llvm::InnerLoopVectorizer Class Reference

InnerLoopVectorizer vectorizes loops which contain only one basic block to a specified vectorization factor (VF). More...

Inheritance diagram for llvm::InnerLoopVectorizer:

Collaboration diagram for llvm::InnerLoopVectorizer:

Public Types
using	VectorParts = SmallVector< Value *, 2 >
	A type for vectorized values in the new loop. More...

Public Member Functions
	InnerLoopVectorizer (Loop *OrigLoop, PredicatedScalarEvolution &PSE, LoopInfo *LI, DominatorTree *DT, const TargetLibraryInfo *TLI, const TargetTransformInfo *TTI, AssumptionCache *AC, OptimizationRemarkEmitter *ORE, ElementCount VecWidth, ElementCount MinProfitableTripCount, unsigned UnrollFactor, LoopVectorizationLegality *LVL, LoopVectorizationCostModel *CM, BlockFrequencyInfo *BFI, ProfileSummaryInfo *PSI, GeneratedRTChecks &RTChecks)
virtual	~InnerLoopVectorizer ()=default
virtual std::pair< BasicBlock *, Value * >	createVectorizedLoopSkeleton ()
	Create a new empty loop that will contain vectorized instructions later on, while the old loop will be used as the scalar remainder. More...
void	fixVectorizedLoop (VPTransformState &State, VPlan &Plan)
	Fix the vectorized code, taking care of header phi's, live-outs, and more. More...
bool	areSafetyChecksAdded ()
void	scalarizeInstruction (const Instruction *Instr, VPReplicateRecipe *RepRecipe, const VPIteration &Instance, bool IfPredicateInstr, VPTransformState &State)
	A helper function to scalarize a single Instruction in the innermost loop. More...
void	packScalarIntoVectorValue (VPValue *Def, const VPIteration &Instance, VPTransformState &State)
	Construct the vector value of a scalarized value V one lane at a time. More...
void	vectorizeInterleaveGroup (const InterleaveGroup< Instruction > *Group, ArrayRef< VPValue * > VPDefs, VPTransformState &State, VPValue *Addr, ArrayRef< VPValue * > StoredValues, VPValue *BlockInMask=nullptr)
	Try to vectorize interleaved access group Group with the base address given in Addr, optionally masking the vector operations if BlockInMask is non-null. More...
void	fixNonInductionPHIs (VPlan &Plan, VPTransformState &State)
	Fix the non-induction PHIs in Plan. More...
bool	useOrderedReductions (const RecurrenceDescriptor &RdxDesc)
	Returns true if the reordering of FP operations is not allowed, but we are able to vectorize with strict in-order reductions for the given RdxDesc. More...
virtual Value *	getBroadcastInstrs (Value *V)
	Create a broadcast instruction. More...
PHINode *	getReductionResumeValue (const RecurrenceDescriptor &RdxDesc)
PHINode *	createInductionResumeValue (PHINode *OrigPhi, const InductionDescriptor &ID, ArrayRef< BasicBlock * > BypassBlocks, std::pair< BasicBlock *, Value * > AdditionalBypass={nullptr, nullptr})
	Create a new phi node for the induction variable OrigPhi to resume iteration count in the scalar epilogue, from where the vectorized loop left off. More...

Protected Types
using	PhiVector = SmallVector< PHINode *, 4 >
	A small list of PHINodes. More...
using	ScalarParts = SmallVector< SmallVector< Value *, 4 >, 2 >
	A type for scalarized values in the new loop. More...

Protected Member Functions
void	fixupIVUsers (PHINode *OrigPhi, const InductionDescriptor &II, Value *VectorTripCount, Value *EndValue, BasicBlock *MiddleBlock, BasicBlock *VectorHeader, VPlan &Plan)
	Set up the values of the IVs correctly when exiting the vector loop. More...
void	fixCrossIterationPHIs (VPTransformState &State)
	Handle all cross-iteration phis in the header. More...
void	fixFixedOrderRecurrence (VPFirstOrderRecurrencePHIRecipe *PhiR, VPTransformState &State)
	Create the exit value of first order recurrences in the middle block and update their users. More...
void	fixReduction (VPReductionPHIRecipe *Phi, VPTransformState &State)
	Create code for the loop exit value of the reduction. More...
void	clearReductionWrapFlags (VPReductionPHIRecipe *PhiR, VPTransformState &State)
	Clear NSW/NUW flags from reduction instructions if necessary. More...
void	sinkScalarOperands (Instruction *PredInst)
	Iteratively sink the scalarized operands of a predicated instruction into the block that was created for it. More...
void	truncateToMinimalBitwidths (VPTransformState &State)
	Shrinks vector element sizes to the smallest bitwidth they can be legally represented as. More...
Value *	getOrCreateTripCount (BasicBlock *InsertBlock)
	Returns (and creates if needed) the original loop trip count. More...
Value *	getOrCreateVectorTripCount (BasicBlock *InsertBlock)
	Returns (and creates if needed) the trip count of the widened loop. More...
Value *	createBitOrPointerCast (Value *V, VectorType *DstVTy, const DataLayout &DL)
	Returns a bitcasted value to the requested vector type. More...
void	emitIterationCountCheck (BasicBlock *Bypass)
	Emit a bypass check to see if the vector trip count is zero, including if it overflows. More...
BasicBlock *	emitSCEVChecks (BasicBlock *Bypass)
	Emit a bypass check to see if all of the SCEV assumptions we've had to make are correct. More...
BasicBlock *	emitMemRuntimeChecks (BasicBlock *Bypass)
	Emit bypass checks to check any memory assumptions we may have made. More...
void	createVectorLoopSkeleton (StringRef Prefix)
	Emit basic blocks (prefixed with Prefix) for the iteration check, vector loop preheader, middle block and scalar preheader. More...
void	createInductionResumeValues (std::pair< BasicBlock *, Value * > AdditionalBypass={nullptr, nullptr})
	Create new phi nodes for the induction variables to resume iteration count in the scalar epilogue, from where the vectorized loop left off. More...
BasicBlock *	completeLoopSkeleton (MDNode *OrigLoopID)
	Complete the loop skeleton by adding debug MDs, creating appropriate conditional branches in the middle block, preparing the builder and running the verifier. More...
void	collectPoisonGeneratingRecipes (VPTransformState &State)
	Collect poison-generating recipes that may generate a poison value that is used after vectorization, even when their operands are not poison. More...
virtual void	printDebugTracesAtStart ()
	Allow subclasses to override and print debug traces before/after vplan execution, when trace information is requested. More...
virtual void	printDebugTracesAtEnd ()

Protected Attributes
Loop *	OrigLoop
	The original loop. More...
PredicatedScalarEvolution &	PSE
	A wrapper around ScalarEvolution used to add runtime SCEV checks. More...
LoopInfo *	LI
	Loop Info. More...
DominatorTree *	DT
	Dominator Tree. More...
AAResults *	AA
	Alias Analysis. More...
const TargetLibraryInfo *	TLI
	Target Library Info. More...
const TargetTransformInfo *	TTI
	Target Transform Info. More...
AssumptionCache *	AC
	Assumption Cache. More...
OptimizationRemarkEmitter *	ORE
	Interface to emit optimization remarks. More...
ElementCount	VF
	The vectorization SIMD factor to use. More...
ElementCount	MinProfitableTripCount
unsigned	UF
	The vectorization unroll factor to use. More...
IRBuilder	Builder
	The builder that we use. More...
BasicBlock *	LoopVectorPreHeader
	The vector-loop preheader. More...
BasicBlock *	LoopScalarPreHeader
	The scalar-loop preheader. More...
BasicBlock *	LoopMiddleBlock
	Middle Block between the vector and the scalar. More...
BasicBlock *	LoopExitBlock
	The unique ExitBlock of the scalar loop if one exists. More...
BasicBlock *	LoopScalarBody
	The scalar loop body. More...
SmallVector< BasicBlock *, 4 >	LoopBypassBlocks
	A list of all bypass blocks. The first block is the entry of the loop. More...
SmallVector< Instruction *, 4 >	PredicatedInstructions
	Store instructions that were predicated. More...
Value *	TripCount = nullptr
	Trip count of the original loop. More...
Value *	VectorTripCount = nullptr
	Trip count of the widened loop (TripCount - TripCount % (VF*UF)) More...
LoopVectorizationLegality *	Legal
	The legality analysis. More...
LoopVectorizationCostModel *	Cost
	The profitablity analysis. More...
bool	AddedSafetyChecks = false
DenseMap< PHINode *, Value * >	IVEndValues
BlockFrequencyInfo *	BFI
	BFI and PSI are used to check for profile guided size optimizations. More...
ProfileSummaryInfo *	PSI
bool	OptForSizeBasedOnProfile
GeneratedRTChecks &	RTChecks
	Structure to hold information about generated runtime checks, responsible for cleaning the checks, if vectorization turns out unprofitable. More...
SmallMapVector< const RecurrenceDescriptor *, PHINode *, 4 >	ReductionResumeValues

Friends
class	LoopVectorizationPlanner

Detailed Description

InnerLoopVectorizer vectorizes loops which contain only one basic block to a specified vectorization factor (VF).

This class performs the widening of scalars into vectors, or multiple scalars. This class also implements the following features:

• It inserts an epilogue loop for handling loops that don't have iteration counts that are known to be a multiple of the vectorization factor.
• It handles the code generation for reduction variables.
• Scalarization (implementation using scalars) of un-vectorizable instructions. InnerLoopVectorizer does not perform any vectorization-legality checks, and relies on the caller to check for the different legality aspects. The InnerLoopVectorizer relies on the LoopVectorizationLegality class to provide information about the induction and reduction variables that were found to a given vectorization factor.

Definition at line 443 of file LoopVectorize.cpp.

Member Typedef Documentation

PhiVector

using llvm::InnerLoopVectorizer::PhiVector = SmallVector<PHINode *, 4>

protected

A small list of PHINodes.

Definition at line 549 of file LoopVectorize.cpp.

ScalarParts

using llvm::InnerLoopVectorizer::ScalarParts = SmallVector<SmallVector<Value *, 4>, 2>

protected

A type for scalarized values in the new loop.

Each value from the original loop, when scalarized, is represented by UF x VF scalar values in the new unrolled loop, where UF is the unroll factor and VF is the vectorization factor.

Definition at line 555 of file LoopVectorize.cpp.

VectorParts

using llvm::InnerLoopVectorizer::VectorParts = SmallVector<Value *, 2>

A type for vectorized values in the new loop.

Each value from the original loop, when vectorized, is represented by UF vector values in the new unrolled loop, where UF is the unroll factor.

Definition at line 490 of file LoopVectorize.cpp.

Constructor & Destructor Documentation

InnerLoopVectorizer()

llvm::InnerLoopVectorizer::InnerLoopVectorizer ( Loop *  OrigLoop, PredicatedScalarEvolution &  PSE, LoopInfo *  LI, DominatorTree *  DT, const TargetLibraryInfo *  TLI, const TargetTransformInfo *  TTI, AssumptionCache *  AC, OptimizationRemarkEmitter *  ORE, ElementCount  VecWidth, ElementCount  MinProfitableTripCount, unsigned  UnrollFactor, LoopVectorizationLegality *  LVL, LoopVectorizationCostModel *  CM, BlockFrequencyInfo *  BFI, ProfileSummaryInfo *  PSI, GeneratedRTChecks &  RTChecks  )

inline

Definition at line 445 of file LoopVectorize.cpp.

References BFI, llvm::LoopBase< BlockT, LoopT >::getHeader(), llvm::IRPass, llvm::UnivariateLinearPolyBase< LeafTy >::isZero(), MinProfitableTripCount, OptForSizeBasedOnProfile, OrigLoop, PSI, and llvm::shouldOptimizeForSize().

~InnerLoopVectorizer()

virtual llvm::InnerLoopVectorizer::~InnerLoopVectorizer ( )

virtualdefault

Member Function Documentation

areSafetyChecksAdded()

bool llvm::InnerLoopVectorizer::areSafetyChecksAdded ( )

inline

Definition at line 485 of file LoopVectorize.cpp.

References AddedSafetyChecks.

Referenced by llvm::LoopVectorizePass::processLoop().

clearReductionWrapFlags()

void InnerLoopVectorizer::clearReductionWrapFlags ( VPReductionPHIRecipe *  PhiR, VPTransformState &  State  )

protected

Clear NSW/NUW flags from reduction instructions if necessary.

Definition at line 4060 of file LoopVectorize.cpp.

References llvm::Add, llvm::VPTransformState::get(), llvm::VPReductionPHIRecipe::getRecurrenceDescriptor(), llvm::RecurrenceDescriptor::getRecurrenceKind(), llvm::SmallPtrSetImpl< PtrType >::insert(), llvm::Mul, llvm::SmallVectorImpl< T >::pop_back_val(), UF, and llvm::VPValue::users().

Referenced by fixReduction().

collectPoisonGeneratingRecipes()

void InnerLoopVectorizer::collectPoisonGeneratingRecipes ( VPTransformState &  State )

protected

Collect poison-generating recipes that may generate a poison value that is used after vectorization, even when their operands are not poison.

Those recipes meet the following conditions:

• Contribute to the address computation of a recipe generating a widen memory load/store (VPWidenMemoryInstructionRecipe or VPInterleaveRecipe).
• Such a widen memory load/store has at least one underlying Instruction that is in a basic block that needs predication and after vectorization the generated instruction won't be predicated.

Definition at line 1015 of file LoopVectorize.cpp.

References llvm::LoopVectorizationLegality::blockNeedsPredication(), llvm::depth_first(), llvm::VPlan::getEntry(), llvm::InterleaveGroup< InstTy >::getMember(), llvm::InterleaveGroup< InstTy >::getNumMembers(), llvm::Instruction::getParent(), I, Legal, and llvm::VPTransformState::Plan.

Referenced by llvm::LoopVectorizationPlanner::executePlan().

completeLoopSkeleton()

BasicBlock * InnerLoopVectorizer::completeLoopSkeleton ( MDNode *  OrigLoopID )

protected

Complete the loop skeleton by adding debug MDs, creating appropriate conditional branches in the middle block, preparing the builder and running the verifier.

Return the preheader of the completed vector loop.

Definition at line 3204 of file LoopVectorize.cpp.

References assert(), Cost, llvm::CmpInst::Create(), DT, llvm::Sched::Fast, llvm::LoopVectorizationCostModel::foldTailByMasking(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), getOrCreateTripCount(), getOrCreateVectorTripCount(), llvm::BasicBlock::getTerminator(), llvm::CmpInst::ICMP_EQ, LoopMiddleBlock, LoopVectorPreHeader, OrigLoop, llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), llvm::Instruction::setDebugLoc(), VectorTripCount, llvm::DominatorTreeBase< NodeT, IsPostDom >::verify(), and VF.

Referenced by llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), and createVectorizedLoopSkeleton().

createBitOrPointerCast()

Value * InnerLoopVectorizer::createBitOrPointerCast ( Value *  V, VectorType *  DstVTy, const DataLayout &  DL  )

protected

Returns a bitcasted value to the requested vector type.

Also handles bitcasts of vector<float> <-> vector<pointer> types.

Definition at line 2911 of file LoopVectorize.cpp.

References assert(), Builder, llvm::IRBuilderBase::CreateBitOrPointerCast(), DL, llvm::FixedVectorType::get(), llvm::Value::getContext(), llvm::Type::getIntNTy(), llvm::Value::getType(), llvm::CastInst::isBitOrNoopPointerCastable(), llvm::Type::isFloatingPointTy(), llvm::Type::isPointerTy(), and VF.

Referenced by vectorizeInterleaveGroup().

createInductionResumeValue()

PHINode * InnerLoopVectorizer::createInductionResumeValue	(	PHINode *	OrigPhi,
		const InductionDescriptor &	ID,
		ArrayRef< BasicBlock * >	BypassBlocks,
		std::pair< BasicBlock *, Value * >	AdditionalBypass = {nullptr, nullptr}
	)

Create a new phi node for the induction variable OrigPhi to resume iteration count in the scalar epilogue, from where the vectorized loop left off.

In cases where the loop skeleton is more complicated (eg. epilogue vectorization) and the resume values can come from an additional bypass block, the AdditionalBypass pair provides information about the bypass block and the end value on the edge from bypass to this loop.

Definition at line 3118 of file LoopVectorize.cpp.

References llvm::PHINode::addIncoming(), assert(), B, BB, llvm::PHINode::Create(), CreateStepValue(), emitTransformedIndex(), llvm::CastInst::getCastOpcode(), llvm::Instruction::getDebugLoc(), llvm::Instruction::getFastMathFlags(), llvm::InductionDescriptor::getInductionBinOp(), getOrCreateVectorTripCount(), llvm::LoopVectorizationLegality::getPrimaryInduction(), llvm::PredicatedScalarEvolution::getSE(), llvm::InductionDescriptor::getStartValue(), llvm::InductionDescriptor::getStep(), llvm::BasicBlock::getTerminator(), llvm::SCEV::getType(), llvm::Value::getType(), IVEndValues, Legal, LoopMiddleBlock, LoopScalarPreHeader, LoopVectorPreHeader, PSE, llvm::Instruction::setDebugLoc(), llvm::PHINode::setIncomingValueForBlock(), llvm::Value::setName(), and VectorTripCount.

Referenced by createInductionResumeValues().

createInductionResumeValues()

void InnerLoopVectorizer::createInductionResumeValues ( std::pair< BasicBlock *, Value * >  AdditionalBypass = {nullptr, nullptr} )

protected

Create new phi nodes for the induction variables to resume iteration count in the scalar epilogue, from where the vectorized loop left off.

In cases where the loop skeleton is more complicated (eg. epilogue vectorization) and the resume values can come from an additional bypass block, the AdditionalBypass pair provides information about the bypass block and the end value on the edge from bypass to this loop.

Definition at line 3183 of file LoopVectorize.cpp.

References assert(), createInductionResumeValue(), llvm::LoopVectorizationLegality::getInductionVars(), Legal, LoopBypassBlocks, LoopScalarPreHeader, and llvm::PHINode::setIncomingValueForBlock().

Referenced by llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), and createVectorizedLoopSkeleton().

createVectorizedLoopSkeleton()

std::pair< BasicBlock *, Value * > InnerLoopVectorizer::createVectorizedLoopSkeleton ( )

virtual

Create a new empty loop that will contain vectorized instructions later on, while the old loop will be used as the scalar remainder.

Control flow is generated around the vectorized (and scalar epilogue) loops consisting of various checks and bypasses. Return the pre-header block of the new loop and the start value for the canonical induction, if it is != 0. The latter is the case when vectorizing the epilogue loop. In the case of epilogue vectorization, this function is overriden to handle the more complex control flow around the loops.

Reimplemented in llvm::InnerLoopAndEpilogueVectorizer.

Definition at line 3240 of file LoopVectorize.cpp.

References completeLoopSkeleton(), createInductionResumeValues(), createVectorLoopSkeleton(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), emitMemRuntimeChecks(), emitSCEVChecks(), llvm::Loop::getLoopID(), llvm::LoopBase< BlockT, LoopT >::getLoopPreheader(), getOrCreateTripCount(), LoopScalarPreHeader, and OrigLoop.

Referenced by llvm::LoopVectorizationPlanner::executePlan().

createVectorLoopSkeleton()

void InnerLoopVectorizer::createVectorLoopSkeleton ( StringRef  Prefix )

protected

Emit basic blocks (prefixed with Prefix) for the iteration check, vector loop preheader, middle block and scalar preheader.

Definition at line 3076 of file LoopVectorize.cpp.

References assert(), Builder, llvm::DominatorTreeBase< NodeT, IsPostDom >::changeImmediateDominator(), Cost, llvm::BranchInst::Create(), DT, llvm::LoopBase< BlockT, LoopT >::getHeader(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), llvm::LoopBase< BlockT, LoopT >::getLoopPreheader(), llvm::BasicBlock::getTerminator(), llvm::IRBuilderBase::getTrue(), llvm::LoopBase< BlockT, LoopT >::getUniqueExitBlock(), LI, LoopExitBlock, LoopMiddleBlock, LoopScalarBody, LoopScalarPreHeader, LoopVectorPreHeader, OrigLoop, llvm::cl::Prefix, llvm::ReplaceInstWithInst(), llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), llvm::Instruction::setDebugLoc(), llvm::SplitBlock(), and VF.

Referenced by llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), and createVectorizedLoopSkeleton().

emitIterationCountCheck()

void InnerLoopVectorizer::emitIterationCountCheck ( BasicBlock *  Bypass )

protected

Emit a bypass check to see if the vector trip count is zero, including if it overflows.

Definition at line 2942 of file LoopVectorize.cpp.

References assert(), Builder, llvm::DominatorTreeBase< NodeT, IsPostDom >::changeImmediateDominator(), Cost, llvm::BranchInst::Create(), llvm::IRBuilderBase::CreateBinaryIntrinsic(), llvm::IRBuilderBase::CreateICmp(), llvm::createStepForVF(), llvm::IRBuilderBase::CreateSub(), DT, llvm::LoopVectorizationCostModel::foldTailByMasking(), llvm::ConstantInt::get(), llvm::IRBuilderBase::getFalse(), llvm::DomTreeNodeBase< NodeT >::getIDom(), llvm::LinearPolySize< LeafTy >::getKnownMinValue(), llvm::DominatorTreeBase< NodeT, IsPostDom >::getNode(), getOrCreateTripCount(), llvm::BasicBlock::getTerminator(), llvm::Value::getType(), llvm::CmpInst::ICMP_ULE, llvm::CmpInst::ICMP_ULT, llvm::LinearPolySize< LeafTy >::isScalable(), LHS, LI, LoopBypassBlocks, LoopExitBlock, LoopVectorPreHeader, MinProfitableTripCount, P, llvm::DominatorTreeBase< NodeT, IsPostDom >::properlyDominates(), llvm::ReplaceInstWithInst(), llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), llvm::SplitBlock(), UF, llvm::APIntOps::umax(), and VF.

emitMemRuntimeChecks()

BasicBlock * InnerLoopVectorizer::emitMemRuntimeChecks ( BasicBlock *  Bypass )

protected

Emit bypass checks to check any memory assumptions we may have made.

Returns the block containing the checks or nullptr if no checks have been added.

Definition at line 3040 of file LoopVectorize.cpp.

References AddedSafetyChecks, assert(), Cost, DEBUG_TYPE, llvm::OptimizationRemarkEmitter::emit(), GeneratedRTChecks::emitMemRuntimeChecks(), EnableVPlanNativePath, llvm::LoopVectorizeHints::FK_Enabled, llvm::LoopVectorizeHints::getForce(), llvm::LoopBase< BlockT, LoopT >::getHeader(), llvm::BasicBlock::getParent(), llvm::Loop::getStartLoc(), llvm::Function::hasOptSize(), llvm::LoopVectorizationCostModel::Hints, LoopBypassBlocks, LoopVectorPreHeader, OptForSizeBasedOnProfile, ORE, OrigLoop, and RTChecks.

Referenced by llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), and createVectorizedLoopSkeleton().

emitSCEVChecks()

BasicBlock * InnerLoopVectorizer::emitSCEVChecks ( BasicBlock *  Bypass )

protected

Emit a bypass check to see if all of the SCEV assumptions we've had to make are correct.

Returns the block containing the checks or nullptr if no checks have been added.

Definition at line 3013 of file LoopVectorize.cpp.

References AddedSafetyChecks, assert(), llvm::DominatorTreeBase< NodeT, IsPostDom >::changeImmediateDominator(), Cost, DT, GeneratedRTChecks::emitSCEVChecks(), llvm::LoopVectorizeHints::FK_Enabled, llvm::LoopVectorizeHints::getForce(), llvm::BasicBlock::getParent(), llvm::Function::hasOptSize(), llvm::LoopVectorizationCostModel::Hints, LoopBypassBlocks, LoopExitBlock, LoopVectorPreHeader, OptForSizeBasedOnProfile, llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), RTChecks, and VF.

Referenced by llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), and createVectorizedLoopSkeleton().

fixCrossIterationPHIs()

void InnerLoopVectorizer::fixCrossIterationPHIs ( VPTransformState &  State )

protected

Handle all cross-iteration phis in the header.

Definition at line 3731 of file LoopVectorize.cpp.

References fixFixedOrderRecurrence(), fixReduction(), llvm::X86ISD::FOR, llvm::VPBlockBase::getEntryBasicBlock(), llvm::VPlan::getVectorLoopRegion(), llvm::VPBasicBlock::phis(), and llvm::VPTransformState::Plan.

Referenced by fixVectorizedLoop().

fixFixedOrderRecurrence()

void InnerLoopVectorizer::fixFixedOrderRecurrence ( VPFirstOrderRecurrencePHIRecipe *  PhiR, VPTransformState &  State  )

protected

Create the exit value of first order recurrences in the middle block and update their users.

Definition at line 3748 of file LoopVectorize.cpp.

References BB, llvm::BasicBlock::begin(), Builder, Cost, llvm::IRBuilderBase::CreateExtractElement(), llvm::IRBuilderBase::CreatePHI(), llvm::IRBuilderBase::CreateSub(), llvm::ConstantInt::get(), llvm::VPTransformState::get(), llvm::VPHeaderPHIRecipe::getBackedgeValue(), llvm::IRBuilderBase::getInt32Ty(), llvm::VPValue::getLiveInIRValue(), llvm::getRuntimeVF(), llvm::VPHeaderPHIRecipe::getStartValue(), llvm::BasicBlock::getTerminator(), llvm::Value::getType(), llvm::VPValue::getUnderlyingValue(), llvm::is_contained(), llvm::ElementCount::isVector(), LoopExitBlock, LoopMiddleBlock, LoopScalarPreHeader, llvm::BasicBlock::phis(), llvm::VPTransformState::Plan, llvm::predecessors(), llvm::VPlan::removeLiveOut(), llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), llvm::PHINode::setIncomingValueForBlock(), llvm::IRBuilderBase::SetInsertPoint(), llvm::Value::setName(), UF, and VF.

Referenced by fixCrossIterationPHIs().

fixNonInductionPHIs()

void InnerLoopVectorizer::fixNonInductionPHIs	(	VPlan &	Plan,
		VPTransformState &	State
	)

Fix the non-induction PHIs in Plan.

Definition at line 4166 of file LoopVectorize.cpp.

References llvm::PHINode::addIncoming(), Builder, llvm::VPTransformState::CFG, llvm::depth_first(), llvm::VPTransformState::get(), llvm::VPlan::getEntry(), llvm::VPWidenPHIRecipe::getIncomingBlock(), llvm::VPWidenPHIRecipe::getIncomingValue(), llvm::VPUser::getNumOperands(), i, P, llvm::IRBuilderBase::SetInsertPoint(), and llvm::VPTransformState::CFGState::VPBB2IRBB.

Referenced by fixVectorizedLoop().

fixReduction()

void InnerLoopVectorizer::fixReduction ( VPReductionPHIRecipe *  Phi, VPTransformState &  State  )

protected

Create code for the loop exit value of the reduction.

Definition at line 3861 of file LoopVectorize.cpp.

References llvm::PHINode::addIncoming(), assert(), llvm::PHINode::blocks(), Builder, llvm::VPTransformState::CFG, clearReductionWrapFlags(), Cost, llvm::PHINode::Create(), llvm::IRBuilderBase::CreateBinOp(), llvm::createMinMaxOp(), llvm::createSelectCmpOp(), llvm::IRBuilderBase::CreateSExt(), llvm::IRBuilderBase::CreateStore(), llvm::createTargetReduction(), llvm::IRBuilderBase::CreateTrunc(), llvm::IRBuilderBase::CreateZExt(), llvm::LoopVectorizationCostModel::foldTailByMasking(), llvm::VPTransformState::get(), llvm::VectorType::get(), llvm::VPHeaderPHIRecipe::getBackedgeValue(), llvm::PHINode::getBasicBlockIndex(), llvm::VPBasicBlock::getEnclosingLoopRegion(), llvm::VPBlockBase::getExitingBasicBlock(), llvm::RecurrenceDescriptor::getFastMathFlags(), llvm::BasicBlock::getFirstInsertionPt(), llvm::PHINode::getIncomingValueForBlock(), llvm::RecurrenceDescriptor::getLoopExitInstr(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), llvm::RecurrenceDescriptor::getOpcode(), llvm::VPRecipeBase::getParent(), llvm::VPReductionPHIRecipe::getRecurrenceDescriptor(), llvm::RecurrenceDescriptor::getRecurrenceKind(), llvm::RecurrenceDescriptor::getRecurrenceStartValue(), llvm::RecurrenceDescriptor::getRecurrenceType(), llvm::VPHeaderPHIRecipe::getStartValue(), llvm::BasicBlock::getTerminator(), llvm::Value::getType(), llvm::VPValue::getUnderlyingValue(), llvm::RecurrenceDescriptor::IntermediateStore, llvm::is_contained(), llvm::VPReductionPHIRecipe::isInLoop(), llvm::VPReductionPHIRecipe::isOrdered(), llvm::LoopVectorizationLegality::isReductionVariable(), llvm::RecurrenceDescriptor::isSelectCmpRecurrenceKind(), llvm::RecurrenceDescriptor::isSigned(), llvm::ElementCount::isVector(), Legal, LoopExitBlock, LoopMiddleBlock, LoopScalarPreHeader, llvm::make_early_inc_range(), OrigLoop, llvm::BasicBlock::phis(), llvm::VPTransformState::Plan, llvm::predecessors(), PreferPredicatedReductionSelect, llvm::TargetTransformInfo::preferPredicatedReductionSelect(), llvm::propagateMetadata(), ReductionResumeValues, llvm::VPlan::removeLiveOut(), llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), llvm::VPTransformState::reset(), llvm::VPTransformState::setDebugLocFromInst(), llvm::IRBuilderBase::setFastMathFlags(), llvm::Instruction::setFastMathFlags(), llvm::PHINode::setIncomingValue(), llvm::IRBuilderBase::SetInsertPoint(), SI, UF, users, llvm::Value::users(), VF, and llvm::VPTransformState::CFGState::VPBB2IRBB.

Referenced by fixCrossIterationPHIs().

fixupIVUsers()

void InnerLoopVectorizer::fixupIVUsers ( PHINode *  OrigPhi, const InductionDescriptor &  II, Value *  VectorTripCount, Value *  EndValue, BasicBlock *  MiddleBlock, BasicBlock *  VectorHeader, VPlan &  Plan  )

protected

Set up the values of the IVs correctly when exiting the vector loop.

Definition at line 3315 of file LoopVectorize.cpp.

References assert(), B, llvm::LoopBase< BlockT, LoopT >::contains(), CreateStepValue(), emitTransformedIndex(), llvm::ConstantInt::get(), llvm::Instruction::getFastMathFlags(), llvm::PHINode::getIncomingValueForBlock(), llvm::InductionDescriptor::getInductionBinOp(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), llvm::PredicatedScalarEvolution::getSE(), llvm::InductionDescriptor::getStartValue(), llvm::InductionDescriptor::getStep(), llvm::BasicBlock::getTerminator(), llvm::SCEV::getType(), llvm::Value::getType(), llvm::LoopBase< BlockT, LoopT >::getUniqueExitBlock(), I, llvm::Type::isIntegerTy(), OrigLoop, PHI, PostInc, PSE, llvm::VPlan::removeLiveOut(), llvm::Value::setName(), llvm::Value::users(), and VectorTripCount.

fixVectorizedLoop()

void InnerLoopVectorizer::fixVectorizedLoop	(	VPTransformState &	State,
		VPlan &	Plan
	)

Fix the vectorized code, taking care of header phi's, live-outs, and more.

Definition at line 3661 of file LoopVectorize.cpp.

References llvm::VPTransformState::CFG, llvm::VPlan::clearLiveOuts(), Cost, EnableVPlanNativePath, fixCrossIterationPHIs(), fixNonInductionPHIs(), llvm::ScalarEvolution::forgetLoop(), llvm::VPBlockBase::getExitingBasicBlock(), llvm::LoopInfoBase< BlockT, LoopT >::getLoopFor(), llvm::PredicatedScalarEvolution::getSE(), llvm::VPlan::getVectorLoopRegion(), llvm::ElementCount::isVector(), LI, OrigLoop, PSE, llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), truncateToMinimalBitwidths(), VF, and llvm::VPTransformState::CFGState::VPBB2IRBB.

Referenced by llvm::LoopVectorizationPlanner::executePlan().

getBroadcastInstrs()

Value * InnerLoopVectorizer::getBroadcastInstrs ( Value *  V )

virtual

Create a broadcast instruction.

This method generates a broadcast instruction (shuffle) for loop invariant values and for the induction value. If this is the induction variable then we extend it to N, N+1, ... this is needed because each iteration in the loop corresponds to a SIMD element.

Definition at line 2261 of file LoopVectorize.cpp.

References Builder, llvm::IRBuilderBase::CreateVectorSplat(), llvm::DominatorTree::dominates(), DT, llvm::Instruction::getParent(), llvm::BasicBlock::getTerminator(), llvm::Loop::isLoopInvariant(), LoopVectorPreHeader, OrigLoop, llvm::IRBuilderBase::SetInsertPoint(), and VF.

getOrCreateTripCount()

Value * InnerLoopVectorizer::getOrCreateTripCount ( BasicBlock *  InsertBlock )

protected

Returns (and creates if needed) the original loop trip count.

Definition at line 2813 of file LoopVectorize.cpp.

References assert(), Builder, llvm::CastInst::CreatePointerCast(), DL, llvm::ScalarEvolution::getAddExpr(), llvm::PredicatedScalarEvolution::getBackedgeTakenCount(), llvm::Module::getDataLayout(), llvm::BasicBlock::getModule(), llvm::ScalarEvolution::getNoopOrZeroExtend(), llvm::ScalarEvolution::getOne(), llvm::Type::getPrimitiveSizeInBits(), llvm::PredicatedScalarEvolution::getSE(), llvm::BasicBlock::getTerminator(), llvm::ScalarEvolution::getTruncateOrNoop(), llvm::SCEV::getType(), llvm::Value::getType(), llvm::ScalarEvolution::getTypeSizeInBits(), llvm::LoopVectorizationLegality::getWidestInductionType(), llvm::Type::isPointerTy(), Legal, PSE, and TripCount.

Referenced by completeLoopSkeleton(), llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), createVectorizedLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), llvm::LoopVectorizationPlanner::executePlan(), and getOrCreateVectorTripCount().

getOrCreateVectorTripCount()

Value * InnerLoopVectorizer::getOrCreateVectorTripCount ( BasicBlock *  InsertBlock )

protected

Returns (and creates if needed) the trip count of the widened loop.

Definition at line 2861 of file LoopVectorize.cpp.

References assert(), Builder, Cost, llvm::IRBuilderBase::CreateAdd(), llvm::IRBuilderBase::CreateICmpEQ(), llvm::IRBuilderBase::CreateSelect(), llvm::createStepForVF(), llvm::IRBuilderBase::CreateSub(), llvm::IRBuilderBase::CreateURem(), llvm::LoopVectorizationCostModel::foldTailByMasking(), llvm::ConstantInt::get(), llvm::LinearPolySize< LeafTy >::getKnownMinValue(), getOrCreateTripCount(), llvm::getRuntimeVF(), llvm::BasicBlock::getTerminator(), llvm::Value::getType(), llvm::isPowerOf2_32(), llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), UF, VectorTripCount, and VF.

Referenced by completeLoopSkeleton(), llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), createInductionResumeValue(), and llvm::LoopVectorizationPlanner::executePlan().

getReductionResumeValue()

PHINode * InnerLoopVectorizer::getReductionResumeValue

(

const RecurrenceDescriptor &

RdxDesc

)

Definition at line 1096 of file LoopVectorize.cpp.

References assert(), and ReductionResumeValues.

Referenced by llvm::LoopVectorizePass::processLoop().

packScalarIntoVectorValue()

void InnerLoopVectorizer::packScalarIntoVectorValue	(	VPValue *	Def,
		const VPIteration &	Instance,
		VPTransformState &	State
	)

Construct the vector value of a scalarized value V one lane at a time.

Definition at line 2506 of file LoopVectorize.cpp.

References llvm::VPTransformState::Builder, Builder, llvm::IRBuilderBase::CreateInsertElement(), llvm::tgtok::Def, llvm::VPTransformState::get(), llvm::VPLane::getAsRuntimeExpr(), llvm::VPIteration::Lane, llvm::VPIteration::Part, llvm::VPTransformState::set(), and VF.

Referenced by llvm::VPReplicateRecipe::execute().

printDebugTracesAtEnd()

virtual void llvm::InnerLoopVectorizer::printDebugTracesAtEnd ( )

inlineprotectedvirtual

Reimplemented in llvm::EpilogueVectorizerEpilogueLoop, and llvm::EpilogueVectorizerMainLoop.

Definition at line 643 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationPlanner::executePlan().

printDebugTracesAtStart()

virtual void llvm::InnerLoopVectorizer::printDebugTracesAtStart ( )

inlineprotectedvirtual

Allow subclasses to override and print debug traces before/after vplan execution, when trace information is requested.

Reimplemented in llvm::EpilogueVectorizerEpilogueLoop, and llvm::EpilogueVectorizerMainLoop.

Definition at line 642 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationPlanner::executePlan().

scalarizeInstruction()

void InnerLoopVectorizer::scalarizeInstruction	(	const Instruction *	Instr,
		VPReplicateRecipe *	RepRecipe,
		const VPIteration &	Instance,
		bool	IfPredicateInstr,
		VPTransformState &	State
	)

A helper function to scalarize a single Instruction in the innermost loop.

Generates a sequence of scalar instances for each lane between MinLane and MaxLane, times each part between MinPart and MaxPart, inclusive. Uses the VPValue operands from RepRecipe instead of Instr's operands.

Definition at line 2756 of file LoopVectorize.cpp.

References AC, llvm::VPTransformState::addNewMetadata(), assert(), llvm::VPTransformState::Builder, llvm::Instruction::clone(), llvm::Instruction::dropPoisonGeneratingFlags(), llvm::enumerate(), llvm::VPTransformState::get(), llvm::Instruction::getDebugLoc(), llvm::VPLane::getFirstLane(), llvm::Value::getName(), llvm::Value::getType(), I, llvm::IRBuilderBase::Insert(), llvm::Type::isAggregateType(), llvm::VPIteration::isFirstIteration(), llvm::vputils::isUniformAfterVectorization(), llvm::Type::isVoidTy(), llvm::VPTransformState::MayGeneratePoisonRecipes, llvm::VPUser::operands(), PredicatedInstructions, llvm::AssumptionCache::registerAssumption(), llvm::VPTransformState::set(), llvm::VPTransformState::setDebugLocFromInst(), llvm::Value::setName(), and llvm::User::setOperand().

Referenced by llvm::VPReplicateRecipe::execute().

sinkScalarOperands()

void InnerLoopVectorizer::sinkScalarOperands ( Instruction *  PredInst )

protected

Iteratively sink the scalarized operands of a predicated instruction into the block that was created for it.

Definition at line 4092 of file LoopVectorize.cpp.

References llvm::all_of(), BB, llvm::SmallVectorImpl< T >::clear(), llvm::SetVector< T, Vector, Set >::empty(), llvm::PHINode::getIncomingValueNumForOperand(), llvm::LoopInfoBase< BlockT, LoopT >::getLoopFor(), llvm::Instruction::getParent(), I, llvm::SetVector< T, Vector, Set >::insert(), LI, llvm::User::op_begin(), llvm::User::op_end(), and llvm::SetVector< T, Vector, Set >::pop_back_val().

truncateToMinimalBitwidths()

void InnerLoopVectorizer::truncateToMinimalBitwidths ( VPTransformState &  State )

protected

Shrinks vector element sizes to the smallest bitwidth they can be legally represented as.

Definition at line 3528 of file LoopVectorize.cpp.

References B, Cost, llvm::SmallPtrSetImpl< PtrType >::count(), llvm::tgtok::Def, llvm::Instruction::eraseFromParent(), llvm::IntegerType::get(), llvm::VPTransformState::get(), llvm::VectorType::get(), llvm::Type::getContext(), llvm::LoopVectorizationCostModel::getMinimalBitwidths(), llvm::User::getOperand(), llvm::VPlan::getVPValue(), llvm::VPTransformState::hasAnyVectorValue(), I, llvm::ARM_PROC::IE, llvm::SmallPtrSetImpl< PtrType >::insert(), llvm_unreachable, llvm::VPTransformState::Plan, llvm::VPTransformState::reset(), SI, smallestIntegerVectorType(), llvm::Value::takeName(), UF, and llvm::Value::use_empty().

Referenced by fixVectorizedLoop().

useOrderedReductions()

bool InnerLoopVectorizer::useOrderedReductions

(

const RecurrenceDescriptor &

RdxDesc

)

Returns true if the reordering of FP operations is not allowed, but we are able to vectorize with strict in-order reductions for the given RdxDesc.

Definition at line 4187 of file LoopVectorize.cpp.

References Cost, and llvm::LoopVectorizationCostModel::useOrderedReductions().

Referenced by llvm::LoopVectorizationCostModel::collectElementTypesForWidening(), llvm::LoopVectorizationCostModel::collectInLoopReductions(), and llvm::VPReductionRecipe::execute().

vectorizeInterleaveGroup()

void InnerLoopVectorizer::vectorizeInterleaveGroup	(	const InterleaveGroup< Instruction > *	Group,
		ArrayRef< VPValue * >	VPDefs,
		VPTransformState &	State,
		VPValue *	Addr,
		ArrayRef< VPValue * >	StoredValues,
		VPValue *	BlockInMask = nullptr
	)

Try to vectorize interleaved access group Group with the base address given in Addr, optionally masking the vector operations if BlockInMask is non-null.

Use State to translate given VPValues to IR values in the vectorized loop.

Definition at line 2556 of file LoopVectorize.cpp.

References llvm::InterleaveGroup< InstTy >::addMetadata(), Addr, assert(), Builder, llvm::concatenateVectors(), Cost, llvm::IRBuilderBase::CreateAlignedLoad(), llvm::IRBuilderBase::CreateAlignedStore(), llvm::IRBuilderBase::CreateBinOp(), llvm::IRBuilderBase::CreateBitCast(), llvm::createBitMaskForGaps(), createBitOrPointerCast(), llvm::IRBuilderBase::CreateGEP(), llvm::createInterleaveMask(), llvm::IRBuilderBase::CreateMaskedLoad(), llvm::IRBuilderBase::CreateMaskedStore(), llvm::createReplicatedMask(), llvm::IRBuilderBase::CreateShuffleVector(), llvm::createStrideMask(), llvm::IRBuilderBase::CreateVectorReverse(), DL, gep, llvm::VPTransformState::get(), llvm::VectorType::get(), llvm::PoisonValue::get(), llvm::InterleaveGroup< InstTy >::getAlign(), llvm::Module::getDataLayout(), llvm::InterleaveGroup< InstTy >::getFactor(), llvm::InterleaveGroup< InstTy >::getIndex(), llvm::InterleaveGroup< InstTy >::getInsertPos(), llvm::IRBuilderBase::getInt32(), llvm::LinearPolySize< LeafTy >::getKnownMinValue(), llvm::getLoadStoreType(), llvm::InterleaveGroup< InstTy >::getMember(), llvm::Instruction::getModule(), llvm::Type::getPointerAddressSpace(), llvm::Type::getPointerTo(), llvm::Value::getType(), i, I, llvm::InterleaveGroup< InstTy >::isReverse(), llvm::LinearPolySize< LeafTy >::isScalable(), llvm::LoopVectorizationCostModel::isScalarEpilogueAllowed(), llvm::InterleaveGroup< InstTy >::requiresScalarEpilogue(), llvm::VPTransformState::set(), llvm::VPTransformState::setDebugLocFromInst(), llvm::Value::stripPointerCasts(), UF, llvm::RegState::Undef, useMaskedInterleavedAccesses(), and VF.

Referenced by llvm::VPInterleaveRecipe::execute().

Friends And Related Function Documentation

LoopVectorizationPlanner

friend class LoopVectorizationPlanner

friend

Definition at line 546 of file LoopVectorize.cpp.

Member Data Documentation

AA

AAResults* llvm::InnerLoopVectorizer::AA

protected

Alias Analysis.

Definition at line 660 of file LoopVectorize.cpp.

AC

AssumptionCache* llvm::InnerLoopVectorizer::AC

protected

Assumption Cache.

Definition at line 669 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationCostModel::collectValuesToIgnore(), getScalarEpilogueLowering(), processLoopInVPlanNativePath(), and scalarizeInstruction().

AddedSafetyChecks

bool llvm::InnerLoopVectorizer::AddedSafetyChecks = false

protected

Definition at line 724 of file LoopVectorize.cpp.

Referenced by areSafetyChecksAdded(), emitMemRuntimeChecks(), and emitSCEVChecks().

BFI

BlockFrequencyInfo* llvm::InnerLoopVectorizer::BFI

protected

BFI and PSI are used to check for profile guided size optimizations.

Definition at line 731 of file LoopVectorize.cpp.

Referenced by getScalarEpilogueLowering(), InnerLoopVectorizer(), and processLoopInVPlanNativePath().

Builder

IRBuilder llvm::InnerLoopVectorizer::Builder

protected

The builder that we use.

Definition at line 685 of file LoopVectorize.cpp.

Referenced by buildScalarSteps(), createBitOrPointerCast(), llvm::VPRecipeBuilder::createBlockInMask(), llvm::VPRecipeBuilder::createEdgeMask(), createVectorLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), llvm::EpilogueVectorizerEpilogueLoop::emitMinimumVectorEpilogueIterCountCheck(), llvm::VPWidenIntOrFpInductionRecipe::execute(), llvm::VPWidenMemoryInstructionRecipe::execute(), llvm::LoopVectorizationPlanner::executePlan(), fixFixedOrderRecurrence(), fixNonInductionPHIs(), fixReduction(), llvm::VPTransformState::get(), getBroadcastInstrs(), getOrCreateTripCount(), getOrCreateVectorTripCount(), getStepVector(), packScalarIntoVectorValue(), and vectorizeInterleaveGroup().

Cost

LoopVectorizationCostModel* llvm::InnerLoopVectorizer::Cost

protected

The profitablity analysis.

Definition at line 721 of file LoopVectorize.cpp.

Referenced by completeLoopSkeleton(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createVectorLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), emitMemRuntimeChecks(), llvm::EpilogueVectorizerEpilogueLoop::emitMinimumVectorEpilogueIterCountCheck(), emitSCEVChecks(), fixFixedOrderRecurrence(), fixReduction(), fixVectorizedLoop(), GeneratedRTChecks::getCost(), getOrCreateVectorTripCount(), llvm::LoopVectorizationCostModel::getVectorCallCost(), llvm::LoopVectorizationCostModel::setCostBasedWideningDecision(), llvm::LoopVectorizationCostModel::setWideningDecision(), truncateToMinimalBitwidths(), useOrderedReductions(), and vectorizeInterleaveGroup().

DT

DominatorTree* llvm::InnerLoopVectorizer::DT

protected

Dominator Tree.

Definition at line 657 of file LoopVectorize.cpp.

Referenced by completeLoopSkeleton(), GeneratedRTChecks::Create(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createVectorLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), GeneratedRTChecks::emitMemRuntimeChecks(), llvm::EpilogueVectorizerEpilogueLoop::emitMinimumVectorEpilogueIterCountCheck(), emitSCEVChecks(), GeneratedRTChecks::emitSCEVChecks(), llvm::LoopVectorizationPlanner::executePlan(), getBroadcastInstrs(), getScalarEpilogueLowering(), and processLoopInVPlanNativePath().

IVEndValues

DenseMap<PHINode *, Value *> llvm::InnerLoopVectorizer::IVEndValues

protected

Definition at line 728 of file LoopVectorize.cpp.

Referenced by createInductionResumeValue().

Legal

LoopVectorizationLegality* llvm::InnerLoopVectorizer::Legal

protected

The legality analysis.

Definition at line 718 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationCostModel::blockNeedsPredicationForAnyReason(), llvm::LoopVectorizationCostModel::canVectorizeReductions(), llvm::LoopVectorizationCostModel::collectElementTypesForWidening(), llvm::LoopVectorizationCostModel::collectInLoopReductions(), collectPoisonGeneratingRecipes(), llvm::LoopVectorizationCostModel::collectValuesToIgnore(), llvm::LoopVectorizationCostModel::computeMaxVF(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createInductionResumeValue(), createInductionResumeValues(), llvm::LoopVectorizationPlanner::executePlan(), fixReduction(), getAddressAccessSCEV(), llvm::LoopVectorizationCostModel::getDivRemSpeculationCost(), getOrCreateTripCount(), llvm::LoopVectorizationCostModel::getSmallestAndWidestTypes(), llvm::LoopVectorizationCostModel::interleavedAccessCanBeWidened(), llvm::LoopVectorizationCostModel::isLegalMaskedLoad(), llvm::LoopVectorizationCostModel::isLegalMaskedStore(), llvm::LoopVectorizationCostModel::isOptimizableIVTruncate(), llvm::LoopVectorizationCostModel::isPredicatedInst(), isStrideMul(), llvm::LoopVectorizationCostModel::memoryInstructionCanBeWidened(), llvm::LoopVectorizationCostModel::runtimeChecksRequired(), llvm::LoopVectorizationCostModel::selectInterleaveCount(), llvm::LoopVectorizationCostModel::setCostBasedWideningDecision(), and llvm::VPRecipeBuilder::tryToCreateWidenRecipe().

LI

LoopInfo* llvm::InnerLoopVectorizer::LI

protected

Loop Info.

Definition at line 654 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationCostModel::calculateRegisterUsage(), collectSupportedLoops(), GeneratedRTChecks::Create(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createVectorLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), GeneratedRTChecks::emitMemRuntimeChecks(), GeneratedRTChecks::emitSCEVChecks(), llvm::VPWidenMemoryInstructionRecipe::execute(), llvm::LoopVectorizationPlanner::executePlan(), fixVectorizedLoop(), getScalarEpilogueLowering(), llvm::LoopVectorizationCostModel::isLegalGatherOrScatter(), processLoopInVPlanNativePath(), and sinkScalarOperands().

LoopBypassBlocks

SmallVector<BasicBlock *, 4> llvm::InnerLoopVectorizer::LoopBypassBlocks

protected

A list of all bypass blocks. The first block is the entry of the loop.

Definition at line 706 of file LoopVectorize.cpp.

Referenced by llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createInductionResumeValues(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), emitMemRuntimeChecks(), llvm::EpilogueVectorizerEpilogueLoop::emitMinimumVectorEpilogueIterCountCheck(), and emitSCEVChecks().

LoopExitBlock

BasicBlock* llvm::InnerLoopVectorizer::LoopExitBlock

protected

The unique ExitBlock of the scalar loop if one exists.

Note that there can be multiple exiting edges reaching this block.

Definition at line 700 of file LoopVectorize.cpp.

Referenced by llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createVectorLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), emitSCEVChecks(), fixFixedOrderRecurrence(), and fixReduction().

LoopMiddleBlock

BasicBlock* llvm::InnerLoopVectorizer::LoopMiddleBlock

protected

Middle Block between the vector and the scalar.

Definition at line 696 of file LoopVectorize.cpp.

Referenced by completeLoopSkeleton(), createInductionResumeValue(), createVectorLoopSkeleton(), fixFixedOrderRecurrence(), and fixReduction().

LoopScalarBody

BasicBlock* llvm::InnerLoopVectorizer::LoopScalarBody

protected

The scalar loop body.

Definition at line 703 of file LoopVectorize.cpp.

Referenced by createVectorLoopSkeleton().

LoopScalarPreHeader

BasicBlock* llvm::InnerLoopVectorizer::LoopScalarPreHeader

protected

The scalar-loop preheader.

Definition at line 693 of file LoopVectorize.cpp.

Referenced by llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createInductionResumeValue(), createInductionResumeValues(), createVectorizedLoopSkeleton(), createVectorLoopSkeleton(), fixFixedOrderRecurrence(), and fixReduction().

LoopVectorPreHeader

BasicBlock* llvm::InnerLoopVectorizer::LoopVectorPreHeader

protected

The vector-loop preheader.

Definition at line 690 of file LoopVectorize.cpp.

Referenced by completeLoopSkeleton(), llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), createInductionResumeValue(), createVectorLoopSkeleton(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), emitMemRuntimeChecks(), GeneratedRTChecks::emitMemRuntimeChecks(), llvm::EpilogueVectorizerEpilogueLoop::emitMinimumVectorEpilogueIterCountCheck(), emitSCEVChecks(), GeneratedRTChecks::emitSCEVChecks(), and getBroadcastInstrs().

MinProfitableTripCount

ElementCount llvm::InnerLoopVectorizer::MinProfitableTripCount

protected

Definition at line 678 of file LoopVectorize.cpp.

Referenced by emitIterationCountCheck(), and InnerLoopVectorizer().

OptForSizeBasedOnProfile

bool llvm::InnerLoopVectorizer::OptForSizeBasedOnProfile

protected

Definition at line 736 of file LoopVectorize.cpp.

Referenced by emitMemRuntimeChecks(), emitSCEVChecks(), and InnerLoopVectorizer().

ORE

OptimizationRemarkEmitter* llvm::InnerLoopVectorizer::ORE

protected

Interface to emit optimization remarks.

Definition at line 672 of file LoopVectorize.cpp.

Referenced by checkMixedPrecision(), collectSupportedLoops(), llvm::LoopVectorizationCostModel::computeMaxVF(), emitMemRuntimeChecks(), llvm::LoopVectorizationPlanner::executePlan(), isExplicitVecOuterLoop(), llvm::LoopVectorizationPlanner::plan(), processLoopInVPlanNativePath(), llvm::reportVectorizationFailure(), llvm::reportVectorizationInfo(), llvm::LoopVectorizationCostModel::runtimeChecksRequired(), and llvm::LoopVectorizationCostModel::selectVectorizationFactor().

OrigLoop

Loop* llvm::InnerLoopVectorizer::OrigLoop

protected

The original loop.

Definition at line 643 of file LoopVectorize.cpp.

Referenced by addUsersInExitBlock(), completeLoopSkeleton(), llvm::VPRecipeBuilder::createBlockInMask(), llvm::VPRecipeBuilder::createEdgeMask(), llvm::EpilogueVectorizerMainLoop::createEpilogueVectorizedLoopSkeleton(), llvm::EpilogueVectorizerEpilogueLoop::createEpilogueVectorizedLoopSkeleton(), CreateStepValue(), createVectorizedLoopSkeleton(), createVectorLoopSkeleton(), createWidenInductionRecipes(), emitMemRuntimeChecks(), llvm::LoopVectorizationPlanner::executePlan(), llvm::VPRecipeBuilder::fixHeaderPhis(), fixReduction(), fixupIVUsers(), fixVectorizedLoop(), getBroadcastInstrs(), InnerLoopVectorizer(), llvm::LoopVectorizationPlanner::plan(), llvm::LoopVectorizationPlanner::planInVPlanNativePath(), llvm::EpilogueVectorizerMainLoop::printDebugTracesAtEnd(), llvm::EpilogueVectorizerEpilogueLoop::printDebugTracesAtEnd(), and llvm::VPRecipeBuilder::tryToCreateWidenRecipe().

PredicatedInstructions

SmallVector<Instruction *, 4> llvm::InnerLoopVectorizer::PredicatedInstructions

protected

Store instructions that were predicated.

Definition at line 709 of file LoopVectorize.cpp.

Referenced by scalarizeInstruction().

PSE

PredicatedScalarEvolution& llvm::InnerLoopVectorizer::PSE

protected

A wrapper around ScalarEvolution used to add runtime SCEV checks.

Applies dynamic knowledge to simplify SCEV expressions and converts them to a more usable form.

Definition at line 651 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationCostModel::computeMaxVF(), createInductionResumeValue(), llvm::LoopVectorizationPlanner::executePlan(), fixupIVUsers(), fixVectorizedLoop(), getAddressAccessSCEV(), getOrCreateTripCount(), llvm::LoopVectorizationCostModel::isMoreProfitable(), processLoopInVPlanNativePath(), llvm::LoopVectorizationCostModel::runtimeChecksRequired(), llvm::LoopVectorizationCostModel::selectInterleaveCount(), and llvm::VPRecipeBuilder::tryToCreateWidenRecipe().

PSI

ProfileSummaryInfo* llvm::InnerLoopVectorizer::PSI

protected

Definition at line 732 of file LoopVectorize.cpp.

Referenced by getScalarEpilogueLowering(), InnerLoopVectorizer(), and processLoopInVPlanNativePath().

ReductionResumeValues

SmallMapVector<const RecurrenceDescriptor *, PHINode *, 4> llvm::InnerLoopVectorizer::ReductionResumeValues

protected

Definition at line 745 of file LoopVectorize.cpp.

Referenced by fixReduction(), and getReductionResumeValue().

RTChecks

GeneratedRTChecks& llvm::InnerLoopVectorizer::RTChecks

protected

Structure to hold information about generated runtime checks, responsible for cleaning the checks, if vectorization turns out unprofitable.

Definition at line 740 of file LoopVectorize.cpp.

Referenced by emitMemRuntimeChecks(), and emitSCEVChecks().

TLI

const TargetLibraryInfo* llvm::InnerLoopVectorizer::TLI

protected

Target Library Info.

Definition at line 663 of file LoopVectorize.cpp.

Referenced by getScalarEpilogueLowering(), llvm::LoopVectorizationCostModel::getVectorCallCost(), llvm::LoopVectorizationCostModel::getVectorIntrinsicCost(), and processLoopInVPlanNativePath().

TripCount

Value* llvm::InnerLoopVectorizer::TripCount = nullptr

protected

Trip count of the original loop.

Definition at line 712 of file LoopVectorize.cpp.

Referenced by llvm::EpilogueVectorizerEpilogueLoop::EpilogueVectorizerEpilogueLoop(), and getOrCreateTripCount().

TTI

const TargetTransformInfo* llvm::InnerLoopVectorizer::TTI

protected

Target Transform Info.

Definition at line 666 of file LoopVectorize.cpp.

Referenced by llvm::LoopVectorizationCostModel::calculateRegisterUsage().

UF

unsigned llvm::InnerLoopVectorizer::UF

protected

The vectorization unroll factor to use.

Each scalar is vectorized to this many different vector instructions.

Definition at line 682 of file LoopVectorize.cpp.

Referenced by clearReductionWrapFlags(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), fixFixedOrderRecurrence(), fixReduction(), getOrCreateVectorTripCount(), truncateToMinimalBitwidths(), and vectorizeInterleaveGroup().

VectorTripCount

Value* llvm::InnerLoopVectorizer::VectorTripCount = nullptr

protected

Trip count of the widened loop (TripCount - TripCount % (VF*UF))

Definition at line 715 of file LoopVectorize.cpp.

Referenced by completeLoopSkeleton(), createInductionResumeValue(), fixupIVUsers(), and getOrCreateVectorTripCount().

VF

ElementCount llvm::InnerLoopVectorizer::VF

protected

The vectorization SIMD factor to use.

Each vector will have this many vector elements.

Definition at line 676 of file LoopVectorize.cpp.

Referenced by areRuntimeChecksProfitable(), llvm::LoopVectorizationPlanner::buildVPlans(), llvm::LoopVectorizationCostModel::calculateRegisterUsage(), llvm::LoopVectorizationCostModel::canTruncateToMinimalBitwidth(), llvm::LoopVectorizationCostModel::collectInstsToScalarize(), llvm::LoopVectorizationCostModel::collectUniformsAndScalars(), completeLoopSkeleton(), GeneratedRTChecks::Create(), createBitOrPointerCast(), llvm::createStepForVF(), createVectorLoopSkeleton(), createWidenInductionRecipes(), emitIterationCountCheck(), llvm::EpilogueVectorizerMainLoop::emitIterationCountCheck(), emitSCEVChecks(), fixFixedOrderRecurrence(), fixReduction(), fixVectorizedLoop(), llvm::VPTransformState::get(), llvm::LoopVectorizationPlanner::getBestPlanFor(), getBroadcastInstrs(), llvm::LoopVectorizationCostModel::getDivRemSpeculationCost(), getOrCreateVectorTripCount(), llvm::getRuntimeVF(), llvm::getRuntimeVFAsFloat(), getStepVector(), llvm::LoopVectorizationCostModel::getVectorCallCost(), llvm::LoopVectorizationCostModel::getVectorIntrinsicCost(), llvm::LoopVectorizationCostModel::getWideningCost(), llvm::LoopVectorizationCostModel::getWideningDecision(), llvm::VPRecipeBuilder::handleReplication(), llvm::LoopVectorizationCostModel::interleavedAccessCanBeWidened(), llvm::LoopVectorizationCostModel::isLegalGatherOrScatter(), llvm::LoopVectorizationCostModel::isOptimizableIVTruncate(), llvm::LoopVectorizationCostModel::isProfitableToScalarize(), llvm::LoopVectorizationCostModel::isScalarAfterVectorization(), llvm::LoopVectorizationCostModel::isScalarWithPredication(), llvm::LoopVectorizationCostModel::isUniformAfterVectorization(), MaybeVectorizeType(), llvm::LoopVectorizationCostModel::memoryInstructionCanBeWidened(), packScalarIntoVectorValue(), llvm::LoopVectorizationPlanner::plan(), llvm::LoopVectorizationPlanner::planInVPlanNativePath(), processLoopInVPlanNativePath(), llvm::LoopVectorizationCostModel::requiresScalarEpilogue(), llvm::LoopVectorizationCostModel::selectInterleaveCount(), llvm::LoopVectorizationCostModel::setCostBasedWideningDecision(), llvm::LoopVectorizationCostModel::setWideningDecision(), llvm::VPRecipeBuilder::tryToCreateWidenRecipe(), and vectorizeInterleaveGroup().

---

The documentation for this class was generated from the following file:

• lib/Transforms/Vectorize/LoopVectorize.cpp