llvm.org GIT mirror llvm / 71589ff lib / Target / X86 / X86Schedule.td
71589ff

Tree @71589ff (Download .tar.gz)

X86Schedule.td @71589ffraw · 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
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
//===-- X86Schedule.td - X86 Scheduling Definitions --------*- tablegen -*-===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// InstrSchedModel annotations for out-of-order CPUs.

// Instructions with folded loads need to read the memory operand immediately,
// but other register operands don't have to be read until the load is ready.
// These operands are marked with ReadAfterLd.
def ReadAfterLd : SchedRead;

// Instructions with both a load and a store folded are modeled as a folded
// load + WriteRMW.
def WriteRMW : SchedWrite;

// Helper to set SchedWrite ExePorts/Latency/ResourceCycles/NumMicroOps.
multiclass X86WriteRes<SchedWrite SchedRW,
                       list<ProcResourceKind> ExePorts,
                       int Lat, list<int> Res, int UOps> {
  def : WriteRes<SchedRW, ExePorts> {
    let Latency = Lat;
    let ResourceCycles = Res;
    let NumMicroOps = UOps;
  }
}

// Most instructions can fold loads, so almost every SchedWrite comes in two
// variants: With and without a folded load.
// An X86FoldableSchedWrite holds a reference to the corresponding SchedWrite
// with a folded load.
class X86FoldableSchedWrite : SchedWrite {
  // The SchedWrite to use when a load is folded into the instruction.
  SchedWrite Folded;
}

// Multiclass that produces a linked pair of SchedWrites.
multiclass X86SchedWritePair {
  // Register-Memory operation.
  def Ld : SchedWrite;
  // Register-Register operation.
  def NAME : X86FoldableSchedWrite {
    let Folded = !cast<SchedWrite>(NAME#"Ld");
  }
}

// Helpers to mark SchedWrites as unsupported.
multiclass X86WriteResUnsupported<SchedWrite SchedRW> {
  let Unsupported = 1 in {
    def : WriteRes<SchedRW, []>;
  }
}
multiclass X86WriteResPairUnsupported<X86FoldableSchedWrite SchedRW> {
  let Unsupported = 1 in {
    def : WriteRes<SchedRW, []>;
    def : WriteRes<SchedRW.Folded, []>;
  }
}

// Multiclass that wraps X86FoldableSchedWrite for each vector width.
class X86SchedWriteWidths<X86FoldableSchedWrite sScl,
                          X86FoldableSchedWrite s128,
                          X86FoldableSchedWrite s256,
                          X86FoldableSchedWrite s512> {
  X86FoldableSchedWrite Scl = sScl; // Scalar float/double operations.
  X86FoldableSchedWrite MMX = sScl; // MMX operations.
  X86FoldableSchedWrite XMM = s128; // XMM operations.
  X86FoldableSchedWrite YMM = s256; // YMM operations.
  X86FoldableSchedWrite ZMM = s512; // ZMM operations.
}

// Multiclass that wraps X86SchedWriteWidths for each fp vector type.
class X86SchedWriteSizes<X86SchedWriteWidths sPS,
                         X86SchedWriteWidths sPD> {
  X86SchedWriteWidths PS = sPS;
  X86SchedWriteWidths PD = sPD;
}

// Multiclass that wraps move/load/store triple for a vector width.
class X86SchedWriteMoveLS<SchedWrite MoveRR,
                          SchedWrite LoadRM,
                          SchedWrite StoreMR> {
  SchedWrite RR = MoveRR;
  SchedWrite RM = LoadRM;
  SchedWrite MR = StoreMR;
}

// Multiclass that wraps X86SchedWriteMoveLS for each vector width.
class X86SchedWriteMoveLSWidths<X86SchedWriteMoveLS sScl,
                                X86SchedWriteMoveLS s128,
                                X86SchedWriteMoveLS s256,
                                X86SchedWriteMoveLS s512> {
  X86SchedWriteMoveLS Scl = sScl; // Scalar float/double operations.
  X86SchedWriteMoveLS MMX = sScl; // MMX operations.
  X86SchedWriteMoveLS XMM = s128; // XMM operations.
  X86SchedWriteMoveLS YMM = s256; // YMM operations.
  X86SchedWriteMoveLS ZMM = s512; // ZMM operations.
}

// Loads, stores, and moves, not folded with other operations.
def WriteLoad    : SchedWrite;
def WriteStore   : SchedWrite;
def WriteStoreNT : SchedWrite;
def WriteMove    : SchedWrite;

// Arithmetic.
defm WriteALU    : X86SchedWritePair; // Simple integer ALU op.
defm WriteADC    : X86SchedWritePair; // Integer ALU + flags op.
def  WriteALURMW : WriteSequence<[WriteALULd, WriteStore]>;
def  WriteADCRMW : WriteSequence<[WriteADCLd, WriteStore]>;
defm WriteIMul   : X86SchedWritePair; // Integer multiplication.
defm WriteIMul64 : X86SchedWritePair; // Integer 64-bit multiplication.
def  WriteIMulH  : SchedWrite;        // Integer multiplication, high part.
def  WriteLEA    : SchedWrite;        // LEA instructions can't fold loads.

def  WriteBSWAP32 : SchedWrite; // Byte Order (Endianness) 32-bit Swap.
def  WriteBSWAP64 : SchedWrite; // Byte Order (Endianness) 64-bit Swap.
defm WriteCMPXCHG : X86SchedWritePair; // Compare and set, compare and swap.
def  WriteCMPXCHGRMW : SchedWrite;     // Compare and set, compare and swap.
def  WriteXCHG    : SchedWrite;        // Compare+Exchange - TODO RMW support.

// Integer division.
defm WriteDiv8   : X86SchedWritePair;
defm WriteDiv16  : X86SchedWritePair;
defm WriteDiv32  : X86SchedWritePair;
defm WriteDiv64  : X86SchedWritePair;
defm WriteIDiv8  : X86SchedWritePair;
defm WriteIDiv16 : X86SchedWritePair;
defm WriteIDiv32 : X86SchedWritePair;
defm WriteIDiv64 : X86SchedWritePair;

defm WriteBSF : X86SchedWritePair; // Bit scan forward.
defm WriteBSR : X86SchedWritePair; // Bit scan reverse.
defm WritePOPCNT : X86SchedWritePair; // Bit population count.
defm WriteLZCNT : X86SchedWritePair; // Leading zero count.
defm WriteTZCNT : X86SchedWritePair; // Trailing zero count.
defm WriteCMOV  : X86SchedWritePair; // Conditional move.
defm WriteCMOV2 : X86SchedWritePair; // Conditional (CF + ZF flag) move.
def  WriteFCMOV : SchedWrite; // X87 conditional move.
def  WriteSETCC : SchedWrite; // Set register based on condition code.
def  WriteSETCCStore : SchedWrite;
def  WriteLAHFSAHF : SchedWrite; // Load/Store flags in AH.
def  WriteBitTest  : SchedWrite; // Bit Test - TODO add memory folding support

// Integer shifts and rotates.
defm WriteShift : X86SchedWritePair;

// Double shift instructions.
def  WriteSHDrri  : SchedWrite;
def  WriteSHDrrcl : SchedWrite;
def  WriteSHDmri  : SchedWrite;
def  WriteSHDmrcl : SchedWrite;

// BMI1 BEXTR/BLS, BMI2 BZHI
defm WriteBEXTR : X86SchedWritePair;
defm WriteBLS   : X86SchedWritePair;
defm WriteBZHI  : X86SchedWritePair;

// Idioms that clear a register, like xorps %xmm0, %xmm0.
// These can often bypass execution ports completely.
def WriteZero : SchedWrite;

// Branches don't produce values, so they have no latency, but they still
// consume resources. Indirect branches can fold loads.
defm WriteJump : X86SchedWritePair;

// Floating point. This covers both scalar and vector operations.
def  WriteFLD0          : SchedWrite;
def  WriteFLD1          : SchedWrite;
def  WriteFLDC          : SchedWrite;
def  WriteFLoad         : SchedWrite;
def  WriteFLoadX        : SchedWrite;
def  WriteFLoadY        : SchedWrite;
def  WriteFMaskedLoad   : SchedWrite;
def  WriteFMaskedLoadY  : SchedWrite;
def  WriteFStore        : SchedWrite;
def  WriteFStoreX       : SchedWrite;
def  WriteFStoreY       : SchedWrite;
def  WriteFStoreNT      : SchedWrite;
def  WriteFStoreNTX     : SchedWrite;
def  WriteFStoreNTY     : SchedWrite;
def  WriteFMaskedStore  : SchedWrite;
def  WriteFMaskedStoreY : SchedWrite;
def  WriteFMove         : SchedWrite;
def  WriteFMoveX        : SchedWrite;
def  WriteFMoveY        : SchedWrite;

defm WriteFAdd    : X86SchedWritePair; // Floating point add/sub.
defm WriteFAddX   : X86SchedWritePair; // Floating point add/sub (XMM).
defm WriteFAddY   : X86SchedWritePair; // Floating point add/sub (YMM).
defm WriteFAddZ   : X86SchedWritePair; // Floating point add/sub (ZMM).
defm WriteFAdd64  : X86SchedWritePair; // Floating point double add/sub.
defm WriteFAdd64X : X86SchedWritePair; // Floating point double add/sub (XMM).
defm WriteFAdd64Y : X86SchedWritePair; // Floating point double add/sub (YMM).
defm WriteFAdd64Z : X86SchedWritePair; // Floating point double add/sub (ZMM).
defm WriteFCmp    : X86SchedWritePair; // Floating point compare.
defm WriteFCmpX   : X86SchedWritePair; // Floating point compare (XMM).
defm WriteFCmpY   : X86SchedWritePair; // Floating point compare (YMM).
defm WriteFCmpZ   : X86SchedWritePair; // Floating point compare (ZMM).
defm WriteFCmp64  : X86SchedWritePair; // Floating point double compare.
defm WriteFCmp64X : X86SchedWritePair; // Floating point double compare (XMM).
defm WriteFCmp64Y : X86SchedWritePair; // Floating point double compare (YMM).
defm WriteFCmp64Z : X86SchedWritePair; // Floating point double compare (ZMM).
defm WriteFCom    : X86SchedWritePair; // Floating point compare to flags.
defm WriteFMul    : X86SchedWritePair; // Floating point multiplication.
defm WriteFMulX   : X86SchedWritePair; // Floating point multiplication (XMM).
defm WriteFMulY   : X86SchedWritePair; // Floating point multiplication (YMM).
defm WriteFMulZ   : X86SchedWritePair; // Floating point multiplication (YMM).
defm WriteFMul64  : X86SchedWritePair; // Floating point double multiplication.
defm WriteFMul64X : X86SchedWritePair; // Floating point double multiplication (XMM).
defm WriteFMul64Y : X86SchedWritePair; // Floating point double multiplication (YMM).
defm WriteFMul64Z : X86SchedWritePair; // Floating point double multiplication (ZMM).
defm WriteFDiv    : X86SchedWritePair; // Floating point division.
defm WriteFDivX   : X86SchedWritePair; // Floating point division (XMM).
defm WriteFDivY   : X86SchedWritePair; // Floating point division (YMM).
defm WriteFDivZ   : X86SchedWritePair; // Floating point division (ZMM).
defm WriteFDiv64  : X86SchedWritePair; // Floating point double division.
defm WriteFDiv64X : X86SchedWritePair; // Floating point double division (XMM).
defm WriteFDiv64Y : X86SchedWritePair; // Floating point double division (YMM).
defm WriteFDiv64Z : X86SchedWritePair; // Floating point double division (ZMM).
defm WriteFSqrt  : X86SchedWritePair; // Floating point square root.
defm WriteFSqrtX : X86SchedWritePair; // Floating point square root (XMM).
defm WriteFSqrtY : X86SchedWritePair; // Floating point square root (YMM).
defm WriteFSqrtZ : X86SchedWritePair; // Floating point square root (ZMM).
defm WriteFSqrt64  : X86SchedWritePair; // Floating point double square root.
defm WriteFSqrt64X : X86SchedWritePair; // Floating point double square root (XMM).
defm WriteFSqrt64Y : X86SchedWritePair; // Floating point double square root (YMM).
defm WriteFSqrt64Z : X86SchedWritePair; // Floating point double square root (ZMM).
defm WriteFSqrt80  : X86SchedWritePair; // Floating point long double square root.
defm WriteFRcp   : X86SchedWritePair; // Floating point reciprocal estimate.
defm WriteFRcpX  : X86SchedWritePair; // Floating point reciprocal estimate (XMM).
defm WriteFRcpY  : X86SchedWritePair; // Floating point reciprocal estimate (YMM).
defm WriteFRcpZ  : X86SchedWritePair; // Floating point reciprocal estimate (ZMM).
defm WriteFRsqrt : X86SchedWritePair; // Floating point reciprocal square root estimate.
defm WriteFRsqrtX: X86SchedWritePair; // Floating point reciprocal square root estimate (XMM).
defm WriteFRsqrtY: X86SchedWritePair; // Floating point reciprocal square root estimate (YMM).
defm WriteFRsqrtZ: X86SchedWritePair; // Floating point reciprocal square root estimate (ZMM).
defm WriteFMA    : X86SchedWritePair; // Fused Multiply Add.
defm WriteFMAX   : X86SchedWritePair; // Fused Multiply Add (XMM).
defm WriteFMAY   : X86SchedWritePair; // Fused Multiply Add (YMM).
defm WriteFMAZ   : X86SchedWritePair; // Fused Multiply Add (ZMM).
defm WriteDPPD   : X86SchedWritePair; // Floating point double dot product.
defm WriteDPPS   : X86SchedWritePair; // Floating point single dot product.
defm WriteDPPSY  : X86SchedWritePair; // Floating point single dot product (YMM).
defm WriteDPPSZ  : X86SchedWritePair; // Floating point single dot product (ZMM).
defm WriteFSign  : X86SchedWritePair; // Floating point fabs/fchs.
defm WriteFRnd   : X86SchedWritePair; // Floating point rounding.
defm WriteFRndY  : X86SchedWritePair; // Floating point rounding (YMM).
defm WriteFRndZ  : X86SchedWritePair; // Floating point rounding (ZMM).
defm WriteFLogic  : X86SchedWritePair; // Floating point and/or/xor logicals.
defm WriteFLogicY : X86SchedWritePair; // Floating point and/or/xor logicals (YMM).
defm WriteFLogicZ : X86SchedWritePair; // Floating point and/or/xor logicals (ZMM).
defm WriteFTest   : X86SchedWritePair; // Floating point TEST instructions.
defm WriteFTestY  : X86SchedWritePair; // Floating point TEST instructions (YMM).
defm WriteFTestZ  : X86SchedWritePair; // Floating point TEST instructions (ZMM).
defm WriteFShuffle  : X86SchedWritePair; // Floating point vector shuffles.
defm WriteFShuffleY : X86SchedWritePair; // Floating point vector shuffles (YMM).
defm WriteFShuffleZ : X86SchedWritePair; // Floating point vector shuffles (ZMM).
defm WriteFVarShuffle  : X86SchedWritePair; // Floating point vector variable shuffles.
defm WriteFVarShuffleY : X86SchedWritePair; // Floating point vector variable shuffles (YMM).
defm WriteFVarShuffleZ : X86SchedWritePair; // Floating point vector variable shuffles (ZMM).
defm WriteFBlend  : X86SchedWritePair; // Floating point vector blends.
defm WriteFBlendY : X86SchedWritePair; // Floating point vector blends (YMM).
defm WriteFBlendZ : X86SchedWritePair; // Floating point vector blends (ZMM).
defm WriteFVarBlend  : X86SchedWritePair; // Fp vector variable blends.
defm WriteFVarBlendY : X86SchedWritePair; // Fp vector variable blends (YMM).
defm WriteFVarBlendZ : X86SchedWritePair; // Fp vector variable blends (YMZMM).

// FMA Scheduling helper class.
class FMASC { X86FoldableSchedWrite Sched = WriteFAdd; }

// Horizontal Add/Sub (float and integer)
defm WriteFHAdd  : X86SchedWritePair;
defm WriteFHAddY : X86SchedWritePair;
defm WriteFHAddZ : X86SchedWritePair;
defm WritePHAdd  : X86SchedWritePair;
defm WritePHAddX : X86SchedWritePair;
defm WritePHAddY : X86SchedWritePair;
defm WritePHAddZ : X86SchedWritePair;

// Vector integer operations.
def  WriteVecLoad         : SchedWrite;
def  WriteVecLoadX        : SchedWrite;
def  WriteVecLoadY        : SchedWrite;
def  WriteVecLoadNT       : SchedWrite;
def  WriteVecLoadNTY      : SchedWrite;
def  WriteVecMaskedLoad   : SchedWrite;
def  WriteVecMaskedLoadY  : SchedWrite;
def  WriteVecStore        : SchedWrite;
def  WriteVecStoreX       : SchedWrite;
def  WriteVecStoreY       : SchedWrite;
def  WriteVecStoreNT      : SchedWrite;
def  WriteVecStoreNTY     : SchedWrite;
def  WriteVecMaskedStore  : SchedWrite;
def  WriteVecMaskedStoreY : SchedWrite;
def  WriteVecMove         : SchedWrite;
def  WriteVecMoveX        : SchedWrite;
def  WriteVecMoveY        : SchedWrite;
def  WriteVecMoveToGpr    : SchedWrite;
def  WriteVecMoveFromGpr  : SchedWrite;

defm WriteVecALU    : X86SchedWritePair; // Vector integer ALU op, no logicals.
defm WriteVecALUX   : X86SchedWritePair; // Vector integer ALU op, no logicals (XMM).
defm WriteVecALUY   : X86SchedWritePair; // Vector integer ALU op, no logicals (YMM).
defm WriteVecALUZ   : X86SchedWritePair; // Vector integer ALU op, no logicals (ZMM).
defm WriteVecLogic  : X86SchedWritePair; // Vector integer and/or/xor logicals.
defm WriteVecLogicX : X86SchedWritePair; // Vector integer and/or/xor logicals (XMM).
defm WriteVecLogicY : X86SchedWritePair; // Vector integer and/or/xor logicals (YMM).
defm WriteVecLogicZ : X86SchedWritePair; // Vector integer and/or/xor logicals (ZMM).
defm WriteVecTest  : X86SchedWritePair; // Vector integer TEST instructions.
defm WriteVecTestY : X86SchedWritePair; // Vector integer TEST instructions (YMM).
defm WriteVecTestZ : X86SchedWritePair; // Vector integer TEST instructions (ZMM).
defm WriteVecShift  : X86SchedWritePair; // Vector integer shifts (default).
defm WriteVecShiftX : X86SchedWritePair; // Vector integer shifts (XMM).
defm WriteVecShiftY : X86SchedWritePair; // Vector integer shifts (YMM).
defm WriteVecShiftZ : X86SchedWritePair; // Vector integer shifts (ZMM).
defm WriteVecShiftImm : X86SchedWritePair; // Vector integer immediate shifts (default).
defm WriteVecShiftImmX: X86SchedWritePair; // Vector integer immediate shifts (XMM).
defm WriteVecShiftImmY: X86SchedWritePair; // Vector integer immediate shifts (YMM).
defm WriteVecShiftImmZ: X86SchedWritePair; // Vector integer immediate shifts (ZMM).
defm WriteVecIMul  : X86SchedWritePair; // Vector integer multiply (default).
defm WriteVecIMulX : X86SchedWritePair; // Vector integer multiply (XMM).
defm WriteVecIMulY : X86SchedWritePair; // Vector integer multiply (YMM).
defm WriteVecIMulZ : X86SchedWritePair; // Vector integer multiply (ZMM).
defm WritePMULLD   : X86SchedWritePair; // Vector PMULLD.
defm WritePMULLDY   : X86SchedWritePair; // Vector PMULLD (YMM).
defm WritePMULLDZ   : X86SchedWritePair; // Vector PMULLD (ZMM).
defm WriteShuffle  : X86SchedWritePair; // Vector shuffles.
defm WriteShuffleX : X86SchedWritePair; // Vector shuffles (XMM).
defm WriteShuffleY : X86SchedWritePair; // Vector shuffles (YMM).
defm WriteShuffleZ : X86SchedWritePair; // Vector shuffles (ZMM).
defm WriteVarShuffle  : X86SchedWritePair; // Vector variable shuffles.
defm WriteVarShuffleX : X86SchedWritePair; // Vector variable shuffles (XMM).
defm WriteVarShuffleY : X86SchedWritePair; // Vector variable shuffles (YMM).
defm WriteVarShuffleZ : X86SchedWritePair; // Vector variable shuffles (ZMM).
defm WriteBlend  : X86SchedWritePair; // Vector blends.
defm WriteBlendY : X86SchedWritePair; // Vector blends (YMM).
defm WriteBlendZ : X86SchedWritePair; // Vector blends (ZMM).
defm WriteVarBlend  : X86SchedWritePair; // Vector variable blends.
defm WriteVarBlendY : X86SchedWritePair; // Vector variable blends (YMM).
defm WriteVarBlendZ : X86SchedWritePair; // Vector variable blends (ZMM).
defm WritePSADBW  : X86SchedWritePair; // Vector PSADBW.
defm WritePSADBWX : X86SchedWritePair; // Vector PSADBW (XMM).
defm WritePSADBWY : X86SchedWritePair; // Vector PSADBW (YMM).
defm WritePSADBWZ : X86SchedWritePair; // Vector PSADBW (ZMM).
defm WriteMPSAD  : X86SchedWritePair; // Vector MPSAD.
defm WriteMPSADY : X86SchedWritePair; // Vector MPSAD (YMM).
defm WriteMPSADZ : X86SchedWritePair; // Vector MPSAD (ZMM).
defm WritePHMINPOS : X86SchedWritePair; // Vector PHMINPOS.

// Vector insert/extract operations.
defm WriteVecInsert : X86SchedWritePair; // Insert gpr to vector element.
def  WriteVecExtract : SchedWrite; // Extract vector element to gpr.
def  WriteVecExtractSt : SchedWrite; // Extract vector element and store.

// MOVMSK operations.
def WriteFMOVMSK    : SchedWrite;
def WriteVecMOVMSK  : SchedWrite;
def WriteVecMOVMSKY : SchedWrite;
def WriteMMXMOVMSK  : SchedWrite;

// Conversion between integer and float.
defm WriteCvtSD2I  : X86SchedWritePair; // Double -> Integer.
defm WriteCvtPD2I  : X86SchedWritePair; // Double -> Integer (XMM).
defm WriteCvtPD2IY : X86SchedWritePair; // Double -> Integer (YMM).
defm WriteCvtPD2IZ : X86SchedWritePair; // Double -> Integer (ZMM).

defm WriteCvtSS2I  : X86SchedWritePair; // Float -> Integer.
defm WriteCvtPS2I  : X86SchedWritePair; // Float -> Integer (XMM).
defm WriteCvtPS2IY : X86SchedWritePair; // Float -> Integer (YMM).
defm WriteCvtPS2IZ : X86SchedWritePair; // Float -> Integer (ZMM).

defm WriteCvtI2SD  : X86SchedWritePair; // Integer -> Double.
defm WriteCvtI2PD  : X86SchedWritePair; // Integer -> Double (XMM).
defm WriteCvtI2PDY : X86SchedWritePair; // Integer -> Double (YMM).
defm WriteCvtI2PDZ : X86SchedWritePair; // Integer -> Double (ZMM).

defm WriteCvtI2SS  : X86SchedWritePair; // Integer -> Float.
defm WriteCvtI2PS  : X86SchedWritePair; // Integer -> Float (XMM).
defm WriteCvtI2PSY : X86SchedWritePair; // Integer -> Float (YMM).
defm WriteCvtI2PSZ : X86SchedWritePair; // Integer -> Float (ZMM).

defm WriteCvtSS2SD  : X86SchedWritePair; // Float -> Double size conversion.
defm WriteCvtPS2PD  : X86SchedWritePair; // Float -> Double size conversion (XMM).
defm WriteCvtPS2PDY : X86SchedWritePair; // Float -> Double size conversion (YMM).
defm WriteCvtPS2PDZ : X86SchedWritePair; // Float -> Double size conversion (ZMM).

defm WriteCvtSD2SS  : X86SchedWritePair; // Double -> Float size conversion.
defm WriteCvtPD2PS  : X86SchedWritePair; // Double -> Float size conversion (XMM).
defm WriteCvtPD2PSY : X86SchedWritePair; // Double -> Float size conversion (YMM).
defm WriteCvtPD2PSZ : X86SchedWritePair; // Double -> Float size conversion (ZMM).

defm WriteCvtPH2PS    : X86SchedWritePair; // Half -> Float size conversion.
defm WriteCvtPH2PSY   : X86SchedWritePair; // Half -> Float size conversion (YMM).
defm WriteCvtPH2PSZ   : X86SchedWritePair; // Half -> Float size conversion (ZMM).

def  WriteCvtPS2PH    : SchedWrite; // // Float -> Half size conversion.
def  WriteCvtPS2PHY   : SchedWrite; // // Float -> Half size conversion (YMM).
def  WriteCvtPS2PHZ   : SchedWrite; // // Float -> Half size conversion (ZMM).
def  WriteCvtPS2PHSt  : SchedWrite; // // Float -> Half + store size conversion.
def  WriteCvtPS2PHYSt : SchedWrite; // // Float -> Half + store size conversion (YMM).
def  WriteCvtPS2PHZSt : SchedWrite; // // Float -> Half + store size conversion (ZMM).

// CRC32 instruction.
defm WriteCRC32 : X86SchedWritePair;

// Strings instructions.
// Packed Compare Implicit Length Strings, Return Mask
defm WritePCmpIStrM : X86SchedWritePair;
// Packed Compare Explicit Length Strings, Return Mask
defm WritePCmpEStrM : X86SchedWritePair;
// Packed Compare Implicit Length Strings, Return Index
defm WritePCmpIStrI : X86SchedWritePair;
// Packed Compare Explicit Length Strings, Return Index
defm WritePCmpEStrI : X86SchedWritePair;

// AES instructions.
defm WriteAESDecEnc : X86SchedWritePair; // Decryption, encryption.
defm WriteAESIMC : X86SchedWritePair; // InvMixColumn.
defm WriteAESKeyGen : X86SchedWritePair; // Key Generation.

// Carry-less multiplication instructions.
defm WriteCLMul : X86SchedWritePair;

// EMMS/FEMMS
def WriteEMMS : SchedWrite;

// Load/store MXCSR
def WriteLDMXCSR : SchedWrite;
def WriteSTMXCSR : SchedWrite;

// Catch-all for expensive system instructions.
def WriteSystem : SchedWrite;

// AVX2.
defm WriteFShuffle256 : X86SchedWritePair; // Fp 256-bit width vector shuffles.
defm WriteFVarShuffle256 : X86SchedWritePair; // Fp 256-bit width variable shuffles.
defm WriteShuffle256 : X86SchedWritePair; // 256-bit width vector shuffles.
defm WriteVarShuffle256 : X86SchedWritePair; // 256-bit width vector variable shuffles.
defm WriteVarVecShift  : X86SchedWritePair; // Variable vector shifts.
defm WriteVarVecShiftY : X86SchedWritePair; // Variable vector shifts (YMM).
defm WriteVarVecShiftZ : X86SchedWritePair; // Variable vector shifts (ZMM).

// Old microcoded instructions that nobody use.
def WriteMicrocoded : SchedWrite;

// Fence instructions.
def WriteFence : SchedWrite;

// Nop, not very useful expect it provides a model for nops!
def WriteNop : SchedWrite;

// Move/Load/Store wrappers.
def WriteFMoveLS
 : X86SchedWriteMoveLS<WriteFMove, WriteFLoad, WriteFStore>;
def WriteFMoveLSX
 : X86SchedWriteMoveLS<WriteFMoveX, WriteFLoadX, WriteFStoreX>;
def WriteFMoveLSY
 : X86SchedWriteMoveLS<WriteFMoveY, WriteFLoadY, WriteFStoreY>;
def SchedWriteFMoveLS
  : X86SchedWriteMoveLSWidths<WriteFMoveLS, WriteFMoveLSX,
                              WriteFMoveLSY, WriteFMoveLSY>;

def WriteFMoveLSNT
 : X86SchedWriteMoveLS<WriteFMove, WriteFLoad, WriteFStoreNT>;
def WriteFMoveLSNTX
 : X86SchedWriteMoveLS<WriteFMove, WriteFLoad, WriteFStoreNTX>;
def WriteFMoveLSNTY
 : X86SchedWriteMoveLS<WriteFMoveY, WriteFLoadY, WriteFStoreNTY>;
def SchedWriteFMoveLSNT
  : X86SchedWriteMoveLSWidths<WriteFMoveLSNT, WriteFMoveLSNTX,
                              WriteFMoveLSNTY, WriteFMoveLSNTY>;

def WriteVecMoveLS
 : X86SchedWriteMoveLS<WriteVecMove, WriteVecLoad, WriteVecStore>;
def WriteVecMoveLSX
 : X86SchedWriteMoveLS<WriteVecMoveX, WriteVecLoadX, WriteVecStoreX>;
def WriteVecMoveLSY
 : X86SchedWriteMoveLS<WriteVecMoveY, WriteVecLoadY, WriteVecStoreY>;
def SchedWriteVecMoveLS
  : X86SchedWriteMoveLSWidths<WriteVecMoveLS, WriteVecMoveLSX,
                              WriteVecMoveLSY, WriteVecMoveLSY>;

def WriteVecMoveLSNT
 : X86SchedWriteMoveLS<WriteVecMove, WriteVecLoadNT, WriteVecStoreNT>;
def WriteVecMoveLSNTX
 : X86SchedWriteMoveLS<WriteVecMoveX, WriteVecLoadNT, WriteVecStoreNT>;
def WriteVecMoveLSNTY
 : X86SchedWriteMoveLS<WriteVecMoveY, WriteVecLoadNTY, WriteVecStoreNTY>;
def SchedWriteVecMoveLSNT
  : X86SchedWriteMoveLSWidths<WriteVecMoveLSNT, WriteVecMoveLSNTX,
                              WriteVecMoveLSNTY, WriteVecMoveLSNTY>;

// Vector width wrappers.
def SchedWriteFAdd
 : X86SchedWriteWidths<WriteFAdd, WriteFAddX, WriteFAddY, WriteFAddZ>;
def SchedWriteFAdd64
 : X86SchedWriteWidths<WriteFAdd64, WriteFAdd64X, WriteFAdd64Y, WriteFAdd64Z>;
def SchedWriteFHAdd
 : X86SchedWriteWidths<WriteFHAdd, WriteFHAdd, WriteFHAddY, WriteFHAddZ>;
def SchedWriteFCmp
 : X86SchedWriteWidths<WriteFCmp, WriteFCmpX, WriteFCmpY, WriteFCmpZ>;
def SchedWriteFCmp64
 : X86SchedWriteWidths<WriteFCmp64, WriteFCmp64X, WriteFCmp64Y, WriteFCmp64Z>;
def SchedWriteFMul
 : X86SchedWriteWidths<WriteFMul, WriteFMulX, WriteFMulY, WriteFMulZ>;
def SchedWriteFMul64
 : X86SchedWriteWidths<WriteFMul64, WriteFMul64X, WriteFMul64Y, WriteFMul64Z>;
def SchedWriteFMA
 : X86SchedWriteWidths<WriteFMA, WriteFMAX, WriteFMAY, WriteFMAZ>;
def SchedWriteDPPD
 : X86SchedWriteWidths<WriteDPPD, WriteDPPD, WriteDPPD, WriteDPPD>;
def SchedWriteDPPS
 : X86SchedWriteWidths<WriteDPPS, WriteDPPS, WriteDPPSY, WriteDPPSZ>;
def SchedWriteFDiv
 : X86SchedWriteWidths<WriteFDiv, WriteFDivX, WriteFDivY, WriteFDivZ>;
def SchedWriteFDiv64
 : X86SchedWriteWidths<WriteFDiv64, WriteFDiv64X, WriteFDiv64Y, WriteFDiv64Z>;
def SchedWriteFSqrt
 : X86SchedWriteWidths<WriteFSqrt, WriteFSqrtX,
                       WriteFSqrtY, WriteFSqrtZ>;
def SchedWriteFSqrt64
 : X86SchedWriteWidths<WriteFSqrt64, WriteFSqrt64X,
                       WriteFSqrt64Y, WriteFSqrt64Z>;
def SchedWriteFRcp
 : X86SchedWriteWidths<WriteFRcp, WriteFRcpX, WriteFRcpY, WriteFRcpZ>;
def SchedWriteFRsqrt
 : X86SchedWriteWidths<WriteFRsqrt, WriteFRsqrtX, WriteFRsqrtY, WriteFRsqrtZ>;
def SchedWriteFRnd
 : X86SchedWriteWidths<WriteFRnd, WriteFRnd, WriteFRndY, WriteFRndZ>;
def SchedWriteFLogic
 : X86SchedWriteWidths<WriteFLogic, WriteFLogic, WriteFLogicY, WriteFLogicZ>;
def SchedWriteFTest
 : X86SchedWriteWidths<WriteFTest, WriteFTest, WriteFTestY, WriteFTestZ>;

def SchedWriteFShuffle
 : X86SchedWriteWidths<WriteFShuffle, WriteFShuffle,
                       WriteFShuffleY, WriteFShuffleZ>;
def SchedWriteFVarShuffle
 : X86SchedWriteWidths<WriteFVarShuffle, WriteFVarShuffle,
                       WriteFVarShuffleY, WriteFVarShuffleZ>;
def SchedWriteFBlend
 : X86SchedWriteWidths<WriteFBlend, WriteFBlend, WriteFBlendY, WriteFBlendZ>;
def SchedWriteFVarBlend
 : X86SchedWriteWidths<WriteFVarBlend, WriteFVarBlend,
                       WriteFVarBlendY, WriteFVarBlendZ>;

def SchedWriteCvtDQ2PD
 : X86SchedWriteWidths<WriteCvtI2SD, WriteCvtI2PD,
                       WriteCvtI2PDY, WriteCvtI2PDZ>;
def SchedWriteCvtDQ2PS
 : X86SchedWriteWidths<WriteCvtI2SS, WriteCvtI2PS,
                       WriteCvtI2PSY, WriteCvtI2PSZ>;
def SchedWriteCvtPD2DQ
 : X86SchedWriteWidths<WriteCvtSD2I, WriteCvtPD2I,
                       WriteCvtPD2IY, WriteCvtPD2IZ>;
def SchedWriteCvtPS2DQ
 : X86SchedWriteWidths<WriteCvtSS2I, WriteCvtPS2I,
                       WriteCvtPS2IY, WriteCvtPS2IZ>;
def SchedWriteCvtPS2PD
 : X86SchedWriteWidths<WriteCvtSS2SD, WriteCvtPS2PD,
                       WriteCvtPS2PDY, WriteCvtPS2PDZ>;
def SchedWriteCvtPD2PS
 : X86SchedWriteWidths<WriteCvtSD2SS, WriteCvtPD2PS,
                       WriteCvtPD2PSY, WriteCvtPD2PSZ>;

def SchedWriteVecALU
 : X86SchedWriteWidths<WriteVecALU, WriteVecALUX, WriteVecALUY, WriteVecALUZ>;
def SchedWritePHAdd
 : X86SchedWriteWidths<WritePHAdd, WritePHAddX, WritePHAddY, WritePHAddZ>;
def SchedWriteVecLogic
 : X86SchedWriteWidths<WriteVecLogic, WriteVecLogicX,
                       WriteVecLogicY, WriteVecLogicZ>;
def SchedWriteVecTest
 : X86SchedWriteWidths<WriteVecTest, WriteVecTest,
                       WriteVecTestY, WriteVecTestZ>;
def SchedWriteVecShift
 : X86SchedWriteWidths<WriteVecShift, WriteVecShiftX,
                       WriteVecShiftY, WriteVecShiftZ>;
def SchedWriteVecShiftImm
 : X86SchedWriteWidths<WriteVecShiftImm, WriteVecShiftImmX,
                       WriteVecShiftImmY, WriteVecShiftImmZ>;
def SchedWriteVarVecShift
 : X86SchedWriteWidths<WriteVarVecShift, WriteVarVecShift,
                       WriteVarVecShiftY, WriteVarVecShiftZ>;
def SchedWriteVecIMul
 : X86SchedWriteWidths<WriteVecIMul, WriteVecIMulX,
                       WriteVecIMulY, WriteVecIMulZ>;
def SchedWritePMULLD
 : X86SchedWriteWidths<WritePMULLD, WritePMULLD,
                       WritePMULLDY, WritePMULLDZ>;
def SchedWriteMPSAD
 : X86SchedWriteWidths<WriteMPSAD, WriteMPSAD,
                       WriteMPSADY, WriteMPSADZ>;
def SchedWritePSADBW
 : X86SchedWriteWidths<WritePSADBW, WritePSADBWX,
                       WritePSADBWY, WritePSADBWZ>;

def SchedWriteShuffle
 : X86SchedWriteWidths<WriteShuffle, WriteShuffleX,
                       WriteShuffleY, WriteShuffleZ>;
def SchedWriteVarShuffle
 : X86SchedWriteWidths<WriteVarShuffle, WriteVarShuffleX,
                       WriteVarShuffleY, WriteVarShuffleZ>;
def SchedWriteBlend
 : X86SchedWriteWidths<WriteBlend, WriteBlend, WriteBlendY, WriteBlendZ>;
def SchedWriteVarBlend
 : X86SchedWriteWidths<WriteVarBlend, WriteVarBlend,
                       WriteVarBlendY, WriteVarBlendZ>;

// Vector size wrappers.
def SchedWriteFAddSizes
 : X86SchedWriteSizes<SchedWriteFAdd, SchedWriteFAdd64>;
def SchedWriteFCmpSizes
 : X86SchedWriteSizes<SchedWriteFCmp, SchedWriteFCmp64>;
def SchedWriteFMulSizes
 : X86SchedWriteSizes<SchedWriteFMul, SchedWriteFMul64>;
def SchedWriteFDivSizes
 : X86SchedWriteSizes<SchedWriteFDiv, SchedWriteFDiv64>;
def SchedWriteFSqrtSizes
 : X86SchedWriteSizes<SchedWriteFSqrt, SchedWriteFSqrt64>;
def SchedWriteFLogicSizes
 : X86SchedWriteSizes<SchedWriteFLogic, SchedWriteFLogic>;
def SchedWriteFShuffleSizes
 : X86SchedWriteSizes<SchedWriteFShuffle, SchedWriteFShuffle>;

//===----------------------------------------------------------------------===//
// Generic Processor Scheduler Models.

// IssueWidth is analogous to the number of decode units. Core and its
// descendents, including Nehalem and SandyBridge have 4 decoders.
// Resources beyond the decoder operate on micro-ops and are bufferred
// so adjacent micro-ops don't directly compete.
//
// MicroOpBufferSize > 1 indicates that RAW dependencies can be
// decoded in the same cycle. The value 32 is a reasonably arbitrary
// number of in-flight instructions.
//
// HighLatency=10 is optimistic. X86InstrInfo::isHighLatencyDef
// indicates high latency opcodes. Alternatively, InstrItinData
// entries may be included here to define specific operand
// latencies. Since these latencies are not used for pipeline hazards,
// they do not need to be exact.
//
// The GenericX86Model contains no instruction schedules
// and disables PostRAScheduler.
class GenericX86Model : SchedMachineModel {
  let IssueWidth = 4;
  let MicroOpBufferSize = 32;
  let LoadLatency = 4;
  let HighLatency = 10;
  let PostRAScheduler = 0;
  let CompleteModel = 0;
}

def GenericModel : GenericX86Model;

// Define a model with the PostRAScheduler enabled.
def GenericPostRAModel : GenericX86Model {
  let PostRAScheduler = 1;
}