llvm.org GIT mirror llvm / testing lib / Analysis / VectorUtils.cpp
testing

Tree @testing (Download .tar.gz)

VectorUtils.cpp @testingraw · 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
//===----------- VectorUtils.cpp - Vectorizer utility functions -----------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines vectorizer utilities.
//
//===----------------------------------------------------------------------===//

#include "llvm/Analysis/VectorUtils.h"
#include "llvm/ADT/EquivalenceClasses.h"
#include "llvm/Analysis/DemandedBits.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/ScalarEvolutionExpressions.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/IR/Value.h"

using namespace llvm;
using namespace llvm::PatternMatch;

/// \brief Identify if the intrinsic is trivially vectorizable.
/// This method returns true if the intrinsic's argument types are all
/// scalars for the scalar form of the intrinsic and all vectors for
/// the vector form of the intrinsic.
bool llvm::isTriviallyVectorizable(Intrinsic::ID ID) {
  switch (ID) {
  case Intrinsic::sqrt:
  case Intrinsic::sin:
  case Intrinsic::cos:
  case Intrinsic::exp:
  case Intrinsic::exp2:
  case Intrinsic::log:
  case Intrinsic::log10:
  case Intrinsic::log2:
  case Intrinsic::fabs:
  case Intrinsic::minnum:
  case Intrinsic::maxnum:
  case Intrinsic::copysign:
  case Intrinsic::floor:
  case Intrinsic::ceil:
  case Intrinsic::trunc:
  case Intrinsic::rint:
  case Intrinsic::nearbyint:
  case Intrinsic::round:
  case Intrinsic::bswap:
  case Intrinsic::bitreverse:
  case Intrinsic::ctpop:
  case Intrinsic::pow:
  case Intrinsic::fma:
  case Intrinsic::fmuladd:
  case Intrinsic::ctlz:
  case Intrinsic::cttz:
  case Intrinsic::powi:
    return true;
  default:
    return false;
  }
}

/// \brief Identifies if the intrinsic has a scalar operand. It check for
/// ctlz,cttz and powi special intrinsics whose argument is scalar.
bool llvm::hasVectorInstrinsicScalarOpd(Intrinsic::ID ID,
                                        unsigned ScalarOpdIdx) {
  switch (ID) {
  case Intrinsic::ctlz:
  case Intrinsic::cttz:
  case Intrinsic::powi:
    return (ScalarOpdIdx == 1);
  default:
    return false;
  }
}

/// \brief Returns intrinsic ID for call.
/// For the input call instruction it finds mapping intrinsic and returns
/// its ID, in case it does not found it return not_intrinsic.
Intrinsic::ID llvm::getVectorIntrinsicIDForCall(const CallInst *CI,
                                                const TargetLibraryInfo *TLI) {
  Intrinsic::ID ID = getIntrinsicForCallSite(CI, TLI);
  if (ID == Intrinsic::not_intrinsic)
    return Intrinsic::not_intrinsic;

  if (isTriviallyVectorizable(ID) || ID == Intrinsic::lifetime_start ||
      ID == Intrinsic::lifetime_end || ID == Intrinsic::assume)
    return ID;
  return Intrinsic::not_intrinsic;
}

/// \brief Find the operand of the GEP that should be checked for consecutive
/// stores. This ignores trailing indices that have no effect on the final
/// pointer.
unsigned llvm::getGEPInductionOperand(const GetElementPtrInst *Gep) {
  const DataLayout &DL = Gep->getModule()->getDataLayout();
  unsigned LastOperand = Gep->getNumOperands() - 1;
  unsigned GEPAllocSize = DL.getTypeAllocSize(Gep->getResultElementType());

  // Walk backwards and try to peel off zeros.
  while (LastOperand > 1 && match(Gep->getOperand(LastOperand), m_Zero())) {
    // Find the type we're currently indexing into.
    gep_type_iterator GEPTI = gep_type_begin(Gep);
    std::advance(GEPTI, LastOperand - 2);

    // If it's a type with the same allocation size as the result of the GEP we
    // can peel off the zero index.
    if (DL.getTypeAllocSize(GEPTI.getIndexedType()) != GEPAllocSize)
      break;
    --LastOperand;
  }

  return LastOperand;
}

/// \brief If the argument is a GEP, then returns the operand identified by
/// getGEPInductionOperand. However, if there is some other non-loop-invariant
/// operand, it returns that instead.
Value *llvm::stripGetElementPtr(Value *Ptr, ScalarEvolution *SE, Loop *Lp) {
  GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(Ptr);
  if (!GEP)
    return Ptr;

  unsigned InductionOperand = getGEPInductionOperand(GEP);

  // Check that all of the gep indices are uniform except for our induction
  // operand.
  for (unsigned i = 0, e = GEP->getNumOperands(); i != e; ++i)
    if (i != InductionOperand &&
        !SE->isLoopInvariant(SE->getSCEV(GEP->getOperand(i)), Lp))
      return Ptr;
  return GEP->getOperand(InductionOperand);
}

/// \brief If a value has only one user that is a CastInst, return it.
Value *llvm::getUniqueCastUse(Value *Ptr, Loop *Lp, Type *Ty) {
  Value *UniqueCast = nullptr;
  for (User *U : Ptr->users()) {
    CastInst *CI = dyn_cast<CastInst>(U);
    if (CI && CI->getType() == Ty) {
      if (!UniqueCast)
        UniqueCast = CI;
      else
        return nullptr;
    }
  }
  return UniqueCast;
}

/// \brief Get the stride of a pointer access in a loop. Looks for symbolic
/// strides "a[i*stride]". Returns the symbolic stride, or null otherwise.
Value *llvm::getStrideFromPointer(Value *Ptr, ScalarEvolution *SE, Loop *Lp) {
  auto *PtrTy = dyn_cast<PointerType>(Ptr->getType());
  if (!PtrTy || PtrTy->isAggregateType())
    return nullptr;

  // Try to remove a gep instruction to make the pointer (actually index at this
  // point) easier analyzable. If OrigPtr is equal to Ptr we are analzying the
  // pointer, otherwise, we are analyzing the index.
  Value *OrigPtr = Ptr;

  // The size of the pointer access.
  int64_t PtrAccessSize = 1;

  Ptr = stripGetElementPtr(Ptr, SE, Lp);
  const SCEV *V = SE->getSCEV(Ptr);

  if (Ptr != OrigPtr)
    // Strip off casts.
    while (const SCEVCastExpr *C = dyn_cast<SCEVCastExpr>(V))
      V = C->getOperand();

  const SCEVAddRecExpr *S = dyn_cast<SCEVAddRecExpr>(V);
  if (!S)
    return nullptr;

  V = S->getStepRecurrence(*SE);
  if (!V)
    return nullptr;

  // Strip off the size of access multiplication if we are still analyzing the
  // pointer.
  if (OrigPtr == Ptr) {
    if (const SCEVMulExpr *M = dyn_cast<SCEVMulExpr>(V)) {
      if (M->getOperand(0)->getSCEVType() != scConstant)
        return nullptr;

      const APInt &APStepVal = cast<SCEVConstant>(M->getOperand(0))->getAPInt();

      // Huge step value - give up.
      if (APStepVal.getBitWidth() > 64)
        return nullptr;

      int64_t StepVal = APStepVal.getSExtValue();
      if (PtrAccessSize != StepVal)
        return nullptr;
      V = M->getOperand(1);
    }
  }

  // Strip off casts.
  Type *StripedOffRecurrenceCast = nullptr;
  if (const SCEVCastExpr *C = dyn_cast<SCEVCastExpr>(V)) {
    StripedOffRecurrenceCast = C->getType();
    V = C->getOperand();
  }

  // Look for the loop invariant symbolic value.
  const SCEVUnknown *U = dyn_cast<SCEVUnknown>(V);
  if (!U)
    return nullptr;

  Value *Stride = U->getValue();
  if (!Lp->isLoopInvariant(Stride))
    return nullptr;

  // If we have stripped off the recurrence cast we have to make sure that we
  // return the value that is used in this loop so that we can replace it later.
  if (StripedOffRecurrenceCast)
    Stride = getUniqueCastUse(Stride, Lp, StripedOffRecurrenceCast);

  return Stride;
}

/// \brief Given a vector and an element number, see if the scalar value is
/// already around as a register, for example if it were inserted then extracted
/// from the vector.
Value *llvm::findScalarElement(Value *V, unsigned EltNo) {
  assert(V->getType()->isVectorTy() && "Not looking at a vector?");
  VectorType *VTy = cast<VectorType>(V->getType());
  unsigned Width = VTy->getNumElements();
  if (EltNo >= Width)  // Out of range access.
    return UndefValue::get(VTy->getElementType());

  if (Constant *C = dyn_cast<Constant>(V))
    return C->getAggregateElement(EltNo);

  if (InsertElementInst *III = dyn_cast<InsertElementInst>(V)) {
    // If this is an insert to a variable element, we don't know what it is.
    if (!isa<ConstantInt>(III->getOperand(2)))
      return nullptr;
    unsigned IIElt = cast<ConstantInt>(III->getOperand(2))->getZExtValue();

    // If this is an insert to the element we are looking for, return the
    // inserted value.
    if (EltNo == IIElt)
      return III->getOperand(1);

    // Otherwise, the insertelement doesn't modify the value, recurse on its
    // vector input.
    return findScalarElement(III->getOperand(0), EltNo);
  }

  if (ShuffleVectorInst *SVI = dyn_cast<ShuffleVectorInst>(V)) {
    unsigned LHSWidth = SVI->getOperand(0)->getType()->getVectorNumElements();
    int InEl = SVI->getMaskValue(EltNo);
    if (InEl < 0)
      return UndefValue::get(VTy->getElementType());
    if (InEl < (int)LHSWidth)
      return findScalarElement(SVI->getOperand(0), InEl);
    return findScalarElement(SVI->getOperand(1), InEl - LHSWidth);
  }

  // Extract a value from a vector add operation with a constant zero.
  Value *Val = nullptr; Constant *Con = nullptr;
  if (match(V, m_Add(m_Value(Val), m_Constant(Con))))
    if (Constant *Elt = Con->getAggregateElement(EltNo))
      if (Elt->isNullValue())
        return findScalarElement(Val, EltNo);

  // Otherwise, we don't know.
  return nullptr;
}

/// \brief Get splat value if the input is a splat vector or return nullptr.
/// This function is not fully general. It checks only 2 cases:
/// the input value is (1) a splat constants vector or (2) a sequence
/// of instructions that broadcast a single value into a vector.
///
const llvm::Value *llvm::getSplatValue(const Value *V) {

  if (auto *C = dyn_cast<Constant>(V))
    if (isa<VectorType>(V->getType()))
      return C->getSplatValue();

  auto *ShuffleInst = dyn_cast<ShuffleVectorInst>(V);
  if (!ShuffleInst)
    return nullptr;
  // All-zero (or undef) shuffle mask elements.
  for (int MaskElt : ShuffleInst->getShuffleMask())
    if (MaskElt != 0 && MaskElt != -1)
      return nullptr;
  // The first shuffle source is 'insertelement' with index 0.
  auto *InsertEltInst =
    dyn_cast<InsertElementInst>(ShuffleInst->getOperand(0));
  if (!InsertEltInst || !isa<ConstantInt>(InsertEltInst->getOperand(2)) ||
      !cast<ConstantInt>(InsertEltInst->getOperand(2))->isZero())
    return nullptr;

  return InsertEltInst->getOperand(1);
}

MapVector<Instruction *, uint64_t>
llvm::computeMinimumValueSizes(ArrayRef<BasicBlock *> Blocks, DemandedBits &DB,
                               const TargetTransformInfo *TTI) {

  // DemandedBits will give us every value's live-out bits. But we want
  // to ensure no extra casts would need to be inserted, so every DAG
  // of connected values must have the same minimum bitwidth.
  EquivalenceClasses<Value *> ECs;
  SmallVector<Value *, 16> Worklist;
  SmallPtrSet<Value *, 4> Roots;
  SmallPtrSet<Value *, 16> Visited;
  DenseMap<Value *, uint64_t> DBits;
  SmallPtrSet<Instruction *, 4> InstructionSet;
  MapVector<Instruction *, uint64_t> MinBWs;

  // Determine the roots. We work bottom-up, from truncs or icmps.
  bool SeenExtFromIllegalType = false;
  for (auto *BB : Blocks)
    for (auto &I : *BB) {
      InstructionSet.insert(&I);

      if (TTI && (isa<ZExtInst>(&I) || isa<SExtInst>(&I)) &&
          !TTI->isTypeLegal(I.getOperand(0)->getType()))
        SeenExtFromIllegalType = true;

      // Only deal with non-vector integers up to 64-bits wide.
      if ((isa<TruncInst>(&I) || isa<ICmpInst>(&I)) &&
          !I.getType()->isVectorTy() &&
          I.getOperand(0)->getType()->getScalarSizeInBits() <= 64) {
        // Don't make work for ourselves. If we know the loaded type is legal,
        // don't add it to the worklist.
        if (TTI && isa<TruncInst>(&I) && TTI->isTypeLegal(I.getType()))
          continue;

        Worklist.push_back(&I);
        Roots.insert(&I);
      }
    }
  // Early exit.
  if (Worklist.empty() || (TTI && !SeenExtFromIllegalType))
    return MinBWs;

  // Now proceed breadth-first, unioning values together.
  while (!Worklist.empty()) {
    Value *Val = Worklist.pop_back_val();
    Value *Leader = ECs.getOrInsertLeaderValue(Val);

    if (Visited.count(Val))
      continue;
    Visited.insert(Val);

    // Non-instructions terminate a chain successfully.
    if (!isa<Instruction>(Val))
      continue;
    Instruction *I = cast<Instruction>(Val);

    // If we encounter a type that is larger than 64 bits, we can't represent
    // it so bail out.
    if (DB.getDemandedBits(I).getBitWidth() > 64)
      return MapVector<Instruction *, uint64_t>();

    uint64_t V = DB.getDemandedBits(I).getZExtValue();
    DBits[Leader] |= V;
    DBits[I] = V;

    // Casts, loads and instructions outside of our range terminate a chain
    // successfully.
    if (isa<SExtInst>(I) || isa<ZExtInst>(I) || isa<LoadInst>(I) ||
        !InstructionSet.count(I))
      continue;

    // Unsafe casts terminate a chain unsuccessfully. We can't do anything
    // useful with bitcasts, ptrtoints or inttoptrs and it'd be unsafe to
    // transform anything that relies on them.
    if (isa<BitCastInst>(I) || isa<PtrToIntInst>(I) || isa<IntToPtrInst>(I) ||
        !I->getType()->isIntegerTy()) {
      DBits[Leader] |= ~0ULL;
      continue;
    }

    // We don't modify the types of PHIs. Reductions will already have been
    // truncated if possible, and inductions' sizes will have been chosen by
    // indvars.
    if (isa<PHINode>(I))
      continue;

    if (DBits[Leader] == ~0ULL)
      // All bits demanded, no point continuing.
      continue;

    for (Value *O : cast<User>(I)->operands()) {
      ECs.unionSets(Leader, O);
      Worklist.push_back(O);
    }
  }

  // Now we've discovered all values, walk them to see if there are
  // any users we didn't see. If there are, we can't optimize that
  // chain.
  for (auto &I : DBits)
    for (auto *U : I.first->users())
      if (U->getType()->isIntegerTy() && DBits.count(U) == 0)
        DBits[ECs.getOrInsertLeaderValue(I.first)] |= ~0ULL;

  for (auto I = ECs.begin(), E = ECs.end(); I != E; ++I) {
    uint64_t LeaderDemandedBits = 0;
    for (auto MI = ECs.member_begin(I), ME = ECs.member_end(); MI != ME; ++MI)
      LeaderDemandedBits |= DBits[*MI];

    uint64_t MinBW = (sizeof(LeaderDemandedBits) * 8) -
                     llvm::countLeadingZeros(LeaderDemandedBits);
    // Round up to a power of 2
    if (!isPowerOf2_64((uint64_t)MinBW))
      MinBW = NextPowerOf2(MinBW);

    // We don't modify the types of PHIs. Reductions will already have been
    // truncated if possible, and inductions' sizes will have been chosen by
    // indvars.
    // If we are required to shrink a PHI, abandon this entire equivalence class.
    bool Abort = false;
    for (auto MI = ECs.member_begin(I), ME = ECs.member_end(); MI != ME; ++MI)
      if (isa<PHINode>(*MI) && MinBW < (*MI)->getType()->getScalarSizeInBits()) {
        Abort = true;
        break;
      }
    if (Abort)
      continue;

    for (auto MI = ECs.member_begin(I), ME = ECs.member_end(); MI != ME; ++MI) {
      if (!isa<Instruction>(*MI))
        continue;
      Type *Ty = (*MI)->getType();
      if (Roots.count(*MI))
        Ty = cast<Instruction>(*MI)->getOperand(0)->getType();
      if (MinBW < Ty->getScalarSizeInBits())
        MinBWs[cast<Instruction>(*MI)] = MinBW;
    }
  }

  return MinBWs;
}

/// \returns \p I after propagating metadata from \p VL.
Instruction *llvm::propagateMetadata(Instruction *Inst, ArrayRef<Value *> VL) {
  Instruction *I0 = cast<Instruction>(VL[0]);
  SmallVector<std::pair<unsigned, MDNode *>, 4> Metadata;
  I0->getAllMetadataOtherThanDebugLoc(Metadata);

  for (auto Kind :
       {LLVMContext::MD_tbaa, LLVMContext::MD_alias_scope,
        LLVMContext::MD_noalias, LLVMContext::MD_fpmath,
        LLVMContext::MD_nontemporal, LLVMContext::MD_invariant_load}) {
    MDNode *MD = I0->getMetadata(Kind);

    for (int J = 1, E = VL.size(); MD && J != E; ++J) {
      const Instruction *IJ = cast<Instruction>(VL[J]);
      MDNode *IMD = IJ->getMetadata(Kind);
      switch (Kind) {
      case LLVMContext::MD_tbaa:
        MD = MDNode::getMostGenericTBAA(MD, IMD);
        break;
      case LLVMContext::MD_alias_scope:
        MD = MDNode::getMostGenericAliasScope(MD, IMD);
        break;
      case LLVMContext::MD_fpmath:
        MD = MDNode::getMostGenericFPMath(MD, IMD);
        break;
      case LLVMContext::MD_noalias:
      case LLVMContext::MD_nontemporal:
      case LLVMContext::MD_invariant_load:
        MD = MDNode::intersect(MD, IMD);
        break;
      default:
        llvm_unreachable("unhandled metadata");
      }
    }

    Inst->setMetadata(Kind, MD);
  }

  return Inst;
}

Constant *llvm::createInterleaveMask(IRBuilder<> &Builder, unsigned VF,
                                     unsigned NumVecs) {
  SmallVector<Constant *, 16> Mask;
  for (unsigned i = 0; i < VF; i++)
    for (unsigned j = 0; j < NumVecs; j++)
      Mask.push_back(Builder.getInt32(j * VF + i));

  return ConstantVector::get(Mask);
}

Constant *llvm::createStrideMask(IRBuilder<> &Builder, unsigned Start,
                                 unsigned Stride, unsigned VF) {
  SmallVector<Constant *, 16> Mask;
  for (unsigned i = 0; i < VF; i++)
    Mask.push_back(Builder.getInt32(Start + i * Stride));

  return ConstantVector::get(Mask);
}

Constant *llvm::createSequentialMask(IRBuilder<> &Builder, unsigned Start,
                                     unsigned NumInts, unsigned NumUndefs) {
  SmallVector<Constant *, 16> Mask;
  for (unsigned i = 0; i < NumInts; i++)
    Mask.push_back(Builder.getInt32(Start + i));

  Constant *Undef = UndefValue::get(Builder.getInt32Ty());
  for (unsigned i = 0; i < NumUndefs; i++)
    Mask.push_back(Undef);

  return ConstantVector::get(Mask);
}

/// A helper function for concatenating vectors. This function concatenates two
/// vectors having the same element type. If the second vector has fewer
/// elements than the first, it is padded with undefs.
static Value *concatenateTwoVectors(IRBuilder<> &Builder, Value *V1,
                                    Value *V2) {
  VectorType *VecTy1 = dyn_cast<VectorType>(V1->getType());
  VectorType *VecTy2 = dyn_cast<VectorType>(V2->getType());
  assert(VecTy1 && VecTy2 &&
         VecTy1->getScalarType() == VecTy2->getScalarType() &&
         "Expect two vectors with the same element type");

  unsigned NumElts1 = VecTy1->getNumElements();
  unsigned NumElts2 = VecTy2->getNumElements();
  assert(NumElts1 >= NumElts2 && "Unexpect the first vector has less elements");

  if (NumElts1 > NumElts2) {
    // Extend with UNDEFs.
    Constant *ExtMask =
        createSequentialMask(Builder, 0, NumElts2, NumElts1 - NumElts2);
    V2 = Builder.CreateShuffleVector(V2, UndefValue::get(VecTy2), ExtMask);
  }

  Constant *Mask = createSequentialMask(Builder, 0, NumElts1 + NumElts2, 0);
  return Builder.CreateShuffleVector(V1, V2, Mask);
}

Value *llvm::concatenateVectors(IRBuilder<> &Builder, ArrayRef<Value *> Vecs) {
  unsigned NumVecs = Vecs.size();
  assert(NumVecs > 1 && "Should be at least two vectors");

  SmallVector<Value *, 8> ResList;
  ResList.append(Vecs.begin(), Vecs.end());
  do {
    SmallVector<Value *, 8> TmpList;
    for (unsigned i = 0; i < NumVecs - 1; i += 2) {
      Value *V0 = ResList[i], *V1 = ResList[i + 1];
      assert((V0->getType() == V1->getType() || i == NumVecs - 2) &&
             "Only the last vector may have a different type");

      TmpList.push_back(concatenateTwoVectors(Builder, V0, V1));
    }

    // Push the last vector if the total number of vectors is odd.
    if (NumVecs % 2 != 0)
      TmpList.push_back(ResList[NumVecs - 1]);

    ResList = TmpList;
    NumVecs = ResList.size();
  } while (NumVecs > 1);

  return ResList[0];
}