LoopPeel.cpp File Reference

#include "llvm/Transforms/Utils/LoopPeel.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/Optional.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/LoopIterator.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/ScalarEvolutionExpressions.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/InstrTypes.h"
#include "llvm/IR/Instruction.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/MDBuilder.h"
#include "llvm/IR/Metadata.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/BasicBlockUtils.h"
#include "llvm/Transforms/Utils/Cloning.h"
#include "llvm/Transforms/Utils/LoopSimplify.h"
#include "llvm/Transforms/Utils/LoopUtils.h"
#include "llvm/Transforms/Utils/UnrollLoop.h"
#include "llvm/Transforms/Utils/ValueMapper.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <limits>

Include dependency graph for LoopPeel.cpp:

Go to the source code of this file.

Macros
#define	DEBUG_TYPE   "loop-peel"

Functions
	STATISTIC (NumPeeled, "Number of loops peeled")
static unsigned	calculateIterationsToInvariance (PHINode *Phi, Loop *L, BasicBlock *BackEdge, SmallDenseMap< PHINode *, unsigned > &IterationsToInvariance)
static unsigned	countToEliminateCompares (Loop &L, unsigned MaxPeelCount, ScalarEvolution &SE)
static void	updateBranchWeights (BasicBlock *Header, BranchInst *LatchBR, uint64_t ExitWeight, uint64_t &FallThroughWeight)
	Update the branch weights of the latch of a peeled-off loop iteration. More...
static void	initBranchWeights (BasicBlock *Header, BranchInst *LatchBR, uint64_t &ExitWeight, uint64_t &FallThroughWeight)
	Initialize the weights. More...
static void	fixupBranchWeights (BasicBlock *Header, BranchInst *LatchBR, uint64_t ExitWeight, uint64_t FallThroughWeight)
	Update the weights of original Latch block after peeling off all iterations. More...
static void	cloneLoopBlocks (Loop *L, unsigned IterNumber, BasicBlock *InsertTop, BasicBlock *InsertBot, SmallVectorImpl< std::pair< BasicBlock *, BasicBlock * >> &ExitEdges, SmallVectorImpl< BasicBlock * > &NewBlocks, LoopBlocksDFS &LoopBlocks, ValueToValueMapTy &VMap, ValueToValueMapTy &LVMap, DominatorTree *DT, LoopInfo *LI, ArrayRef< MDNode * > LoopLocalNoAliasDeclScopes)
	Clones the body of the loop L, putting it between InsertTop and InsertBot. More...

Variables
static cl::opt< unsigned >	UnrollPeelCount ("unroll-peel-count", cl::Hidden, cl::desc("Set the unroll peeling count, for testing purposes"))
static cl::opt< bool >	UnrollAllowPeeling ("unroll-allow-peeling", cl::init(true), cl::Hidden, cl::desc("Allows loops to be peeled when the dynamic " "trip count is known to be low."))
static cl::opt< bool >	UnrollAllowLoopNestsPeeling ("unroll-allow-loop-nests-peeling", cl::init(false), cl::Hidden, cl::desc("Allows loop nests to be peeled."))
static cl::opt< unsigned >	UnrollPeelMaxCount ("unroll-peel-max-count", cl::init(7), cl::Hidden, cl::desc("Max average trip count which will cause loop peeling."))
static cl::opt< unsigned >	UnrollForcePeelCount ("unroll-force-peel-count", cl::init(0), cl::Hidden, cl::desc("Force a peel count regardless of profiling information."))
static const char *	PeeledCountMetaData = "llvm.loop.peeled.count"
static const unsigned	InfiniteIterationsToInvariance

Macro Definition Documentation

DEBUG_TYPE

#define DEBUG_TYPE   "loop-peel"

Definition at line 50 of file LoopPeel.cpp.

Function Documentation

calculateIterationsToInvariance()

static unsigned calculateIterationsToInvariance ( PHINode *  Phi, Loop *  L, BasicBlock *  BackEdge, SmallDenseMap< PHINode *, unsigned > &  IterationsToInvariance  )

static

Definition at line 130 of file LoopPeel.cpp.

References assert(), llvm::DenseMapBase< SmallDenseMap< KeyT, ValueT, 4, DenseMapInfo< KeyT >, llvm::detail::DenseMapPair< KeyT, ValueT > >, KeyT, ValueT, DenseMapInfo< KeyT >, llvm::detail::DenseMapPair< KeyT, ValueT > >::end(), llvm::DenseMapBase< SmallDenseMap< KeyT, ValueT, 4, DenseMapInfo< KeyT >, llvm::detail::DenseMapPair< KeyT, ValueT > >, KeyT, ValueT, DenseMapInfo< KeyT >, llvm::detail::DenseMapPair< KeyT, ValueT > >::find(), llvm::LoopBase< BlockT, LoopT >::getHeader(), llvm::PHINode::getIncomingValueForBlock(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), llvm::Instruction::getParent(), I, InfiniteIterationsToInvariance, and llvm::Loop::isLoopInvariant().

Referenced by llvm::computePeelCount().

cloneLoopBlocks()

static void cloneLoopBlocks ( Loop *  L, unsigned  IterNumber, BasicBlock *  InsertTop, BasicBlock *  InsertBot, SmallVectorImpl< std::pair< BasicBlock *, BasicBlock * >> &  ExitEdges, SmallVectorImpl< BasicBlock * > &  NewBlocks, LoopBlocksDFS &  LoopBlocks, ValueToValueMapTy &  VMap, ValueToValueMapTy &  LVMap, DominatorTree *  DT, LoopInfo *  LI, ArrayRef< MDNode * >  LoopLocalNoAliasDeclScopes  )

static

Clones the body of the loop L, putting it between InsertTop and InsertBot.

Parameters

	IterNumber	The serial number of the iteration currently being peeled off.
	ExitEdges	The exit edges of the original loop.
[out]	NewBlocks	A list of the blocks in the newly created clone
[out]	VMap	The value map between the loop and the new clone.
	LoopBlocks	A helper for DFS-traversal of the loop.
	LVMap	A value-map that maps instructions from the original loop to instructions in the last peeled-off iteration.

Definition at line 493 of file LoopPeel.cpp.

References llvm::LoopBase< BlockT, LoopT >::addBasicBlockToLoop(), llvm::DominatorTreeBase< NodeT, IsPostDom >::addNewBlock(), BB, llvm::LoopBlocksDFS::beginRPO(), llvm::DominatorTreeBase< NodeT, IsPostDom >::changeImmediateDominator(), llvm::cloneAndAdaptNoAliasScopes(), llvm::CloneBasicBlock(), llvm::cloneLoop(), llvm::numbers::e, llvm::LoopBlocksDFS::endRPO(), llvm::MipsISD::Ext, F, llvm::DomTreeNodeBase< NodeT >::getBlock(), llvm::BasicBlock::getContext(), llvm::LoopBase< BlockT, LoopT >::getHeader(), llvm::DomTreeNodeBase< NodeT >::getIDom(), llvm::PHINode::getIncomingValueForBlock(), llvm::LoopInfoBase< BlockT, LoopT >::getLoopFor(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), llvm::LoopBase< BlockT, LoopT >::getLoopPreheader(), llvm::DominatorTreeBase< NodeT, IsPostDom >::getNode(), llvm::BranchInst::getNumSuccessors(), llvm::BasicBlock::getParent(), llvm::LoopBase< BlockT, LoopT >::getParentLoop(), llvm::BranchInst::getSuccessor(), llvm::BasicBlock::getTerminator(), I, llvm::Instruction::setSuccessor(), and llvm::BranchInst::setSuccessor().

Referenced by llvm::peelLoop().

countToEliminateCompares()

static unsigned countToEliminateCompares ( Loop &  L, unsigned  MaxPeelCount, ScalarEvolution &  SE  )

static

Definition at line 177 of file LoopPeel.cpp.

References assert(), BB, llvm::LoopBase< BlockT, LoopT >::blocks(), llvm::SCEVAddRecExpr::evaluateAtIteration(), llvm::ScalarEvolution::evaluatePredicate(), llvm::ScalarEvolution::getAddExpr(), llvm::ScalarEvolution::getConstant(), llvm::CmpInst::getInversePredicate(), llvm::SCEVAddRecExpr::getLoop(), llvm::LoopBase< BlockT, LoopT >::getLoopLatch(), llvm::ScalarEvolution::getMonotonicPredicateType(), llvm::ScalarEvolution::getSCEV(), llvm::SCEVAddRecExpr::getStepRecurrence(), llvm::CmpInst::getSwappedPredicate(), llvm::SCEV::getType(), llvm::SCEVNAryExpr::hasNoSelfWrap(), llvm::SCEVAddRecExpr::isAffine(), llvm::ICmpInst::isEquality(), llvm::ScalarEvolution::isKnownPredicate(), llvm::Loop::isLoopSimplifyForm(), llvm::PatternMatch::m_ICmp(), llvm::PatternMatch::m_Value(), llvm::PatternMatch::match(), llvm::max(), and std::swap().

Referenced by llvm::computePeelCount().

fixupBranchWeights()

static void fixupBranchWeights ( BasicBlock *  Header, BranchInst *  LatchBR, uint64_t  ExitWeight, uint64_t  FallThroughWeight  )

static

Update the weights of original Latch block after peeling off all iterations.

Parameters

Header	The header block.
LatchBR	The latch branch.
ExitWeight	The weight of the edge from Latch to Exit.
FallThroughWeight	The weight of the edge from Latch to Header.

Definition at line 466 of file LoopPeel.cpp.

References llvm::MDBuilder::createBranchWeights(), llvm::Value::getContext(), llvm::BranchInst::getSuccessor(), and llvm::Instruction::setMetadata().

Referenced by llvm::peelLoop().

initBranchWeights()

static void initBranchWeights ( BasicBlock *  Header, BranchInst *  LatchBR, uint64_t &  ExitWeight, uint64_t &  FallThroughWeight  )

static

Initialize the weights.

Parameters

	Header	The header block.
	LatchBR	The latch branch.
[out]	ExitWeight	The weight of the edge from Latch to Exit.
[out]	FallThroughWeight	The weight of the edge from Latch to Header.

Definition at line 449 of file LoopPeel.cpp.

References llvm::Instruction::extractProfMetadata(), and llvm::BranchInst::getSuccessor().

Referenced by llvm::peelLoop().

STATISTIC()

STATISTIC	(	NumPeeled	,
		"Number of loops peeled"
	)

updateBranchWeights()

static void updateBranchWeights ( BasicBlock *  Header, BranchInst *  LatchBR, uint64_t  ExitWeight, uint64_t &  FallThroughWeight  )

static

Update the branch weights of the latch of a peeled-off loop iteration.

This sets the branch weights for the latch of the recently peeled off loop iteration correctly. Let F is a weight of the edge from latch to header. Let E is a weight of the edge from latch to exit. F/(F+E) is a probability to go to loop and E/(F+E) is a probability to go to exit. Then, Estimated TripCount = F / E. For I-th (counting from 0) peeled off iteration we set the the weights for the peeled latch as (TC - I, 1). It gives us reasonable distribution, The probability to go to exit 1/(TC-I) increases. At the same time the estimated trip count of remaining loop reduces by I. To avoid dealing with division rounding we can just multiple both part of weights to E and use weight as (F - I * E, E).

Parameters

	Header	The copy of the header block that belongs to next iteration.
	LatchBR	The copy of the latch branch that belongs to this iteration.
[in,out]	FallThroughWeight	The weight of the edge from latch to header before peeling (in) and after peeled off one iteration (out).

Definition at line 425 of file LoopPeel.cpp.

References llvm::MDBuilder::createBranchWeights(), llvm::Value::getContext(), llvm::BranchInst::getSuccessor(), and llvm::Instruction::setMetadata().

Referenced by llvm::peelLoop().

Variable Documentation

InfiniteIterationsToInvariance

const unsigned InfiniteIterationsToInvariance

static

Initial value:

=
    std::numeric_limits<unsigned>::max()

Definition at line 81 of file LoopPeel.cpp.

Referenced by calculateIterationsToInvariance(), and llvm::computePeelCount().

PeeledCountMetaData

const char* PeeledCountMetaData = "llvm.loop.peeled.count"

static

Definition at line 76 of file LoopPeel.cpp.

Referenced by llvm::computePeelCount(), and llvm::peelLoop().

UnrollAllowLoopNestsPeeling

cl::opt<bool> UnrollAllowLoopNestsPeeling("unroll-allow-loop-nests-peeling", cl::init(false), cl::Hidden, cl::desc("Allows loop nests to be peeled."))

static

Referenced by llvm::gatherPeelingPreferences().

UnrollAllowPeeling

cl::opt<bool> UnrollAllowPeeling("unroll-allow-peeling", cl::init(true), cl::Hidden, cl::desc("Allows loops to be peeled when the dynamic " "trip count is known to be low."))

static

Referenced by llvm::gatherPeelingPreferences().

UnrollForcePeelCount

cl::opt<unsigned> UnrollForcePeelCount("unroll-force-peel-count", cl::init(0), cl::Hidden, cl::desc("Force a peel count regardless of profiling information."))

static

Referenced by llvm::computePeelCount().

UnrollPeelCount

cl::opt<unsigned> UnrollPeelCount("unroll-peel-count", cl::Hidden, cl::desc("Set the unroll peeling count, for testing purposes"))

static

Referenced by llvm::gatherPeelingPreferences().

UnrollPeelMaxCount

cl::opt<unsigned> UnrollPeelMaxCount("unroll-peel-max-count", cl::init(7), cl::Hidden, cl::desc("Max average trip count which will cause loop peeling."))

static

Referenced by llvm::computePeelCount().