bool InterleavedAccess::lowerInterleavedStore(
StoreInst *SI, SmallVector<Instruction *, 32> &DeadInsts) {
if (!SI->isSimple())
return false;
ShuffleVectorInst *SVI = dyn_cast<ShuffleVectorInst>(SI->getValueOperand());
if (!SVI || !SVI->hasOneUse())
return false;
// Check if the shufflevector is RE-interleave shuffle.
unsigned Factor;
unsigned OpNumElts = SVI->getOperand(0)->getType()->getVectorNumElements();
if (!isReInterleaveMask(SVI->getShuffleMask(), Factor, MaxFactor, OpNumElts))
return false;
LLVM_DEBUG(dbgs() << "IA: Found an interleaved store: " << *SI << "\n");
// Try to create target specific intrinsics to replace the store and shuffle.
if (!TLI->lowerInterleavedStore(SI, SVI, Factor))
return false;
// Already have a new target specific interleaved store. Erase the old store.
DeadInsts.push_back(SI);
DeadInsts.push_back(SVI);
return true;
}
bool InterleavedAccess::lowerInterleavedLoad(
LoadInst *LI, SmallVector<Instruction *, 32> &DeadInsts) {
if (!LI->isSimple())
return false;
SmallVector<ShuffleVectorInst *, 4> Shuffles;
// Check if all users of this load are shufflevectors.
for (auto UI = LI->user_begin(), E = LI->user_end(); UI != E; UI++) {
ShuffleVectorInst *SVI = dyn_cast<ShuffleVectorInst>(*UI);
if (!SVI || !isa<UndefValue>(SVI->getOperand(1)))
return false;
Shuffles.push_back(SVI);
}
if (Shuffles.empty())
return false;
unsigned Factor, Index;
// Check if the first shufflevector is DE-interleave shuffle.
if (!isDeInterleaveMask(Shuffles[0]->getShuffleMask(), Factor, Index))
return false;
// Holds the corresponding index for each DE-interleave shuffle.
SmallVector<unsigned, 4> Indices;
Indices.push_back(Index);
Type *VecTy = Shuffles[0]->getType();
// Check if other shufflevectors are also DE-interleaved of the same type
// and factor as the first shufflevector.
for (unsigned i = 1; i < Shuffles.size(); i++) {
if (Shuffles[i]->getType() != VecTy)
return false;
if (!isDeInterleaveMaskOfFactor(Shuffles[i]->getShuffleMask(), Factor,
Index))
return false;
Indices.push_back(Index);
}
DEBUG(dbgs() << "IA: Found an interleaved load: " << *LI << "\n");
// Try to create target specific intrinsics to replace the load and shuffles.
if (!TLI->lowerInterleavedLoad(LI, Shuffles, Indices, Factor))
return false;
for (auto SVI : Shuffles)
DeadInsts.push_back(SVI);
DeadInsts.push_back(LI);
return true;
}
// Returns true if the shuffle is extracting a contiguous range of values from
// LHS, for example:
// +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
// Input: |AA|BB|CC|DD|EE|FF|GG|HH|II|JJ|KK|LL|MM|NN|OO|PP|
// Shuffles to: |EE|FF|GG|HH|
// +--+--+--+--+
static bool isShuffleExtractingFromLHS(ShuffleVectorInst &SVI,
SmallVector<int, 16> &Mask) {
unsigned LHSElems =
cast<VectorType>(SVI.getOperand(0)->getType())->getNumElements();
unsigned MaskElems = Mask.size();
unsigned BegIdx = Mask.front();
unsigned EndIdx = Mask.back();
if (BegIdx > EndIdx || EndIdx >= LHSElems || EndIdx - BegIdx != MaskElems - 1)
return false;
for (unsigned I = 0; I != MaskElems; ++I)
if (static_cast<unsigned>(Mask[I]) != BegIdx + I)
return false;
return true;
}
static bool isByteSwap64(ShuffleVectorInst &SI, SmallVector<int, 16>&RefMasks)
{
RefMasks.clear();
unsigned VWidth = cast<VectorType>(SI.getType())->getNumElements();
VectorType *LHS = cast<VectorType>(SI.getOperand(0)->getType());
VectorType *RHS = cast<VectorType>(SI.getOperand(1)->getType());
IntegerType *IT = dyn_cast<IntegerType>(LHS->getElementType());
//When Element Type is not IntegerType or the Result's element number
//can't be divided by 8, return false
//TODO:Need to check all masks are all constants.
if (IT == nullptr
|| ! IT->isIntegerTy(8)
|| VWidth % 8 != 0) {
return false;
}
SmallVector<int, 16> Masks(SI.getShuffleMask());
bool isByteSwap = true;
for (unsigned i = 0; i < VWidth / 8; ++i) {
unsigned base = Masks[i * 8];
if (base % 8 != 7) {
isByteSwap = false;
break;
}
for (unsigned j = 1; j < 8; ++j) {
if (base - Masks[i * 8 + j] != j) {
isByteSwap = false;
break;
}
}
if (isByteSwap) {
RefMasks.push_back(base / 8);
} else {
break;
}
}
if (!isByteSwap) {
RefMasks.clear();
}
return isByteSwap;
}
bool Scalarizer::visitShuffleVectorInst(ShuffleVectorInst &SVI) {
VectorType *VT = dyn_cast<VectorType>(SVI.getType());
if (!VT)
return false;
unsigned NumElems = VT->getNumElements();
Scatterer Op0 = scatter(&SVI, SVI.getOperand(0));
Scatterer Op1 = scatter(&SVI, SVI.getOperand(1));
ValueVector Res;
Res.resize(NumElems);
for (unsigned I = 0; I < NumElems; ++I) {
int Selector = SVI.getMaskValue(I);
if (Selector < 0)
Res[I] = UndefValue::get(VT->getElementType());
else if (unsigned(Selector) < Op0.size())
Res[I] = Op0[Selector];
else
Res[I] = Op1[Selector - Op0.size()];
}
gather(&SVI, Res);
return true;
}
static bool matchVectorSplittingReduction(const ExtractElementInst *ReduxRoot,
unsigned &Opcode, Type *&Ty) {
if (!EnableReduxCost)
return false;
// Need to extract the first element.
ConstantInt *CI = dyn_cast<ConstantInt>(ReduxRoot->getOperand(1));
unsigned Idx = ~0u;
if (CI)
Idx = CI->getZExtValue();
if (Idx != 0)
return false;
BinaryOperator *RdxStart = dyn_cast<BinaryOperator>(ReduxRoot->getOperand(0));
if (!RdxStart)
return false;
unsigned RdxOpcode = RdxStart->getOpcode();
Type *VecTy = ReduxRoot->getOperand(0)->getType();
unsigned NumVecElems = VecTy->getVectorNumElements();
if (!isPowerOf2_32(NumVecElems))
return false;
// We look for a sequence of shuffles and adds like the following matching one
// fadd, shuffle vector pair at a time.
//
// %rdx.shuf = shufflevector <4 x float> %rdx, <4 x float> undef,
// <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
// %bin.rdx = fadd <4 x float> %rdx, %rdx.shuf
// %rdx.shuf7 = shufflevector <4 x float> %bin.rdx, <4 x float> undef,
// <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
// %bin.rdx8 = fadd <4 x float> %bin.rdx, %rdx.shuf7
// %r = extractelement <4 x float> %bin.rdx8, i32 0
unsigned MaskStart = 1;
Value *RdxOp = RdxStart;
SmallVector<int, 32> ShuffleMask(NumVecElems, 0);
unsigned NumVecElemsRemain = NumVecElems;
while (NumVecElemsRemain - 1) {
// Check for the right reduction operation.
BinaryOperator *BinOp;
if (!(BinOp = dyn_cast<BinaryOperator>(RdxOp)))
return false;
if (BinOp->getOpcode() != RdxOpcode)
return false;
Value *NextRdxOp;
ShuffleVectorInst *Shuffle;
std::tie(NextRdxOp, Shuffle) = getShuffleAndOtherOprd(BinOp);
// Check the current reduction operation and the shuffle use the same value.
if (Shuffle == nullptr)
return false;
if (Shuffle->getOperand(0) != NextRdxOp)
return false;
// Check that shuffle masks matches.
for (unsigned j = 0; j != MaskStart; ++j)
ShuffleMask[j] = MaskStart + j;
// Fill the rest of the mask with -1 for undef.
std::fill(&ShuffleMask[MaskStart], ShuffleMask.end(), -1);
SmallVector<int, 16> Mask = Shuffle->getShuffleMask();
if (ShuffleMask != Mask)
return false;
RdxOp = NextRdxOp;
NumVecElemsRemain /= 2;
MaskStart *= 2;
}
Opcode = RdxOpcode;
Ty = VecTy;
return true;
}
static bool matchPairwiseReductionAtLevel(const BinaryOperator *BinOp,
unsigned Level, unsigned NumLevels) {
// Match one level of pairwise operations.
// %rdx.shuf.0.0 = shufflevector <4 x float> %rdx, <4 x float> undef,
// <4 x i32> <i32 0, i32 2 , i32 undef, i32 undef>
// %rdx.shuf.0.1 = shufflevector <4 x float> %rdx, <4 x float> undef,
// <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
// %bin.rdx.0 = fadd <4 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
if (BinOp == nullptr)
return false;
assert(BinOp->getType()->isVectorTy() && "Expecting a vector type");
unsigned Opcode = BinOp->getOpcode();
Value *L = BinOp->getOperand(0);
Value *R = BinOp->getOperand(1);
ShuffleVectorInst *LS = dyn_cast<ShuffleVectorInst>(L);
if (!LS && Level)
return false;
ShuffleVectorInst *RS = dyn_cast<ShuffleVectorInst>(R);
if (!RS && Level)
return false;
// On level 0 we can omit one shufflevector instruction.
if (!Level && !RS && !LS)
return false;
// Shuffle inputs must match.
Value *NextLevelOpL = LS ? LS->getOperand(0) : nullptr;
Value *NextLevelOpR = RS ? RS->getOperand(0) : nullptr;
Value *NextLevelOp = nullptr;
if (NextLevelOpR && NextLevelOpL) {
// If we have two shuffles their operands must match.
if (NextLevelOpL != NextLevelOpR)
return false;
NextLevelOp = NextLevelOpL;
} else if (Level == 0 && (NextLevelOpR || NextLevelOpL)) {
// On the first level we can omit the shufflevector <0, undef,...>. So the
// input to the other shufflevector <1, undef> must match with one of the
// inputs to the current binary operation.
// Example:
// %NextLevelOpL = shufflevector %R, <1, undef ...>
// %BinOp = fadd %NextLevelOpL, %R
if (NextLevelOpL && NextLevelOpL != R)
return false;
else if (NextLevelOpR && NextLevelOpR != L)
return false;
NextLevelOp = NextLevelOpL ? R : L;
} else
return false;
// Check that the next levels binary operation exists and matches with the
// current one.
BinaryOperator *NextLevelBinOp = nullptr;
if (Level + 1 != NumLevels) {
if (!(NextLevelBinOp = dyn_cast<BinaryOperator>(NextLevelOp)))
return false;
else if (NextLevelBinOp->getOpcode() != Opcode)
return false;
}
// Shuffle mask for pairwise operation must match.
if (matchPairwiseShuffleMask(LS, true, Level)) {
if (!matchPairwiseShuffleMask(RS, false, Level))
return false;
} else if (matchPairwiseShuffleMask(RS, true, Level)) {
if (!matchPairwiseShuffleMask(LS, false, Level))
return false;
} else
return false;
if (++Level == NumLevels)
return true;
// Match next level.
return matchPairwiseReductionAtLevel(NextLevelBinOp, Level, NumLevels);
}
Instruction *InstCombiner::visitShuffleVectorInst(ShuffleVectorInst &SVI) {
Value *LHS = SVI.getOperand(0);
Value *RHS = SVI.getOperand(1);
SmallVector<int, 16> Mask = getShuffleMask(&SVI);
bool MadeChange = false;
// Undefined shuffle mask -> undefined value.
if (isa<UndefValue>(SVI.getOperand(2)))
return ReplaceInstUsesWith(SVI, UndefValue::get(SVI.getType()));
unsigned VWidth = cast<VectorType>(SVI.getType())->getNumElements();
APInt UndefElts(VWidth, 0);
APInt AllOnesEltMask(APInt::getAllOnesValue(VWidth));
if (Value *V = SimplifyDemandedVectorElts(&SVI, AllOnesEltMask, UndefElts)) {
if (V != &SVI)
return ReplaceInstUsesWith(SVI, V);
LHS = SVI.getOperand(0);
RHS = SVI.getOperand(1);
MadeChange = true;
}
unsigned LHSWidth = cast<VectorType>(LHS->getType())->getNumElements();
// Canonicalize shuffle(x ,x,mask) -> shuffle(x, undef,mask')
// Canonicalize shuffle(undef,x,mask) -> shuffle(x, undef,mask').
if (LHS == RHS || isa<UndefValue>(LHS)) {
if (isa<UndefValue>(LHS) && LHS == RHS) {
// shuffle(undef,undef,mask) -> undef.
Value* result = (VWidth == LHSWidth)
? LHS : UndefValue::get(SVI.getType());
return ReplaceInstUsesWith(SVI, result);
}
// Remap any references to RHS to use LHS.
std::vector<Constant*> Elts;
for (unsigned i = 0, e = LHSWidth; i != VWidth; ++i) {
if (Mask[i] < 0)
Elts.push_back(UndefValue::get(Type::getInt32Ty(SVI.getContext())));
else {
if ((Mask[i] >= (int)e && isa<UndefValue>(RHS)) ||
(Mask[i] < (int)e && isa<UndefValue>(LHS))) {
Mask[i] = -1; // Turn into undef.
Elts.push_back(UndefValue::get(Type::getInt32Ty(SVI.getContext())));
} else {
Mask[i] = Mask[i] % e; // Force to LHS.
Elts.push_back(ConstantInt::get(Type::getInt32Ty(SVI.getContext()),
Mask[i]));
}
}
}
SVI.setOperand(0, SVI.getOperand(1));
SVI.setOperand(1, UndefValue::get(RHS->getType()));
SVI.setOperand(2, ConstantVector::get(Elts));
LHS = SVI.getOperand(0);
RHS = SVI.getOperand(1);
MadeChange = true;
}
if (VWidth == LHSWidth) {
// Analyze the shuffle, are the LHS or RHS and identity shuffles?
bool isLHSID = true, isRHSID = true;
for (unsigned i = 0, e = Mask.size(); i != e; ++i) {
if (Mask[i] < 0) continue; // Ignore undef values.
// Is this an identity shuffle of the LHS value?
isLHSID &= (Mask[i] == (int)i);
// Is this an identity shuffle of the RHS value?
isRHSID &= (Mask[i]-e == i);
}
// Eliminate identity shuffles.
if (isLHSID) return ReplaceInstUsesWith(SVI, LHS);
if (isRHSID) return ReplaceInstUsesWith(SVI, RHS);
}
// If the LHS is a shufflevector itself, see if we can combine it with this
// one without producing an unusual shuffle.
// Cases that might be simplified:
// 1.
// x1=shuffle(v1,v2,mask1)
// x=shuffle(x1,undef,mask)
// ==>
// x=shuffle(v1,undef,newMask)
// newMask[i] = (mask[i] < x1.size()) ? mask1[mask[i]] : -1
// 2.
// x1=shuffle(v1,undef,mask1)
// x=shuffle(x1,x2,mask)
// where v1.size() == mask1.size()
// ==>
// x=shuffle(v1,x2,newMask)
// newMask[i] = (mask[i] < x1.size()) ? mask1[mask[i]] : mask[i]
// 3.
// x2=shuffle(v2,undef,mask2)
// x=shuffle(x1,x2,mask)
// where v2.size() == mask2.size()
// ==>
// x=shuffle(x1,v2,newMask)
// newMask[i] = (mask[i] < x1.size())
// ? mask[i] : mask2[mask[i]-x1.size()]+x1.size()
// 4.
// x1=shuffle(v1,undef,mask1)
// x2=shuffle(v2,undef,mask2)
// x=shuffle(x1,x2,mask)
// where v1.size() == v2.size()
// ==>
// x=shuffle(v1,v2,newMask)
// newMask[i] = (mask[i] < x1.size())
// ? mask1[mask[i]] : mask2[mask[i]-x1.size()]+v1.size()
//
// Here we are really conservative:
// we are absolutely afraid of producing a shuffle mask not in the input
// program, because the code gen may not be smart enough to turn a merged
// shuffle into two specific shuffles: it may produce worse code. As such,
// we only merge two shuffles if the result is either a splat or one of the
// input shuffle masks. In this case, merging the shuffles just removes
// one instruction, which we know is safe. This is good for things like
// turning: (splat(splat)) -> splat, or
// merge(V[0..n], V[n+1..2n]) -> V[0..2n]
ShuffleVectorInst* LHSShuffle = dyn_cast<ShuffleVectorInst>(LHS);
ShuffleVectorInst* RHSShuffle = dyn_cast<ShuffleVectorInst>(RHS);
if (LHSShuffle)
if (!isa<UndefValue>(LHSShuffle->getOperand(1)) && !isa<UndefValue>(RHS))
LHSShuffle = NULL;
if (RHSShuffle)
if (!isa<UndefValue>(RHSShuffle->getOperand(1)))
RHSShuffle = NULL;
if (!LHSShuffle && !RHSShuffle)
return MadeChange ? &SVI : 0;
Value* LHSOp0 = NULL;
Value* LHSOp1 = NULL;
Value* RHSOp0 = NULL;
unsigned LHSOp0Width = 0;
unsigned RHSOp0Width = 0;
if (LHSShuffle) {
LHSOp0 = LHSShuffle->getOperand(0);
LHSOp1 = LHSShuffle->getOperand(1);
LHSOp0Width = cast<VectorType>(LHSOp0->getType())->getNumElements();
}
if (RHSShuffle) {
RHSOp0 = RHSShuffle->getOperand(0);
RHSOp0Width = cast<VectorType>(RHSOp0->getType())->getNumElements();
}
Value* newLHS = LHS;
Value* newRHS = RHS;
if (LHSShuffle) {
// case 1
if (isa<UndefValue>(RHS)) {
newLHS = LHSOp0;
newRHS = LHSOp1;
}
// case 2 or 4
else if (LHSOp0Width == LHSWidth) {
newLHS = LHSOp0;
}
}
// case 3 or 4
if (RHSShuffle && RHSOp0Width == LHSWidth) {
newRHS = RHSOp0;
}
// case 4
if (LHSOp0 == RHSOp0) {
newLHS = LHSOp0;
newRHS = NULL;
}
if (newLHS == LHS && newRHS == RHS)
return MadeChange ? &SVI : 0;
SmallVector<int, 16> LHSMask;
SmallVector<int, 16> RHSMask;
if (newLHS != LHS) {
LHSMask = getShuffleMask(LHSShuffle);
}
if (RHSShuffle && newRHS != RHS) {
RHSMask = getShuffleMask(RHSShuffle);
}
unsigned newLHSWidth = (newLHS != LHS) ? LHSOp0Width : LHSWidth;
SmallVector<int, 16> newMask;
bool isSplat = true;
int SplatElt = -1;
// Create a new mask for the new ShuffleVectorInst so that the new
// ShuffleVectorInst is equivalent to the original one.
for (unsigned i = 0; i < VWidth; ++i) {
int eltMask;
if (Mask[i] == -1) {
// This element is an undef value.
eltMask = -1;
} else if (Mask[i] < (int)LHSWidth) {
// This element is from left hand side vector operand.
//
// If LHS is going to be replaced (case 1, 2, or 4), calculate the
// new mask value for the element.
if (newLHS != LHS) {
eltMask = LHSMask[Mask[i]];
// If the value selected is an undef value, explicitly specify it
// with a -1 mask value.
if (eltMask >= (int)LHSOp0Width && isa<UndefValue>(LHSOp1))
eltMask = -1;
}
else
eltMask = Mask[i];
} else {
// This element is from right hand side vector operand
//
// If the value selected is an undef value, explicitly specify it
// with a -1 mask value. (case 1)
if (isa<UndefValue>(RHS))
eltMask = -1;
// If RHS is going to be replaced (case 3 or 4), calculate the
// new mask value for the element.
else if (newRHS != RHS) {
eltMask = RHSMask[Mask[i]-LHSWidth];
// If the value selected is an undef value, explicitly specify it
// with a -1 mask value.
if (eltMask >= (int)RHSOp0Width) {
assert(isa<UndefValue>(RHSShuffle->getOperand(1))
&& "should have been check above");
eltMask = -1;
}
}
else
eltMask = Mask[i]-LHSWidth;
// If LHS's width is changed, shift the mask value accordingly.
// If newRHS == NULL, i.e. LHSOp0 == RHSOp0, we want to remap any
// references to RHSOp0 to LHSOp0, so we don't need to shift the mask.
if (eltMask >= 0 && newRHS != NULL)
eltMask += newLHSWidth;
}
// Check if this could still be a splat.
if (eltMask >= 0) {
if (SplatElt >= 0 && SplatElt != eltMask)
isSplat = false;
SplatElt = eltMask;
}
newMask.push_back(eltMask);
}
// If the result mask is equal to one of the original shuffle masks,
// or is a splat, do the replacement.
if (isSplat || newMask == LHSMask || newMask == RHSMask || newMask == Mask) {
SmallVector<Constant*, 16> Elts;
Type *Int32Ty = Type::getInt32Ty(SVI.getContext());
for (unsigned i = 0, e = newMask.size(); i != e; ++i) {
if (newMask[i] < 0) {
Elts.push_back(UndefValue::get(Int32Ty));
} else {
Elts.push_back(ConstantInt::get(Int32Ty, newMask[i]));
}
}
if (newRHS == NULL)
newRHS = UndefValue::get(newLHS->getType());
return new ShuffleVectorInst(newLHS, newRHS, ConstantVector::get(Elts));
}
return MadeChange ? &SVI : 0;
}
static ReductionKind matchPairwiseReductionAtLevel(Instruction *I,
unsigned Level,
unsigned NumLevels) {
// Match one level of pairwise operations.
// %rdx.shuf.0.0 = shufflevector <4 x float> %rdx, <4 x float> undef,
// <4 x i32> <i32 0, i32 2 , i32 undef, i32 undef>
// %rdx.shuf.0.1 = shufflevector <4 x float> %rdx, <4 x float> undef,
// <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
// %bin.rdx.0 = fadd <4 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
if (!I)
return RK_None;
assert(I->getType()->isVectorTy() && "Expecting a vector type");
Optional<ReductionData> RD = getReductionData(I);
if (!RD)
return RK_None;
ShuffleVectorInst *LS = dyn_cast<ShuffleVectorInst>(RD->LHS);
if (!LS && Level)
return RK_None;
ShuffleVectorInst *RS = dyn_cast<ShuffleVectorInst>(RD->RHS);
if (!RS && Level)
return RK_None;
// On level 0 we can omit one shufflevector instruction.
if (!Level && !RS && !LS)
return RK_None;
// Shuffle inputs must match.
Value *NextLevelOpL = LS ? LS->getOperand(0) : nullptr;
Value *NextLevelOpR = RS ? RS->getOperand(0) : nullptr;
Value *NextLevelOp = nullptr;
if (NextLevelOpR && NextLevelOpL) {
// If we have two shuffles their operands must match.
if (NextLevelOpL != NextLevelOpR)
return RK_None;
NextLevelOp = NextLevelOpL;
} else if (Level == 0 && (NextLevelOpR || NextLevelOpL)) {
// On the first level we can omit the shufflevector <0, undef,...>. So the
// input to the other shufflevector <1, undef> must match with one of the
// inputs to the current binary operation.
// Example:
// %NextLevelOpL = shufflevector %R, <1, undef ...>
// %BinOp = fadd %NextLevelOpL, %R
if (NextLevelOpL && NextLevelOpL != RD->RHS)
return RK_None;
else if (NextLevelOpR && NextLevelOpR != RD->LHS)
return RK_None;
NextLevelOp = NextLevelOpL ? RD->RHS : RD->LHS;
} else
return RK_None;
// Check that the next levels binary operation exists and matches with the
// current one.
if (Level + 1 != NumLevels) {
Optional<ReductionData> NextLevelRD =
getReductionData(cast<Instruction>(NextLevelOp));
if (!NextLevelRD || !RD->hasSameData(*NextLevelRD))
return RK_None;
}
// Shuffle mask for pairwise operation must match.
if (matchPairwiseShuffleMask(LS, /*IsLeft=*/true, Level)) {
if (!matchPairwiseShuffleMask(RS, /*IsLeft=*/false, Level))
return RK_None;
} else if (matchPairwiseShuffleMask(RS, /*IsLeft=*/true, Level)) {
if (!matchPairwiseShuffleMask(LS, /*IsLeft=*/false, Level))
return RK_None;
} else {
return RK_None;
}
if (++Level == NumLevels)
return RD->Kind;
// Match next level.
return matchPairwiseReductionAtLevel(cast<Instruction>(NextLevelOp), Level,
NumLevels);
}
bool InterleavedAccess::lowerInterleavedLoad(
LoadInst *LI, SmallVector<Instruction *, 32> &DeadInsts) {
if (!LI->isSimple())
return false;
SmallVector<ShuffleVectorInst *, 4> Shuffles;
SmallVector<ExtractElementInst *, 4> Extracts;
// Check if all users of this load are shufflevectors. If we encounter any
// users that are extractelement instructions, we save them to later check if
// they can be modifed to extract from one of the shufflevectors instead of
// the load.
for (auto UI = LI->user_begin(), E = LI->user_end(); UI != E; UI++) {
auto *Extract = dyn_cast<ExtractElementInst>(*UI);
if (Extract && isa<ConstantInt>(Extract->getIndexOperand())) {
Extracts.push_back(Extract);
continue;
}
ShuffleVectorInst *SVI = dyn_cast<ShuffleVectorInst>(*UI);
if (!SVI || !isa<UndefValue>(SVI->getOperand(1)))
return false;
Shuffles.push_back(SVI);
}
if (Shuffles.empty())
return false;
unsigned Factor, Index;
// Check if the first shufflevector is DE-interleave shuffle.
if (!isDeInterleaveMask(Shuffles[0]->getShuffleMask(), Factor, Index,
MaxFactor))
return false;
// Holds the corresponding index for each DE-interleave shuffle.
SmallVector<unsigned, 4> Indices;
Indices.push_back(Index);
Type *VecTy = Shuffles[0]->getType();
// Check if other shufflevectors are also DE-interleaved of the same type
// and factor as the first shufflevector.
for (unsigned i = 1; i < Shuffles.size(); i++) {
if (Shuffles[i]->getType() != VecTy)
return false;
if (!isDeInterleaveMaskOfFactor(Shuffles[i]->getShuffleMask(), Factor,
Index))
return false;
Indices.push_back(Index);
}
// Try and modify users of the load that are extractelement instructions to
// use the shufflevector instructions instead of the load.
if (!tryReplaceExtracts(Extracts, Shuffles))
return false;
LLVM_DEBUG(dbgs() << "IA: Found an interleaved load: " << *LI << "\n");
// Try to create target specific intrinsics to replace the load and shuffles.
if (!TLI->lowerInterleavedLoad(LI, Shuffles, Indices, Factor))
return false;
for (auto SVI : Shuffles)
DeadInsts.push_back(SVI);
DeadInsts.push_back(LI);
return true;
}
Instruction *InstCombiner::visitShuffleVectorInst(ShuffleVectorInst &SVI) {
Value *LHS = SVI.getOperand(0);
Value *RHS = SVI.getOperand(1);
SmallVector<int, 16> Mask = SVI.getShuffleMask();
Type *Int32Ty = Type::getInt32Ty(SVI.getContext());
bool MadeChange = false;
// Undefined shuffle mask -> undefined value.
if (isa<UndefValue>(SVI.getOperand(2)))
return replaceInstUsesWith(SVI, UndefValue::get(SVI.getType()));
unsigned VWidth = cast<VectorType>(SVI.getType())->getNumElements();
APInt UndefElts(VWidth, 0);
APInt AllOnesEltMask(APInt::getAllOnesValue(VWidth));
if (Value *V = SimplifyDemandedVectorElts(&SVI, AllOnesEltMask, UndefElts)) {
if (V != &SVI)
return replaceInstUsesWith(SVI, V);
LHS = SVI.getOperand(0);
RHS = SVI.getOperand(1);
MadeChange = true;
}
unsigned LHSWidth = cast<VectorType>(LHS->getType())->getNumElements();
// Canonicalize shuffle(x ,x,mask) -> shuffle(x, undef,mask')
// Canonicalize shuffle(undef,x,mask) -> shuffle(x, undef,mask').
if (LHS == RHS || isa<UndefValue>(LHS)) {
if (isa<UndefValue>(LHS) && LHS == RHS) {
// shuffle(undef,undef,mask) -> undef.
Value *Result = (VWidth == LHSWidth)
? LHS : UndefValue::get(SVI.getType());
return replaceInstUsesWith(SVI, Result);
}
// Remap any references to RHS to use LHS.
SmallVector<Constant*, 16> Elts;
for (unsigned i = 0, e = LHSWidth; i != VWidth; ++i) {
if (Mask[i] < 0) {
Elts.push_back(UndefValue::get(Int32Ty));
continue;
}
if ((Mask[i] >= (int)e && isa<UndefValue>(RHS)) ||
(Mask[i] < (int)e && isa<UndefValue>(LHS))) {
Mask[i] = -1; // Turn into undef.
Elts.push_back(UndefValue::get(Int32Ty));
} else {
Mask[i] = Mask[i] % e; // Force to LHS.
Elts.push_back(ConstantInt::get(Int32Ty, Mask[i]));
}
}
SVI.setOperand(0, SVI.getOperand(1));
SVI.setOperand(1, UndefValue::get(RHS->getType()));
SVI.setOperand(2, ConstantVector::get(Elts));
LHS = SVI.getOperand(0);
RHS = SVI.getOperand(1);
MadeChange = true;
}
if (VWidth == LHSWidth) {
// Analyze the shuffle, are the LHS or RHS and identity shuffles?
bool isLHSID, isRHSID;
recognizeIdentityMask(Mask, isLHSID, isRHSID);
// Eliminate identity shuffles.
if (isLHSID) return replaceInstUsesWith(SVI, LHS);
if (isRHSID) return replaceInstUsesWith(SVI, RHS);
}
if (isa<UndefValue>(RHS) && CanEvaluateShuffled(LHS, Mask)) {
Value *V = EvaluateInDifferentElementOrder(LHS, Mask);
return replaceInstUsesWith(SVI, V);
}
// SROA generates shuffle+bitcast when the extracted sub-vector is bitcast to
// a non-vector type. We can instead bitcast the original vector followed by
// an extract of the desired element:
//
// %sroa = shufflevector <16 x i8> %in, <16 x i8> undef,
// <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// %1 = bitcast <4 x i8> %sroa to i32
// Becomes:
// %bc = bitcast <16 x i8> %in to <4 x i32>
// %ext = extractelement <4 x i32> %bc, i32 0
//
// If the shuffle is extracting a contiguous range of values from the input
// vector then each use which is a bitcast of the extracted size can be
// replaced. This will work if the vector types are compatible, and the begin
// index is aligned to a value in the casted vector type. If the begin index
// isn't aligned then we can shuffle the original vector (keeping the same
// vector type) before extracting.
//
// This code will bail out if the target type is fundamentally incompatible
// with vectors of the source type.
//
// Example of <16 x i8>, target type i32:
// Index range [4,8): v-----------v Will work.
// +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
// <16 x i8>: | | | | | | | | | | | | | | | | |
// <4 x i32>: | | | | |
// +-----------+-----------+-----------+-----------+
// Index range [6,10): ^-----------^ Needs an extra shuffle.
// Target type i40: ^--------------^ Won't work, bail.
if (isShuffleExtractingFromLHS(SVI, Mask)) {
Value *V = LHS;
unsigned MaskElems = Mask.size();
unsigned BegIdx = Mask.front();
VectorType *SrcTy = cast<VectorType>(V->getType());
unsigned VecBitWidth = SrcTy->getBitWidth();
unsigned SrcElemBitWidth = DL.getTypeSizeInBits(SrcTy->getElementType());
assert(SrcElemBitWidth && "vector elements must have a bitwidth");
unsigned SrcNumElems = SrcTy->getNumElements();
SmallVector<BitCastInst *, 8> BCs;
DenseMap<Type *, Value *> NewBCs;
for (User *U : SVI.users())
if (BitCastInst *BC = dyn_cast<BitCastInst>(U))
if (!BC->use_empty())
// Only visit bitcasts that weren't previously handled.
BCs.push_back(BC);
for (BitCastInst *BC : BCs) {
Type *TgtTy = BC->getDestTy();
unsigned TgtElemBitWidth = DL.getTypeSizeInBits(TgtTy);
if (!TgtElemBitWidth)
continue;
unsigned TgtNumElems = VecBitWidth / TgtElemBitWidth;
bool VecBitWidthsEqual = VecBitWidth == TgtNumElems * TgtElemBitWidth;
bool BegIsAligned = 0 == ((SrcElemBitWidth * BegIdx) % TgtElemBitWidth);
if (!VecBitWidthsEqual)
continue;
if (!VectorType::isValidElementType(TgtTy))
continue;
VectorType *CastSrcTy = VectorType::get(TgtTy, TgtNumElems);
if (!BegIsAligned) {
// Shuffle the input so [0,NumElements) contains the output, and
// [NumElems,SrcNumElems) is undef.
SmallVector<Constant *, 16> ShuffleMask(SrcNumElems,
UndefValue::get(Int32Ty));
for (unsigned I = 0, E = MaskElems, Idx = BegIdx; I != E; ++Idx, ++I)
ShuffleMask[I] = ConstantInt::get(Int32Ty, Idx);
V = Builder->CreateShuffleVector(V, UndefValue::get(V->getType()),
ConstantVector::get(ShuffleMask),
SVI.getName() + ".extract");
BegIdx = 0;
}
unsigned SrcElemsPerTgtElem = TgtElemBitWidth / SrcElemBitWidth;
assert(SrcElemsPerTgtElem);
BegIdx /= SrcElemsPerTgtElem;
bool BCAlreadyExists = NewBCs.find(CastSrcTy) != NewBCs.end();
auto *NewBC =
BCAlreadyExists
? NewBCs[CastSrcTy]
: Builder->CreateBitCast(V, CastSrcTy, SVI.getName() + ".bc");
if (!BCAlreadyExists)
NewBCs[CastSrcTy] = NewBC;
auto *Ext = Builder->CreateExtractElement(
NewBC, ConstantInt::get(Int32Ty, BegIdx), SVI.getName() + ".extract");
// The shufflevector isn't being replaced: the bitcast that used it
// is. InstCombine will visit the newly-created instructions.
replaceInstUsesWith(*BC, Ext);
MadeChange = true;
}
}
// If the LHS is a shufflevector itself, see if we can combine it with this
// one without producing an unusual shuffle.
// Cases that might be simplified:
// 1.
// x1=shuffle(v1,v2,mask1)
// x=shuffle(x1,undef,mask)
// ==>
// x=shuffle(v1,undef,newMask)
// newMask[i] = (mask[i] < x1.size()) ? mask1[mask[i]] : -1
// 2.
// x1=shuffle(v1,undef,mask1)
// x=shuffle(x1,x2,mask)
// where v1.size() == mask1.size()
// ==>
// x=shuffle(v1,x2,newMask)
// newMask[i] = (mask[i] < x1.size()) ? mask1[mask[i]] : mask[i]
// 3.
// x2=shuffle(v2,undef,mask2)
// x=shuffle(x1,x2,mask)
// where v2.size() == mask2.size()
// ==>
// x=shuffle(x1,v2,newMask)
// newMask[i] = (mask[i] < x1.size())
// ? mask[i] : mask2[mask[i]-x1.size()]+x1.size()
// 4.
// x1=shuffle(v1,undef,mask1)
// x2=shuffle(v2,undef,mask2)
// x=shuffle(x1,x2,mask)
// where v1.size() == v2.size()
// ==>
// x=shuffle(v1,v2,newMask)
// newMask[i] = (mask[i] < x1.size())
// ? mask1[mask[i]] : mask2[mask[i]-x1.size()]+v1.size()
//
// Here we are really conservative:
// we are absolutely afraid of producing a shuffle mask not in the input
// program, because the code gen may not be smart enough to turn a merged
// shuffle into two specific shuffles: it may produce worse code. As such,
// we only merge two shuffles if the result is either a splat or one of the
// input shuffle masks. In this case, merging the shuffles just removes
// one instruction, which we know is safe. This is good for things like
// turning: (splat(splat)) -> splat, or
// merge(V[0..n], V[n+1..2n]) -> V[0..2n]
ShuffleVectorInst* LHSShuffle = dyn_cast<ShuffleVectorInst>(LHS);
ShuffleVectorInst* RHSShuffle = dyn_cast<ShuffleVectorInst>(RHS);
if (LHSShuffle)
if (!isa<UndefValue>(LHSShuffle->getOperand(1)) && !isa<UndefValue>(RHS))
LHSShuffle = nullptr;
if (RHSShuffle)
if (!isa<UndefValue>(RHSShuffle->getOperand(1)))
RHSShuffle = nullptr;
if (!LHSShuffle && !RHSShuffle)
return MadeChange ? &SVI : nullptr;
Value* LHSOp0 = nullptr;
Value* LHSOp1 = nullptr;
Value* RHSOp0 = nullptr;
unsigned LHSOp0Width = 0;
unsigned RHSOp0Width = 0;
if (LHSShuffle) {
LHSOp0 = LHSShuffle->getOperand(0);
LHSOp1 = LHSShuffle->getOperand(1);
LHSOp0Width = cast<VectorType>(LHSOp0->getType())->getNumElements();
}
if (RHSShuffle) {
RHSOp0 = RHSShuffle->getOperand(0);
RHSOp0Width = cast<VectorType>(RHSOp0->getType())->getNumElements();
}
Value* newLHS = LHS;
Value* newRHS = RHS;
if (LHSShuffle) {
// case 1
if (isa<UndefValue>(RHS)) {
newLHS = LHSOp0;
newRHS = LHSOp1;
}
// case 2 or 4
else if (LHSOp0Width == LHSWidth) {
newLHS = LHSOp0;
}
}
// case 3 or 4
if (RHSShuffle && RHSOp0Width == LHSWidth) {
newRHS = RHSOp0;
}
// case 4
if (LHSOp0 == RHSOp0) {
newLHS = LHSOp0;
newRHS = nullptr;
}
if (newLHS == LHS && newRHS == RHS)
return MadeChange ? &SVI : nullptr;
SmallVector<int, 16> LHSMask;
SmallVector<int, 16> RHSMask;
if (newLHS != LHS)
LHSMask = LHSShuffle->getShuffleMask();
if (RHSShuffle && newRHS != RHS)
RHSMask = RHSShuffle->getShuffleMask();
unsigned newLHSWidth = (newLHS != LHS) ? LHSOp0Width : LHSWidth;
SmallVector<int, 16> newMask;
bool isSplat = true;
int SplatElt = -1;
// Create a new mask for the new ShuffleVectorInst so that the new
// ShuffleVectorInst is equivalent to the original one.
for (unsigned i = 0; i < VWidth; ++i) {
int eltMask;
if (Mask[i] < 0) {
// This element is an undef value.
eltMask = -1;
} else if (Mask[i] < (int)LHSWidth) {
// This element is from left hand side vector operand.
//
// If LHS is going to be replaced (case 1, 2, or 4), calculate the
// new mask value for the element.
if (newLHS != LHS) {
eltMask = LHSMask[Mask[i]];
// If the value selected is an undef value, explicitly specify it
// with a -1 mask value.
if (eltMask >= (int)LHSOp0Width && isa<UndefValue>(LHSOp1))
eltMask = -1;
} else
eltMask = Mask[i];
} else {
// This element is from right hand side vector operand
//
// If the value selected is an undef value, explicitly specify it
// with a -1 mask value. (case 1)
if (isa<UndefValue>(RHS))
eltMask = -1;
// If RHS is going to be replaced (case 3 or 4), calculate the
// new mask value for the element.
else if (newRHS != RHS) {
eltMask = RHSMask[Mask[i]-LHSWidth];
// If the value selected is an undef value, explicitly specify it
// with a -1 mask value.
if (eltMask >= (int)RHSOp0Width) {
assert(isa<UndefValue>(RHSShuffle->getOperand(1))
&& "should have been check above");
eltMask = -1;
}
} else
eltMask = Mask[i]-LHSWidth;
// If LHS's width is changed, shift the mask value accordingly.
// If newRHS == NULL, i.e. LHSOp0 == RHSOp0, we want to remap any
// references from RHSOp0 to LHSOp0, so we don't need to shift the mask.
// If newRHS == newLHS, we want to remap any references from newRHS to
// newLHS so that we can properly identify splats that may occur due to
// obfuscation across the two vectors.
if (eltMask >= 0 && newRHS != nullptr && newLHS != newRHS)
eltMask += newLHSWidth;
}
// Check if this could still be a splat.
if (eltMask >= 0) {
if (SplatElt >= 0 && SplatElt != eltMask)
isSplat = false;
SplatElt = eltMask;
}
newMask.push_back(eltMask);
}
// If the result mask is equal to one of the original shuffle masks,
// or is a splat, do the replacement.
if (isSplat || newMask == LHSMask || newMask == RHSMask || newMask == Mask) {
SmallVector<Constant*, 16> Elts;
for (unsigned i = 0, e = newMask.size(); i != e; ++i) {
if (newMask[i] < 0) {
Elts.push_back(UndefValue::get(Int32Ty));
} else {
Elts.push_back(ConstantInt::get(Int32Ty, newMask[i]));
}
}
if (!newRHS)
newRHS = UndefValue::get(newLHS->getType());
return new ShuffleVectorInst(newLHS, newRHS, ConstantVector::get(Elts));
}
// If the result mask is an identity, replace uses of this instruction with
// corresponding argument.
bool isLHSID, isRHSID;
recognizeIdentityMask(newMask, isLHSID, isRHSID);
if (isLHSID && VWidth == LHSOp0Width) return replaceInstUsesWith(SVI, newLHS);
if (isRHSID && VWidth == RHSOp0Width) return replaceInstUsesWith(SVI, newRHS);
return MadeChange ? &SVI : nullptr;
}