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

use super::super::*;
use sixty_four::{i64x2, u64x2};

#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct u32x2(u32, u32);
#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct i32x2(i32, i32);
#[repr(simd)]
#[cfg_attr(feature = "with-serde", derive(Serialize, Deserialize))]
#[derive(Debug, Copy)]
pub struct bool32ix2(i32, i32);

#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct f32x2(f32, f32);
#[repr(simd)]
#[cfg_attr(feature = "with-serde", derive(Serialize, Deserialize))]
#[derive(Debug, Copy)]
pub struct bool32fx2(i32, i32);

#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct u16x4(u16, u16, u16, u16);
#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct i16x4(i16, i16, i16, i16);
#[repr(simd)]
#[cfg_attr(feature = "with-serde", derive(Serialize, Deserialize))]
#[derive(Debug, Copy)]
pub struct bool16ix4(i16, i16, i16, i16);

#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct u8x8(u8, u8, u8, u8,
            u8, u8, u8, u8);
#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct i8x8(i8, i8, i8, i8,
                i8, i8, i8, i8);
#[repr(simd)]
#[cfg_attr(feature = "with-serde", derive(Serialize, Deserialize))]
#[derive(Debug, Copy)]
pub struct bool8ix8(i8, i8, i8, i8,
                    i8, i8, i8, i8);

#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct i64x1(i64);
#[repr(simd)]
#[derive(Debug, Copy, Clone)]
pub struct u64x1(u64);

macro_rules! half_bools {
    ($($ty: ty, $as_u: ty, $elem: ty, $all: ident ($min: ident), $any: ident ($max: ident);)*) => {
        $(
            impl $ty {
                #[inline]
                pub fn $all(self) -> bool {
                    unsafe {
                        let t: $as_u = bitcast(self);
                        let y = $min(t, mem::uninitialized());
                        let y32: u32x2 = bitcast(y);
                        y32.0 == 0xFFFFFFFF
                    }
                }
                #[inline]
                pub fn $any(self) -> bool {
                    unsafe {
                        let t: $as_u = bitcast(self);
                        let y = $max(t, mem::uninitialized());
                        let y32: u32x2 = bitcast(y);
                        y32.0 != 0
                    }
                }
            }

            impl Clone for $ty {
                #[inline] fn clone(&self) -> Self {
                    *self
                }
            }

            unsafe impl Simd for $ty {
                type Bool = $ty;
                type Elem = $elem;
            }

            )*
    }
}

half_bools! {
    bool32fx2, u32x2, i32, bool32fx2_all(arm_vpmin_u32), bool32fx2_any(arm_vpmax_u32);
    bool8ix8, u8x8, i8, bool8ix8_all(arm_vpmin_u8), bool8ix8_any(arm_vpmax_u8);
    bool16ix4, u16x4, i16, bool16ix4_all(arm_vpmin_u16), bool16ix4_any(arm_vpmax_u16);
    bool32ix2, u32x2, f32, bool32ix2_all(arm_vpmin_u32), bool32ix2_any(arm_vpmax_u32);
}

macro_rules! half_simd {
    ($($ty: ty, $elem: ty, $bool_ty: ty;)*) => {
        $(
            unsafe impl Simd for $ty {
                type Bool = $bool_ty;
                type Elem = $elem;
            }
            )*
    }
}

half_simd! {
    f32x2, f32, bool32fx2;
    u32x2, u32, bool32ix2;
    i32x2, i32, bool32ix2;
    u16x4, u16, bool16ix4;
    i16x4, i16, bool16ix4;
    u8x8, u8, bool8ix8;
    i8x8, i8, bool8ix8;
}

#[allow(dead_code)]
extern "platform-intrinsic" {
    fn arm_vhadd_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vhadd_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vhadd_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vhadd_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vhadd_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vhadd_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vhaddq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vhaddq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vhaddq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vhaddq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vhaddq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vhaddq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vrhadd_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vrhadd_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vrhadd_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vrhadd_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vrhadd_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vrhadd_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vrhaddq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vrhaddq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vrhaddq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vrhaddq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vrhaddq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vrhaddq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vqadd_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vqadd_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vqadd_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vqadd_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vqadd_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vqadd_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vqadd_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vqadd_u64(x: u64x1, y: u64x1) -> u64x1;
    fn arm_vqaddq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vqaddq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vqaddq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vqaddq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vqaddq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vqaddq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vqaddq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vqaddq_u64(x: u64x2, y: u64x2) -> u64x2;
    fn arm_vraddhn_s16(x: i16x8, y: i16x8) -> i8x8;
    fn arm_vraddhn_u16(x: u16x8, y: u16x8) -> u8x8;
    fn arm_vraddhn_s32(x: i32x4, y: i32x4) -> i16x4;
    fn arm_vraddhn_u32(x: u32x4, y: u32x4) -> u16x4;
    fn arm_vraddhn_s64(x: i64x2, y: i64x2) -> i32x2;
    fn arm_vraddhn_u64(x: u64x2, y: u64x2) -> u32x2;
    fn arm_vfma_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vfmaq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vqdmulh_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vqdmulh_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vqdmulhq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vqdmulhq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vqrdmulh_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vqrdmulh_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vqrdmulhq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vqrdmulhq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vmull_s8(x: i8x8, y: i8x8) -> i16x8;
    fn arm_vmull_u8(x: u8x8, y: u8x8) -> u16x8;
    fn arm_vmull_s16(x: i16x4, y: i16x4) -> i32x4;
    fn arm_vmull_u16(x: u16x4, y: u16x4) -> u32x4;
    fn arm_vmull_s32(x: i32x2, y: i32x2) -> i64x2;
    fn arm_vmull_u32(x: u32x2, y: u32x2) -> u64x2;
    fn arm_vqdmullq_s8(x: i8x8, y: i8x8) -> i16x8;
    fn arm_vqdmullq_s16(x: i16x4, y: i16x4) -> i32x4;
    fn arm_vhsub_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vhsub_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vhsub_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vhsub_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vhsub_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vhsub_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vhsubq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vhsubq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vhsubq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vhsubq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vhsubq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vhsubq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vqsub_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vqsub_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vqsub_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vqsub_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vqsub_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vqsub_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vqsub_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vqsub_u64(x: u64x1, y: u64x1) -> u64x1;
    fn arm_vqsubq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vqsubq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vqsubq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vqsubq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vqsubq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vqsubq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vqsubq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vqsubq_u64(x: u64x2, y: u64x2) -> u64x2;
    fn arm_vrsubhn_s16(x: i16x8, y: i16x8) -> i8x8;
    fn arm_vrsubhn_u16(x: u16x8, y: u16x8) -> u8x8;
    fn arm_vrsubhn_s32(x: i32x4, y: i32x4) -> i16x4;
    fn arm_vrsubhn_u32(x: u32x4, y: u32x4) -> u16x4;
    fn arm_vrsubhn_s64(x: i64x2, y: i64x2) -> i32x2;
    fn arm_vrsubhn_u64(x: u64x2, y: u64x2) -> u32x2;
    fn arm_vabd_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vabd_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vabd_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vabd_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vabd_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vabd_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vabd_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vabdq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vabdq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vabdq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vabdq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vabdq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vabdq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vabdq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vmax_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vmax_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vmax_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vmax_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vmax_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vmax_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vmax_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vmaxq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vmaxq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vmaxq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vmaxq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vmaxq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vmaxq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vmaxq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vmin_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vmin_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vmin_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vmin_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vmin_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vmin_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vmin_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vminq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vminq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vminq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vminq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vminq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vminq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vminq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vshl_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vshl_u8(x: u8x8, y: i8x8) -> u8x8;
    fn arm_vshl_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vshl_u16(x: u16x4, y: i16x4) -> u16x4;
    fn arm_vshl_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vshl_u32(x: u32x2, y: i32x2) -> u32x2;
    fn arm_vshl_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vshl_u64(x: u64x1, y: i64x1) -> u64x1;
    fn arm_vshlq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vshlq_u8(x: u8x16, y: i8x16) -> u8x16;
    fn arm_vshlq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vshlq_u16(x: u16x8, y: i16x8) -> u16x8;
    fn arm_vshlq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vshlq_u32(x: u32x4, y: i32x4) -> u32x4;
    fn arm_vshlq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vshlq_u64(x: u64x2, y: i64x2) -> u64x2;
    fn arm_vqshl_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vqshl_u8(x: u8x8, y: i8x8) -> u8x8;
    fn arm_vqshl_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vqshl_u16(x: u16x4, y: i16x4) -> u16x4;
    fn arm_vqshl_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vqshl_u32(x: u32x2, y: i32x2) -> u32x2;
    fn arm_vqshl_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vqshl_u64(x: u64x1, y: i64x1) -> u64x1;
    fn arm_vqshlq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vqshlq_u8(x: u8x16, y: i8x16) -> u8x16;
    fn arm_vqshlq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vqshlq_u16(x: u16x8, y: i16x8) -> u16x8;
    fn arm_vqshlq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vqshlq_u32(x: u32x4, y: i32x4) -> u32x4;
    fn arm_vqshlq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vqshlq_u64(x: u64x2, y: i64x2) -> u64x2;
    fn arm_vrshl_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vrshl_u8(x: u8x8, y: i8x8) -> u8x8;
    fn arm_vrshl_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vrshl_u16(x: u16x4, y: i16x4) -> u16x4;
    fn arm_vrshl_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vrshl_u32(x: u32x2, y: i32x2) -> u32x2;
    fn arm_vrshl_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vrshl_u64(x: u64x1, y: i64x1) -> u64x1;
    fn arm_vrshlq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vrshlq_u8(x: u8x16, y: i8x16) -> u8x16;
    fn arm_vrshlq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vrshlq_u16(x: u16x8, y: i16x8) -> u16x8;
    fn arm_vrshlq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vrshlq_u32(x: u32x4, y: i32x4) -> u32x4;
    fn arm_vrshlq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vrshlq_u64(x: u64x2, y: i64x2) -> u64x2;
    fn arm_vqrshl_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vqrshl_u8(x: u8x8, y: i8x8) -> u8x8;
    fn arm_vqrshl_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vqrshl_u16(x: u16x4, y: i16x4) -> u16x4;
    fn arm_vqrshl_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vqrshl_u32(x: u32x2, y: i32x2) -> u32x2;
    fn arm_vqrshl_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vqrshl_u64(x: u64x1, y: i64x1) -> u64x1;
    fn arm_vqrshlq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vqrshlq_u8(x: u8x16, y: i8x16) -> u8x16;
    fn arm_vqrshlq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vqrshlq_u16(x: u16x8, y: i16x8) -> u16x8;
    fn arm_vqrshlq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vqrshlq_u32(x: u32x4, y: i32x4) -> u32x4;
    fn arm_vqrshlq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vqrshlq_u64(x: u64x2, y: i64x2) -> u64x2;
    fn arm_vqshrun_n_s16(x: i16x8, y: u32) -> i8x8;
    fn arm_vqshrun_n_s32(x: i32x4, y: u32) -> i16x4;
    fn arm_vqshrun_n_s64(x: i64x2, y: u32) -> i32x2;
    fn arm_vqrshrun_n_s16(x: i16x8, y: u32) -> i8x8;
    fn arm_vqrshrun_n_s32(x: i32x4, y: u32) -> i16x4;
    fn arm_vqrshrun_n_s64(x: i64x2, y: u32) -> i32x2;
    fn arm_vqshrn_n_s16(x: i16x8, y: u32) -> i8x8;
    fn arm_vqshrn_n_u16(x: u16x8, y: u32) -> u8x8;
    fn arm_vqshrn_n_s32(x: i32x4, y: u32) -> i16x4;
    fn arm_vqshrn_n_u32(x: u32x4, y: u32) -> u16x4;
    fn arm_vqshrn_n_s64(x: i64x2, y: u32) -> i32x2;
    fn arm_vqshrn_n_u64(x: u64x2, y: u32) -> u32x2;
    fn arm_vrshrn_n_s16(x: i16x8, y: u32) -> i8x8;
    fn arm_vrshrn_n_u16(x: u16x8, y: u32) -> u8x8;
    fn arm_vrshrn_n_s32(x: i32x4, y: u32) -> i16x4;
    fn arm_vrshrn_n_u32(x: u32x4, y: u32) -> u16x4;
    fn arm_vrshrn_n_s64(x: i64x2, y: u32) -> i32x2;
    fn arm_vrshrn_n_u64(x: u64x2, y: u32) -> u32x2;
    fn arm_vqrshrn_n_s16(x: i16x8, y: u32) -> i8x8;
    fn arm_vqrshrn_n_u16(x: u16x8, y: u32) -> u8x8;
    fn arm_vqrshrn_n_s32(x: i32x4, y: u32) -> i16x4;
    fn arm_vqrshrn_n_u32(x: u32x4, y: u32) -> u16x4;
    fn arm_vqrshrn_n_s64(x: i64x2, y: u32) -> i32x2;
    fn arm_vqrshrn_n_u64(x: u64x2, y: u32) -> u32x2;
    fn arm_vsri_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vsri_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vsri_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vsri_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vsri_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vsri_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vsri_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vsri_u64(x: u64x1, y: u64x1) -> u64x1;
    fn arm_vsriq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vsriq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vsriq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vsriq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vsriq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vsriq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vsriq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vsriq_u64(x: u64x2, y: u64x2) -> u64x2;
    fn arm_vsli_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vsli_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vsli_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vsli_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vsli_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vsli_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vsli_s64(x: i64x1, y: i64x1) -> i64x1;
    fn arm_vsli_u64(x: u64x1, y: u64x1) -> u64x1;
    fn arm_vsliq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vsliq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vsliq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vsliq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vsliq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vsliq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vsliq_s64(x: i64x2, y: i64x2) -> i64x2;
    fn arm_vsliq_u64(x: u64x2, y: u64x2) -> u64x2;
    fn arm_vvqmovn_s16(x: i16x8) -> i8x8;
    fn arm_vvqmovn_u16(x: u16x8) -> u8x8;
    fn arm_vvqmovn_s32(x: i32x4) -> i16x4;
    fn arm_vvqmovn_u32(x: u32x4) -> u16x4;
    fn arm_vvqmovn_s64(x: i64x2) -> i32x2;
    fn arm_vvqmovn_u64(x: u64x2) -> u32x2;
    fn arm_vabs_s8(x: i8x8) -> i8x8;
    fn arm_vabs_s16(x: i16x4) -> i16x4;
    fn arm_vabs_s32(x: i32x2) -> i32x2;
    fn arm_vabsq_s8(x: i8x16) -> i8x16;
    fn arm_vabsq_s16(x: i16x8) -> i16x8;
    fn arm_vabsq_s32(x: i32x4) -> i32x4;
    fn arm_vabs_f32(x: f32x2) -> f32x2;
    fn arm_vabsq_f32(x: f32x4) -> f32x4;
    fn arm_vqabs_s8(x: i8x8) -> i8x8;
    fn arm_vqabs_s16(x: i16x4) -> i16x4;
    fn arm_vqabs_s32(x: i32x2) -> i32x2;
    fn arm_vqabsq_s8(x: i8x16) -> i8x16;
    fn arm_vqabsq_s16(x: i16x8) -> i16x8;
    fn arm_vqabsq_s32(x: i32x4) -> i32x4;
    fn arm_vqneg_s8(x: i8x8) -> i8x8;
    fn arm_vqneg_s16(x: i16x4) -> i16x4;
    fn arm_vqneg_s32(x: i32x2) -> i32x2;
    fn arm_vqnegq_s8(x: i8x16) -> i8x16;
    fn arm_vqnegq_s16(x: i16x8) -> i16x8;
    fn arm_vqnegq_s32(x: i32x4) -> i32x4;
    fn arm_vclz_s8(x: i8x8) -> i8x8;
    fn arm_vclz_u8(x: u8x8) -> u8x8;
    fn arm_vclz_s16(x: i16x4) -> i16x4;
    fn arm_vclz_u16(x: u16x4) -> u16x4;
    fn arm_vclz_s32(x: i32x2) -> i32x2;
    fn arm_vclz_u32(x: u32x2) -> u32x2;
    fn arm_vclzq_s8(x: i8x16) -> i8x16;
    fn arm_vclzq_u8(x: u8x16) -> u8x16;
    fn arm_vclzq_s16(x: i16x8) -> i16x8;
    fn arm_vclzq_u16(x: u16x8) -> u16x8;
    fn arm_vclzq_s32(x: i32x4) -> i32x4;
    fn arm_vclzq_u32(x: u32x4) -> u32x4;
    fn arm_vcls_s8(x: i8x8) -> i8x8;
    fn arm_vcls_u8(x: u8x8) -> u8x8;
    fn arm_vcls_s16(x: i16x4) -> i16x4;
    fn arm_vcls_u16(x: u16x4) -> u16x4;
    fn arm_vcls_s32(x: i32x2) -> i32x2;
    fn arm_vcls_u32(x: u32x2) -> u32x2;
    fn arm_vclsq_s8(x: i8x16) -> i8x16;
    fn arm_vclsq_u8(x: u8x16) -> u8x16;
    fn arm_vclsq_s16(x: i16x8) -> i16x8;
    fn arm_vclsq_u16(x: u16x8) -> u16x8;
    fn arm_vclsq_s32(x: i32x4) -> i32x4;
    fn arm_vclsq_u32(x: u32x4) -> u32x4;
    fn arm_vcnt_s8(x: i8x8) -> i8x8;
    fn arm_vcnt_u8(x: u8x8) -> u8x8;
    fn arm_vcntq_s8(x: i8x16) -> i8x16;
    fn arm_vcntq_u8(x: u8x16) -> u8x16;
    fn arm_vrecpe_u32(x: u32x2) -> u32x2;
    fn arm_vrecpe_f32(x: f32x2) -> f32x2;
    fn arm_vrecpeq_u32(x: u32x4) -> u32x4;
    fn arm_vrecpeq_f32(x: f32x4) -> f32x4;
    fn arm_vrecps_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vrecpsq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vsqrt_f32(x: f32x2) -> f32x2;
    fn arm_vsqrtq_f32(x: f32x4) -> f32x4;
    fn arm_vrsqrte_u32(x: u32x2) -> u32x2;
    fn arm_vrsqrte_f32(x: f32x2) -> f32x2;
    fn arm_vrsqrteq_u32(x: u32x4) -> u32x4;
    fn arm_vrsqrteq_f32(x: f32x4) -> f32x4;
    fn arm_vrsqrts_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vrsqrtsq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vbsl_s8(x: u8x8, y: i8x8) -> i8x8;
    fn arm_vbsl_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vbsl_s16(x: u16x4, y: i16x4) -> i16x4;
    fn arm_vbsl_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vbsl_s32(x: u32x2, y: i32x2) -> i32x2;
    fn arm_vbsl_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vbsl_s64(x: u64x1, y: i64x1) -> i64x1;
    fn arm_vbsl_u64(x: u64x1, y: u64x1) -> u64x1;
    fn arm_vbslq_s8(x: u8x16, y: i8x16) -> i8x16;
    fn arm_vbslq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vbslq_s16(x: u16x8, y: i16x8) -> i16x8;
    fn arm_vbslq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vbslq_s32(x: u32x4, y: i32x4) -> i32x4;
    fn arm_vbslq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vbslq_s64(x: u64x2, y: i64x2) -> i64x2;
    fn arm_vbslq_u64(x: u64x2, y: u64x2) -> u64x2;
    fn arm_vpadd_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vpadd_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vpadd_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vpadd_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vpadd_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vpadd_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vpadd_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vpaddl_s16(x: i8x8) -> i16x4;
    fn arm_vpaddl_u16(x: u8x8) -> u16x4;
    fn arm_vpaddl_s32(x: i16x4) -> i32x2;
    fn arm_vpaddl_u32(x: u16x4) -> u32x2;
    fn arm_vpaddl_s64(x: i32x2) -> i64x1;
    fn arm_vpaddl_u64(x: u32x2) -> u64x1;
    fn arm_vpaddlq_s16(x: i8x16) -> i16x8;
    fn arm_vpaddlq_u16(x: u8x16) -> u16x8;
    fn arm_vpaddlq_s32(x: i16x8) -> i32x4;
    fn arm_vpaddlq_u32(x: u16x8) -> u32x4;
    fn arm_vpaddlq_s64(x: i32x4) -> i64x2;
    fn arm_vpaddlq_u64(x: u32x4) -> u64x2;
    fn arm_vpadal_s16(x: i16x4, y: i8x8) -> i16x4;
    fn arm_vpadal_u16(x: u16x4, y: u8x8) -> u16x4;
    fn arm_vpadal_s32(x: i32x2, y: i16x4) -> i32x2;
    fn arm_vpadal_u32(x: u32x2, y: u16x4) -> u32x2;
    fn arm_vpadal_s64(x: i64x1, y: i32x2) -> i64x1;
    fn arm_vpadal_u64(x: u64x1, y: u32x2) -> u64x1;
    fn arm_vpadalq_s16(x: i16x8, y: i8x16) -> i16x8;
    fn arm_vpadalq_u16(x: u16x8, y: u8x16) -> u16x8;
    fn arm_vpadalq_s32(x: i32x4, y: i16x8) -> i32x4;
    fn arm_vpadalq_u32(x: u32x4, y: u16x8) -> u32x4;
    fn arm_vpadalq_s64(x: i64x2, y: i32x4) -> i64x2;
    fn arm_vpadalq_u64(x: u64x2, y: u32x4) -> u64x2;
    fn arm_vpmax_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vpmax_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vpmax_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vpmax_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vpmax_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vpmax_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vpmax_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vpmin_s8(x: i8x8, y: i8x8) -> i8x8;
    fn arm_vpmin_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vpmin_s16(x: i16x4, y: i16x4) -> i16x4;
    fn arm_vpmin_u16(x: u16x4, y: u16x4) -> u16x4;
    fn arm_vpmin_s32(x: i32x2, y: i32x2) -> i32x2;
    fn arm_vpmin_u32(x: u32x2, y: u32x2) -> u32x2;
    fn arm_vpmin_f32(x: f32x2, y: f32x2) -> f32x2;
    fn arm_vpminq_s8(x: i8x16, y: i8x16) -> i8x16;
    fn arm_vpminq_u8(x: u8x16, y: u8x16) -> u8x16;
    fn arm_vpminq_s16(x: i16x8, y: i16x8) -> i16x8;
    fn arm_vpminq_u16(x: u16x8, y: u16x8) -> u16x8;
    fn arm_vpminq_s32(x: i32x4, y: i32x4) -> i32x4;
    fn arm_vpminq_u32(x: u32x4, y: u32x4) -> u32x4;
    fn arm_vpminq_f32(x: f32x4, y: f32x4) -> f32x4;
    fn arm_vtbl1_s8(x: i8x8, y: u8x8) -> i8x8;
    fn arm_vtbl1_u8(x: u8x8, y: u8x8) -> u8x8;
    fn arm_vtbx1_s8(x: i8x8, y: i8x8, z: u8x8) -> i8x8;
    fn arm_vtbx1_u8(x: u8x8, y: u8x8, z: u8x8) -> u8x8;
    fn arm_vtbl2_s8(x: (i8x8, i8x8), y: u8x8) -> i8x8;
    fn arm_vtbl2_u8(x: (u8x8, u8x8), y: u8x8) -> u8x8;
    fn arm_vtbx2_s8(x: (i8x8, i8x8), y: u8x8) -> i8x8;
    fn arm_vtbx2_u8(x: (u8x8, u8x8), y: u8x8) -> u8x8;
    fn arm_vtbl3_s8(x: (i8x8, i8x8, i8x8), y: u8x8) -> i8x8;
    fn arm_vtbl3_u8(x: (u8x8, u8x8, u8x8), y: u8x8) -> u8x8;
    fn arm_vtbx3_s8(x: i8x8, y: (i8x8, i8x8, i8x8), z: u8x8) -> i8x8;
    fn arm_vtbx3_u8(x: u8x8, y: (u8x8, u8x8, u8x8), z: u8x8) -> u8x8;
    fn arm_vtbl4_s8(x: (i8x8, i8x8, i8x8, i8x8), y: u8x8) -> i8x8;
    fn arm_vtbl4_u8(x: (u8x8, u8x8, u8x8, u8x8), y: u8x8) -> u8x8;
    fn arm_vtbx4_s8(x: i8x8, y: (i8x8, i8x8, i8x8, i8x8), z: u8x8) -> i8x8;
    fn arm_vtbx4_u8(x: u8x8, y: (u8x8, u8x8, u8x8, u8x8), z: u8x8) -> u8x8;
}


impl u8x8 {
    #[inline]
    pub fn table_lookup_1(self, t0: u8x8) -> u8x8 {
        unsafe {arm_vtbl1_u8(t0, self)}
    }
    #[inline]
    pub fn table_lookup_2(self, t0: u8x8, t1: u8x8) -> u8x8 {
        unsafe {arm_vtbl2_u8((t0, t1), self)}
    }
    #[inline]
    pub fn table_lookup_3(self, t0: u8x8, t1: u8x8, t2: u8x8) -> u8x8 {
        unsafe {arm_vtbl3_u8((t0, t1, t2), self)}
    }
    #[inline]
    pub fn table_lookup_4(self, t0: u8x8, t1: u8x8, t2: u8x8, t3: u8x8) -> u8x8 {
        unsafe {arm_vtbl4_u8((t0, t1, t2, t3), self)}
    }
}

#[doc(hidden)]
pub mod common {
    use super::super::super::*;
    use super::*;
    use core::mem;

    #[inline]
    pub fn f32x4_sqrt(x: f32x4) -> f32x4 {
        unsafe {super::arm_vsqrtq_f32(x)}
    }
    #[inline]
    pub fn f32x4_approx_rsqrt(x: f32x4) -> f32x4 {
        unsafe {super::arm_vrsqrteq_f32(x)}
    }
    #[inline]
    pub fn f32x4_approx_reciprocal(x: f32x4) -> f32x4 {
        unsafe {super::arm_vrecpeq_f32(x)}
    }
    #[inline]
    pub fn f32x4_max(x: f32x4, y: f32x4) -> f32x4 {
        unsafe {super::arm_vmaxq_f32(x, y)}
    }
    #[inline]
    pub fn f32x4_min(x: f32x4, y: f32x4) -> f32x4 {
        unsafe {super::arm_vminq_f32(x, y)}
    }

    macro_rules! bools {
        ($($ty: ty, $as_u: ty, $shuffle_fn: ident, $lo_idxs: expr, $hi_idxs: expr, $all: ident ($min: ident), $any: ident ($max: ident);)*) => {
            $(
                #[inline]
                pub fn $all(x: $ty) -> bool {
                    unsafe {
                        let t: $as_u = bitcast(x);
                        let lo = $shuffle_fn(t, t, $lo_idxs);
                        let hi = $shuffle_fn(t, t, $hi_idxs);
                        let x = super::$min(lo, hi);
                        let y = super::$min(x, mem::uninitialized());
                        let y32: u32x2 = bitcast(y);
                        y32.0 == 0xFFFFFFFF
                    }
                }
                #[inline]
                pub fn $any(x: $ty) -> bool {
                    unsafe {
                        let t: $as_u = bitcast(x);
                        let lo = $shuffle_fn(t, t, $lo_idxs);
                        let hi = $shuffle_fn(t, t, $hi_idxs);
                        let x = super::$max(lo, hi);
                        let y = super::$max(x, mem::uninitialized());
                        let y32: u32x2 = bitcast(y);
                        y32.0 != 0
                    }
                }
                )*
        }
    }

    bools! {
        bool32fx4, u32x4, simd_shuffle2, [0, 1], [2, 3], bool32fx4_all(arm_vpmin_u32), bool32fx4_any(arm_vpmax_u32);
        bool8ix16, u8x16, simd_shuffle8, [0, 1, 2, 3, 4, 5, 6, 7], [8, 9, 10, 11, 12, 13, 14, 15], bool8ix16_all(arm_vpmin_u8), bool8ix16_any(arm_vpmax_u8);
        bool16ix8, u16x8, simd_shuffle4, [0, 1, 2, 3], [4, 5, 6, 7], bool16ix8_all(arm_vpmin_u16), bool16ix8_any(arm_vpmax_u16);
        bool32ix4, u32x4, simd_shuffle2, [0, 1], [2, 3], bool32ix4_all(arm_vpmin_u32), bool32ix4_any(arm_vpmax_u32);
    }
}