
                   &out[27]);
  double_butterfly(temp0[26], cospi21_v, temp0[21], cospi11_v, &out[21],
                   &out[11]);
  double_butterfly(temp0[25], cospi13_v, temp0[22], cospi19_v, &out[13],
                   &out[19]);
  double_butterfly(temp0[24], cospi29_v, temp0[23], cospi3_v, &out[29],
                   &out[3]);

  if (pass == 0) {
    for (i = 0; i < 32; i++) {
      out[i] = sub_round_shift(out[i]);
    }
  }
}

void vpx_fdct32x32_rd_vsx(const int16_t *input, tran_low_t *out, int stride) {
  int16x8_t temp0[32];
  int16x8_t temp1[32];
  int16x8_t temp2[32];
  int16x8_t temp3[32];
  int16x8_t temp4[32];
  int16x8_t temp5[32];
  int16x8_t temp6[32];

  // Process in 8x32 columns.
  load(input, stride, temp0);
  fdct32_vsx(temp0, temp1, 0);

  load(input + 8, stride, temp0);
  fdct32_vsx(temp0, temp2, 0);

  load(input + 16, stride, temp0);
  fdct32_vsx(temp0, temp3, 0);

  load(input + 24, stride, temp0);
  fdct32_vsx(temp0, temp4, 0);

  // Generate the top row by munging the first set of 8 from each one
  // together.
  transpose_8x8(&temp1[0], &temp0[0]);
  transpose_8x8(&temp2[0], &temp0[8]);
  transpose_8x8(&temp3[0], &temp0[16]);
  transpose_8x8(&temp4[0], &temp0[24]);

  fdct32_vsx(temp0, temp5, 1);

  transpose_8x8(&temp5[0], &temp6[0]);
  transpose_8x8(&temp5[8], &temp6[8]);
  transpose_8x8(&temp5[16], &temp6[16]);
  transpose_8x8(&temp5[24], &temp6[24]);

  store(out, temp6);

  // Second row of 8x32.
  transpose_8x8(&temp1[8], &temp0[0]);
  transpose_8x8(&temp2[8], &temp0[8]);
  transpose_8x8(&temp3[8], &temp0[16]);
  transpose_8x8(&temp4[8], &temp0[24]);

  fdct32_vsx(temp0, temp5, 1);

  transpose_8x8(&temp5[0], &temp6[0]);
  transpose_8x8(&temp5[8], &temp6[8]);
  transpose_8x8(&temp5[16], &temp6[16]);
  transpose_8x8(&temp5[24], &temp6[24]);

  store(out + 8 * 32, temp6);

  // Third row of 8x32
  transpose_8x8(&temp1[16], &temp0[0]);
  transpose_8x8(&temp2[16], &temp0[8]);
  transpose_8x8(&temp3[16], &temp0[16]);
  transpose_8x8(&temp4[16], &temp0[24]);

  fdct32_vsx(temp0, temp5, 1);

  transpose_8x8(&temp5[0], &temp6[0]);
  transpose_8x8(&temp5[8], &temp6[8]);
  transpose_8x8(&temp5[16], &temp6[16]);
  transpose_8x8(&temp5[24], &temp6[24]);

  store(out + 16 * 32, temp6);

  // Final row of 8x32.
  transpose_8x8(&temp1[24], &temp0[0]);
  transpose_8x8(&temp2[24], &temp0[8]);
  transpose_8x8(&temp3[24], &temp0[16]);
  transpose_8x8(&temp4[24], &temp0[24]);

  fdct32_vsx(temp0, temp5, 1);

  transpose_8x8(&temp5[0], &temp6[0]);
  transpose_8x8(&temp5[8], &temp6[8]);
  transpose_8x8(&temp5[16], &temp6[16]);
  transpose_8x8(&temp5[24], &temp6[24]);

  store(out + 24 * 32, temp6);
}