diff options
author | Dimitry Andric <dim@FreeBSD.org> | 2022-02-18 22:41:20 +0000 |
---|---|---|
committer | Dimitry Andric <dim@FreeBSD.org> | 2022-05-14 11:45:57 +0000 |
commit | d56accc7c3dcc897489b6a07834763a03b9f3d68 (patch) | |
tree | 918f41a708218122215937f4ab4e68b1a942da68 /contrib/llvm-project/llvm/lib/Transforms | |
parent | 8885dff6cec52378084211fdd9366a73833eceee (diff) | |
parent | 7eff647615f93a9aaff1997e1880b195dc3aabe6 (diff) | |
download | src-d56accc7c3dcc897489b6a07834763a03b9f3d68.tar.gz src-d56accc7c3dcc897489b6a07834763a03b9f3d68.zip |
Merge llvm-project release/14.x llvmorg-14.0.0-rc1-74-g4dc3cb8e3255
This updates llvm, clang, compiler-rt, libc++, libunwind, lld, lldb and
openmp to llvmorg-14.0.0-rc1-74-g4dc3cb8e3255.
PR: 261742
MFC after: 2 weeks
Diffstat (limited to 'contrib/llvm-project/llvm/lib/Transforms')
5 files changed, 88 insertions, 222 deletions
diff --git a/contrib/llvm-project/llvm/lib/Transforms/IPO/AttributorAttributes.cpp b/contrib/llvm-project/llvm/lib/Transforms/IPO/AttributorAttributes.cpp index 2d88e329e093..4e4f768ed2cb 100644 --- a/contrib/llvm-project/llvm/lib/Transforms/IPO/AttributorAttributes.cpp +++ b/contrib/llvm-project/llvm/lib/Transforms/IPO/AttributorAttributes.cpp @@ -236,7 +236,8 @@ static Value *constructPointer(Type *ResTy, Type *PtrElemTy, Value *Ptr, } // Ensure the result has the requested type. - Ptr = IRB.CreateBitOrPointerCast(Ptr, ResTy, Ptr->getName() + ".cast"); + Ptr = IRB.CreatePointerBitCastOrAddrSpaceCast(Ptr, ResTy, + Ptr->getName() + ".cast"); LLVM_DEBUG(dbgs() << "Constructed pointer: " << *Ptr << "\n"); return Ptr; @@ -2691,40 +2692,38 @@ struct AAUndefinedBehaviorImpl : public AAUndefinedBehavior { return true; }; - auto InspectReturnInstForUB = - [&](Value &V, const SmallSetVector<ReturnInst *, 4> RetInsts) { - // Check if a return instruction always cause UB or not - // Note: It is guaranteed that the returned position of the anchor - // scope has noundef attribute when this is called. - // We also ensure the return position is not "assumed dead" - // because the returned value was then potentially simplified to - // `undef` in AAReturnedValues without removing the `noundef` - // attribute yet. - - // When the returned position has noundef attriubte, UB occur in the - // following cases. - // (1) Returned value is known to be undef. - // (2) The value is known to be a null pointer and the returned - // position has nonnull attribute (because the returned value is - // poison). - bool FoundUB = false; - if (isa<UndefValue>(V)) { - FoundUB = true; - } else { - if (isa<ConstantPointerNull>(V)) { - auto &NonNullAA = A.getAAFor<AANonNull>( - *this, IRPosition::returned(*getAnchorScope()), - DepClassTy::NONE); - if (NonNullAA.isKnownNonNull()) - FoundUB = true; - } - } + auto InspectReturnInstForUB = [&](Instruction &I) { + auto &RI = cast<ReturnInst>(I); + // Either we stopped and the appropriate action was taken, + // or we got back a simplified return value to continue. + Optional<Value *> SimplifiedRetValue = + stopOnUndefOrAssumed(A, RI.getReturnValue(), &I); + if (!SimplifiedRetValue.hasValue() || !SimplifiedRetValue.getValue()) + return true; - if (FoundUB) - for (ReturnInst *RI : RetInsts) - KnownUBInsts.insert(RI); - return true; - }; + // Check if a return instruction always cause UB or not + // Note: It is guaranteed that the returned position of the anchor + // scope has noundef attribute when this is called. + // We also ensure the return position is not "assumed dead" + // because the returned value was then potentially simplified to + // `undef` in AAReturnedValues without removing the `noundef` + // attribute yet. + + // When the returned position has noundef attriubte, UB occurs in the + // following cases. + // (1) Returned value is known to be undef. + // (2) The value is known to be a null pointer and the returned + // position has nonnull attribute (because the returned value is + // poison). + if (isa<ConstantPointerNull>(*SimplifiedRetValue)) { + auto &NonNullAA = A.getAAFor<AANonNull>( + *this, IRPosition::returned(*getAnchorScope()), DepClassTy::NONE); + if (NonNullAA.isKnownNonNull()) + KnownUBInsts.insert(&I); + } + + return true; + }; bool UsedAssumedInformation = false; A.checkForAllInstructions(InspectMemAccessInstForUB, *this, @@ -2747,8 +2746,9 @@ struct AAUndefinedBehaviorImpl : public AAUndefinedBehavior { auto &RetPosNoUndefAA = A.getAAFor<AANoUndef>(*this, ReturnIRP, DepClassTy::NONE); if (RetPosNoUndefAA.isKnownNoUndef()) - A.checkForAllReturnedValuesAndReturnInsts(InspectReturnInstForUB, - *this); + A.checkForAllInstructions(InspectReturnInstForUB, *this, + {Instruction::Ret}, UsedAssumedInformation, + /* CheckBBLivenessOnly */ true); } } @@ -6749,8 +6749,8 @@ struct AAPrivatizablePtrArgument final : public AAPrivatizablePtrImpl { Type *PrivPtrType = PrivType->getPointerTo(); if (Base->getType() != PrivPtrType) - Base = BitCastInst::CreateBitOrPointerCast(Base, PrivPtrType, "", - ACS.getInstruction()); + Base = BitCastInst::CreatePointerBitCastOrAddrSpaceCast( + Base, PrivPtrType, "", ACS.getInstruction()); // Traverse the type, build GEPs and loads. if (auto *PrivStructType = dyn_cast<StructType>(PrivType)) { @@ -6817,14 +6817,16 @@ struct AAPrivatizablePtrArgument final : public AAPrivatizablePtrImpl { Function &ReplacementFn, Function::arg_iterator ArgIt) { BasicBlock &EntryBB = ReplacementFn.getEntryBlock(); Instruction *IP = &*EntryBB.getFirstInsertionPt(); - Instruction *AI = new AllocaInst(PrivatizableType.getValue(), 0, + const DataLayout &DL = IP->getModule()->getDataLayout(); + unsigned AS = DL.getAllocaAddrSpace(); + Instruction *AI = new AllocaInst(PrivatizableType.getValue(), AS, Arg->getName() + ".priv", IP); createInitialization(PrivatizableType.getValue(), *AI, ReplacementFn, ArgIt->getArgNo(), *IP); if (AI->getType() != Arg->getType()) - AI = - BitCastInst::CreateBitOrPointerCast(AI, Arg->getType(), "", IP); + AI = BitCastInst::CreatePointerBitCastOrAddrSpaceCast( + AI, Arg->getType(), "", IP); Arg->replaceAllUsesWith(AI); for (CallInst *CI : TailCalls) diff --git a/contrib/llvm-project/llvm/lib/Transforms/IPO/OpenMPOpt.cpp b/contrib/llvm-project/llvm/lib/Transforms/IPO/OpenMPOpt.cpp index 2d765fb6ce6d..520b6ebf9e74 100644 --- a/contrib/llvm-project/llvm/lib/Transforms/IPO/OpenMPOpt.cpp +++ b/contrib/llvm-project/llvm/lib/Transforms/IPO/OpenMPOpt.cpp @@ -1458,7 +1458,6 @@ private: case Intrinsic::nvvm_barrier0_and: case Intrinsic::nvvm_barrier0_or: case Intrinsic::nvvm_barrier0_popc: - case Intrinsic::amdgcn_s_barrier: return true; default: break; diff --git a/contrib/llvm-project/llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp b/contrib/llvm-project/llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp index 8f5933b7bd71..ddc747a2ca29 100644 --- a/contrib/llvm-project/llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp +++ b/contrib/llvm-project/llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp @@ -655,10 +655,13 @@ Value *InferAddressSpacesImpl::cloneInstructionWithNewAddressSpace( case Instruction::IntToPtr: { assert(isNoopPtrIntCastPair(cast<Operator>(I), *DL, TTI)); Value *Src = cast<Operator>(I->getOperand(0))->getOperand(0); - assert(Src->getType()->getPointerAddressSpace() == NewAddrSpace); - if (Src->getType() != NewPtrType) - return new BitCastInst(Src, NewPtrType); - return Src; + if (Src->getType() == NewPtrType) + return Src; + + // If we had a no-op inttoptr/ptrtoint pair, we may still have inferred a + // source address space from a generic pointer source need to insert a cast + // back. + return CastInst::CreatePointerBitCastOrAddrSpaceCast(Src, NewPtrType); } default: llvm_unreachable("Unexpected opcode"); diff --git a/contrib/llvm-project/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/contrib/llvm-project/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp index 3290439ecd07..21c16f07e237 100644 --- a/contrib/llvm-project/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp +++ b/contrib/llvm-project/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp @@ -1701,6 +1701,11 @@ public: private: unsigned NumPredStores = 0; + /// Convenience function that returns the value of vscale_range iff + /// vscale_range.min == vscale_range.max or otherwise returns the value + /// returned by the corresponding TLI method. + Optional<unsigned> getVScaleForTuning() const; + /// \return An upper bound for the vectorization factors for both /// fixed and scalable vectorization, where the minimum-known number of /// elements is a power-of-2 larger than zero. If scalable vectorization is @@ -5600,6 +5605,18 @@ ElementCount LoopVectorizationCostModel::getMaximizedVFForTarget( return MaxVF; } +Optional<unsigned> LoopVectorizationCostModel::getVScaleForTuning() const { + if (TheFunction->hasFnAttribute(Attribute::VScaleRange)) { + auto Attr = TheFunction->getFnAttribute(Attribute::VScaleRange); + auto Min = Attr.getVScaleRangeMin(); + auto Max = Attr.getVScaleRangeMax(); + if (Max && Min == Max) + return Max; + } + + return TTI.getVScaleForTuning(); +} + bool LoopVectorizationCostModel::isMoreProfitable( const VectorizationFactor &A, const VectorizationFactor &B) const { InstructionCost CostA = A.Cost; @@ -5624,7 +5641,7 @@ bool LoopVectorizationCostModel::isMoreProfitable( // Improve estimate for the vector width if it is scalable. unsigned EstimatedWidthA = A.Width.getKnownMinValue(); unsigned EstimatedWidthB = B.Width.getKnownMinValue(); - if (Optional<unsigned> VScale = TTI.getVScaleForTuning()) { + if (Optional<unsigned> VScale = getVScaleForTuning()) { if (A.Width.isScalable()) EstimatedWidthA *= VScale.getValue(); if (B.Width.isScalable()) @@ -5673,7 +5690,7 @@ VectorizationFactor LoopVectorizationCostModel::selectVectorizationFactor( #ifndef NDEBUG unsigned AssumedMinimumVscale = 1; - if (Optional<unsigned> VScale = TTI.getVScaleForTuning()) + if (Optional<unsigned> VScale = getVScaleForTuning()) AssumedMinimumVscale = VScale.getValue(); unsigned Width = Candidate.Width.isScalable() @@ -5885,8 +5902,20 @@ LoopVectorizationCostModel::selectEpilogueVectorizationFactor( return Result; } + // If MainLoopVF = vscale x 2, and vscale is expected to be 4, then we know + // the main loop handles 8 lanes per iteration. We could still benefit from + // vectorizing the epilogue loop with VF=4. + ElementCount EstimatedRuntimeVF = MainLoopVF; + if (MainLoopVF.isScalable()) { + EstimatedRuntimeVF = ElementCount::getFixed(MainLoopVF.getKnownMinValue()); + if (Optional<unsigned> VScale = getVScaleForTuning()) + EstimatedRuntimeVF *= VScale.getValue(); + } + for (auto &NextVF : ProfitableVFs) - if (ElementCount::isKnownLT(NextVF.Width, MainLoopVF) && + if (((!NextVF.Width.isScalable() && MainLoopVF.isScalable() && + ElementCount::isKnownLT(NextVF.Width, EstimatedRuntimeVF)) || + ElementCount::isKnownLT(NextVF.Width, MainLoopVF)) && (Result.Width.isScalar() || isMoreProfitable(NextVF, Result)) && LVP.hasPlanWithVF(NextVF.Width)) Result = NextVF; diff --git a/contrib/llvm-project/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp b/contrib/llvm-project/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp index 15b349f53fd9..25bf69729c70 100644 --- a/contrib/llvm-project/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp +++ b/contrib/llvm-project/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp @@ -471,36 +471,17 @@ static bool isValidForAlternation(unsigned Opcode) { return true; } -static InstructionsState getSameOpcode(ArrayRef<Value *> VL, - unsigned BaseIndex = 0); - -/// Checks if the provided operands of 2 cmp instructions are compatible, i.e. -/// compatible instructions or constants, or just some other regular values. -static bool areCompatibleCmpOps(Value *BaseOp0, Value *BaseOp1, Value *Op0, - Value *Op1) { - return (isConstant(BaseOp0) && isConstant(Op0)) || - (isConstant(BaseOp1) && isConstant(Op1)) || - (!isa<Instruction>(BaseOp0) && !isa<Instruction>(Op0) && - !isa<Instruction>(BaseOp1) && !isa<Instruction>(Op1)) || - getSameOpcode({BaseOp0, Op0}).getOpcode() || - getSameOpcode({BaseOp1, Op1}).getOpcode(); -} - /// \returns analysis of the Instructions in \p VL described in /// InstructionsState, the Opcode that we suppose the whole list /// could be vectorized even if its structure is diverse. static InstructionsState getSameOpcode(ArrayRef<Value *> VL, - unsigned BaseIndex) { + unsigned BaseIndex = 0) { // Make sure these are all Instructions. if (llvm::any_of(VL, [](Value *V) { return !isa<Instruction>(V); })) return InstructionsState(VL[BaseIndex], nullptr, nullptr); bool IsCastOp = isa<CastInst>(VL[BaseIndex]); bool IsBinOp = isa<BinaryOperator>(VL[BaseIndex]); - bool IsCmpOp = isa<CmpInst>(VL[BaseIndex]); - CmpInst::Predicate BasePred = - IsCmpOp ? cast<CmpInst>(VL[BaseIndex])->getPredicate() - : CmpInst::BAD_ICMP_PREDICATE; unsigned Opcode = cast<Instruction>(VL[BaseIndex])->getOpcode(); unsigned AltOpcode = Opcode; unsigned AltIndex = BaseIndex; @@ -533,57 +514,6 @@ static InstructionsState getSameOpcode(ArrayRef<Value *> VL, continue; } } - } else if (IsCmpOp && isa<CmpInst>(VL[Cnt])) { - auto *BaseInst = cast<Instruction>(VL[BaseIndex]); - auto *Inst = cast<Instruction>(VL[Cnt]); - Type *Ty0 = BaseInst->getOperand(0)->getType(); - Type *Ty1 = Inst->getOperand(0)->getType(); - if (Ty0 == Ty1) { - Value *BaseOp0 = BaseInst->getOperand(0); - Value *BaseOp1 = BaseInst->getOperand(1); - Value *Op0 = Inst->getOperand(0); - Value *Op1 = Inst->getOperand(1); - CmpInst::Predicate CurrentPred = - cast<CmpInst>(VL[Cnt])->getPredicate(); - CmpInst::Predicate SwappedCurrentPred = - CmpInst::getSwappedPredicate(CurrentPred); - // Check for compatible operands. If the corresponding operands are not - // compatible - need to perform alternate vectorization. - if (InstOpcode == Opcode) { - if (BasePred == CurrentPred && - areCompatibleCmpOps(BaseOp0, BaseOp1, Op0, Op1)) - continue; - if (BasePred == SwappedCurrentPred && - areCompatibleCmpOps(BaseOp0, BaseOp1, Op1, Op0)) - continue; - if (E == 2 && - (BasePred == CurrentPred || BasePred == SwappedCurrentPred)) - continue; - auto *AltInst = cast<CmpInst>(VL[AltIndex]); - CmpInst::Predicate AltPred = AltInst->getPredicate(); - Value *AltOp0 = AltInst->getOperand(0); - Value *AltOp1 = AltInst->getOperand(1); - // Check if operands are compatible with alternate operands. - if (AltPred == CurrentPred && - areCompatibleCmpOps(AltOp0, AltOp1, Op0, Op1)) - continue; - if (AltPred == SwappedCurrentPred && - areCompatibleCmpOps(AltOp0, AltOp1, Op1, Op0)) - continue; - } - if (BaseIndex == AltIndex) { - assert(isValidForAlternation(Opcode) && - isValidForAlternation(InstOpcode) && - "Cast isn't safe for alternation, logic needs to be updated!"); - AltIndex = Cnt; - continue; - } - auto *AltInst = cast<CmpInst>(VL[AltIndex]); - CmpInst::Predicate AltPred = AltInst->getPredicate(); - if (BasePred == CurrentPred || BasePred == SwappedCurrentPred || - AltPred == CurrentPred || AltPred == SwappedCurrentPred) - continue; - } } else if (InstOpcode == Opcode || InstOpcode == AltOpcode) continue; return InstructionsState(VL[BaseIndex], nullptr, nullptr); @@ -4424,41 +4354,9 @@ void BoUpSLP::buildTree_rec(ArrayRef<Value *> VL, unsigned Depth, LLVM_DEBUG(dbgs() << "SLP: added a ShuffleVector op.\n"); // Reorder operands if reordering would enable vectorization. - auto *CI = dyn_cast<CmpInst>(VL0); - if (isa<BinaryOperator>(VL0) || CI) { + if (isa<BinaryOperator>(VL0)) { ValueList Left, Right; - if (!CI || all_of(VL, [](Value *V) { - return cast<CmpInst>(V)->isCommutative(); - })) { - reorderInputsAccordingToOpcode(VL, Left, Right, *DL, *SE, *this); - } else { - CmpInst::Predicate P0 = CI->getPredicate(); - CmpInst::Predicate AltP0 = cast<CmpInst>(S.AltOp)->getPredicate(); - CmpInst::Predicate AltP0Swapped = CmpInst::getSwappedPredicate(AltP0); - Value *BaseOp0 = VL0->getOperand(0); - Value *BaseOp1 = VL0->getOperand(1); - // Collect operands - commute if it uses the swapped predicate or - // alternate operation. - for (Value *V : VL) { - auto *Cmp = cast<CmpInst>(V); - Value *LHS = Cmp->getOperand(0); - Value *RHS = Cmp->getOperand(1); - CmpInst::Predicate CurrentPred = CI->getPredicate(); - CmpInst::Predicate CurrentPredSwapped = - CmpInst::getSwappedPredicate(CurrentPred); - if (P0 == AltP0 || P0 == AltP0Swapped) { - if ((P0 == CurrentPred && - !areCompatibleCmpOps(BaseOp0, BaseOp1, LHS, RHS)) || - (P0 == CurrentPredSwapped && - !areCompatibleCmpOps(BaseOp0, BaseOp1, RHS, LHS))) - std::swap(LHS, RHS); - } else if (!areCompatibleCmpOps(BaseOp0, BaseOp1, LHS, RHS)) { - std::swap(LHS, RHS); - } - Left.push_back(LHS); - Right.push_back(RHS); - } - } + reorderInputsAccordingToOpcode(VL, Left, Right, *DL, *SE, *this); TE->setOperand(0, Left); TE->setOperand(1, Right); buildTree_rec(Left, Depth + 1, {TE, 0}); @@ -5390,8 +5288,7 @@ InstructionCost BoUpSLP::getEntryCost(const TreeEntry *E, ((Instruction::isBinaryOp(E->getOpcode()) && Instruction::isBinaryOp(E->getAltOpcode())) || (Instruction::isCast(E->getOpcode()) && - Instruction::isCast(E->getAltOpcode())) || - (isa<CmpInst>(VL0) && isa<CmpInst>(E->getAltOp()))) && + Instruction::isCast(E->getAltOpcode()))) && "Invalid Shuffle Vector Operand"); InstructionCost ScalarCost = 0; if (NeedToShuffleReuses) { @@ -5439,14 +5336,6 @@ InstructionCost BoUpSLP::getEntryCost(const TreeEntry *E, VecCost = TTI->getArithmeticInstrCost(E->getOpcode(), VecTy, CostKind); VecCost += TTI->getArithmeticInstrCost(E->getAltOpcode(), VecTy, CostKind); - } else if (auto *CI0 = dyn_cast<CmpInst>(VL0)) { - VecCost = TTI->getCmpSelInstrCost(E->getOpcode(), ScalarTy, - Builder.getInt1Ty(), - CI0->getPredicate(), CostKind, VL0); - VecCost += TTI->getCmpSelInstrCost( - E->getOpcode(), ScalarTy, Builder.getInt1Ty(), - cast<CmpInst>(E->getAltOp())->getPredicate(), CostKind, - E->getAltOp()); } else { Type *Src0SclTy = E->getMainOp()->getOperand(0)->getType(); Type *Src1SclTy = E->getAltOp()->getOperand(0)->getType(); @@ -5463,29 +5352,6 @@ InstructionCost BoUpSLP::getEntryCost(const TreeEntry *E, E->Scalars, E->ReorderIndices, E->ReuseShuffleIndices, [E](Instruction *I) { assert(E->isOpcodeOrAlt(I) && "Unexpected main/alternate opcode"); - if (auto *CI0 = dyn_cast<CmpInst>(E->getMainOp())) { - auto *AltCI0 = cast<CmpInst>(E->getAltOp()); - auto *CI = cast<CmpInst>(I); - CmpInst::Predicate P0 = CI0->getPredicate(); - CmpInst::Predicate AltP0 = AltCI0->getPredicate(); - CmpInst::Predicate AltP0Swapped = - CmpInst::getSwappedPredicate(AltP0); - CmpInst::Predicate CurrentPred = CI->getPredicate(); - CmpInst::Predicate CurrentPredSwapped = - CmpInst::getSwappedPredicate(CurrentPred); - if (P0 == AltP0 || P0 == AltP0Swapped) { - // Alternate cmps have same/swapped predicate as main cmps but - // different order of compatible operands. - return !( - (P0 == CurrentPred && - areCompatibleCmpOps(CI0->getOperand(0), CI0->getOperand(1), - I->getOperand(0), I->getOperand(1))) || - (P0 == CurrentPredSwapped && - areCompatibleCmpOps(CI0->getOperand(0), CI0->getOperand(1), - I->getOperand(1), I->getOperand(0)))); - } - return CurrentPred != P0 && CurrentPredSwapped != P0; - } return I->getOpcode() == E->getAltOpcode(); }, Mask); @@ -6968,12 +6834,11 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { ((Instruction::isBinaryOp(E->getOpcode()) && Instruction::isBinaryOp(E->getAltOpcode())) || (Instruction::isCast(E->getOpcode()) && - Instruction::isCast(E->getAltOpcode())) || - (isa<CmpInst>(VL0) && isa<CmpInst>(E->getAltOp()))) && + Instruction::isCast(E->getAltOpcode()))) && "Invalid Shuffle Vector Operand"); Value *LHS = nullptr, *RHS = nullptr; - if (Instruction::isBinaryOp(E->getOpcode()) || isa<CmpInst>(VL0)) { + if (Instruction::isBinaryOp(E->getOpcode())) { setInsertPointAfterBundle(E); LHS = vectorizeTree(E->getOperand(0)); RHS = vectorizeTree(E->getOperand(1)); @@ -6993,15 +6858,6 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { static_cast<Instruction::BinaryOps>(E->getOpcode()), LHS, RHS); V1 = Builder.CreateBinOp( static_cast<Instruction::BinaryOps>(E->getAltOpcode()), LHS, RHS); - } else if (auto *CI0 = dyn_cast<CmpInst>(VL0)) { - V0 = Builder.CreateCmp(CI0->getPredicate(), LHS, RHS); - auto *AltCI = cast<CmpInst>(E->getAltOp()); - CmpInst::Predicate AltPred = AltCI->getPredicate(); - unsigned AltIdx = - std::distance(E->Scalars.begin(), find(E->Scalars, AltCI)); - if (AltCI->getOperand(0) != E->getOperand(0)[AltIdx]) - AltPred = CmpInst::getSwappedPredicate(AltPred); - V1 = Builder.CreateCmp(AltPred, LHS, RHS); } else { V0 = Builder.CreateCast( static_cast<Instruction::CastOps>(E->getOpcode()), LHS, VecTy); @@ -7026,29 +6882,6 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { E->Scalars, E->ReorderIndices, E->ReuseShuffleIndices, [E](Instruction *I) { assert(E->isOpcodeOrAlt(I) && "Unexpected main/alternate opcode"); - if (auto *CI0 = dyn_cast<CmpInst>(E->getMainOp())) { - auto *AltCI0 = cast<CmpInst>(E->getAltOp()); - auto *CI = cast<CmpInst>(I); - CmpInst::Predicate P0 = CI0->getPredicate(); - CmpInst::Predicate AltP0 = AltCI0->getPredicate(); - CmpInst::Predicate AltP0Swapped = - CmpInst::getSwappedPredicate(AltP0); - CmpInst::Predicate CurrentPred = CI->getPredicate(); - CmpInst::Predicate CurrentPredSwapped = - CmpInst::getSwappedPredicate(CurrentPred); - if (P0 == AltP0 || P0 == AltP0Swapped) { - // Alternate cmps have same/swapped predicate as main cmps but - // different order of compatible operands. - return !( - (P0 == CurrentPred && - areCompatibleCmpOps(CI0->getOperand(0), CI0->getOperand(1), - I->getOperand(0), I->getOperand(1))) || - (P0 == CurrentPredSwapped && - areCompatibleCmpOps(CI0->getOperand(0), CI0->getOperand(1), - I->getOperand(1), I->getOperand(0)))); - } - return CurrentPred != P0 && CurrentPredSwapped != P0; - } return I->getOpcode() == E->getAltOpcode(); }, Mask, &OpScalars, &AltScalars); |