llvm.org GIT mirror llvm / release_27 lib / Target / ARM / ARMScheduleV7.td
release_27

Tree @release_27 (Download .tar.gz)

ARMScheduleV7.td @release_27raw · history · blame

  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
//===- ARMScheduleV7.td - ARM v7 Scheduling Definitions ----*- tablegen -*-===//
// 
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
// 
//===----------------------------------------------------------------------===//
//
// This file defines the itinerary class data for the ARM v7 processors.
//
//===----------------------------------------------------------------------===//

//
// Scheduling information derived from "Cortex-A8 Technical Reference Manual".
//
// Dual issue pipeline represented by FU_Pipe0 | FU_Pipe1
//
def CortexA8Itineraries : ProcessorItineraries<[

  // Two fully-pipelined integer ALU pipelines
  //
  // No operand cycles
  InstrItinData<IIC_iALUx    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>]>,
  //
  // Binary Instructions that produce a result
  InstrItinData<IIC_iALUi    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 2]>,
  InstrItinData<IIC_iALUr    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 2, 2]>,
  InstrItinData<IIC_iALUsi   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 2, 1]>,
  InstrItinData<IIC_iALUsr   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 2, 1, 1]>,
  //
  // Unary Instructions that produce a result
  InstrItinData<IIC_iUNAr    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 2]>,
  InstrItinData<IIC_iUNAsi   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1]>,
  InstrItinData<IIC_iUNAsr   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1, 1]>,
  //
  // Compare instructions
  InstrItinData<IIC_iCMPi    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2]>,
  InstrItinData<IIC_iCMPr    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 2]>,
  InstrItinData<IIC_iCMPsi   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1]>,
  InstrItinData<IIC_iCMPsr   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1, 1]>,
  //
  // Move instructions, unconditional
  InstrItinData<IIC_iMOVi    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [1]>,
  InstrItinData<IIC_iMOVr    , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [1, 1]>,
  InstrItinData<IIC_iMOVsi   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [1, 1]>,
  InstrItinData<IIC_iMOVsr   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [1, 1, 1]>,
  //
  // Move instructions, conditional
  InstrItinData<IIC_iCMOVi   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2]>,
  InstrItinData<IIC_iCMOVr   , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1]>,
  InstrItinData<IIC_iCMOVsi  , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1]>,
  InstrItinData<IIC_iCMOVsr  , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>], [2, 1, 1]>,

  // Integer multiply pipeline
  // Result written in E5, but that is relative to the last cycle of multicycle,
  // so we use 6 for those cases
  //
  InstrItinData<IIC_iMUL16   , [InstrStage<1, [FU_Pipe0]>], [5, 1, 1]>,
  InstrItinData<IIC_iMAC16   , [InstrStage<1, [FU_Pipe1], 0>, 
                                InstrStage<2, [FU_Pipe0]>], [6, 1, 1, 4]>,
  InstrItinData<IIC_iMUL32   , [InstrStage<1, [FU_Pipe1], 0>, 
                                InstrStage<2, [FU_Pipe0]>], [6, 1, 1]>,
  InstrItinData<IIC_iMAC32   , [InstrStage<1, [FU_Pipe1], 0>, 
                                InstrStage<2, [FU_Pipe0]>], [6, 1, 1, 4]>,
  InstrItinData<IIC_iMUL64   , [InstrStage<2, [FU_Pipe1], 0>, 
                                InstrStage<3, [FU_Pipe0]>], [6, 6, 1, 1]>,
  InstrItinData<IIC_iMAC64   , [InstrStage<2, [FU_Pipe1], 0>, 
                                InstrStage<3, [FU_Pipe0]>], [6, 6, 1, 1]>,
  
  // Integer load pipeline
  //
  // loads have an extra cycle of latency, but are fully pipelined
  // use FU_Issue to enforce the 1 load/store per cycle limit
  //
  // Immediate offset
  InstrItinData<IIC_iLoadi   , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 1]>,
  //
  // Register offset
  InstrItinData<IIC_iLoadr   , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 1, 1]>,
  //
  // Scaled register offset, issues over 2 cycles
  InstrItinData<IIC_iLoadsi  , [InstrStage<2, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0], 0>,
                                InstrStage<1, [FU_Pipe1]>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [4, 1, 1]>,
  //
  // Immediate offset with update
  InstrItinData<IIC_iLoadiu  , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 2, 1]>,
  //
  // Register offset with update
  InstrItinData<IIC_iLoadru  , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 2, 1, 1]>,
  //
  // Scaled register offset with update, issues over 2 cycles
  InstrItinData<IIC_iLoadsiu , [InstrStage<2, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0], 0>,
                                InstrStage<1, [FU_Pipe1]>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [4, 3, 1, 1]>,
  //
  // Load multiple
  InstrItinData<IIC_iLoadm   , [InstrStage<2, [FU_Issue], 0>,
                                InstrStage<2, [FU_Pipe0], 0>,
                                InstrStage<2, [FU_Pipe1]>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>]>,

  // Integer store pipeline
  //
  // use FU_Issue to enforce the 1 load/store per cycle limit
  //
  // Immediate offset
  InstrItinData<IIC_iStorei  , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 1]>,
  //
  // Register offset
  InstrItinData<IIC_iStorer  , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 1, 1]>,
  //
  // Scaled register offset, issues over 2 cycles
  InstrItinData<IIC_iStoresi , [InstrStage<2, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0], 0>,
                                InstrStage<1, [FU_Pipe1]>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 1, 1]>,
  //
  // Immediate offset with update
  InstrItinData<IIC_iStoreiu , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [2, 3, 1]>,
  //
  // Register offset with update
  InstrItinData<IIC_iStoreru  , [InstrStage<1, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [2, 3, 1, 1]>,
  //
  // Scaled register offset with update, issues over 2 cycles
  InstrItinData<IIC_iStoresiu, [InstrStage<2, [FU_Issue], 0>,
                                InstrStage<1, [FU_Pipe0], 0>,
                                InstrStage<1, [FU_Pipe1]>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>], [3, 3, 1, 1]>,
  //
  // Store multiple
  InstrItinData<IIC_iStorem  , [InstrStage<2, [FU_Issue], 0>,
                                InstrStage<2, [FU_Pipe0], 0>,
                                InstrStage<2, [FU_Pipe1]>,
                                InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                                InstrStage<1, [FU_LdSt0]>]>,
  
  // Branch
  //
  // no delay slots, so the latency of a branch is unimportant
  InstrItinData<IIC_Br      , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>]>,

  // VFP
  // Issue through integer pipeline, and execute in NEON unit. We assume
  // RunFast mode so that NFP pipeline is used for single-precision when
  // possible.
  //
  // FP Special Register to Integer Register File Move
  InstrItinData<IIC_fpSTAT , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                              InstrStage<1, [FU_NLSPipe]>]>,
  //
  // Single-precision FP Unary
  InstrItinData<IIC_fpUNA32 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [7, 1]>,
  //
  // Double-precision FP Unary
  InstrItinData<IIC_fpUNA64 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<4, [FU_NPipe], 0>,
                               InstrStage<4, [FU_NLSPipe]>], [4, 1]>,
  //
  // Single-precision FP Compare
  InstrItinData<IIC_fpCMP32 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [1, 1]>,
  //
  // Double-precision FP Compare
  InstrItinData<IIC_fpCMP64 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<4, [FU_NPipe], 0>,
                               InstrStage<4, [FU_NLSPipe]>], [4, 1]>,
  //
  // Single to Double FP Convert
  InstrItinData<IIC_fpCVTSD , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<7, [FU_NPipe], 0>,
                               InstrStage<7, [FU_NLSPipe]>], [7, 1]>,
  //
  // Double to Single FP Convert
  InstrItinData<IIC_fpCVTDS , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<5, [FU_NPipe], 0>,
                               InstrStage<5, [FU_NLSPipe]>], [5, 1]>,
  //
  // Single-Precision FP to Integer Convert
  InstrItinData<IIC_fpCVTSI , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [7, 1]>,
  //
  // Double-Precision FP to Integer Convert
  InstrItinData<IIC_fpCVTDI , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<8, [FU_NPipe], 0>,
                               InstrStage<8, [FU_NLSPipe]>], [8, 1]>,
  //
  // Integer to Single-Precision FP Convert
  InstrItinData<IIC_fpCVTIS , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [7, 1]>,
  //
  // Integer to Double-Precision FP Convert
  InstrItinData<IIC_fpCVTID , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<8, [FU_NPipe], 0>,
                               InstrStage<8, [FU_NLSPipe]>], [8, 1]>,
  //
  // Single-precision FP ALU
  InstrItinData<IIC_fpALU32 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [7, 1, 1]>,
  //
  // Double-precision FP ALU
  InstrItinData<IIC_fpALU64 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<9, [FU_NPipe], 0>,
                               InstrStage<9, [FU_NLSPipe]>], [9, 1, 1]>,
  //
  // Single-precision FP Multiply
  InstrItinData<IIC_fpMUL32 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [7, 1, 1]>,
  //
  // Double-precision FP Multiply
  InstrItinData<IIC_fpMUL64 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<11, [FU_NPipe], 0>,
                               InstrStage<11, [FU_NLSPipe]>], [11, 1, 1]>,
  //
  // Single-precision FP MAC
  InstrItinData<IIC_fpMAC32 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [7, 2, 1, 1]>,
  //
  // Double-precision FP MAC
  InstrItinData<IIC_fpMAC64 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<19, [FU_NPipe], 0>,
                               InstrStage<19, [FU_NLSPipe]>], [19, 2, 1, 1]>,
  //
  // Single-precision FP DIV
  InstrItinData<IIC_fpDIV32 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<20, [FU_NPipe], 0>,
                               InstrStage<20, [FU_NLSPipe]>], [20, 1, 1]>,
  //
  // Double-precision FP DIV
  InstrItinData<IIC_fpDIV64 , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<29, [FU_NPipe], 0>,
                               InstrStage<29, [FU_NLSPipe]>], [29, 1, 1]>,
  //
  // Single-precision FP SQRT
  InstrItinData<IIC_fpSQRT32, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<19, [FU_NPipe], 0>,
                               InstrStage<19, [FU_NLSPipe]>], [19, 1]>,
  //
  // Double-precision FP SQRT
  InstrItinData<IIC_fpSQRT64, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<29, [FU_NPipe], 0>,
                               InstrStage<29, [FU_NLSPipe]>], [29, 1]>,
  //
  // Single-precision FP Load
  // use FU_Issue to enforce the 1 load/store per cycle limit
  InstrItinData<IIC_fpLoad32, [InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // Double-precision FP Load
  // use FU_Issue to enforce the 1 load/store per cycle limit
  InstrItinData<IIC_fpLoad64, [InstrStage<2, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0], 0>,
                               InstrStage<1, [FU_Pipe1]>,
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // FP Load Multiple
  // use FU_Issue to enforce the 1 load/store per cycle limit
  InstrItinData<IIC_fpLoadm,  [InstrStage<3, [FU_Issue], 0>, 
                               InstrStage<2, [FU_Pipe0], 0>,
                               InstrStage<2, [FU_Pipe1]>,
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // Single-precision FP Store
  // use FU_Issue to enforce the 1 load/store per cycle limit
  InstrItinData<IIC_fpStore32,[InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // Double-precision FP Store
  // use FU_Issue to enforce the 1 load/store per cycle limit
  InstrItinData<IIC_fpStore64,[InstrStage<2, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0], 0>,
                               InstrStage<1, [FU_Pipe1]>,
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // FP Store Multiple
  // use FU_Issue to enforce the 1 load/store per cycle limit
  InstrItinData<IIC_fpStorem, [InstrStage<3, [FU_Issue], 0>, 
                               InstrStage<2, [FU_Pipe0], 0>,
                               InstrStage<2, [FU_Pipe1]>,
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,

  // NEON
  // Issue through integer pipeline, and execute in NEON unit.
  //
  // VLD1
  InstrItinData<IIC_VLD1,     [InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // VLD2
  InstrItinData<IIC_VLD2,     [InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 2, 1]>,
  //
  // VLD3
  InstrItinData<IIC_VLD3,     [InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 2, 2, 1]>,
  //
  // VLD4
  InstrItinData<IIC_VLD4,     [InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 2, 2, 2, 1]>,
  //
  // VST
  InstrItinData<IIC_VST,      [InstrStage<1, [FU_Issue], 0>, 
                               InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_LdSt0], 0>,
                               InstrStage<1, [FU_NLSPipe]>]>,
  //
  // Double-register FP Unary
  InstrItinData<IIC_VUNAD,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [5, 2]>,
  //
  // Quad-register FP Unary
  // Result written in N5, but that is relative to the last cycle of multicycle,
  // so we use 6 for those cases
  InstrItinData<IIC_VUNAQ,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [6, 2]>,
  //
  // Double-register FP Binary
  InstrItinData<IIC_VBIND,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [5, 2, 2]>,
  //
  // Quad-register FP Binary
  // Result written in N5, but that is relative to the last cycle of multicycle,
  // so we use 6 for those cases
  InstrItinData<IIC_VBINQ,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [6, 2, 2]>,
  //
  // Move Immediate
  InstrItinData<IIC_VMOVImm,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3]>,
  //
  // Double-register Permute Move
  InstrItinData<IIC_VMOVD,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 1]>,
  //
  // Quad-register Permute Move
  // Result written in N2, but that is relative to the last cycle of multicycle,
  // so we use 3 for those cases
  InstrItinData<IIC_VMOVQ,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 1]>,
  //
  // Integer to Single-precision Move
  InstrItinData<IIC_VMOVIS ,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 1]>,
  //
  // Integer to Double-precision Move
  InstrItinData<IIC_VMOVID ,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 1, 1]>,
  //
  // Single-precision to Integer Move
  InstrItinData<IIC_VMOVSI ,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [20, 1]>,
  //
  // Double-precision to Integer Move
  InstrItinData<IIC_VMOVDI ,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [20, 20, 1]>,
  //
  // Integer to Lane Move
  InstrItinData<IIC_VMOVISL , [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 1, 1]>,
  //
  // Double-register Permute
  InstrItinData<IIC_VPERMD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 2, 1, 1]>,
  //
  // Quad-register Permute
  // Result written in N2, but that is relative to the last cycle of multicycle,
  // so we use 3 for those cases
  InstrItinData<IIC_VPERMQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 3, 1, 1]>,
  //
  // Quad-register Permute (3 cycle issue)
  // Result written in N2, but that is relative to the last cycle of multicycle,
  // so we use 4 for those cases
  InstrItinData<IIC_VPERMQ3,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>,
                               InstrStage<1, [FU_NPipe], 0>,
                               InstrStage<2, [FU_NLSPipe]>], [4, 4, 1, 1]>,
  //
  // Double-register FP Multiple-Accumulate
  InstrItinData<IIC_VMACD,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [9, 2, 2, 3]>,
  //
  // Quad-register FP Multiple-Accumulate
  // Result written in N9, but that is relative to the last cycle of multicycle,
  // so we use 10 for those cases
  InstrItinData<IIC_VMACQ,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [10, 2, 2, 3]>,
  //
  // Double-register Reciprical Step
  InstrItinData<IIC_VRECSD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [9, 2, 2]>,
  //
  // Quad-register Reciprical Step
  InstrItinData<IIC_VRECSQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [10, 2, 2]>,
  //
  // Double-register Integer Count
  InstrItinData<IIC_VCNTiD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3, 2, 2]>,
  //
  // Quad-register Integer Count
  // Result written in N3, but that is relative to the last cycle of multicycle,
  // so we use 4 for those cases
  InstrItinData<IIC_VCNTiQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [4, 2, 2]>,
  //
  // Double-register Integer Unary
  InstrItinData<IIC_VUNAiD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 2]>,
  //
  // Quad-register Integer Unary
  InstrItinData<IIC_VUNAiQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 2]>,
  //
  // Double-register Integer Q-Unary
  InstrItinData<IIC_VQUNAiD,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 1]>,
  //
  // Quad-register Integer CountQ-Unary
  InstrItinData<IIC_VQUNAiQ,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 1]>,
  //
  // Double-register Integer Binary
  InstrItinData<IIC_VBINiD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3, 2, 2]>,
  //
  // Quad-register Integer Binary
  InstrItinData<IIC_VBINiQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3, 2, 2]>,
  //
  // Double-register Integer Binary (4 cycle)
  InstrItinData<IIC_VBINi4D,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 2, 1]>,
  //
  // Quad-register Integer Binary (4 cycle)
  InstrItinData<IIC_VBINi4Q,  [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 2, 1]>,
  //
  // Double-register Integer Subtract
  InstrItinData<IIC_VSUBiD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3, 2, 1]>,
  //
  // Quad-register Integer Subtract
  InstrItinData<IIC_VSUBiQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3, 2, 1]>,
  //
  // Double-register Integer Shift
  InstrItinData<IIC_VSHLiD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [3, 1, 1]>,
  //
  // Quad-register Integer Shift
  InstrItinData<IIC_VSHLiQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [4, 1, 1]>,
  //
  // Double-register Integer Shift (4 cycle)
  InstrItinData<IIC_VSHLi4D,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [4, 1, 1]>,
  //
  // Quad-register Integer Shift (4 cycle)
  InstrItinData<IIC_VSHLi4Q,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [5, 1, 1]>,
  //
  // Double-register Integer Pair Add Long
  InstrItinData<IIC_VPALiD,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [6, 3, 2, 1]>,
  //
  // Quad-register Integer Pair Add Long
  InstrItinData<IIC_VPALiQ,   [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [7, 3, 2, 1]>,
  //
  // Double-register Integer Multiply (.8, .16)
  InstrItinData<IIC_VMULi16D, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [6, 2, 2]>,
  //
  // Double-register Integer Multiply (.32)
  InstrItinData<IIC_VMULi32D, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [7, 2, 1]>,
  //
  // Quad-register Integer Multiply (.8, .16)
  InstrItinData<IIC_VMULi16Q, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [7, 2, 2]>,
  //
  // Quad-register Integer Multiply (.32)
  InstrItinData<IIC_VMULi32Q, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>,
                               InstrStage<2, [FU_NLSPipe], 0>,
                               InstrStage<3, [FU_NPipe]>], [9, 2, 1]>,
  //
  // Double-register Integer Multiply-Accumulate (.8, .16)
  InstrItinData<IIC_VMACi16D, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>], [6, 2, 2, 3]>,
  //
  // Double-register Integer Multiply-Accumulate (.32)
  InstrItinData<IIC_VMACi32D, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [7, 2, 1, 3]>,
  //
  // Quad-register Integer Multiply-Accumulate (.8, .16)
  InstrItinData<IIC_VMACi16Q, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NPipe]>], [7, 2, 2, 3]>,
  //
  // Quad-register Integer Multiply-Accumulate (.32)
  InstrItinData<IIC_VMACi32Q, [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NPipe]>,
                               InstrStage<2, [FU_NLSPipe], 0>,
                               InstrStage<3, [FU_NPipe]>], [9, 2, 1, 3]>,
  //
  // Double-register VEXT
  InstrItinData<IIC_VEXTD,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>], [2, 1, 1]>,
  //
  // Quad-register VEXT
  InstrItinData<IIC_VEXTQ,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 1, 1]>,
  //
  // VTB
  InstrItinData<IIC_VTB1,     [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 2, 1]>,
  InstrItinData<IIC_VTB2,     [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 2, 2, 1]>,
  InstrItinData<IIC_VTB3,     [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>,
                               InstrStage<1, [FU_NPipe], 0>,
                               InstrStage<2, [FU_NLSPipe]>], [4, 2, 2, 3, 1]>,
  InstrItinData<IIC_VTB4,     [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>,
                               InstrStage<1, [FU_NPipe], 0>,
                               InstrStage<2, [FU_NLSPipe]>], [4, 2, 2, 3, 3, 1]>,
  //
  // VTBX
  InstrItinData<IIC_VTBX1,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 1, 2, 1]>,
  InstrItinData<IIC_VTBX2,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<2, [FU_NLSPipe]>], [3, 1, 2, 2, 1]>,
  InstrItinData<IIC_VTBX3,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>,
                               InstrStage<1, [FU_NPipe], 0>,
                               InstrStage<2, [FU_NLSPipe]>], [4, 1, 2, 2, 3, 1]>,
  InstrItinData<IIC_VTBX4,    [InstrStage<1, [FU_Pipe0, FU_Pipe1]>,
                               InstrStage<1, [FU_NLSPipe]>,
                               InstrStage<1, [FU_NPipe], 0>,
                               InstrStage<2, [FU_NLSPipe]>], [4, 1, 2, 2, 3, 3, 1]>
]>;