Drive the analysis of memory accesses in the loop. More...

#include "llvm/Analysis/LoopAccessAnalysis.h"

Public Member Functions
	LoopAccessInfo (Loop L, ScalarEvolution SE, const TargetTransformInfo TTI, const TargetLibraryInfo TLI, AAResults AA, DominatorTree DT, LoopInfo *LI)

bool	canVectorizeMemory () const
	Return true we can analyze the memory accesses in the loop and there are no memory dependence cycles.

bool	hasConvergentOp () const
	Return true if there is a convergent operation in the loop.

const RuntimePointerChecking *	getRuntimePointerChecking () const

unsigned	getNumRuntimePointerChecks () const
	Number of memchecks required to prove independence of otherwise may-alias pointers.

bool	isInvariant (Value *V) const
	Returns true if value `V` is loop invariant.

unsigned	getNumStores () const

unsigned	getNumLoads () const

const OptimizationRemarkAnalysis *	getReport () const
	The diagnostics report generated for the analysis.

const MemoryDepChecker &	getDepChecker () const
	the Memory Dependence Checker which can determine the loop-independent and loop-carried dependences between memory accesses.

SmallVector< Instruction *, 4 >	getInstructionsForAccess (Value *Ptr, bool isWrite) const
	Return the list of instructions that use `Ptr` to read or write memory.

const DenseMap< Value , const SCEV > &	getSymbolicStrides () const
	If an access has a symbolic strides, this maps the pointer value to the stride symbol.

void	print (raw_ostream &OS, unsigned Depth=0) const
	Print the information about the memory accesses in the loop.

bool	hasStoreStoreDependenceInvolvingLoopInvariantAddress () const
	Return true if the loop has memory dependence involving two stores to an invariant address, else return false.

bool	hasLoadStoreDependenceInvolvingLoopInvariantAddress () const
	Return true if the loop has memory dependence involving a load and a store to an invariant address, else return false.

ArrayRef< StoreInst * >	getStoresToInvariantAddresses () const
	Return the list of stores to invariant addresses.

const PredicatedScalarEvolution &	getPSE () const
	Used to add runtime SCEV checks.

Static Public Member Functions
static bool	blockNeedsPredication (BasicBlock BB, Loop TheLoop, DominatorTree *DT)
	Return true if the block BB needs to be predicated in order for the loop to be vectorized.

Detailed Description

Drive the analysis of memory accesses in the loop.

This class is responsible for analyzing the memory accesses of a loop. It collects the accesses and then its main helper the AccessAnalysis class finds and categorizes the dependences in buildDependenceSets.

For memory dependences that can be analyzed at compile time, it determines whether the dependence is part of cycle inhibiting vectorization. This work is delegated to the MemoryDepChecker class.

For memory dependences that cannot be determined at compile time, it generates run-time checks to prove independence. This is done by AccessAnalysis::canCheckPtrAtRT and the checks are maintained by the RuntimePointerCheck class.

If pointers can wrap or can't be expressed as affine AddRec expressions by ScalarEvolution, we will generate run-time checks by emitting a SCEVUnionPredicate.

Checks for both memory dependences and the SCEV predicates contained in the PSE must be emitted in order for the results of this analysis to be valid.

Definition at line 622 of file LoopAccessAnalysis.h.

Constructor & Destructor Documentation

◆ LoopAccessInfo()

LoopAccessInfo::LoopAccessInfo	(	Loop *	L,
		ScalarEvolution *	SE,
		const TargetTransformInfo *	TTI,
		const TargetLibraryInfo *	TLI,
		AAResults *	AA,
		DominatorTree *	DT,
		LoopInfo *	LI
	)

Definition at line 2990 of file LoopAccessAnalysis.cpp.

References llvm::details::FixedOrScalableQuantity< LeafTy, ValueTy >::getFixedValue(), llvm::TargetTransformInfo::getRegisterBitWidth(), llvm::details::FixedOrScalableQuantity< LeafTy, ValueTy >::isNonZero(), llvm::TargetTransformInfo::RGK_FixedWidthVector, and llvm::TargetTransformInfo::RGK_ScalableVector.

Member Function Documentation

◆ blockNeedsPredication()

bool LoopAccessInfo::blockNeedsPredication	(	BasicBlock *	BB,
		Loop *	TheLoop,
		DominatorTree *	DT
	)

static

Return true if the block BB needs to be predicated in order for the loop to be vectorized.

Definition at line 2760 of file LoopAccessAnalysis.cpp.

References assert(), llvm::LoopBase< BlockT, LoopT >::contains(), llvm::DominatorTree::dominates(), and llvm::LoopBase< BlockT, LoopT >::getLoopLatch().

Referenced by llvm::LoopVectorizationLegality::blockNeedsPredication().

◆ canVectorizeMemory()

bool llvm::LoopAccessInfo::canVectorizeMemory ( ) const

inline

Return true we can analyze the memory accesses in the loop and there are no memory dependence cycles.

Note that for dependences between loads & stores with uniform addresses, hasStoreStoreDependenceInvolvingLoopInvariantAddress and hasLoadStoreDependenceInvolvingLoopInvariantAddress also need to be checked.

Definition at line 634 of file LoopAccessAnalysis.h.

◆ getDepChecker()

const MemoryDepChecker & llvm::LoopAccessInfo::getDepChecker ( ) const

inline

the Memory Dependence Checker which can determine the loop-independent and loop-carried dependences between memory accesses.

Definition at line 668 of file LoopAccessAnalysis.h.

Referenced by llvm::LoopVersioning::annotateLoopWithNoAlias(), llvm::LoopVectorizationLegality::getMaxSafeVectorWidthInBits(), llvm::LoopVectorizationLegality::isSafeForAnyVectorWidth(), and print().

◆ getInstructionsForAccess()

SmallVector< Instruction *, 4 > llvm::LoopAccessInfo::getInstructionsForAccess	(	Value *	Ptr,
		bool	isWrite
	)		const

inline

Return the list of instructions that use Ptr to read or write memory.

Definition at line 672 of file LoopAccessAnalysis.h.

References Ptr.

◆ getNumLoads()

unsigned llvm::LoopAccessInfo::getNumLoads ( ) const

inline

Definition at line 660 of file LoopAccessAnalysis.h.

Referenced by llvm::LoopVectorizationLegality::getNumLoads().

◆ getNumRuntimePointerChecks()

unsigned llvm::LoopAccessInfo::getNumRuntimePointerChecks ( ) const

inline

Number of memchecks required to prove independence of otherwise may-alias pointers.

Definition at line 647 of file LoopAccessAnalysis.h.

◆ getNumStores()

unsigned llvm::LoopAccessInfo::getNumStores ( ) const

inline

Definition at line 659 of file LoopAccessAnalysis.h.

Referenced by llvm::LoopVectorizationLegality::getNumStores().

◆ getPSE()

const PredicatedScalarEvolution & llvm::LoopAccessInfo::getPSE ( ) const

inline

Used to add runtime SCEV checks.

Simplifies SCEV expressions and converts them to a more usable form. All SCEV expressions during the analysis should be re-written (and therefore simplified) according to PSE. A user of LoopAccessAnalysis will need to emit the runtime checks associated with this predicate.

Definition at line 708 of file LoopAccessAnalysis.h.

Referenced by canTailPredicateLoop().

◆ getReport()

const OptimizationRemarkAnalysis * llvm::LoopAccessInfo::getReport ( ) const

inline

The diagnostics report generated for the analysis.

E.g. why we couldn't analyze the loop.

Definition at line 664 of file LoopAccessAnalysis.h.

◆ getRuntimePointerChecking()

const RuntimePointerChecking * llvm::LoopAccessInfo::getRuntimePointerChecking ( ) const

inline

Definition at line 641 of file LoopAccessAnalysis.h.

Referenced by llvm::LoopVectorizationLegality::canVectorize(), llvm::LoopVectorizationPlanner::executePlan(), llvm::LoopVectorizationLegality::getRuntimePointerChecking(), llvm::LoopVersioning::prepareNoAliasMetadata(), and llvm::LoopVersioning::versionLoop().

◆ getStoresToInvariantAddresses()

ArrayRef< StoreInst * > llvm::LoopAccessInfo::getStoresToInvariantAddresses ( ) const

inline

Return the list of stores to invariant addresses.

Definition at line 699 of file LoopAccessAnalysis.h.

◆ getSymbolicStrides()

const DenseMap< Value *, const SCEV * > & llvm::LoopAccessInfo::getSymbolicStrides ( ) const

inline

If an access has a symbolic strides, this maps the pointer value to the stride symbol.

Definition at line 679 of file LoopAccessAnalysis.h.

Referenced by llvm::InterleavedAccessInfo::analyzeInterleaving(), llvm::LoopVectorizationLegality::isConsecutivePtr(), and llvm::LoopVectorizationCostModel::runtimeChecksRequired().

◆ hasConvergentOp()

bool llvm::LoopAccessInfo::hasConvergentOp ( ) const

inline

Return true if there is a convergent operation in the loop.

There may still be reported runtime pointer checks that would be required, but it is not legal to insert them.

Definition at line 639 of file LoopAccessAnalysis.h.

◆ hasLoadStoreDependenceInvolvingLoopInvariantAddress()

bool llvm::LoopAccessInfo::hasLoadStoreDependenceInvolvingLoopInvariantAddress ( ) const

inline

Return true if the loop has memory dependence involving a load and a store to an invariant address, else return false.

Definition at line 694 of file LoopAccessAnalysis.h.

◆ hasStoreStoreDependenceInvolvingLoopInvariantAddress()

bool llvm::LoopAccessInfo::hasStoreStoreDependenceInvolvingLoopInvariantAddress ( ) const

inline

Return true if the loop has memory dependence involving two stores to an invariant address, else return false.

Definition at line 688 of file LoopAccessAnalysis.h.

◆ isInvariant()

bool LoopAccessInfo::isInvariant ( Value * V ) const

Returns true if value V is loop invariant.

Definition at line 2789 of file LoopAccessAnalysis.cpp.

References llvm::ScalarEvolution::getSCEV(), llvm::PredicatedScalarEvolution::getSE(), llvm::ScalarEvolution::isLoopInvariant(), and llvm::ScalarEvolution::isSCEVable().

Referenced by llvm::LoopVectorizationLegality::isInvariant().

◆ print()

void LoopAccessInfo::print	(	raw_ostream &	OS,
		unsigned	Depth = `0`
	)		const

Print the information about the memory accesses in the loop.

Definition at line 3018 of file LoopAccessAnalysis.cpp.

References llvm::Depth, getDepChecker(), llvm::MemoryDepChecker::getDependences(), llvm::MemoryDepChecker::getMaxSafeVectorWidthInBits(), llvm::MemoryDepChecker::getMemoryInstructions(), llvm::PredicatedScalarEvolution::getPredicate(), llvm::raw_ostream::indent(), llvm::MemoryDepChecker::isSafeForAnyVectorWidth(), OS, llvm::PredicatedScalarEvolution::print(), llvm::MemoryDepChecker::Dependence::print(), and llvm::SCEVPredicate::print().

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

include/llvm/Analysis/LoopAccessAnalysis.h
lib/Analysis/LoopAccessAnalysis.cpp

Public Member Functions

Static Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ LoopAccessInfo()

Member Function Documentation

◆ blockNeedsPredication()

◆ canVectorizeMemory()

◆ getDepChecker()

◆ getInstructionsForAccess()

◆ getNumLoads()

◆ getNumRuntimePointerChecks()

◆ getNumStores()

◆ getPSE()

◆ getReport()

◆ getRuntimePointerChecking()

◆ getStoresToInvariantAddresses()

◆ getSymbolicStrides()

◆ hasConvergentOp()

◆ hasLoadStoreDependenceInvolvingLoopInvariantAddress()

◆ hasStoreStoreDependenceInvolvingLoopInvariantAddress()

◆ isInvariant()

◆ print()