Google Git Google Git
Sign in
llvm/llvm-project/56ee698b48ab31be816dbdfd4bea8c3761663dbe/./llvm/lib/Analysis/LoopAccessAnalysis.cpp
blob: 00b67ed4b63f02de42c97f020544a4c44e09e330 [file]
//===- LoopAccessAnalysis.cpp - Loop Access Analysis Implementation --------==//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// The implementation for the loop memory dependence that was originally
// developed for the loop vectorizer.
//
//===----------------------------------------------------------------------===//
#include "llvm/Analysis/LoopAccessAnalysis.h"
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/EquivalenceClasses.h"
#include "llvm/ADT/PointerIntPair.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/AliasSetTracker.h"
#include "llvm/Analysis/AssumeBundleQueries.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/LoopAnalysisManager.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/LoopIterator.h"
#include "llvm/Analysis/MemoryLocation.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/ScalarEvolutionExpressions.h"
#include "llvm/Analysis/ScalarEvolutionPatternMatch.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/Analysis/VectorUtils.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DebugLoc.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/DiagnosticInfo.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/IntrinsicInst.h"
#include "llvm/IR/PassManager.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/Value.h"
#include "llvm/IR/ValueHandle.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <iterator>
#include <utility>
#include <variant>
#include <vector>
using namespace llvm;
using namespace llvm::SCEVPatternMatch;
#define DEBUG_TYPE "loop-accesses"
static cl::opt<ElementCount, true>
VectorizationFactor("force-vector-width", cl::Hidden,
cl::desc("Sets the SIMD width. Zero is autoselect."),
cl::location(VectorizerParams::VectorizationFactor));
ElementCount VectorizerParams::VectorizationFactor;
static cl::opt<unsigned, true>
VectorizationInterleave("force-vector-interleave", cl::Hidden,
cl::desc("Sets the vectorization interleave count. "
"Zero is autoselect."),
cl::location(
VectorizerParams::VectorizationInterleave));
unsigned VectorizerParams::VectorizationInterleave;
static cl::opt<unsigned, true> RuntimeMemoryCheckThreshold(
"runtime-memory-check-threshold", cl::Hidden,
cl::desc("When performing memory disambiguation checks at runtime do not "
"generate more than this number of comparisons (default = 8)."),
cl::location(VectorizerParams::RuntimeMemoryCheckThreshold), cl::init(8));
unsigned VectorizerParams::RuntimeMemoryCheckThreshold;
/// The maximum iterations used to merge memory checks
static cl::opt<unsigned> MemoryCheckMergeThreshold(
"memory-check-merge-threshold", cl::Hidden,
cl::desc("Maximum number of comparisons done when trying to merge "
"runtime memory checks. (default = 100)"),
cl::init(100));
/// Maximum SIMD width.
const unsigned VectorizerParams::MaxVectorWidth = 64;
/// We collect dependences up to this threshold.
static cl::opt<unsigned>
MaxDependences("max-dependences", cl::Hidden,
cl::desc("Maximum number of dependences collected by "
"loop-access analysis (default = 100)"),
cl::init(100));
/// This enables versioning on the strides of symbolically striding memory
/// accesses in code like the following.
/// for (i = 0; i < N; ++i)
/// A[i * Stride1] += B[i * Stride2] ...
///
/// Will be roughly translated to
/// if (Stride1 == 1 && Stride2 == 1) {
/// for (i = 0; i < N; i+=4)
/// A[i:i+3] += ...
/// } else
/// ...
static cl::opt<bool> EnableMemAccessVersioning(
"enable-mem-access-versioning", cl::init(true), cl::Hidden,
cl::desc("Enable symbolic stride memory access versioning"));
/// Enable store-to-load forwarding conflict detection. This option can
/// be disabled for correctness testing.
static cl::opt<bool> EnableForwardingConflictDetection(
"store-to-load-forwarding-conflict-detection", cl::Hidden,
cl::desc("Enable conflict detection in loop-access analysis"),
cl::init(true));
static cl::opt<unsigned> MaxForkedSCEVDepth(
"max-forked-scev-depth", cl::Hidden,
cl::desc("Maximum recursion depth when finding forked SCEVs (default = 5)"),
cl::init(5));
static cl::opt<bool> SpeculateUnitStride(
"laa-speculate-unit-stride", cl::Hidden,
cl::desc("Speculate that non-constant strides are unit in LAA"),
cl::init(true));
static cl::opt<bool, true> HoistRuntimeChecks(
"hoist-runtime-checks", cl::Hidden,
cl::desc(
"Hoist inner loop runtime memory checks to outer loop if possible"),
cl::location(VectorizerParams::HoistRuntimeChecks), cl::init(true));
bool VectorizerParams::HoistRuntimeChecks;
bool VectorizerParams::isInterleaveForced() {
return ::VectorizationInterleave.getNumOccurrences() > 0;
}
const SCEV *
llvm::replaceSymbolicStrideSCEV(PredicatedScalarEvolution &PSE,
const SymbolicStrideMap &PtrToStride,
Value *Ptr) {
const SCEV *OrigSCEV = PSE.getSCEV(Ptr);
// If there is an entry in the map return the SCEV of the pointer with the
// symbolic stride replaced by one.
const SCEVUnknown *StrideSCEV = PtrToStride.lookup(Ptr);
if (!StrideSCEV)
// For a non-symbolic stride, just return the original expression.
return OrigSCEV;
ScalarEvolution *SE = PSE.getSE();
const SCEV *CT = SE->getOne(StrideSCEV->getType());
PSE.addPredicate(*SE->getEqualPredicate(StrideSCEV, CT));
const SCEV *Expr = PSE.getSCEV(Ptr);
LLVM_DEBUG(dbgs() << "LAA: Replacing SCEV: " << *OrigSCEV
<< " by: " << *Expr << "\n");
return Expr;
}
RuntimeCheckingPtrGroup::RuntimeCheckingPtrGroup(
unsigned Index, const RuntimePointerChecking &RtCheck)
: High(RtCheck.Pointers[Index].End), Low(RtCheck.Pointers[Index].Start),
AddressSpace(RtCheck.Pointers[Index]
.PointerValue->getType()
->getPointerAddressSpace()),
NeedsFreeze(RtCheck.Pointers[Index].NeedsFreeze) {
Members.push_back(Index);
}
/// Returns \p A + \p B, if it is guaranteed not to unsigned wrap. Otherwise
/// return nullptr. \p A and \p B must have the same type.
static const SCEV *addSCEVNoOverflow(const SCEV *A, const SCEV *B,
ScalarEvolution &SE) {
if (!SE.willNotOverflow(Instruction::Add, /*IsSigned=*/false, A, B))
return nullptr;
return SE.getAddExpr(A, B);
}
/// Returns \p A * \p B, if it is guaranteed not to unsigned wrap. Otherwise
/// return nullptr. \p A and \p B must have the same type.
static const SCEV *mulSCEVNoOverflow(const SCEV *A, const SCEV *B,
ScalarEvolution &SE) {
if (!SE.willNotOverflow(Instruction::Mul, /*IsSigned=*/false, A, B))
return nullptr;
return SE.getMulExpr(A, B);
}
/// Return true, if evaluating \p AR at \p MaxBTC cannot wrap, because \p AR at
/// \p MaxBTC is guaranteed inbounds of the accessed object.
static bool evaluatePtrAddRecAtMaxBTCWillNotWrap(
const SCEVAddRecExpr *AR, const SCEV *MaxBTC, const SCEV *EltSize,
ScalarEvolution &SE, const DataLayout &DL, DominatorTree *DT,
AssumptionCache *AC,
std::optional<ScalarEvolution::LoopGuards> &LoopGuards) {
auto *PointerBase = SE.getPointerBase(AR->getStart());
auto *StartPtr = dyn_cast<SCEVUnknown>(PointerBase);
if (!StartPtr)
return false;
const Loop *L = AR->getLoop();
bool CheckForNonNull;
Value *StartPtrV = StartPtr->getValue();
// We can ignore frees, as the fact that an object of a certain size existed
// at the location *at some point* is sufficient to derive the nowrap fact.
uint64_t DerefBytes = StartPtrV->getPointerDereferenceableBytes(
DL, CheckForNonNull, /*CanBeFreed=*/nullptr);
// If the deref size is only known when the pointer is non-null, ignore it
// here and fall back to a dereferenceable assumption below.
if (DerefBytes && CheckForNonNull)
DerefBytes = 0;
const SCEV *Step = AR->getStepRecurrence(SE);
Type *WiderTy = SE.getWiderType(MaxBTC->getType(), Step->getType());
const SCEV *DerefBytesSCEV = SE.getConstant(WiderTy, DerefBytes);
// Check if we have a suitable dereferencable assumption we can use.
Instruction *CtxI = &*L->getHeader()->getFirstNonPHIIt();
if (BasicBlock *LoopPred = L->getLoopPredecessor()) {
if (isa<UncondBrInst, CondBrInst>(LoopPred->getTerminator()))
CtxI = LoopPred->getTerminator();
}
getKnowledgeForValue(
StartPtrV, Attribute::Dereferenceable, *AC,
[&](RetainedKnowledge RK, Instruction *Assume, auto) {
if (!isValidAssumeForContext(Assume, CtxI, DT))
return false;
const SCEV *DerefRKSCEV = SE.getSCEV(RK.IRArgValue);
Type *CommonTy =
SE.getWiderType(DerefBytesSCEV->getType(), DerefRKSCEV->getType());
DerefBytesSCEV = SE.getNoopOrZeroExtend(DerefBytesSCEV, CommonTy);
DerefRKSCEV = SE.getNoopOrZeroExtend(DerefRKSCEV, CommonTy);
DerefBytesSCEV = SE.getUMaxExpr(DerefBytesSCEV, DerefRKSCEV);
// Continue with other assumptions.
return false;
});
if (DerefBytesSCEV->isZero())
return false;
bool IsKnownNonNegative = SE.isKnownNonNegative(Step);
if (!IsKnownNonNegative && !SE.isKnownNegative(Step))
return false;
WiderTy = SE.getWiderType(WiderTy, DerefBytesSCEV->getType());
Step = SE.getNoopOrSignExtend(Step, WiderTy);
MaxBTC = SE.getNoopOrZeroExtend(MaxBTC, WiderTy);
// For the computations below, make sure they don't unsigned wrap.
// FIXME: for a negative step the lowest accessed address is not
// AR->getStart() but AR->evaluateAtIteration(MaxBTC, SE); the check below
// therefore compares StartPtr against the highest accessed address instead
// of the lowest.
if (!SE.isKnownPredicate(CmpInst::ICMP_UGE, AR->getStart(), StartPtr))
return false;
const SCEV *StartOffset = SE.getNoopOrZeroExtend(
SE.getMinusSCEV(AR->getStart(), StartPtr), WiderTy);
if (!LoopGuards)
LoopGuards.emplace(ScalarEvolution::LoopGuards::collect(AR->getLoop(), SE));
MaxBTC = SE.applyLoopGuards(MaxBTC, *LoopGuards);
const SCEV *AbsStep = SE.getAbsExpr(Step, /*IsNSW=*/false);
// Total distance (in bytes) between the first and the last
// accessed pointer.
const SCEV *DistToLastIter = mulSCEVNoOverflow(MaxBTC, AbsStep, SE);
if (!DistToLastIter) {
// Re-try with constant max backedge-taken count if using the symbolic one
// failed.
MaxBTC = SE.getConstantMaxBackedgeTakenCount(AR->getLoop());
if (isa<SCEVCouldNotCompute>(MaxBTC))
return false;
MaxBTC = SE.getNoopOrZeroExtend(MaxBTC, WiderTy);
DistToLastIter = mulSCEVNoOverflow(MaxBTC, AbsStep, SE);
if (!DistToLastIter)
return false;
}
// Total length in bytes of the accessed range (from the first accessed
// byte through the end of the last access).
const SCEV *AccessedBytes = addSCEVNoOverflow(
DistToLastIter, SE.getNoopOrZeroExtend(EltSize, WiderTy), SE);
if (!AccessedBytes)
return false;
// Compute MaxOffset per direction: exclusive upper offset of the
// accessed range.
const SCEV *MaxOffset;
if (IsKnownNonNegative) {
MaxOffset = addSCEVNoOverflow(StartOffset, AccessedBytes, SE);
if (!MaxOffset)
return false;
DerefBytesSCEV = SE.applyLoopGuards(DerefBytesSCEV, *LoopGuards);
} else {
// FIXME: two independent off-by-EltSize bugs on this branch:
// 1. StartOffset here is actually the HIGHEST offset, because it is
// computed from AR->getStart() rather than
// AR->evaluateAtIteration(MaxBTC, SE) (see FIXME above).
// 2. The lower check is over-strict by EltSize and the upper is
// under-counted by EltSize.
assert(SE.isKnownNegative(Step) && "must be known negative");
if (!SE.isKnownPredicate(CmpInst::ICMP_SGE, StartOffset, AccessedBytes))
return false;
MaxOffset = StartOffset;
}
// MaxOffset must not exceed the deref-region end.
return SE.isKnownPredicate(CmpInst::ICMP_ULE, MaxOffset, DerefBytesSCEV);
}
/// Return true if \p S is known to be monotonically non-decreasing
/// (in the unsigned sense, without unsigned wrap) across iterations of \p L.
static bool isKnownNonDecreasingInLoop(const SCEV *S, const Loop *L,
ScalarEvolution &SE) {
if (SE.isLoopInvariant(S, L))
return true;
switch (S->getSCEVType()) {
case scUDivExpr: {
// Non-decreasing in the numerator when the divisor is loop-invariant.
const auto *UDiv = cast<SCEVUDivExpr>(S);
return SE.isLoopInvariant(UDiv->getRHS(), L) &&
isKnownNonDecreasingInLoop(UDiv->getLHS(), L, SE);
}
case scAddRecExpr: {
auto *AR = cast<SCEVAddRecExpr>(S);
assert(AR->getLoop() == L &&
"trying to check for AddRec in different loop");
return SE.getMonotonicPredicateType(AR, ICmpInst::ICMP_UGE) ==
ScalarEvolution::MonotonicPredicateType::MonotonicallyIncreasing;
}
default:
return false;
}
}
/// Try to bound a loop-variant pointer that is not an affine AddRec.
///
/// If the offset is provably monotonically non-decreasing the accessed range is
/// bounded by the offset's value at the first iteration (via
/// SplitIntoInitAndPostInc) and last iteration (via getSCEVAtScope). The
/// returned range is half-open: \p EltSizeSCEV is added to the address of the
/// last accessed element to form the end.
///
/// Returns {nullptr, nullptr} if no such bound can be formed.
static std::pair<const SCEV *, const SCEV *>
getNonAffineMonotonicBounds(const Loop *Lp, const SCEV *PtrExpr,
const SCEV *EltSizeSCEV, ScalarEvolution *SE) {
const auto *PtrAdd = dyn_cast<SCEVAddExpr>(PtrExpr);
if (!PtrAdd || !PtrAdd->hasNoUnsignedWrap())
return {nullptr, nullptr};
const SCEV *Base = *find_if(PtrAdd->operands(), [](const auto &Op) {
return Op->getType()->isPointerTy();
});
if (isa<SCEVCouldNotCompute>(Base) || !SE->isLoopInvariant(Base, Lp))
return {nullptr, nullptr};
const SCEV *Offset = SE->getMinusSCEV(PtrExpr, Base);
if (isa<SCEVCouldNotCompute>(Offset) ||
!isKnownNonDecreasingInLoop(Offset, Lp, *SE))
return {nullptr, nullptr};
const SCEV *OffStart = SE->SplitIntoInitAndPostInc(Lp, Offset).first;
const SCEV *OffEnd = SE->getSCEVAtScope(Offset, Lp->getParentLoop());
if (isa<SCEVCouldNotCompute>(OffStart) || isa<SCEVCouldNotCompute>(OffEnd) ||
!SE->isLoopInvariant(OffStart, Lp) || !SE->isLoopInvariant(OffEnd, Lp))
return {nullptr, nullptr};
return {SE->getAddExpr(Base, OffStart),
SE->getAddExpr(Base, OffEnd, EltSizeSCEV)};
}
std::pair<const SCEV *, const SCEV *> llvm::getStartAndEndForAccess(
const Loop *Lp, const SCEV *PtrExpr, Type *AccessTy, const SCEV *BTC,
const SCEV *MaxBTC, ScalarEvolution *SE,
DenseMap<std::pair<const SCEV *, const SCEV *>,
std::pair<const SCEV *, const SCEV *>> *PointerBounds,
DominatorTree *DT, AssumptionCache *AC,
std::optional<ScalarEvolution::LoopGuards> &LoopGuards) {
auto &DL = Lp->getHeader()->getDataLayout();
Type *IdxTy = DL.getIndexType(PtrExpr->getType());
const SCEV *EltSizeSCEV = SE->getStoreSizeOfExpr(IdxTy, AccessTy);
// Delegate to the SCEV-based overload, passing through the cache.
return getStartAndEndForAccess(Lp, PtrExpr, EltSizeSCEV, BTC, MaxBTC, SE,
PointerBounds, DT, AC, LoopGuards);
}
std::pair<const SCEV *, const SCEV *> llvm::getStartAndEndForAccess(
const Loop *Lp, const SCEV *PtrExpr, const SCEV *EltSizeSCEV,
const SCEV *BTC, const SCEV *MaxBTC, ScalarEvolution *SE,
DenseMap<std::pair<const SCEV *, const SCEV *>,
std::pair<const SCEV *, const SCEV *>> *PointerBounds,
DominatorTree *DT, AssumptionCache *AC,
std::optional<ScalarEvolution::LoopGuards> &LoopGuards) {
std::pair<const SCEV *, const SCEV *> *PtrBoundsPair;
if (PointerBounds) {
auto [Iter, Ins] = PointerBounds->insert(
{{PtrExpr, EltSizeSCEV},
{SE->getCouldNotCompute(), SE->getCouldNotCompute()}});
if (!Ins)
return Iter->second;
PtrBoundsPair = &Iter->second;
}
// ScStart is the lowest accessed address; ScEnd is the highest one plus the
// size of the accessed element.
const SCEV *ScStart;
const SCEV *ScEnd;
auto &DL = Lp->getHeader()->getDataLayout();
if (SE->isLoopInvariant(PtrExpr, Lp)) {
ScStart = PtrExpr;
ScEnd = SE->getAddExpr(PtrExpr, EltSizeSCEV);
} else if (auto *AR = dyn_cast<SCEVAddRecExpr>(PtrExpr)) {
const SCEV *Step = AR->getStepRecurrence(*SE);
// The address of the last accessed element, if it can be computed
// precisely.
const SCEV *LastAddr = nullptr;
if (!isa<SCEVCouldNotCompute>(BTC)) {
// Evaluating AR at an exact BTC is safe: LAA separately checks that
// accesses cannot wrap in the loop. If evaluating AR at BTC wraps, then
// the loop either triggers UB when executing a memory access with a
// poison pointer or the wrapping/poisoned pointer is not used.
LastAddr = AR->evaluateAtIteration(BTC, *SE);
} else if (evaluatePtrAddRecAtMaxBTCWillNotWrap(
AR, MaxBTC, EltSizeSCEV, *SE, DL, DT, AC, LoopGuards)) {
LastAddr = AR->evaluateAtIteration(MaxBTC, *SE);
}
const SCEV *Start = AR->getStart();
Type *PtrTy = AR->getType();
if (SE->isKnownNegative(Step)) {
ScStart =
LastAddr
? LastAddr
: SE->getSCEV(ConstantExpr::getIntToPtr(
Constant::getNullValue(DL.getIndexType(PtrTy)), PtrTy));
ScEnd = SE->getAddExpr(Start, EltSizeSCEV);
} else if (SE->isKnownNonNegative(Step)) {
ScStart = Start;
// The highest address for the type saturates; adding EltSize to it would
// wrap to the start of the address space.
if (LastAddr)
ScEnd = SE->getAddExpr(LastAddr, EltSizeSCEV);
else
ScEnd = SE->getSCEV(ConstantExpr::getIntToPtr(
Constant::getAllOnesValue(DL.getIndexType(PtrTy)), PtrTy));
} else {
if (!LastAddr)
return {SE->getCouldNotCompute(), SE->getCouldNotCompute()};
// Fallback case: the step is not constant, but we can still
// get the upper and lower bounds of the interval by using min/max
// expressions.
ScStart = SE->getUMinExpr(Start, LastAddr);
ScEnd = SE->getAddExpr(SE->getUMaxExpr(Start, LastAddr), EltSizeSCEV);
}
} else {
// The pointer is loop-variant but not an affine AddRec. Try to form a
// tight bound for a monotonic offset (see getNonAffineMonotonicBounds).
std::tie(ScStart, ScEnd) =
getNonAffineMonotonicBounds(Lp, PtrExpr, EltSizeSCEV, SE);
if (!ScStart)
return {SE->getCouldNotCompute(), SE->getCouldNotCompute()};
}
assert(SE->isLoopInvariant(ScStart, Lp) && "ScStart needs to be invariant");
assert(SE->isLoopInvariant(ScEnd, Lp) && "ScEnd needs to be invariant");
std::pair<const SCEV *, const SCEV *> Res = {ScStart, ScEnd};
if (PointerBounds)
*PtrBoundsPair = Res;
return Res;
}
/// Calculate Start and End points of memory access using
/// getStartAndEndForAccess.
bool RuntimePointerChecking::insert(Loop *Lp, Value *Ptr, const SCEV *PtrExpr,
Type *AccessTy, bool WritePtr,
unsigned DepSetId, unsigned ASId,
PredicatedScalarEvolution &PSE,
bool NeedsFreeze) {
const SCEV *SymbolicMaxBTC = PSE.getSymbolicMaxBackedgeTakenCount();
const SCEV *BTC = PSE.getBackedgeTakenCount();
const auto &[ScStart, ScEnd] = getStartAndEndForAccess(
Lp, PtrExpr, AccessTy, BTC, SymbolicMaxBTC, PSE.getSE(),
&DC.getPointerBounds(), DC.getDT(), DC.getAC(), LoopGuards);
if (isa<SCEVCouldNotCompute>(ScStart) || isa<SCEVCouldNotCompute>(ScEnd))
return false;
Pointers.emplace_back(Ptr, ScStart, ScEnd, WritePtr, DepSetId, ASId, PtrExpr,
NeedsFreeze);
return true;
}
bool RuntimePointerChecking::tryToCreateDiffCheck(
const RuntimeCheckingPtrGroup &CGI, const RuntimeCheckingPtrGroup &CGJ) {
// If either group contains multiple different pointers, bail out.
// TODO: Support multiple pointers by using the minimum or maximum pointer,
// depending on src & sink.
if (CGI.Members.size() != 1 || CGJ.Members.size() != 1)
return false;
const PointerInfo *Src = &Pointers[CGI.Members[0]];
const PointerInfo *Sink = &Pointers[CGJ.Members[0]];
// If either pointer is read and written, multiple checks may be needed. Bail
// out.
if (!DC.getOrderForAccess(Src->PointerValue, !Src->IsWritePtr).empty() ||
!DC.getOrderForAccess(Sink->PointerValue, !Sink->IsWritePtr).empty())
return false;
ArrayRef<unsigned> AccSrc =
DC.getOrderForAccess(Src->PointerValue, Src->IsWritePtr);
ArrayRef<unsigned> AccSink =
DC.getOrderForAccess(Sink->PointerValue, Sink->IsWritePtr);
// If either pointer is accessed multiple times, there may not be a clear
// src/sink relation. Bail out for now.
if (AccSrc.size() != 1 || AccSink.size() != 1)
return false;
// If the sink is accessed before src, swap src/sink.
if (AccSink[0] < AccSrc[0])
std::swap(Src, Sink);
const SCEVConstant *Step;
const SCEV *SrcStart;
const SCEV *SinkStart;
const Loop *InnerLoop = DC.getInnermostLoop();
if (!match(Src->Expr,
m_scev_AffineAddRec(m_SCEV(SrcStart), m_SCEVConstant(Step),
m_SpecificLoop(InnerLoop))) ||
!match(Sink->Expr,
m_scev_AffineAddRec(m_SCEV(SinkStart), m_scev_Specific(Step),
m_SpecificLoop(InnerLoop))))
return false;
SmallVector<Instruction *, 4> SrcInsts =
DC.getInstructionsForAccess(Src->PointerValue, Src->IsWritePtr);
SmallVector<Instruction *, 4> SinkInsts =
DC.getInstructionsForAccess(Sink->PointerValue, Sink->IsWritePtr);
Type *SrcTy = getLoadStoreType(SrcInsts[0]);
Type *DstTy = getLoadStoreType(SinkInsts[0]);
if (isa<ScalableVectorType>(SrcTy) || isa<ScalableVectorType>(DstTy))
return false;
const DataLayout &DL = InnerLoop->getHeader()->getDataLayout();
unsigned AllocSize =
std::max(DL.getTypeAllocSize(SrcTy), DL.getTypeAllocSize(DstTy));
// Only matching constant steps matching the AllocSize are supported at the
// moment. This simplifies the difference computation. Can be extended in the
// future.
if (Step->getAPInt().abs() != AllocSize)
return false;
// When counting down, the dependence distance needs to be swapped.
if (Step->getValue()->isNegative())
std::swap(SinkStart, SrcStart);
const SCEV *SinkStartInt = SE->getPtrToAddrExpr(SinkStart);
const SCEV *SrcStartInt = SE->getPtrToAddrExpr(SrcStart);
if (isa<SCEVCouldNotCompute>(SinkStartInt) ||
isa<SCEVCouldNotCompute>(SrcStartInt))
return false;
// If the start values for both Src and Sink also vary according to an outer
// loop, then it's probably better to avoid creating diff checks because
// they may not be hoisted. We should instead let llvm::addRuntimeChecks
// do the expanded full range overlap checks, which can be hoisted.
if (HoistRuntimeChecks && InnerLoop->getParentLoop() &&
isa<SCEVAddRecExpr>(SinkStartInt) && isa<SCEVAddRecExpr>(SrcStartInt)) {
auto *SrcStartAR = cast<SCEVAddRecExpr>(SrcStartInt);
auto *SinkStartAR = cast<SCEVAddRecExpr>(SinkStartInt);
const Loop *StartARLoop = SrcStartAR->getLoop();
if (StartARLoop == SinkStartAR->getLoop() &&
StartARLoop == InnerLoop->getParentLoop() &&
// If the diff check would already be loop invariant (due to the
// recurrences being the same), then we prefer to keep the diff checks
// because they are cheaper.
SrcStartAR->getStepRecurrence(*SE) !=
SinkStartAR->getStepRecurrence(*SE)) {
LLVM_DEBUG(dbgs() << "LAA: Not creating diff runtime check, since these "
"cannot be hoisted out of the outer loop\n");
return false;
}
}
LLVM_DEBUG(dbgs() << "LAA: Creating diff runtime check for:\n"
<< "SrcStart: " << *SrcStartInt << '\n'
<< "SinkStartInt: " << *SinkStartInt << '\n');
DiffChecks.emplace_back(SrcStartInt, SinkStartInt, AllocSize,
Src->NeedsFreeze || Sink->NeedsFreeze);
return true;
}
SmallVector<RuntimePointerCheck, 4> RuntimePointerChecking::generateChecks() {
SmallVector<RuntimePointerCheck, 4> Checks;
for (unsigned I = 0; I < CheckingGroups.size(); ++I) {
for (unsigned J = I + 1; J < CheckingGroups.size(); ++J) {
const RuntimeCheckingPtrGroup &CGI = CheckingGroups[I];
const RuntimeCheckingPtrGroup &CGJ = CheckingGroups[J];
if (needsChecking(CGI, CGJ)) {
CanUseDiffCheck = CanUseDiffCheck && tryToCreateDiffCheck(CGI, CGJ);
Checks.emplace_back(&CGI, &CGJ);
}
}
}
return Checks;
}
void RuntimePointerChecking::generateChecks(
MemoryDepChecker::DepCandidates &DepCands) {
assert(Checks.empty() && "Checks is not empty");
groupChecks(DepCands);
Checks = generateChecks();
}
bool RuntimePointerChecking::needsChecking(
const RuntimeCheckingPtrGroup &M, const RuntimeCheckingPtrGroup &N) const {
for (const auto &I : M.Members)
for (const auto &J : N.Members)
if (needsChecking(I, J))
return true;
return false;
}
/// Compare \p I and \p J and return the minimum.
/// Return nullptr in case we couldn't find an answer.
static const SCEV *getMinFromExprs(const SCEV *I, const SCEV *J,
ScalarEvolution *SE) {
std::optional<APInt> Diff = SE->computeConstantDifference(J, I);
if (!Diff)
return nullptr;
return Diff->isNegative() ? J : I;
}
bool RuntimeCheckingPtrGroup::addPointer(
unsigned Index, const RuntimePointerChecking &RtCheck) {
return addPointer(
Index, RtCheck.Pointers[Index].Start, RtCheck.Pointers[Index].End,
RtCheck.Pointers[Index].PointerValue->getType()->getPointerAddressSpace(),
RtCheck.Pointers[Index].NeedsFreeze, *RtCheck.SE);
}
bool RuntimeCheckingPtrGroup::addPointer(unsigned Index, const SCEV *Start,
const SCEV *End, unsigned AS,
bool NeedsFreeze,
ScalarEvolution &SE) {
assert(AddressSpace == AS &&
"all pointers in a checking group must be in the same address space");
// Compare the starts and ends with the known minimum and maximum
// of this set. We need to know how we compare against the min/max
// of the set in order to be able to emit memchecks.
const SCEV *Min0 = getMinFromExprs(Start, Low, &SE);
if (!Min0)
return false;
const SCEV *Min1 = getMinFromExprs(End, High, &SE);
if (!Min1)
return false;
// Update the low bound expression if we've found a new min value.
if (Min0 == Start)
Low = Start;
// Update the high bound expression if we've found a new max value.
if (Min1 != End)
High = End;
Members.push_back(Index);
this->NeedsFreeze |= NeedsFreeze;
return true;
}
void RuntimePointerChecking::groupChecks(
MemoryDepChecker::DepCandidates &DepCands) {
// We build the groups from dependency candidates equivalence classes
// because:
// - We know that pointers in the same equivalence class share
// the same underlying object and therefore there is a chance
// that we can compare pointers
// - We wouldn't be able to merge two pointers for which we need
// to emit a memcheck. The classes in DepCands are already
// conveniently built such that no two pointers in the same
// class need checking against each other.
// We use the following (greedy) algorithm to construct the groups
// For every pointer in the equivalence class:
// For each existing group:
// - if the difference between this pointer and the min/max bounds
// of the group is a constant, then make the pointer part of the
// group and update the min/max bounds of that group as required.
CheckingGroups.clear();
// If we need to check two pointers to the same underlying object
// with a non-constant difference, we shouldn't perform any pointer
// grouping with those pointers. This is because we can easily get
// into cases where the resulting check would return false, even when
// the accesses are safe.
//
// The following example shows this:
// for (i = 0; i < 1000; ++i)
// a[5000 + i * m] = a[i] + a[i + 9000]
//
// Here grouping gives a check of (5000, 5000 + 1000 * m) against
// (0, 10000) which is always false. However, if m is 1, there is no
// dependence. Not grouping the checks for a[i] and a[i + 9000] allows
// us to perform an accurate check in this case.
//
// In the above case, we have a non-constant distance and an Unknown
// dependence between accesses to the same underlying object, and could retry
// with runtime checks without dependency information being available. In this
// case we will use the fallback path and create separate checking groups for
// accesses not present in DepCands.
unsigned TotalComparisons = 0;
DenseMap<Value *, SmallVector<unsigned>> PositionMap;
for (unsigned Index = 0; Index < Pointers.size(); ++Index)
PositionMap[Pointers[Index].PointerValue].push_back(Index);
// We need to keep track of what pointers we've already seen so we
// don't process them twice.
SmallSet<unsigned, 2> Seen;
// Go through all equivalence classes, get the "pointer check groups"
// and add them to the overall solution. We use the order in which accesses
// appear in 'Pointers' to enforce determinism.
for (unsigned I = 0; I < Pointers.size(); ++I) {
// We've seen this pointer before, and therefore already processed
// its equivalence class.
if (Seen.contains(I))
continue;
MemoryDepChecker::MemAccessInfo Access(Pointers[I].PointerValue,
Pointers[I].IsWritePtr);
// If there is no entry in the dependency partition, there are no potential
// accesses to merge; simply add a new pointer checking group.
if (!DepCands.contains(Access)) {
CheckingGroups.push_back(RuntimeCheckingPtrGroup(I, *this));
continue;
}
SmallVector<RuntimeCheckingPtrGroup, 2> Groups;
// Because DepCands is constructed by visiting accesses in the order in
// which they appear in alias sets (which is deterministic) and the
// iteration order within an equivalence class member is only dependent on
// the order in which unions and insertions are performed on the
// equivalence class, the iteration order is deterministic.
for (auto M : DepCands.members(Access)) {
auto PointerI = PositionMap.find(M.getPointer());
// If we can't find the pointer in PositionMap that means we can't
// generate a memcheck for it.
if (PointerI == PositionMap.end())
continue;
for (unsigned Pointer : PointerI->second) {
bool Merged = false;
// Mark this pointer as seen.
Seen.insert(Pointer);
// Go through all the existing sets and see if we can find one
// which can include this pointer.
for (RuntimeCheckingPtrGroup &Group : Groups) {
// Don't perform more than a certain amount of comparisons.
// This should limit the cost of grouping the pointers to something
// reasonable. If we do end up hitting this threshold, the algorithm
// will create separate groups for all remaining pointers.
if (TotalComparisons > MemoryCheckMergeThreshold)
break;
TotalComparisons++;
if (Group.addPointer(Pointer, *this)) {
Merged = true;
break;
}
}
if (!Merged)
// We couldn't add this pointer to any existing set or the threshold
// for the number of comparisons has been reached. Create a new group
// to hold the current pointer.
Groups.emplace_back(Pointer, *this);
}
}
// We've computed the grouped checks for this partition.
// Save the results and continue with the next one.
llvm::append_range(CheckingGroups, Groups);
}
}
bool RuntimePointerChecking::arePointersInSamePartition(
const SmallVectorImpl<int> &PtrToPartition, unsigned PtrIdx1,
unsigned PtrIdx2) {
return (PtrToPartition[PtrIdx1] != -1 &&
PtrToPartition[PtrIdx1] == PtrToPartition[PtrIdx2]);
}
bool RuntimePointerChecking::needsChecking(unsigned I, unsigned J) const {
const PointerInfo &PointerI = Pointers[I];
const PointerInfo &PointerJ = Pointers[J];
// No need to check if two readonly pointers intersect.
if (!PointerI.IsWritePtr && !PointerJ.IsWritePtr)
return false;
// Only need to check pointers between two different dependency sets.
if (PointerI.DependencySetId == PointerJ.DependencySetId)
return false;
// Only need to check pointers in the same alias set.
return PointerI.AliasSetId == PointerJ.AliasSetId;
}
/// Assign each RuntimeCheckingPtrGroup pointer an index for stable UTC output.
static DenseMap<const RuntimeCheckingPtrGroup *, unsigned>
getPtrToIdxMap(ArrayRef<RuntimeCheckingPtrGroup> CheckingGroups) {
DenseMap<const RuntimeCheckingPtrGroup *, unsigned> PtrIndices;
for (const auto &[Idx, CG] : enumerate(CheckingGroups))
PtrIndices[&CG] = Idx;
return PtrIndices;
}
void RuntimePointerChecking::printChecks(
raw_ostream &OS, const SmallVectorImpl<RuntimePointerCheck> &Checks,
unsigned Depth) const {
unsigned N = 0;
auto PtrIndices = getPtrToIdxMap(CheckingGroups);
for (const auto &[Check1, Check2] : Checks) {
const auto &First = Check1->Members, &Second = Check2->Members;
OS.indent(Depth) << "Check " << N++ << ":\n";
OS.indent(Depth + 2) << "Comparing group GRP" << PtrIndices.at(Check1)
<< ":\n";
for (unsigned K : First)
OS.indent(Depth + 2) << *Pointers[K].PointerValue << "\n";
OS.indent(Depth + 2) << "Against group GRP" << PtrIndices.at(Check2)
<< ":\n";
for (unsigned K : Second)
OS.indent(Depth + 2) << *Pointers[K].PointerValue << "\n";
}
}
void RuntimePointerChecking::print(raw_ostream &OS, unsigned Depth) const {
OS.indent(Depth) << "Run-time memory checks:\n";
printChecks(OS, Checks, Depth);
OS.indent(Depth) << "Grouped accesses:\n";
auto PtrIndices = getPtrToIdxMap(CheckingGroups);
for (const auto &CG : CheckingGroups) {
OS.indent(Depth + 2) << "Group GRP" << PtrIndices.at(&CG) << ":\n";
OS.indent(Depth + 4) << "(Low: " << *CG.Low << " High: " << *CG.High
<< ")\n";
for (unsigned Member : CG.Members) {
OS.indent(Depth + 6) << "Member: " << *Pointers[Member].Expr << "\n";
}
}
}
namespace {
/// Analyses memory accesses in a loop.
///
/// Checks whether run time pointer checks are needed and builds sets for data
/// dependence checking.
class AccessAnalysis {
public:
using MemAccessInfo =
PointerIntPair<Value * /* AccessPtr */, 1, bool /* IsWrite */>;
AccessAnalysis(const Loop *TheLoop, AAResults *AA, const LoopInfo *LI,
DominatorTree &DT, MemoryDepChecker::DepCandidates &DA,
PredicatedScalarEvolution &PSE,
SmallPtrSetImpl<MDNode *> &LoopAliasScopes)
: TheLoop(TheLoop), BAA(*AA), AST(BAA), LI(LI), DT(DT), DepCands(DA),
PSE(PSE), LoopAliasScopes(LoopAliasScopes) {
// We're analyzing dependences across loop iterations.
BAA.enableCrossIterationMode();
}
/// Register a load and whether it is only read from.
void addLoad(const MemoryLocation &Loc, Type *AccessTy, bool IsReadOnly) {
Value *Ptr = const_cast<Value *>(Loc.Ptr);
AST.add(adjustLoc(Loc));
Accesses[MemAccessInfo(Ptr, false)].insert(AccessTy);
if (IsReadOnly)
ReadOnlyPtr.insert(Ptr);
}
/// Register a store.
void addStore(const MemoryLocation &Loc, Type *AccessTy) {
Value *Ptr = const_cast<Value *>(Loc.Ptr);
AST.add(adjustLoc(Loc));
Accesses[MemAccessInfo(Ptr, true)].insert(AccessTy);
}
/// Check if we can emit a run-time no-alias check for \p Access.
///
/// Returns true if we can emit a run-time no alias check for \p Access.
/// If we can check this access, this also adds it to a dependence set and
/// adds a run-time to check for it to \p RtCheck. If \p Assume is true,
/// we will attempt to use additional run-time checks in order to get
/// the bounds of the pointer.
bool createCheckForAccess(RuntimePointerChecking &RtCheck,
MemAccessInfo Access, Type *AccessTy,
const SymbolicStrideMap &Strides,
DenseMap<Value *, unsigned> &DepSetId,
Loop *TheLoop, unsigned &RunningDepId,
unsigned ASId, bool Assume);
/// Check whether we can check the pointers at runtime for
/// non-intersection.
///
/// Returns true if we need no check or if we do and we can generate them
/// (i.e. the pointers have computable bounds). A return value of false means
/// we couldn't analyze and generate runtime checks for all pointers in the
/// loop, but if \p AllowPartial is set then we will have checks for those
/// pointers we could analyze. \p DepChecker is used to remove unknown
/// dependences from DepCands.
bool canCheckPtrAtRT(RuntimePointerChecking &RtCheck, Loop *TheLoop,
const SymbolicStrideMap &Strides,
Value *&UncomputablePtr, bool AllowPartial,
const MemoryDepChecker &DepChecker);
/// Goes over all memory accesses, checks whether a RT check is needed
/// and builds sets of dependent accesses.
void buildDependenceSets();
/// Initial processing of memory accesses determined that we need to
/// perform dependency checking.
///
/// Note that this can later be cleared if we retry memcheck analysis without
/// dependency checking (i.e. ShouldRetryWithRuntimeChecks).
bool isDependencyCheckNeeded() const { return !CheckDeps.empty(); }
/// We decided that no dependence analysis would be used. Reset the state.
void resetDepChecks(MemoryDepChecker &DepChecker) {
CheckDeps.clear();
DepChecker.clearDependences();
}
ArrayRef<MemAccessInfo> getDependenciesToCheck() const { return CheckDeps; }
private:
using PtrAccessMap = MapVector<MemAccessInfo, SmallSetVector<Type *, 1>>;
/// Adjust the MemoryLocation so that it represents accesses to this
/// location across all iterations, rather than a single one.
MemoryLocation adjustLoc(MemoryLocation Loc) const {
// The accessed location varies within the loop, but remains within the
// underlying object.
Loc.Size = LocationSize::beforeOrAfterPointer();
Loc.AATags.Scope = adjustAliasScopeList(Loc.AATags.Scope);
Loc.AATags.NoAlias = adjustAliasScopeList(Loc.AATags.NoAlias);
return Loc;
}
/// Drop alias scopes that are only valid within a single loop iteration.
MDNode *adjustAliasScopeList(MDNode *ScopeList) const {
if (!ScopeList)
return nullptr;
// For the sake of simplicity, drop the whole scope list if any scope is
// iteration-local.
if (any_of(ScopeList->operands(), [&](Metadata *Scope) {
return LoopAliasScopes.contains(cast<MDNode>(Scope));
}))
return nullptr;
return ScopeList;
}
/// Map of all accesses. Values are the types used to access memory pointed to
/// by the pointer.
PtrAccessMap Accesses;
/// The loop being checked.
const Loop *TheLoop;
/// List of accesses that need a further dependence check.
SmallVector<MemAccessInfo, 8> CheckDeps;
/// Set of pointers that are read only.
SmallPtrSet<Value*, 16> ReadOnlyPtr;
/// Batched alias analysis results.
BatchAAResults BAA;
/// An alias set tracker to partition the access set by underlying object and
//intrinsic property (such as TBAA metadata).
AliasSetTracker AST;
/// The LoopInfo of the loop being checked.
const LoopInfo *LI;
/// The dominator tree of the function.
DominatorTree &DT;
/// Sets of potentially dependent accesses - members of one set share an
/// underlying pointer. The set "CheckDeps" identfies which sets really need a
/// dependence check.
MemoryDepChecker::DepCandidates &DepCands;
/// Initial processing of memory accesses determined that we may need
/// to add memchecks. Perform the analysis to determine the necessary checks.
///
/// Note that, this is different from isDependencyCheckNeeded. When we retry
/// memcheck analysis without dependency checking
/// (i.e. ShouldRetryWithRuntimeChecks), isDependencyCheckNeeded is
/// cleared while this remains set if we have potentially dependent accesses.
bool IsRTCheckAnalysisNeeded = false;
/// The SCEV predicate containing all the SCEV-related assumptions.
PredicatedScalarEvolution &PSE;
DenseMap<Value *, SmallVector<const Value *, 16>> UnderlyingObjects;
/// Alias scopes that are declared inside the loop, and as such not valid
/// across iterations.
SmallPtrSetImpl<MDNode *> &LoopAliasScopes;
};
} // end anonymous namespace
std::optional<int64_t>
llvm::getStrideFromAddRec(const SCEVAddRecExpr *AR, const Loop *Lp,
Type *AccessTy, Value *Ptr,
PredicatedScalarEvolution &PSE) {
if (isa<ScalableVectorType>(AccessTy)) {
LLVM_DEBUG(dbgs() << "LAA: Bad stride - Scalable object: " << *AccessTy
<< "\n");
return std::nullopt;
}
// The access function must stride over the innermost loop.
if (Lp != AR->getLoop()) {
LLVM_DEBUG({
dbgs() << "LAA: Bad stride - Not striding over innermost loop ";
if (Ptr)
dbgs() << *Ptr << " ";
dbgs() << "SCEV: " << *AR << "\n";
});
return std::nullopt;
}
// Check the step is constant.
const SCEV *Step = AR->getStepRecurrence(*PSE.getSE());
// Calculate the pointer stride and check if it is constant.
const APInt *APStepVal;
if (!match(Step, m_scev_APInt(APStepVal))) {
LLVM_DEBUG({
dbgs() << "LAA: Bad stride - Not a constant strided ";
if (Ptr)
dbgs() << *Ptr << " ";
dbgs() << "SCEV: " << *AR << "\n";
});
return std::nullopt;
}
const auto &DL = Lp->getHeader()->getDataLayout();
TypeSize AllocSize = DL.getTypeAllocSize(AccessTy);
int64_t Size = AllocSize.getFixedValue();
// Huge step value - give up.
std::optional<int64_t> StepVal = APStepVal->trySExtValue();
if (!StepVal)
return std::nullopt;
// Strided access.
return *StepVal % Size ? std::nullopt : std::make_optional(*StepVal / Size);
}
/// Check whether \p AR is a non-wrapping AddRec. If \p Ptr is not nullptr, use
/// information from the IR pointer value to determine no-wrap. If \p Predicates
/// is not nullptr add no-wrap assumptions if needed.
static bool
isNoWrap(PredicatedScalarEvolution &PSE, const SCEVAddRecExpr *AR, Value *Ptr,
Type *AccessTy, const Loop *L, const DominatorTree &DT,
std::optional<int64_t> Stride = std::nullopt,
SmallVectorImpl<const SCEVPredicate *> *Predicates = nullptr) {
// FIXME: This should probably only return true for NUW.
if (any(AR->getNoWrapFlags()))
return true;
// An nusw getelementptr that is an AddRec cannot wrap. If it would wrap,
// the distance between the previously accessed location and the wrapped
// location will be larger than half the pointer index type space. In that
// case, the GEP would be poison and any memory access dependent on it would
// be immediate UB when executed.
if (auto *GEP = dyn_cast_if_present<GetElementPtrInst>(Ptr);
GEP && GEP->hasNoUnsignedSignedWrap()) {
// For the above reasoning to apply, the pointer must be dereferenced in
// every iteration.
if (L->getHeader() == L->getLoopLatch() ||
any_of(GEP->users(), [L, &DT, GEP](User *U) {
if (getLoadStorePointerOperand(U) != GEP)
return false;
BasicBlock *UserBB = cast<Instruction>(U)->getParent();
if (!L->contains(UserBB))
return false;
return !LoopAccessInfo::blockNeedsPredication(UserBB, L, &DT);
}))
return true;
}
if (!Stride)
Stride = getStrideFromAddRec(AR, L, AccessTy, Ptr, PSE);
if (Stride) {
// If the null pointer is undefined, then a access sequence which would
// otherwise access it can be assumed not to unsigned wrap. Note that this
// assumes the object in memory is aligned to the natural alignment.
unsigned AddrSpace = AR->getType()->getPointerAddressSpace();
if (!NullPointerIsDefined(L->getHeader()->getParent(), AddrSpace) &&
(Stride == 1 || Stride == -1))
return true;
}
if (Ptr && Predicates) {
ScalarEvolution &SE = *PSE.getSE();
SCEVWrapPredicate::IncrementWrapFlags Flags =
SCEVWrapPredicate::IncrementNUSW;
Predicates->push_back(SE.getWrapPredicate(AR, Flags));
LLVM_DEBUG(dbgs() << "LAA: Pointer may wrap:\n"
<< "LAA: Pointer: " << *Ptr << "\n"
<< "LAA: SCEV: " << *AR << "\n"
<< "LAA: Added an overflow assumption\n");
return true;
}
return false;
}
static void visitPointers(Value *StartPtr, const Loop &InnermostLoop,
function_ref<void(Value *)> AddPointer) {
SmallPtrSet<Value *, 8> Visited;
SmallVector<Value *> WorkList;
WorkList.push_back(StartPtr);
while (!WorkList.empty()) {
Value *Ptr = WorkList.pop_back_val();
if (!Visited.insert(Ptr).second)
continue;
auto *PN = dyn_cast<PHINode>(Ptr);
// SCEV does not look through non-header PHIs inside the loop. Such phis
// can be analyzed by adding separate accesses for each incoming pointer
// value.
if (PN && InnermostLoop.contains(PN->getParent()) &&
PN->getParent() != InnermostLoop.getHeader()) {
llvm::append_range(WorkList, PN->incoming_values());
} else
AddPointer(Ptr);
}
}
// Walk back through the IR for a pointer, looking for a select like the
// following:
//
// %offset = select i1 %cmp, i64 %a, i64 %b
// %addr = getelementptr double, double* %base, i64 %offset
// %ld = load double, double* %addr, align 8
//
// We won't be able to form a single SCEVAddRecExpr from this since the
// address for each loop iteration depends on %cmp. We could potentially
// produce multiple valid SCEVAddRecExprs, though, and check all of them for
// memory safety/aliasing if needed.
//
// If we encounter some IR we don't yet handle, or something obviously fine
// like a constant, then we just add the SCEV for that term to the list passed
// in by the caller. If we have a node that may potentially yield a valid
// SCEVAddRecExpr then we decompose it into parts and build the SCEV terms
// ourselves before adding to the list.
static void findForkedSCEVs(
ScalarEvolution *SE, const Loop *L, Value *Ptr,
SmallVectorImpl<PointerIntPair<const SCEV *, 1, bool>> &ScevList,
unsigned Depth) {
// If our Value is a SCEVAddRecExpr, loop invariant, not an instruction, or
// we've exceeded our limit on recursion, just return whatever we have
// regardless of whether it can be used for a forked pointer or not, along
// with an indication of whether it might be a poison or undef value.
const SCEV *Scev = SE->getSCEV(Ptr);
if (isa<SCEVAddRecExpr>(Scev) || L->isLoopInvariant(Ptr) ||
!isa<Instruction>(Ptr) || Depth == 0) {
ScevList.emplace_back(Scev, !isGuaranteedNotToBeUndefOrPoison(Ptr));
return;
}
Depth--;
auto UndefPoisonCheck = [](PointerIntPair<const SCEV *, 1, bool> S) {
return get<1>(S);
};
auto GetBinOpExpr = [&SE](unsigned Opcode, const SCEV *L,
const SCEV *R) -> const SCEV * {
switch (Opcode) {
case Instruction::Add:
return SE->getAddExpr(L, R);
case Instruction::Sub:
return SE->getMinusSCEV(L, R);
default:
llvm_unreachable("Unexpected binary operator when walking ForkedPtrs");
}
};
Instruction *I = cast<Instruction>(Ptr);
unsigned Opcode = I->getOpcode();
switch (Opcode) {
case Instruction::GetElementPtr: {
auto *GEP = cast<GetElementPtrInst>(I);
Type *SourceTy = GEP->getSourceElementType();
// We only handle base + single offset GEPs here for now.
// Not dealing with preexisting gathers yet, so no vectors.
if (I->getNumOperands() != 2 || SourceTy->isVectorTy()) {
ScevList.emplace_back(Scev, !isGuaranteedNotToBeUndefOrPoison(GEP));
break;
}
SmallVector<PointerIntPair<const SCEV *, 1, bool>, 2> BaseScevs;
SmallVector<PointerIntPair<const SCEV *, 1, bool>, 2> OffsetScevs;
findForkedSCEVs(SE, L, I->getOperand(0), BaseScevs, Depth);
findForkedSCEVs(SE, L, I->getOperand(1), OffsetScevs, Depth);
// See if we need to freeze our fork...
bool NeedsFreeze = any_of(BaseScevs, UndefPoisonCheck) ||
any_of(OffsetScevs, UndefPoisonCheck);
// Check that we only have a single fork, on either the base or the offset.
// Copy the SCEV across for the one without a fork in order to generate
// the full SCEV for both sides of the GEP.
if (OffsetScevs.size() == 2 && BaseScevs.size() == 1)
BaseScevs.push_back(BaseScevs[0]);
else if (BaseScevs.size() == 2 && OffsetScevs.size() == 1)
OffsetScevs.push_back(OffsetScevs[0]);
else {
ScevList.emplace_back(Scev, NeedsFreeze);
break;
}
Type *IntPtrTy = SE->getEffectiveSCEVType(GEP->getPointerOperandType());
// Find the size of the type being pointed to. We only have a single
// index term (guarded above) so we don't need to index into arrays or
// structures, just get the size of the scalar value.
const SCEV *Size = SE->getSizeOfExpr(IntPtrTy, SourceTy);
for (auto [B, O] : zip(BaseScevs, OffsetScevs)) {
const SCEV *Base = get<0>(B);
const SCEV *Offset = get<0>(O);
// Scale up the offsets by the size of the type, then add to the bases.
const SCEV *Scaled =
SE->getMulExpr(Size, SE->getTruncateOrSignExtend(Offset, IntPtrTy));
ScevList.emplace_back(SE->getAddExpr(Base, Scaled), NeedsFreeze);
}
break;
}
case Instruction::Select: {
SmallVector<PointerIntPair<const SCEV *, 1, bool>, 2> ChildScevs;
// A select means we've found a forked pointer, but we currently only
// support a single select per pointer so if there's another behind this
// then we just bail out and return the generic SCEV.
findForkedSCEVs(SE, L, I->getOperand(1), ChildScevs, Depth);
findForkedSCEVs(SE, L, I->getOperand(2), ChildScevs, Depth);
if (ChildScevs.size() == 2)
append_range(ScevList, ChildScevs);
else
ScevList.emplace_back(Scev, !isGuaranteedNotToBeUndefOrPoison(Ptr));
break;
}
case Instruction::PHI: {
SmallVector<PointerIntPair<const SCEV *, 1, bool>, 2> ChildScevs;
// A phi means we've found a forked pointer, but we currently only
// support a single phi per pointer so if there's another behind this
// then we just bail out and return the generic SCEV.
if (I->getNumOperands() == 2) {
findForkedSCEVs(SE, L, I->getOperand(0), ChildScevs, Depth);
findForkedSCEVs(SE, L, I->getOperand(1), ChildScevs, Depth);
}
if (ChildScevs.size() == 2)
append_range(ScevList, ChildScevs);
else
ScevList.emplace_back(Scev, !isGuaranteedNotToBeUndefOrPoison(Ptr));
break;
}
case Instruction::Add:
case Instruction::Sub: {
SmallVector<PointerIntPair<const SCEV *, 1, bool>> LScevs;
SmallVector<PointerIntPair<const SCEV *, 1, bool>> RScevs;
findForkedSCEVs(SE, L, I->getOperand(0), LScevs, Depth);
findForkedSCEVs(SE, L, I->getOperand(1), RScevs, Depth);
// See if we need to freeze our fork...
bool NeedsFreeze =
any_of(LScevs, UndefPoisonCheck) || any_of(RScevs, UndefPoisonCheck);
// Check that we only have a single fork, on either the left or right side.
// Copy the SCEV across for the one without a fork in order to generate
// the full SCEV for both sides of the BinOp.
if (LScevs.size() == 2 && RScevs.size() == 1)
RScevs.push_back(RScevs[0]);
else if (RScevs.size() == 2 && LScevs.size() == 1)
LScevs.push_back(LScevs[0]);
else {
ScevList.emplace_back(Scev, NeedsFreeze);
break;
}
for (auto [L, R] : zip(LScevs, RScevs))
ScevList.emplace_back(GetBinOpExpr(Opcode, get<0>(L), get<0>(R)),
NeedsFreeze);
break;
}
default:
// Just return the current SCEV if we haven't handled the instruction yet.
LLVM_DEBUG(dbgs() << "ForkedPtr unhandled instruction: " << *I << "\n");
ScevList.emplace_back(Scev, !isGuaranteedNotToBeUndefOrPoison(Ptr));
break;
}
}
bool AccessAnalysis::createCheckForAccess(RuntimePointerChecking &RtCheck,
MemAccessInfo Access, Type *AccessTy,
const SymbolicStrideMap &StridesMap,
DenseMap<Value *, unsigned> &DepSetId,
Loop *TheLoop, unsigned &RunningDepId,
unsigned ASId, bool Assume) {
Value *Ptr = Access.getPointer();
ScalarEvolution *SE = PSE.getSE();
const DataLayout &DL = TheLoop->getHeader()->getDataLayout();
assert(SE->isSCEVable(Ptr->getType()) && "Value is not SCEVable!");
SmallVector<PointerIntPair<const SCEV *, 1, bool>> RTCheckPtrs;
findForkedSCEVs(SE, TheLoop, Ptr, RTCheckPtrs, MaxForkedSCEVDepth);
assert(!RTCheckPtrs.empty() &&
"Must have some runtime-check pointer candidates");
// RTCheckPtrs must have size 2 if there are forked pointers. Otherwise, there
// are no forked pointers; replaceSymbolicStridesSCEV in this case.
auto IsLoopInvariantOrAR =
[&SE, &TheLoop](const PointerIntPair<const SCEV *, 1, bool> &P) {
return SE->isLoopInvariant(P.getPointer(), TheLoop) ||
isa<SCEVAddRecExpr>(P.getPointer());
};
if (RTCheckPtrs.size() == 2 && all_of(RTCheckPtrs, IsLoopInvariantOrAR)) {
LLVM_DEBUG(dbgs() << "LAA: Found forked pointer: " << *Ptr << "\n";
for (const auto &[Idx, Q] : enumerate(RTCheckPtrs)) dbgs()
<< "\t(" << Idx << ") " << *Q.getPointer() << "\n");
} else {
RTCheckPtrs = {{replaceSymbolicStrideSCEV(PSE, StridesMap, Ptr), false}};
}
/// Check whether all pointers can participate in a runtime bounds check. They
/// must either be invariant or non-wrapping affine AddRecs.
SmallVector<const SCEVPredicate *> Predicates;
for (auto &P : RTCheckPtrs) {
// The bounds for loop-invariant pointer is trivial.
if (SE->isLoopInvariant(P.getPointer(), TheLoop))
continue;
const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(P.getPointer());
if (!AR && Assume)
AR = PSE.getAsAddRec(Ptr, &Predicates);
if (!AR || !AR->isAffine()) {
// Check if bounds for non-affine monotonic expressions can be formed.
const SCEV *EltSizeSCEV = SE->getStoreSizeOfExpr(
DL.getIndexType(P.getPointer()->getType()), AccessTy);
if (!Assume ||
!getNonAffineMonotonicBounds(TheLoop, P.getPointer(), EltSizeSCEV, SE)
.first)
return false;
continue;
}
// If there's only one option for Ptr, commit the predicates collected by
// getAsAddRec and look Ptr up again afterwards: the lookup below reads the
// assumptions back from PSE, so they need to be committed first.
if (RTCheckPtrs.size() == 1) {
PSE.addPredicates(Predicates);
Predicates.clear();
if (auto *StrideAR = dyn_cast<SCEVAddRecExpr>(
replaceSymbolicStrideSCEV(PSE, StridesMap, Ptr)))
AR = StrideAR;
P.setPointer(AR);
}
if (!isNoWrap(PSE, AR, RTCheckPtrs.size() == 1 ? Ptr : nullptr, AccessTy,
TheLoop, DT, /*Stride=*/std::nullopt,
Assume ? &Predicates : nullptr))
return false;
}
PSE.addPredicates(Predicates);
// Remember the number of pointers inserted so far, to remove the pointers of
// this access again if the bounds of any of them cannot be computed, to avoid
// partial inserts.
unsigned NumPointers = RtCheck.Pointers.size();
for (const auto &[PtrExpr, NeedsFreeze] : RTCheckPtrs) {
// The id of the dependence set.
unsigned DepId;
if (DepCands.contains(Access)) {
Value *Leader = DepCands.getLeaderValue(Access).getPointer();
unsigned &LeaderId = DepSetId[Leader];
if (!LeaderId)
LeaderId = RunningDepId++;
DepId = LeaderId;
} else
// Each access has its own dependence set.
DepId = RunningDepId++;
bool IsWrite = Access.getInt();
if (!RtCheck.insert(TheLoop, Ptr, PtrExpr, AccessTy, IsWrite, DepId, ASId,
PSE, NeedsFreeze)) {
RtCheck.Pointers.truncate(NumPointers);
return false;
}
LLVM_DEBUG(dbgs() << "LAA: Found a runtime check ptr:" << *Ptr << '\n');
}
return true;
}
bool AccessAnalysis::canCheckPtrAtRT(RuntimePointerChecking &RtCheck,
Loop *TheLoop,
const SymbolicStrideMap &StridesMap,
Value *&UncomputablePtr, bool AllowPartial,
const MemoryDepChecker &DepChecker) {
// Find pointers with computable bounds. We are going to use this information
// to place a runtime bound check.
bool CanDoRT = true;
bool MayNeedRTCheck = false;
if (!IsRTCheckAnalysisNeeded) return true;
if (auto *Deps = DepChecker.getDependences()) {
// If there are unknown dependences, this means runtime checks are needed to
// ensure there's no overlap between accesses to the same underlying object.
// Remove the equivalence classes containing both source and destination
// accesses from DepCands. This ensures runtime checks will be generated
// between those accesses and prevents them from being grouped together.
for (const auto &Dep : *Deps) {
if (Dep.Type != MemoryDepChecker::Dependence::Unknown) {
assert(MemoryDepChecker::Dependence::isSafeForVectorization(Dep.Type) ==
MemoryDepChecker::VectorizationSafetyStatus::Safe &&
"Should only skip safe dependences");
continue;
}
Instruction *Src = Dep.getSource(DepChecker);
Instruction *Dst = Dep.getDestination(DepChecker);
DepCands.eraseClass({getPointerOperand(Src), Src->mayWriteToMemory()});
DepCands.eraseClass({getPointerOperand(Dst), Dst->mayWriteToMemory()});
}
} else {
CheckDeps.clear();
DepCands = {};
}
// We assign a consecutive id to access from different alias sets.
// Accesses between different groups doesn't need to be checked.
unsigned ASId = 0;
for (const auto &AS : AST) {
int NumReadPtrChecks = 0;
int NumWritePtrChecks = 0;
bool CanDoAliasSetRT = true;
++ASId;
auto ASPointers = AS.getPointers();
// We assign consecutive id to access from different dependence sets.
// Accesses within the same set don't need a runtime check.
unsigned RunningDepId = 1;
DenseMap<Value *, unsigned> DepSetId;
SmallVector<std::pair<MemAccessInfo, Type *>, 4> Retries;
// First, count how many write and read accesses are in the alias set. Also
// collect MemAccessInfos for later.
SmallVector<MemAccessInfo, 4> AccessInfos;
for (const Value *ConstPtr : ASPointers) {
Value *Ptr = const_cast<Value *>(ConstPtr);
bool IsWrite = Accesses.contains(MemAccessInfo(Ptr, true));
if (IsWrite)
++NumWritePtrChecks;
else
++NumReadPtrChecks;
AccessInfos.emplace_back(Ptr, IsWrite);
}
// We do not need runtime checks for this alias set, if there are no writes
// or a single write and no reads.
if (NumWritePtrChecks == 0 ||
(NumWritePtrChecks == 1 && NumReadPtrChecks == 0)) {
assert((ASPointers.size() <= 1 ||
all_of(ASPointers,
[this](const Value *Ptr) {
MemAccessInfo AccessWrite(const_cast<Value *>(Ptr),
true);
return !DepCands.contains(AccessWrite);
})) &&
"Can only skip updating CanDoRT below, if all entries in AS "
"are reads or there is at most 1 entry");
continue;
}
for (auto &Access : AccessInfos) {
for (const auto &AccessTy : Accesses[Access]) {
if (!createCheckForAccess(RtCheck, Access, AccessTy, StridesMap,
DepSetId, TheLoop, RunningDepId, ASId,
false)) {
LLVM_DEBUG(dbgs() << "LAA: Can't find bounds for ptr:"
<< *Access.getPointer() << '\n');
Retries.emplace_back(Access, AccessTy);
CanDoAliasSetRT = false;
}
}
}
// Note that this function computes CanDoRT and MayNeedRTCheck
// independently. For example CanDoRT=false, MayNeedRTCheck=false means that
// we have a pointer for which we couldn't find the bounds but we don't
// actually need to emit any checks so it does not matter.
//
// We need runtime checks for this alias set, if there are at least 2
// dependence sets (in which case RunningDepId > 2) or if we need to re-try
// any bound checks (because in that case the number of dependence sets is
// incomplete).
bool NeedsAliasSetRTCheck = RunningDepId > 2 || !Retries.empty();
// We need to perform run-time alias checks, but some pointers had bounds
// that couldn't be checked.
if (NeedsAliasSetRTCheck && !CanDoAliasSetRT) {
// Reset the CanDoSetRt flag and retry all accesses that have failed.
// We know that we need these checks, so we can now be more aggressive
// and add further checks if required (overflow checks).
CanDoAliasSetRT = true;
for (const auto &[Access, AccessTy] : Retries) {
if (!createCheckForAccess(RtCheck, Access, AccessTy, StridesMap,
DepSetId, TheLoop, RunningDepId, ASId,
/*Assume=*/true)) {
CanDoAliasSetRT = false;
UncomputablePtr = Access.getPointer();
if (!AllowPartial)
break;
}
}
}
CanDoRT &= CanDoAliasSetRT;
MayNeedRTCheck |= NeedsAliasSetRTCheck;
++ASId;
}
// If the pointers that we would use for the bounds comparison have different
// address spaces, assume the values aren't directly comparable, so we can't
// use them for the runtime check. We also have to assume they could
// overlap. In the future there should be metadata for whether address spaces
// are disjoint.
unsigned NumPointers = RtCheck.Pointers.size();
for (unsigned i = 0; i < NumPointers; ++i) {
for (unsigned j = i + 1; j < NumPointers; ++j) {
// Only need to check pointers between two different dependency sets.
if (RtCheck.Pointers[i].DependencySetId ==
RtCheck.Pointers[j].DependencySetId)
continue;
// Only need to check pointers in the same alias set.
if (RtCheck.Pointers[i].AliasSetId != RtCheck.Pointers[j].AliasSetId)
continue;
Value *PtrI = RtCheck.Pointers[i].PointerValue;
Value *PtrJ = RtCheck.Pointers[j].PointerValue;
unsigned ASi = PtrI->getType()->getPointerAddressSpace();
unsigned ASj = PtrJ->getType()->getPointerAddressSpace();
if (ASi != ASj) {
LLVM_DEBUG(
dbgs() << "LAA: Runtime check would require comparison between"
" different address spaces\n");
return false;
}
}
}
if (MayNeedRTCheck && (CanDoRT || AllowPartial))
RtCheck.generateChecks(DepCands);
LLVM_DEBUG(dbgs() << "LAA: We need to do " << RtCheck.getNumberOfChecks()
<< " pointer comparisons.\n");
// If we can do run-time checks, but there are no checks, no runtime checks
// are needed. This can happen when all pointers point to the same underlying
// object for example.
RtCheck.Need = CanDoRT ? RtCheck.getNumberOfChecks() != 0 : MayNeedRTCheck;
bool CanDoRTIfNeeded = !RtCheck.Need || CanDoRT;
assert(CanDoRTIfNeeded == (CanDoRT || !MayNeedRTCheck) &&
"CanDoRTIfNeeded depends on RtCheck.Need");
if (!CanDoRTIfNeeded && !AllowPartial)
RtCheck.reset();
return CanDoRTIfNeeded;
}
void AccessAnalysis::buildDependenceSets() {
// We process the set twice: first we process read-write pointers, last we
// process read-only pointers. This allows us to skip dependence tests for
// read-only pointers.
LLVM_DEBUG(dbgs() << "LAA: Processing memory accesses...\n");
LLVM_DEBUG(dbgs() << " AST: "; AST.dump());
LLVM_DEBUG(dbgs() << "LAA: Accesses(" << Accesses.size() << "):\n");
LLVM_DEBUG({
for (const auto &[A, _] : Accesses)
dbgs() << "\t" << *A.getPointer() << " ("
<< (A.getInt()
? "write"
: (ReadOnlyPtr.contains(A.getPointer()) ? "read-only"
: "read"))
<< ")\n";
});
// The AliasSetTracker has nicely partitioned our pointers by metadata
// compatibility and potential for underlying-object overlap. As a result, we
// only need to check for potential pointer dependencies within each alias
// set.
for (const auto &AS : AST) {
bool AliasSetHasWrite = false;
// Map of (pointer to underlying objects, accessed address space) to last
// access encountered.
using UnderlyingObjToAccessMap =
DenseMap<std::pair<const Value *, unsigned>, MemAccessInfo>;
UnderlyingObjToAccessMap ObjToLastAccess;
// Set of access to check after all writes have been processed.
PtrAccessMap DeferredAccesses;
// Iterate over each alias set twice, once to process read/write pointers,
// and then to process read-only pointers.
auto ProcessAccesses = [&](bool UseDeferred) {
PtrAccessMap &S = UseDeferred ? DeferredAccesses : Accesses;
// Note that both the alias-set tracker and the alias sets themselves used
// ordered collections internally and so the iteration order here is
// deterministic.
for (const Value *ConstPtr : AS.getPointers()) {
Value *Ptr = const_cast<Value *>(ConstPtr);
// For a single memory access in AliasSetTracker, Accesses may contain
// both read and write, and they both need to be handled for CheckDeps.
for (auto [AccessPtr, IsWrite] : S.keys()) {
if (AccessPtr != Ptr)
continue;
// If we're using the deferred access set, then it contains only
// reads.
bool IsReadOnlyPtr = ReadOnlyPtr.contains(Ptr) && !IsWrite;
if (UseDeferred && !IsReadOnlyPtr)
continue;
// Otherwise, the pointer must be in the PtrAccessSet, either as a
// read or a write.
assert(((IsReadOnlyPtr && UseDeferred) || IsWrite ||
S.contains(MemAccessInfo(Ptr, false))) &&
"Alias-set pointer not in the access set?");
MemAccessInfo Access(Ptr, IsWrite);
DepCands.insert(Access);
// Memorize read-only pointers for later processing and skip them in
// the first round (they need to be checked after we have seen all
// write pointers). Note: we also mark pointer that are not
// consecutive as "read-only" pointers (so that we check
// "a[b[i]] +="). Hence, we need the second check for "!IsWrite".
if (!UseDeferred && IsReadOnlyPtr) {
// We only use the pointer keys, the types vector values don't
// matter.
DeferredAccesses.insert({Access, {}});
continue;
}
// If this is a write - check other reads and writes for conflicts. If
// this is a read only check other writes for conflicts (but only if
// there is no other write to the ptr - this is an optimization to
// catch "a[i] = a[i] + " without having to do a dependence check).
if ((IsWrite || IsReadOnlyPtr) && AliasSetHasWrite) {
CheckDeps.push_back(Access);
IsRTCheckAnalysisNeeded = true;
}
if (IsWrite)
AliasSetHasWrite = true;
// Create sets of pointers connected by a shared alias set and
// underlying object.
SmallVector<const Value *, 16> &UOs = UnderlyingObjects[Ptr];
UOs = {};
::getUnderlyingObjects(Ptr, UOs, LI);
LLVM_DEBUG(dbgs()
<< "Underlying objects for pointer " << *Ptr << "\n");
for (const Value *UnderlyingObj : UOs) {
// nullptr never alias, don't join sets for pointer that have "null"
// in their UnderlyingObjects list.
if (isa<ConstantPointerNull>(UnderlyingObj) &&
!NullPointerIsDefined(
TheLoop->getHeader()->getParent(),
UnderlyingObj->getType()->getPointerAddressSpace()))
continue;
auto [It, Inserted] = ObjToLastAccess.try_emplace(
{UnderlyingObj,
cast<PointerType>(Ptr->getType())->getAddressSpace()},
Access);
if (!Inserted) {
DepCands.unionSets(Access, It->second);
It->second = Access;
}
LLVM_DEBUG(dbgs() << " " << *UnderlyingObj << "\n");
}
}
}
};
ProcessAccesses(false);
ProcessAccesses(true);
}
}
/// Check whether the access through \p Ptr has a constant stride.
std::optional<int64_t>
llvm::getPtrStride(PredicatedScalarEvolution &PSE, Type *AccessTy, Value *Ptr,
const Loop *Lp, const DominatorTree &DT,
const SymbolicStrideMap &StridesMap, bool ShouldCheckWrap,
SmallVectorImpl<const SCEVPredicate *> *Predicates) {
const SCEV *PtrScev = replaceSymbolicStrideSCEV(PSE, StridesMap, Ptr);
if (PSE.getSE()->isLoopInvariant(PtrScev, Lp))
return 0;
assert(Ptr->getType()->isPointerTy() && "Unexpected non-ptr");
const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(PtrScev);
if (Predicates && !AR) {
AR = PSE.getSE()->convertSCEVToAddRecWithPredicates(PtrScev, Lp,
*Predicates);
}
if (!AR) {
LLVM_DEBUG(dbgs() << "LAA: Bad stride - Not an AddRecExpr pointer " << *Ptr
<< " SCEV: " << *PtrScev << "\n");
return std::nullopt;
}
std::optional<int64_t> Stride =
getStrideFromAddRec(AR, Lp, AccessTy, Ptr, PSE);
if (!ShouldCheckWrap || !Stride)
return Stride;
if (isNoWrap(PSE, AR, Ptr, AccessTy, Lp, DT, Stride, Predicates))
return Stride;
LLVM_DEBUG(
dbgs() << "LAA: Bad stride - Pointer may wrap in the address space "
<< *Ptr << " SCEV: " << *AR << "\n");
return std::nullopt;
}
/// Check whether the access through \p Ptr has a constant stride.
std::optional<int64_t> llvm::getPtrStride(PredicatedScalarEvolution &PSE,
Type *AccessTy, Value *Ptr,
const Loop *Lp,
const DominatorTree &DT,
const SymbolicStrideMap &StridesMap,
bool Assume, bool ShouldCheckWrap) {
SmallVector<const SCEVPredicate *> Predicates;
std::optional<int64_t> Stride =
getPtrStride(PSE, AccessTy, Ptr, Lp, DT, StridesMap, ShouldCheckWrap,
Assume ? &Predicates : nullptr);
PSE.addPredicates(Predicates);
return Stride;
}
std::optional<int64_t> llvm::getPointersDiff(Type *ElemTyA, Value *PtrA,
Type *ElemTyB, Value *PtrB,
const DataLayout &DL,
ScalarEvolution &SE,
bool StrictCheck, bool CheckType) {
assert(PtrA && PtrB && "Expected non-nullptr pointers.");
// Make sure that A and B are different pointers.
if (PtrA == PtrB)
return 0;
// Make sure that the element types are the same if required.
if (CheckType && ElemTyA != ElemTyB)
return std::nullopt;
unsigned ASA = PtrA->getType()->getPointerAddressSpace();
unsigned ASB = PtrB->getType()->getPointerAddressSpace();
// Check that the address spaces match.
if (ASA != ASB)
return std::nullopt;
unsigned IdxWidth = DL.getIndexSizeInBits(ASA);
APInt OffsetA(IdxWidth, 0), OffsetB(IdxWidth, 0);
const Value *PtrA1 = PtrA->stripAndAccumulateConstantOffsets(
DL, OffsetA, /*AllowNonInbounds=*/true);
const Value *PtrB1 = PtrB->stripAndAccumulateConstantOffsets(
DL, OffsetB, /*AllowNonInbounds=*/true);
std::optional<int64_t> Val;
if (PtrA1 == PtrB1) {
// Retrieve the address space again as pointer stripping now tracks through
// `addrspacecast`.
ASA = cast<PointerType>(PtrA1->getType())->getAddressSpace();
ASB = cast<PointerType>(PtrB1->getType())->getAddressSpace();
// Check that the address spaces match and that the pointers are valid.
if (ASA != ASB)
return std::nullopt;
IdxWidth = DL.getIndexSizeInBits(ASA);
OffsetA = OffsetA.sextOrTrunc(IdxWidth);
OffsetB = OffsetB.sextOrTrunc(IdxWidth);
OffsetB -= OffsetA;
Val = OffsetB.trySExtValue();
} else {
// Otherwise compute the distance with SCEV between the base pointers.
const SCEV *PtrSCEVA = SE.getSCEV(PtrA);
const SCEV *PtrSCEVB = SE.getSCEV(PtrB);
std::optional<APInt> Diff =
SE.computeConstantDifference(PtrSCEVB, PtrSCEVA);
if (!Diff)
return std::nullopt;
Val = Diff->trySExtValue();
}
if (!Val)
return std::nullopt;
int64_t Size = DL.getTypeStoreSize(ElemTyA);
int64_t Dist = *Val / Size;
// Ensure that the calculated distance matches the type-based one after all
// the bitcasts removal in the provided pointers.
if (!StrictCheck || Dist * Size == Val)
return Dist;
return std::nullopt;
}
bool llvm::sortPtrAccesses(ArrayRef<Value *> VL, Type *ElemTy,
const DataLayout &DL, ScalarEvolution &SE,
SmallVectorImpl<unsigned> &SortedIndices) {
assert(llvm::all_of(
VL, [](const Value *V) { return V->getType()->isPointerTy(); }) &&
"Expected list of pointer operands.");
// Walk over the pointers, and map each of them to an offset relative to
// first pointer in the array.
Value *Ptr0 = VL[0];
using DistOrdPair = std::pair<int64_t, unsigned>;
auto Compare = llvm::less_first();
std::set<DistOrdPair, decltype(Compare)> Offsets(Compare);
Offsets.emplace(0, 0);
bool IsConsecutive = true;
for (auto [Idx, Ptr] : drop_begin(enumerate(VL))) {
std::optional<int64_t> Diff =
getPointersDiff(ElemTy, Ptr0, ElemTy, Ptr, DL, SE,
/*StrictCheck=*/true);
if (!Diff)
return false;
// Check if the pointer with the same offset is found.
int64_t Offset = *Diff;
auto [It, IsInserted] = Offsets.emplace(Offset, Idx);
if (!IsInserted)
return false;
// Consecutive order if the inserted element is the last one.
IsConsecutive &= std::next(It) == Offsets.end();
}
SortedIndices.clear();
if (!IsConsecutive) {
// Fill SortedIndices array only if it is non-consecutive.
SortedIndices.resize(VL.size());
for (auto [Idx, Off] : enumerate(Offsets))
SortedIndices[Idx] = Off.second;
}
return true;
}
/// Returns true if the memory operations \p A and \p B are consecutive.
bool llvm::isConsecutiveAccess(Value *A, Value *B, const DataLayout &DL,
ScalarEvolution &SE, bool CheckType) {
Value *PtrA = getLoadStorePointerOperand(A);
Value *PtrB = getLoadStorePointerOperand(B);
if (!PtrA || !PtrB)
return false;
Type *ElemTyA = getLoadStoreType(A);
Type *ElemTyB = getLoadStoreType(B);
std::optional<int64_t> Diff =
getPointersDiff(ElemTyA, PtrA, ElemTyB, PtrB, DL, SE,
/*StrictCheck=*/true, CheckType);
return Diff == 1;
}
void MemoryDepChecker::addAccess(StoreInst *SI) {
visitPointers(SI->getPointerOperand(), *InnermostLoop,
[this, SI](Value *Ptr) {
Accesses[MemAccessInfo(Ptr, true)].push_back(AccessIdx);
InstMap.push_back(SI);
++AccessIdx;
});
}
void MemoryDepChecker::addAccess(LoadInst *LI) {
visitPointers(LI->getPointerOperand(), *InnermostLoop,
[this, LI](Value *Ptr) {
Accesses[MemAccessInfo(Ptr, false)].push_back(AccessIdx);
InstMap.push_back(LI);
++AccessIdx;
});
}
MemoryDepChecker::VectorizationSafetyStatus
MemoryDepChecker::Dependence::isSafeForVectorization(DepType Type) {
switch (Type) {
case NoDep:
case Forward:
case BackwardVectorizable:
return VectorizationSafetyStatus::Safe;
case Unknown:
return VectorizationSafetyStatus::PossiblySafeWithRtChecks;
case ForwardButPreventsForwarding:
case Backward:
case BackwardVectorizableButPreventsForwarding:
case IndirectUnsafe:
case InvariantUnsafe:
return VectorizationSafetyStatus::Unsafe;
}
llvm_unreachable("unexpected DepType!");
}
bool MemoryDepChecker::Dependence::isBackward() const {
switch (Type) {
case NoDep:
case Forward:
case ForwardButPreventsForwarding:
case Unknown:
case IndirectUnsafe:
case InvariantUnsafe:
return false;
case BackwardVectorizable:
case Backward:
case BackwardVectorizableButPreventsForwarding:
return true;
}
llvm_unreachable("unexpected DepType!");
}
bool MemoryDepChecker::Dependence::isPossiblyBackward() const {
return isBackward() || Type == Unknown || Type == IndirectUnsafe ||
Type == InvariantUnsafe;
}
bool MemoryDepChecker::Dependence::isForward() const {
switch (Type) {
case Forward:
case ForwardButPreventsForwarding:
return true;
case NoDep:
case Unknown:
case BackwardVectorizable:
case Backward:
case BackwardVectorizableButPreventsForwarding:
case IndirectUnsafe:
case InvariantUnsafe:
return false;
}
llvm_unreachable("unexpected DepType!");
}
bool MemoryDepChecker::couldPreventStoreLoadForward(uint64_t Distance,
uint64_t TypeByteSize,
unsigned CommonStride) {
// If loads occur at a distance that is not a multiple of a feasible vector
// factor store-load forwarding does not take place.
// Positive dependences might cause troubles because vectorizing them might
// prevent store-load forwarding making vectorized code run a lot slower.
// a[i] = a[i-3] ^ a[i-8];
// The stores to a[i:i+1] don't align with the stores to a[i-3:i-2] and
// hence on your typical architecture store-load forwarding does not take
// place. Vectorizing in such cases does not make sense.
// Store-load forwarding distance.
// Maximum vector factor.
uint64_t MaxVFWithoutSLForwardIssuesPowerOf2 =
std::min(VectorizerParams::MaxVectorWidth * TypeByteSize,
MaxStoreLoadForwardSafeDistanceInBits);
// Compute the smallest VF at which the store and load would be misaligned
// and recent enough to still be in the store buffer.
for (uint64_t VF = 2 * TypeByteSize;
VF <= MaxVFWithoutSLForwardIssuesPowerOf2; VF *= 2) {
if (isStoreLoadForwardingConflict(Distance, VF, TypeByteSize, VF)) {
MaxVFWithoutSLForwardIssuesPowerOf2 = (VF >> 1);
break;
}
}
if (MaxVFWithoutSLForwardIssuesPowerOf2 < 2 * TypeByteSize) {
LLVM_DEBUG(
dbgs() << "LAA: Distance " << Distance
<< " that could cause a store-load forwarding conflict\n");
return true;
}
if (CommonStride &&
MaxVFWithoutSLForwardIssuesPowerOf2 <
MaxStoreLoadForwardSafeDistanceInBits &&
MaxVFWithoutSLForwardIssuesPowerOf2 !=
VectorizerParams::MaxVectorWidth * TypeByteSize) {
uint64_t MaxVF =
bit_floor(MaxVFWithoutSLForwardIssuesPowerOf2 / CommonStride);
uint64_t MaxVFInBits = MaxVF * TypeByteSize * 8;
MaxStoreLoadForwardSafeDistanceInBits =
std::min(MaxStoreLoadForwardSafeDistanceInBits, MaxVFInBits);
if (MaxVF < 2) {
LLVM_DEBUG(
dbgs() << "LAA: strided access with Distance " << Distance
<< " that could cause a store-load forwarding conflict\n");
return true;
}
}
return false;
}
void MemoryDepChecker::mergeInStatus(VectorizationSafetyStatus S) {
if (Status < S)
Status = S;
}
/// Given a dependence-distance \p Dist between two memory accesses, that have
/// strides in the same direction whose absolute value of the maximum stride is
/// given in \p MaxStride, in a loop whose maximum backedge taken count is \p
/// MaxBTC, check if it is possible to prove statically that the dependence
/// distance is larger than the range that the accesses will travel through the
/// execution of the loop. If so, return true; false otherwise. This is useful
/// for example in loops such as the following (PR31098):
///
/// for (i = 0; i < D; ++i) {
/// = out[i];
/// out[i+D] =
/// }
static bool isSafeDependenceDistance(const DataLayout &DL, ScalarEvolution &SE,
const SCEV &MaxBTC, const SCEV &Dist,
uint64_t MaxStride) {
// If we can prove that
// (**) |Dist| > MaxBTC * Step
// where Step is the absolute stride of the memory accesses in bytes,
// then there is no dependence.
//
// Rationale:
// We basically want to check if the absolute distance (|Dist/Step|)
// is >= the loop iteration count (or > MaxBTC).
// This is equivalent to the Strong SIV Test (Practical Dependence Testing,
// Section 4.2.1); Note, that for vectorization it is sufficient to prove
// that the dependence distance is >= VF; This is checked elsewhere.
// But in some cases we can prune dependence distances early, and
// even before selecting the VF, and without a runtime test, by comparing
// the distance against the loop iteration count. Since the vectorized code
// will be executed only if LoopCount >= VF, proving distance >= LoopCount
// also guarantees that distance >= VF.
//
const SCEV *Step = SE.getConstant(MaxBTC.getType(), MaxStride);
const SCEV *Product = SE.getMulExpr(&MaxBTC, Step);
const SCEV *CastedDist = &Dist;
const SCEV *CastedProduct = Product;
uint64_t DistTypeSizeBits = DL.getTypeSizeInBits(Dist.getType());
uint64_t ProductTypeSizeBits = DL.getTypeSizeInBits(Product->getType());
// The dependence distance can be positive/negative, so we sign extend Dist;
// The multiplication of the absolute stride in bytes and the
// backedgeTakenCount is non-negative, so we zero extend Product.
if (DistTypeSizeBits > ProductTypeSizeBits)
CastedProduct = SE.getZeroExtendExpr(Product, Dist.getType());
else
CastedDist = SE.getNoopOrSignExtend(&Dist, Product->getType());
// Is Dist - (MaxBTC * Step) > 0 ?
// (If so, then we have proven (**) because |Dist| >= Dist)
const SCEV *Minus = SE.getMinusSCEV(CastedDist, CastedProduct);
if (SE.isKnownPositive(Minus))
return true;
// Second try: Is -Dist - (MaxBTC * Step) > 0 ?
// (If so, then we have proven (**) because |Dist| >= -1*Dist)
const SCEV *NegDist = SE.getNegativeSCEV(CastedDist);
Minus = SE.getMinusSCEV(NegDist, CastedProduct);
return SE.isKnownPositive(Minus);
}
/// Check the dependence for two accesses with the same stride \p Stride.
/// \p Distance is the positive distance in bytes, and \p TypeByteSize is type
/// size in bytes.
///
/// \returns true if they are independent.
static bool areStridedAccessesIndependent(uint64_t Distance, uint64_t Stride,
uint64_t TypeByteSize) {
assert(Stride > 1 && "The stride must be greater than 1");
assert(TypeByteSize > 0 && "The type size in byte must be non-zero");
assert(Distance > 0 && "The distance must be non-zero");
// Skip if the distance is not multiple of type byte size.
if (Distance % TypeByteSize)
return false;
// No dependence if the distance is not multiple of the stride.
// E.g.
// for (i = 0; i < 1024 ; i += 4)
// A[i+2] = A[i] + 1;
//
// Two accesses in memory (distance is 2, stride is 4):
// | A[0] | | | | A[4] | | | |
// | | | A[2] | | | | A[6] | |
//
// E.g.
// for (i = 0; i < 1024 ; i += 3)
// A[i+4] = A[i] + 1;
//
// Two accesses in memory (distance is 4, stride is 3):
// | A[0] | | | A[3] | | | A[6] | | |
// | | | | | A[4] | | | A[7] | |
return Distance % Stride;
}
bool MemoryDepChecker::areAccessesCompletelyBeforeOrAfter(const SCEV *Src,
Type *SrcTy,
const SCEV *Sink,
Type *SinkTy) {
const SCEV *BTC = PSE.getBackedgeTakenCount();
const SCEV *SymbolicMaxBTC = PSE.getSymbolicMaxBackedgeTakenCount();
ScalarEvolution &SE = *PSE.getSE();
const auto &[SrcStart_, SrcEnd_] =
getStartAndEndForAccess(InnermostLoop, Src, SrcTy, BTC, SymbolicMaxBTC,
&SE, &PointerBounds, DT, AC, LoopGuards);
if (isa<SCEVCouldNotCompute>(SrcStart_) || isa<SCEVCouldNotCompute>(SrcEnd_))
return false;
const auto &[SinkStart_, SinkEnd_] =
getStartAndEndForAccess(InnermostLoop, Sink, SinkTy, BTC, SymbolicMaxBTC,
&SE, &PointerBounds, DT, AC, LoopGuards);
if (isa<SCEVCouldNotCompute>(SinkStart_) ||
isa<SCEVCouldNotCompute>(SinkEnd_))
return false;
if (!LoopGuards)
LoopGuards.emplace(ScalarEvolution::LoopGuards::collect(InnermostLoop, SE));
auto SrcEnd = SE.applyLoopGuards(SrcEnd_, *LoopGuards);
auto SinkStart = SE.applyLoopGuards(SinkStart_, *LoopGuards);
if (SE.isKnownPredicate(CmpInst::ICMP_ULE, SrcEnd, SinkStart))
return true;
auto SinkEnd = SE.applyLoopGuards(SinkEnd_, *LoopGuards);
auto SrcStart = SE.applyLoopGuards(SrcStart_, *LoopGuards);
return SE.isKnownPredicate(CmpInst::ICMP_ULE, SinkEnd, SrcStart);
}
std::variant<MemoryDepChecker::Dependence::DepType,
MemoryDepChecker::DepDistanceStrideAndSizeInfo>
MemoryDepChecker::getDependenceDistanceStrideAndSize(
const AccessAnalysis::MemAccessInfo &A, Instruction *AInst,
const AccessAnalysis::MemAccessInfo &B, Instruction *BInst) {
const auto &DL = InnermostLoop->getHeader()->getDataLayout();
auto &SE = *PSE.getSE();
const auto &[APtr, AIsWrite] = A;
const auto &[BPtr, BIsWrite] = B;
// Two reads are independent.
if (!AIsWrite && !BIsWrite)
return MemoryDepChecker::Dependence::NoDep;
Type *ATy = getLoadStoreType(AInst);
Type *BTy = getLoadStoreType(BInst);
// We cannot check pointers in different address spaces.
if (APtr->getType()->getPointerAddressSpace() !=
BPtr->getType()->getPointerAddressSpace())
return MemoryDepChecker::Dependence::Unknown;
SmallVector<const SCEVPredicate *> Predicates;
std::optional<int64_t> StrideAPtr =
getPtrStride(PSE, ATy, APtr, InnermostLoop, *DT, SymbolicStrides,
/*ShouldCheckWrap=*/true, &Predicates);
std::optional<int64_t> StrideBPtr =
getPtrStride(PSE, BTy, BPtr, InnermostLoop, *DT, SymbolicStrides,
/*ShouldCheckWrap=*/true, &Predicates);
PSE.addPredicates(Predicates);
const SCEV *Src = PSE.getSCEV(APtr);
const SCEV *Sink = PSE.getSCEV(BPtr);
// If the induction step is negative we have to invert source and sink of the
// dependence when measuring the distance between them. We should not swap
// AIsWrite with BIsWrite, as their uses expect them in program order.
if (StrideAPtr && *StrideAPtr < 0) {
std::swap(Src, Sink);
std::swap(AInst, BInst);
std::swap(ATy, BTy);
std::swap(StrideAPtr, StrideBPtr);
}
const SCEV *Dist = SE.getMinusSCEV(Sink, Src);
LLVM_DEBUG(dbgs() << "LAA: Src Scev: " << *Src << "Sink Scev: " << *Sink
<< "\n");
LLVM_DEBUG(dbgs() << "LAA: Distance for " << *AInst << " to " << *BInst
<< ": " << *Dist << "\n");
// Need accesses with constant strides and the same direction for further
// dependence analysis. We don't want to vectorize "A[B[i]] += ..." and
// similar code or pointer arithmetic that could wrap in the address space.
// If either Src or Sink are not strided (i.e. not a non-wrapping AddRec) and
// not loop-invariant (stride will be 0 in that case), we cannot analyze the
// dependence further and also cannot generate runtime checks.
if (!StrideAPtr || !StrideBPtr) {
LLVM_DEBUG(dbgs() << "Pointer access with non-constant stride\n");
return MemoryDepChecker::Dependence::IndirectUnsafe;
}
int64_t StrideAPtrInt = *StrideAPtr;
int64_t StrideBPtrInt = *StrideBPtr;
LLVM_DEBUG(dbgs() << "LAA: Src induction step: " << StrideAPtrInt
<< " Sink induction step: " << StrideBPtrInt << "\n");
// At least Src or Sink are loop invariant and the other is strided or
// invariant.
if (!StrideAPtrInt || !StrideBPtrInt) {
// If both are loop-invariant and access the same location, we cannot
// vectorize.
if (!StrideAPtrInt && !StrideBPtrInt && Dist->isZero())
return MemoryDepChecker::Dependence::InvariantUnsafe;
// Otherwise, we can generate a runtime check to disambiguate the accesses.
return MemoryDepChecker::Dependence::Unknown;
}
// Both Src and Sink have a constant stride, check if they are in the same
// direction.
if ((StrideAPtrInt > 0) != (StrideBPtrInt > 0)) {
LLVM_DEBUG(
dbgs() << "Pointer access with strides in different directions\n");
return MemoryDepChecker::Dependence::Unknown;
}
TypeSize AStoreSz = DL.getTypeStoreSize(ATy);
TypeSize BStoreSz = DL.getTypeStoreSize(BTy);
// If store sizes are not the same, set TypeByteSize to zero, so we can check
// it in the caller isDependent.
uint64_t ASz = DL.getTypeAllocSize(ATy);
uint64_t BSz = DL.getTypeAllocSize(BTy);
uint64_t TypeByteSize = (AStoreSz == BStoreSz) ? BSz : 0;
uint64_t StrideAScaled = AbsoluteValue(StrideAPtrInt) * ASz;
uint64_t StrideBScaled = AbsoluteValue(StrideBPtrInt) * BSz;
uint64_t MaxStride = std::max(StrideAScaled, StrideBScaled);
std::optional<uint64_t> CommonStride;
if (StrideAScaled == StrideBScaled)
CommonStride = StrideAScaled;
// TODO: Historically, we didn't retry with runtime checks when (unscaled)
// strides were different but there is no inherent reason to.
if (!isa<SCEVConstant>(Dist))
ShouldRetryWithRuntimeChecks |= StrideAPtrInt == StrideBPtrInt;
// If distance is a SCEVCouldNotCompute, return Unknown immediately.
if (isa<SCEVCouldNotCompute>(Dist)) {
LLVM_DEBUG(dbgs() << "LAA: Uncomputable distance.\n");
return Dependence::Unknown;
}
return DepDistanceStrideAndSizeInfo(Dist, MaxStride, CommonStride,
TypeByteSize, AIsWrite, BIsWrite);
}
MemoryDepChecker::Dependence::DepType
MemoryDepChecker::isDependent(const MemAccessInfo &A, unsigned AIdx,
const MemAccessInfo &B, unsigned BIdx) {
assert(AIdx < BIdx && "Must pass arguments in program order");
// Check if we can prove that Sink only accesses memory after Src's end or
// vice versa. The helper is used to perform the checks only on the exit paths
// where it helps to improve the analysis result.
auto CheckCompletelyBeforeOrAfter = [&]() {
auto *APtr = A.getPointer();
auto *BPtr = B.getPointer();
Type *ATy = getLoadStoreType(InstMap[AIdx]);
Type *BTy = getLoadStoreType(InstMap[BIdx]);
const SCEV *Src = PSE.getSCEV(APtr);
const SCEV *Sink = PSE.getSCEV(BPtr);
return areAccessesCompletelyBeforeOrAfter(Src, ATy, Sink, BTy);
};
// Get the dependence distance, stride, type size and what access writes for
// the dependence between A and B.
auto Res =
getDependenceDistanceStrideAndSize(A, InstMap[AIdx], B, InstMap[BIdx]);
if (std::holds_alternative<Dependence::DepType>(Res)) {
if (std::get<Dependence::DepType>(Res) == Dependence::Unknown &&
CheckCompletelyBeforeOrAfter())
return Dependence::NoDep;
return std::get<Dependence::DepType>(Res);
}
auto &[Dist, MaxStride, CommonStride, TypeByteSize, AIsWrite, BIsWrite] =
std::get<DepDistanceStrideAndSizeInfo>(Res);
bool HasSameSize = TypeByteSize > 0;
ScalarEvolution &SE = *PSE.getSE();
auto &DL = InnermostLoop->getHeader()->getDataLayout();
// If the distance between the acecsses is larger than their maximum absolute
// stride multiplied by the symbolic maximum backedge taken count (which is an
// upper bound of the number of iterations), the accesses are independet, i.e.
// they are far enough appart that accesses won't access the same location
// across all loop ierations.
if (HasSameSize &&
isSafeDependenceDistance(
DL, SE, *(PSE.getSymbolicMaxBackedgeTakenCount()), *Dist, MaxStride))
return Dependence::NoDep;
const APInt *APDist = nullptr;
uint64_t ConstDist = 0;
if (match(Dist, m_scev_APInt(APDist))) {
std::optional<uint64_t> Val = APDist->abs().tryZExtValue();
if (!Val) {
LLVM_DEBUG(dbgs() << "LAA: Constant distance does not fit in 64 bits.\n");
return Dependence::Unknown;
}
ConstDist = *Val;
}
// Attempt to prove strided accesses independent.
if (APDist) {
// If the distance between accesses and their strides are known constants,
// check whether the accesses interlace each other.
if (ConstDist > 0 && CommonStride && CommonStride > 1 && HasSameSize &&
areStridedAccessesIndependent(ConstDist, *CommonStride, TypeByteSize)) {
LLVM_DEBUG(dbgs() << "LAA: Strided accesses are independent\n");
return Dependence::NoDep;
}
} else {
if (!LoopGuards)
LoopGuards.emplace(
ScalarEvolution::LoopGuards::collect(InnermostLoop, SE));
Dist = SE.applyLoopGuards(Dist, *LoopGuards);
}
// Negative distances are not plausible dependencies.
if (SE.isKnownNonPositive(Dist)) {
if (SE.isKnownNonNegative(Dist)) {
// Equal-sized accesses to the same location are forward.
if (HasSameSize)
return Dependence::Forward;
if (CommonStride) {
// For mixed sizes, CommonStride is asserted to cover both accesses when
// computed in getDependenceDistanceStrideAndSize, so different
// iterations cannot overlap.
[[maybe_unused]] uint64_t ASz =
DL.getTypeAllocSize(getLoadStoreType(InstMap[AIdx]));
[[maybe_unused]] uint64_t BSz =
DL.getTypeAllocSize(getLoadStoreType(InstMap[BIdx]));
assert(*CommonStride >= std::max(ASz, BSz) &&
"Invariant from getDependenceDistanceStrideAndSize broken!");
return Dependence::Forward;
}
LLVM_DEBUG(dbgs() << "LAA: possibly zero dependence difference but "
"different type sizes\n");
return Dependence::Unknown;
}
bool IsTrueDataDependence = (AIsWrite && !BIsWrite);
// Check if the first access writes to a location that is read in a later
// iteration, where the distance between them is not a multiple of a vector
// factor and relatively small.
//
// NOTE: There is no need to update MaxSafeVectorWidthInBits after call to
// couldPreventStoreLoadForward, even if it changed MinDepDistBytes, since a
// forward dependency will allow vectorization using any width.
if (IsTrueDataDependence && EnableForwardingConflictDetection) {
if (!ConstDist) {
return CheckCompletelyBeforeOrAfter() ? Dependence::NoDep
: Dependence::Unknown;
}
if (!HasSameSize ||
couldPreventStoreLoadForward(ConstDist, TypeByteSize)) {
LLVM_DEBUG(
dbgs() << "LAA: Forward but may prevent st->ld forwarding\n");
return Dependence::ForwardButPreventsForwarding;
}
}
LLVM_DEBUG(dbgs() << "LAA: Dependence is negative\n");
return Dependence::Forward;
}
std::optional<int64_t> MinDistanceOpt =
SE.getSignedRangeMin(Dist).trySExtValue();
if (!MinDistanceOpt) {
LLVM_DEBUG(dbgs() << "LAA: Minimum distance does not fit in 64 bits.\n");
return Dependence::Unknown;
}
int64_t MinDistance = *MinDistanceOpt;
// Below we only handle strictly positive distances.
if (MinDistance <= 0) {
return CheckCompletelyBeforeOrAfter() ? Dependence::NoDep
: Dependence::Unknown;
}
if (!HasSameSize) {
if (CheckCompletelyBeforeOrAfter())
return Dependence::NoDep;
LLVM_DEBUG(dbgs() << "LAA: ReadWrite-Write positive dependency with "
"different type sizes\n");
return Dependence::Unknown;
}
// Bail out early if passed-in parameters make vectorization not feasible.
unsigned MinForcedFactor =
std::max(1U, VectorizerParams::VectorizationFactor.getKnownMinValue());
unsigned ForcedUnroll = (VectorizerParams::VectorizationInterleave ?
VectorizerParams::VectorizationInterleave : 1);
// The minimum number of iterations for a vectorized/unrolled version.
unsigned MinNumIter = std::max(MinForcedFactor * ForcedUnroll, 2U);
// It's not vectorizable if the distance is smaller than the minimum distance
// needed for a vectroized/unrolled version. Vectorizing one iteration in
// front needs MaxStride. Vectorizing the last iteration needs TypeByteSize.
// (No need to plus the last gap distance).
//
// E.g. Assume one char is 1 byte in memory and one int is 4 bytes.
// foo(int *A) {
// int *B = (int *)((char *)A + 14);
// for (i = 0 ; i < 1024 ; i += 2)
// B[i] = A[i] + 1;
// }
//
// Two accesses in memory (stride is 4 * 2):
// | A[0] | | A[2] | | A[4] | | A[6] | |
// | B[0] | | B[2] | | B[4] |
//
// MinDistance needs for vectorizing iterations except the last iteration:
// 4 * 2 * (MinNumIter - 1). MinDistance needs for the last iteration: 4.
// So the minimum distance needed is: 4 * 2 * (MinNumIter - 1) + 4.
//
// If MinNumIter is 2, it is vectorizable as the minimum distance needed is
// 12, which is less than distance.
//
// If MinNumIter is 4 (Say if a user forces the vectorization factor to be 4),
// the minimum distance needed is 28, which is greater than distance. It is
// not safe to do vectorization.
//
// We use MaxStride (maximum of src and sink strides) to get a conservative
// lower bound on the MinDistanceNeeded in case of different strides.
// We know that Dist is positive, but it may not be constant. Use the signed
// minimum for computations below, as this ensures we compute the closest
// possible dependence distance.
uint64_t MinDistanceNeeded = MaxStride * (MinNumIter - 1) + TypeByteSize;
if (MinDistanceNeeded > static_cast<uint64_t>(MinDistance)) {
if (!ConstDist) {
// For non-constant distances, we checked the lower bound of the
// dependence distance and the distance may be larger at runtime (and safe
// for vectorization). Classify it as Unknown, so we re-try with runtime
// checks, unless we can prove both accesses cannot overlap.
return CheckCompletelyBeforeOrAfter() ? Dependence::NoDep
: Dependence::Unknown;
}
LLVM_DEBUG(dbgs() << "LAA: Failure because of positive minimum distance "
<< MinDistance << '\n');
return Dependence::Backward;
}
// Unsafe if the minimum distance needed is greater than smallest dependence
// distance distance.
if (MinDistanceNeeded > MinDepDistBytes) {
LLVM_DEBUG(dbgs() << "LAA: Failure because it needs at least "
<< MinDistanceNeeded << " size in bytes\n");
return Dependence::Backward;
}
MinDepDistBytes =
std::min(static_cast<uint64_t>(MinDistance), MinDepDistBytes);
bool IsTrueDataDependence = (!AIsWrite && BIsWrite);
if (IsTrueDataDependence && EnableForwardingConflictDetection && ConstDist &&
couldPreventStoreLoadForward(MinDistance, TypeByteSize, *CommonStride))
return Dependence::BackwardVectorizableButPreventsForwarding;
uint64_t MaxVF = MinDepDistBytes / MaxStride;
LLVM_DEBUG(dbgs() << "LAA: Positive min distance " << MinDistance
<< " with max VF = " << MaxVF << '\n');
uint64_t MaxVFInBits = MaxVF * TypeByteSize * 8;
if (!ConstDist && MaxVFInBits < MaxTargetVectorWidthInBits) {
// For non-constant distances, we checked the lower bound of the dependence
// distance and the distance may be larger at runtime (and safe for
// vectorization). Classify it as Unknown, so we re-try with runtime checks,
// unless we can prove both accesses cannot overlap.
return CheckCompletelyBeforeOrAfter() ? Dependence::NoDep
: Dependence::Unknown;
}
if (CheckCompletelyBeforeOrAfter())
return Dependence::NoDep;
MaxSafeVectorWidthInBits = std::min(MaxSafeVectorWidthInBits, MaxVFInBits);
return Dependence::BackwardVectorizable;
}
bool MemoryDepChecker::areDepsSafe(const DepCandidates &DepCands,
ArrayRef<MemAccessInfo> CheckDeps) {
MinDepDistBytes = -1;
SmallPtrSet<MemAccessInfo, 8> Visited;
for (MemAccessInfo CurAccess : CheckDeps) {
if (Visited.contains(CurAccess))
continue;
// Check accesses within this set.
EquivalenceClasses<MemAccessInfo>::member_iterator AI =
DepCands.findLeader(CurAccess);
EquivalenceClasses<MemAccessInfo>::member_iterator AE =
DepCands.member_end();
// Check every access pair.
while (AI != AE) {
Visited.insert(*AI);
bool AIIsWrite = AI->getInt();
// Reads from the same pointer don't create extra hazards, but multiple
// stores do (WAW), so start from AI for writes and next(AI) for reads.
EquivalenceClasses<MemAccessInfo>::member_iterator OI =
(AIIsWrite ? AI : std::next(AI));
while (OI != AE) {
// Check every accessing instruction pair in program order.
auto &Acc = Accesses[*AI];
for (std::vector<unsigned>::iterator I1 = Acc.begin(), I1E = Acc.end();
I1 != I1E; ++I1)
// When checking for WAW (OI == AI) caused by multiple writes to the
// same pointer, start I2 at the next access past I1 to avoid
// self-comparison.
for (std::vector<unsigned>::iterator
I2 = (OI == AI ? std::next(I1) : Accesses[*OI].begin()),
I2E = (OI == AI ? I1E : Accesses[*OI].end());
I2 != I2E; ++I2) {
auto A = std::make_pair(&*AI, *I1);
auto B = std::make_pair(&*OI, *I2);
assert(*I1 != *I2);
if (*I1 > *I2)
std::swap(A, B);
Dependence::DepType Type =
isDependent(*A.first, A.second, *B.first, B.second);
mergeInStatus(Dependence::isSafeForVectorization(Type));
// Gather dependences unless we accumulated MaxDependences
// dependences. In that case return as soon as we find the first
// unsafe dependence. This puts a limit on this quadratic
// algorithm.
if (RecordDependences) {
if (Type != Dependence::NoDep)
Dependences.emplace_back(A.second, B.second, Type);
if (Dependences.size() >= MaxDependences) {
RecordDependences = false;
Dependences.clear();
LLVM_DEBUG(dbgs()
<< "Too many dependences, stopped recording\n");
}
}
if (!RecordDependences && !isSafeForVectorization())
return false;
}
++OI;
}
++AI;
}
}
LLVM_DEBUG(dbgs() << "Total Dependences: " << Dependences.size() << "\n");
return isSafeForVectorization();
}
SmallVector<Instruction *, 4>
MemoryDepChecker::getInstructionsForAccess(Value *Ptr, bool IsWrite) const {
MemAccessInfo Access(Ptr, IsWrite);
auto I = Accesses.find(Access);
SmallVector<Instruction *, 4> Insts;
if (I != Accesses.end()) {
transform(I->second, std::back_inserter(Insts),
[&](unsigned Idx) { return this->InstMap[Idx]; });
}
return Insts;
}
const char *MemoryDepChecker::Dependence::DepName[] = {
"NoDep",
"Unknown",
"IndirectUnsafe",
"InvariantUnsafe",
"Forward",
"ForwardButPreventsForwarding",
"Backward",
"BackwardVectorizable",
"BackwardVectorizableButPreventsForwarding"};
void MemoryDepChecker::Dependence::print(
raw_ostream &OS, unsigned Depth,
const SmallVectorImpl<Instruction *> &Instrs) const {
OS.indent(Depth) << DepName[Type] << ":\n";
OS.indent(Depth + 2) << *Instrs[Source] << " -> \n";
OS.indent(Depth + 2) << *Instrs[Destination] << "\n";
}
bool LoopAccessInfo::canAnalyzeLoop() {
// We need to have a loop header.
LLVM_DEBUG(dbgs() << "\nLAA: Checking a loop in '"
<< TheLoop->getHeader()->getParent()->getName() << "' from "
<< TheLoop->getLocStr() << "\n");
// We can only analyze innermost loops.
if (!TheLoop->isInnermost()) {
LLVM_DEBUG(dbgs() << "LAA: loop is not the innermost loop\n");
recordAnalysis("NotInnerMostLoop") << "loop is not the innermost loop";
return false;
}
// We must have a single backedge.
if (TheLoop->getNumBackEdges() != 1) {
LLVM_DEBUG(
dbgs() << "LAA: loop control flow is not understood by analyzer\n");
recordAnalysis("CFGNotUnderstood")
<< "loop control flow is not understood by analyzer";
return false;
}
// ScalarEvolution needs to be able to find the symbolic max backedge taken
// count, which is an upper bound on the number of loop iterations. The loop
// may execute fewer iterations, if it exits via an uncountable exit.
const SCEV *ExitCount = PSE->getSymbolicMaxBackedgeTakenCount();
if (isa<SCEVCouldNotCompute>(ExitCount)) {
recordAnalysis("CantComputeNumberOfIterations")
<< "could not determine number of loop iterations";
LLVM_DEBUG(dbgs() << "LAA: SCEV could not compute the loop exit count.\n");
return false;
}
LLVM_DEBUG(dbgs() << "LAA: Found an analyzable loop: "
<< TheLoop->getHeader()->getName() << "\n");
return true;
}
bool LoopAccessInfo::analyzeLoop(AAResults *AA, const LoopInfo *LI,
const TargetLibraryInfo *TLI,
DominatorTree *DT) {
// Holds the Load and Store instructions.
SmallVector<LoadInst *, 16> Loads;
SmallVector<StoreInst *, 16> Stores;
SmallPtrSet<MDNode *, 8> LoopAliasScopes;
// Holds all the different accesses in the loop.
unsigned NumReads = 0;
unsigned NumReadWrites = 0;
bool HasComplexMemInst = false;
// A runtime check is only legal to insert if there are no convergent calls.
HasConvergentOp = false;
PtrRtChecking->Pointers.clear();
PtrRtChecking->Need = false;
const bool IsAnnotatedParallel = TheLoop->isAnnotatedParallel();
const bool EnableMemAccessVersioningOfLoop =
EnableMemAccessVersioning &&
!TheLoop->getHeader()->getParent()->hasOptSize();
// Traverse blocks in fixed RPOT order, regardless of their storage in the
// loop info, as it may be arbitrary.
LoopBlocksRPO RPOT(TheLoop);
RPOT.perform(LI);
// Don't return early as soon as we found a memory access that cannot be
// vectorize - HasConvergentOp must still be computed as it is part of LAI's
// public API (used by LoopDistribute).
for (BasicBlock *BB : RPOT) {
// Scan the BB and collect legal loads and stores. Also detect any
// convergent instructions.
for (Instruction &I : *BB) {
if (auto *Call = dyn_cast<CallBase>(&I)) {
if (Call->isConvergent())
HasConvergentOp = true;
}
// Unsafe to vectorize and we already found a convergent operation, can
// early return now.
if (HasComplexMemInst && HasConvergentOp)
return false;
// Already unsafe to vectorize; keep scanning for convergent ops.
if (HasComplexMemInst)
continue;
// Record alias scopes defined inside the loop.
if (auto *Decl = dyn_cast<NoAliasScopeDeclInst>(&I))
for (Metadata *Op : Decl->getScopeList()->operands())
LoopAliasScopes.insert(cast<MDNode>(Op));
// Many math library functions read the rounding mode. We will only
// vectorize a loop if it contains known function calls that don't set
// the flag. Therefore, it is safe to ignore this read from memory.
auto *Call = dyn_cast<CallInst>(&I);
if (Call && getVectorIntrinsicIDForCall(Call, TLI))
continue;
// If this is a load, save it. If this instruction can read from memory
// but is not a load, we only allow it if it's a call to a function with a
// vector mapping and no pointer arguments.
if (I.mayReadFromMemory()) {
auto hasPointerArgs = [](CallBase *CB) {
return any_of(CB->args(), [](Value const *Arg) {
return Arg->getType()->isPointerTy();
});
};
// If the function has an explicit vectorized counterpart, and does not
// take output/input pointers, we can safely assume that it can be
// vectorized.
if (Call && !Call->isNoBuiltin() && Call->getCalledFunction() &&
!hasPointerArgs(Call) && !VFDatabase::getMappings(*Call).empty())
continue;
auto *Ld = dyn_cast<LoadInst>(&I);
if (!Ld) {
recordAnalysis("CantVectorizeInstruction", &I)
<< "instruction cannot be vectorized";
HasComplexMemInst = true;
continue;
}
if (!Ld->isSimple() && !IsAnnotatedParallel) {
recordAnalysis("NonSimpleLoad", Ld)
<< "read with atomic ordering or volatile read";
LLVM_DEBUG(dbgs() << "LAA: Found a non-simple load.\n");
HasComplexMemInst = true;
continue;
}
Loads.push_back(Ld);
DepChecker->addAccess(Ld);
if (EnableMemAccessVersioningOfLoop)
collectStridedAccess(Ld);
continue;
}
// Save 'store' instructions. Abort if other instructions write to memory.
if (I.mayWriteToMemory()) {
auto *St = dyn_cast<StoreInst>(&I);
if (!St) {
recordAnalysis("CantVectorizeInstruction", &I)
<< "instruction cannot be vectorized";
HasComplexMemInst = true;
continue;
}
if (!St->isSimple() && !IsAnnotatedParallel) {
recordAnalysis("NonSimpleStore", St)
<< "write with atomic ordering or volatile write";
LLVM_DEBUG(dbgs() << "LAA: Found a non-simple store.\n");
HasComplexMemInst = true;
continue;
}
Stores.push_back(St);
DepChecker->addAccess(St);
if (EnableMemAccessVersioningOfLoop)
collectStridedAccess(St);
}
} // Next instr.
} // Next block.
if (HasComplexMemInst)
return false;
// Now we have two lists that hold the loads and the stores.
// Next, we find the pointers that they use.
// Check if we see any stores. If there are no stores, then we don't
// care if the pointers are *restrict*.
if (!Stores.size()) {
LLVM_DEBUG(dbgs() << "LAA: Found a read-only loop!\n");
return true;
}
MemoryDepChecker::DepCandidates DepCands;
AccessAnalysis Accesses(TheLoop, AA, LI, *DT, DepCands, *PSE,
LoopAliasScopes);
// Holds the analyzed pointers. We don't want to call getUnderlyingObjects
// multiple times on the same object. If the ptr is accessed twice, once
// for read and once for write, it will only appear once (on the write
// list). This is okay, since we are going to check for conflicts between
// writes and between reads and writes, but not between reads and reads.
SmallSet<std::pair<Value *, Type *>, 16> Seen;
// Record uniform store addresses to identify if we have multiple stores
// to the same address.
SmallPtrSet<Value *, 16> UniformStores;
for (StoreInst *ST : Stores) {
Value *Ptr = ST->getPointerOperand();
if (isInvariant(Ptr)) {
// Record store instructions to loop invariant addresses
StoresToInvariantAddresses.push_back(ST);
HasStoreStoreDependenceInvolvingLoopInvariantAddress |=
!UniformStores.insert(Ptr).second;
}
// If we did *not* see this pointer before, insert it to the read-write
// list. At this phase it is only a 'write' list.
Type *AccessTy = getLoadStoreType(ST);
if (Seen.insert({Ptr, AccessTy}).second) {
++NumReadWrites;
MemoryLocation Loc = MemoryLocation::get(ST);
// The TBAA metadata could have a control dependency on the predication
// condition, so we cannot rely on it when determining whether or not we
// need runtime pointer checks.
if (blockNeedsPredication(ST->getParent(), TheLoop, DT))
Loc.AATags.TBAA = nullptr;
// Expand forked pointers (i.e., a phi of multiple strided pointers) into
// all alternatives.
visitPointers(const_cast<Value *>(Loc.Ptr), *TheLoop,
[&Accesses, AccessTy, Loc](Value *Ptr) {
MemoryLocation NewLoc = Loc.getWithNewPtr(Ptr);
Accesses.addStore(NewLoc, AccessTy);
});
}
}
if (IsAnnotatedParallel) {
LLVM_DEBUG(
dbgs() << "LAA: A loop annotated parallel, ignore memory dependency "
<< "checks.\n");
return true;
}
for (LoadInst *LD : Loads) {
Value *Ptr = LD->getPointerOperand();
// If we did *not* see this pointer before, insert it to the read list. If
// we *did* see it before, then it is already in the read-write list. This
// allows us to vectorize expressions such as A[i] += x; Because the address
// of A[i] is a read-write pointer. This only works if the index of A[i] is
// strictly monotonic, which we approximate (conservatively) via
// getPtrStride. If the address is unknown (e.g. A[B[i]]) then we may read,
// modify, and write overlapping words. Note that "zero stride" is unsafe
// and is being handled below.
bool IsReadOnlyPtr = false;
Type *AccessTy = getLoadStoreType(LD);
if (Seen.insert({Ptr, AccessTy}).second ||
!getPtrStride(*PSE, AccessTy, Ptr, TheLoop, *DT, SymbolicStrides, false,
true)) {
++NumReads;
IsReadOnlyPtr = true;
}
// See if there is an unsafe dependency between a load to a uniform address and
// store to the same uniform address.
if (UniformStores.contains(Ptr)) {
LLVM_DEBUG(dbgs() << "LAA: Found an unsafe dependency between a uniform "
"load and uniform store to the same address!\n");
HasLoadStoreDependenceInvolvingLoopInvariantAddress = true;
}
MemoryLocation Loc = MemoryLocation::get(LD);
// The TBAA metadata could have a control dependency on the predication
// condition, so we cannot rely on it when determining whether or not we
// need runtime pointer checks.
if (blockNeedsPredication(LD->getParent(), TheLoop, DT))
Loc.AATags.TBAA = nullptr;
// Expand forked pointers (i.e., a phi of multiple strided pointers) into
// all alternatives.
visitPointers(const_cast<Value *>(Loc.Ptr), *TheLoop,
[&Accesses, AccessTy, Loc, IsReadOnlyPtr](Value *Ptr) {
MemoryLocation NewLoc = Loc.getWithNewPtr(Ptr);
Accesses.addLoad(NewLoc, AccessTy, IsReadOnlyPtr);
});
}
// If we write (or read-write) to a single destination and there are no other
// reads in this loop then is it safe to vectorize: the vectorized stores
// preserve ordering via replication or order-preserving @llvm.masked.scatter.
if (NumReadWrites == 1 && NumReads == 0) {
LLVM_DEBUG(dbgs() << "LAA: Found a write-only loop!\n");
return true;
}
// Build dependence sets and check whether we need a runtime pointer bounds
// check.
Accesses.buildDependenceSets();
// Find pointers with computable bounds. We are going to use this information
// to place a runtime bound check.
Value *UncomputablePtr = nullptr;
HasCompletePtrRtChecking =
Accesses.canCheckPtrAtRT(*PtrRtChecking, TheLoop, SymbolicStrides,
UncomputablePtr, AllowPartial, getDepChecker());
if (!HasCompletePtrRtChecking) {
const auto *I = dyn_cast_or_null<Instruction>(UncomputablePtr);
recordAnalysis("CantIdentifyArrayBounds", I)
<< "cannot identify array bounds";
LLVM_DEBUG(dbgs() << "LAA: We can't vectorize because we can't find "
<< "the array bounds.\n");
return false;
}
LLVM_DEBUG(
dbgs() << "LAA: May be able to perform a memory runtime check if needed.\n");
bool DepsAreSafe = true;
if (Accesses.isDependencyCheckNeeded()) {
LLVM_DEBUG(dbgs() << "LAA: Checking memory dependencies\n");
DepsAreSafe =
DepChecker->areDepsSafe(DepCands, Accesses.getDependenciesToCheck());
if (!DepsAreSafe && DepChecker->shouldRetryWithRuntimeChecks()) {
LLVM_DEBUG(dbgs() << "LAA: Retrying with memory checks\n");
PtrRtChecking->reset();
PtrRtChecking->Need = true;
UncomputablePtr = nullptr;
HasCompletePtrRtChecking = Accesses.canCheckPtrAtRT(
*PtrRtChecking, TheLoop, SymbolicStrides, UncomputablePtr,
AllowPartial, getDepChecker());
// Check that we found the bounds for the pointer.
if (!HasCompletePtrRtChecking) {
auto *I = dyn_cast_or_null<Instruction>(UncomputablePtr);
recordAnalysis("CantCheckMemDepsAtRunTime", I)
<< "cannot check memory dependencies at runtime";
LLVM_DEBUG(dbgs() << "LAA: Can't vectorize with memory checks\n");
return false;
}
// Clear the dependency checks. They are no longer needed.
Accesses.resetDepChecks(*DepChecker);
DepsAreSafe = true;
}
}
// Update the invariant address dependence flags based on dependences found
// by the dep checker. Even if dependences were not recorded (too many to
// track), any InvariantUnsafe dep would still have set the status to Unsafe
if (const auto *Deps = DepChecker->getDependences()) {
for (const auto &Dep : *Deps) {
if (Dep.Type != MemoryDepChecker::Dependence::InvariantUnsafe)
continue;
Instruction *Src = Dep.getSource(*DepChecker);
Instruction *Dst = Dep.getDestination(*DepChecker);
if (isa<LoadInst>(Src) != isa<LoadInst>(Dst)) {
HasLoadStoreDependenceInvolvingLoopInvariantAddress = true;
} else {
assert(isa<StoreInst>(Src) && isa<StoreInst>(Dst) &&
"Expected both to be stores");
HasStoreStoreDependenceInvolvingLoopInvariantAddress = true;
}
}
}
if (HasConvergentOp) {
recordAnalysis("CantInsertRuntimeCheckWithConvergent")
<< "cannot add control dependency to convergent operation";
LLVM_DEBUG(dbgs() << "LAA: We can't vectorize because a runtime check "
"would be needed with a convergent operation\n");
return false;
}
if (DepsAreSafe) {
LLVM_DEBUG(
dbgs() << "LAA: No unsafe dependent memory operations in loop. We"
<< (PtrRtChecking->Need ? "" : " don't")
<< " need runtime memory checks.\n");
return true;
}
emitUnsafeDependenceRemark();
return false;
}
void LoopAccessInfo::emitUnsafeDependenceRemark() {
const auto *Deps = getDepChecker().getDependences();
if (!Deps)
return;
const auto *Found =
llvm::find_if(*Deps, [](const MemoryDepChecker::Dependence &D) {
return MemoryDepChecker::Dependence::isSafeForVectorization(D.Type) !=
MemoryDepChecker::VectorizationSafetyStatus::Safe;
});
if (Found == Deps->end())
return;
MemoryDepChecker::Dependence Dep = *Found;
LLVM_DEBUG(dbgs() << "LAA: unsafe dependent memory operations in loop\n");
// Emit remark for first unsafe dependence
bool HasForcedDistribution =
getBooleanLoopAttribute(TheLoop, "llvm.loop.distribute.enable");
const std::string Info =
HasForcedDistribution
? "unsafe dependent memory operations in loop."
: "unsafe dependent memory operations in loop. Use "
"#pragma clang loop distribute(enable) to allow loop distribution "
"to attempt to isolate the offending operations into a separate "
"loop";
OptimizationRemarkAnalysis &R =
recordAnalysis("UnsafeDep", Dep.getDestination(getDepChecker())) << Info;
switch (Dep.Type) {
case MemoryDepChecker::Dependence::NoDep:
case MemoryDepChecker::Dependence::Forward:
case MemoryDepChecker::Dependence::BackwardVectorizable:
llvm_unreachable("Unexpected dependence");
case MemoryDepChecker::Dependence::Backward:
R << "\nBackward loop carried data dependence.";
break;
case MemoryDepChecker::Dependence::ForwardButPreventsForwarding:
R << "\nForward loop carried data dependence that prevents "
"store-to-load forwarding.";
break;
case MemoryDepChecker::Dependence::BackwardVectorizableButPreventsForwarding:
R << "\nBackward loop carried data dependence that prevents "
"store-to-load forwarding.";
break;
case MemoryDepChecker::Dependence::IndirectUnsafe:
R << "\nUnsafe indirect dependence.";
break;
case MemoryDepChecker::Dependence::InvariantUnsafe:
R << "\nUnsafe dependence on loop-invariant address.";
break;
case MemoryDepChecker::Dependence::Unknown:
R << "\nUnknown data dependence.";
break;
}
if (Instruction *I = Dep.getSource(getDepChecker())) {
DebugLoc SourceLoc = I->getDebugLoc();
if (auto *DD = dyn_cast_or_null<Instruction>(getPointerOperand(I)))
SourceLoc = DD->getDebugLoc();
if (SourceLoc)
R << " Memory location is the same as accessed at "
<< ore::NV("Location", SourceLoc);
}
}
bool LoopAccessInfo::blockNeedsPredication(const BasicBlock *BB,
const Loop *TheLoop,
const DominatorTree *DT) {
assert(TheLoop->contains(BB) && "Unknown block used");
// Blocks that do not dominate the latch need predication.
const BasicBlock *Latch = TheLoop->getLoopLatch();
assert(Latch && "Loop expected to have a single latch.");
return !DT->dominates(BB, Latch);
}
OptimizationRemarkAnalysis &
LoopAccessInfo::recordAnalysis(StringRef RemarkName, const Instruction *I) {
assert(!Report && "Multiple reports generated");
const BasicBlock *CodeRegion = TheLoop->getHeader();
DebugLoc DL = TheLoop->getStartLoc();
if (I) {
CodeRegion = I->getParent();
// If there is no debug location attached to the instruction, revert back to
// using the loop's.
if (I->getDebugLoc())
DL = I->getDebugLoc();
}
Report = std::make_unique<OptimizationRemarkAnalysis>(DEBUG_TYPE, RemarkName,
DL, CodeRegion);
return *Report;
}
bool LoopAccessInfo::isInvariant(Value *V) const {
auto *SE = PSE->getSE();
if (TheLoop->isLoopInvariant(V))
return true;
if (!SE->isSCEVable(V->getType()))
return false;
const SCEV *S = SE->getSCEV(V);
return SE->isLoopInvariant(S, TheLoop);
}
/// If \p Ptr is a GEP, which has a loop-variant operand, return that operand.
/// Otherwise, return \p Ptr.
static Value *getLoopVariantGEPOperand(Value *Ptr, ScalarEvolution *SE,
Loop *Lp) {
auto *GEP = dyn_cast<GetElementPtrInst>(Ptr);
if (!GEP)
return Ptr;
Value *V = Ptr;
for (const Use &U : GEP->operands()) {
if (!SE->isLoopInvariant(SE->getSCEV(U), Lp)) {
if (V == Ptr)
V = U;
else
// There must be exactly one loop-variant operand.
return Ptr;
}
}
return V;
}
/// Get the stride of a pointer access in a loop. Looks for symbolic
/// strides "a[i*stride]". Returns the symbolic stride, or null otherwise.
static const SCEV *getStrideFromPointer(Value *Ptr, ScalarEvolution *SE, Loop *Lp) {
auto *PtrTy = dyn_cast<PointerType>(Ptr->getType());
if (!PtrTy)
return nullptr;
// Try to remove a gep instruction to make the pointer (actually index at this
// point) easier analyzable. If OrigPtr is equal to Ptr we are analyzing the
// pointer, otherwise, we are analyzing the index.
Value *OrigPtr = Ptr;
Ptr = getLoopVariantGEPOperand(Ptr, SE, Lp);
const SCEV *V = SE->getSCEV(Ptr);
if (Ptr != OrigPtr)
// Strip off casts.
while (auto *C = dyn_cast<SCEVIntegralCastExpr>(V))
V = C->getOperand();
if (!match(V, m_scev_AffineAddRec(m_SCEV(), m_SCEV(V), m_SpecificLoop(Lp))))
return nullptr;
// Note that the restriction after this loop invariant check are only
// profitability restrictions.
if (!SE->isLoopInvariant(V, Lp))
return nullptr;
// Look for the loop invariant symbolic value.
if (isa<SCEVUnknown>(V))
return V;
// Look through multiplies that scale a stride by a constant.
match(V, m_scev_Mul(m_SCEVConstant(), m_SCEV(V)));
if (auto *C = dyn_cast<SCEVIntegralCastExpr>(V))
if (isa<SCEVUnknown>(C->getOperand()))
return V;
return nullptr;
}
void LoopAccessInfo::collectStridedAccess(Value *MemAccess) {
Value *Ptr = getLoadStorePointerOperand(MemAccess);
if (!Ptr)
return;
// Note: getStrideFromPointer is a *profitability* heuristic. We
// could broaden the scope of values returned here - to anything
// which happens to be loop invariant and contributes to the
// computation of an interesting IV - but we chose not to as we
// don't have a cost model here, and broadening the scope exposes
// far too many unprofitable cases.
const SCEV *StrideExpr = getStrideFromPointer(Ptr, PSE->getSE(), TheLoop);
if (!StrideExpr)
return;
if (match(StrideExpr, m_scev_UndefOrPoison()))
return;
LLVM_DEBUG(dbgs() << "LAA: Found a strided access that is a candidate for "
"versioning:");
LLVM_DEBUG(dbgs() << " Ptr: " << *Ptr << " Stride: " << *StrideExpr << "\n");
if (!SpeculateUnitStride) {
LLVM_DEBUG(dbgs() << " Chose not to due to -laa-speculate-unit-stride\n");
return;
}
// Avoid adding the "Stride == 1" predicate when we know that
// Stride >= Trip-Count. Such a predicate will effectively optimize a single
// or zero iteration loop, as Trip-Count <= Stride == 1.
//
// TODO: We are currently not making a very informed decision on when it is
// beneficial to apply stride versioning. It might make more sense that the
// users of this analysis (such as the vectorizer) will trigger it, based on
// their specific cost considerations; For example, in cases where stride
// versioning does not help resolving memory accesses/dependences, the
// vectorizer should evaluate the cost of the runtime test, and the benefit
// of various possible stride specializations, considering the alternatives
// of using gather/scatters (if available).
const SCEV *MaxBTC = PSE->getSymbolicMaxBackedgeTakenCount();
// Match the types so we can compare the stride and the MaxBTC.
// The Stride can be positive/negative, so we sign extend Stride;
// The backedgeTakenCount is non-negative, so we zero extend MaxBTC.
const DataLayout &DL = TheLoop->getHeader()->getDataLayout();
uint64_t StrideTypeSizeBits = DL.getTypeSizeInBits(StrideExpr->getType());
uint64_t BETypeSizeBits = DL.getTypeSizeInBits(MaxBTC->getType());
const SCEV *CastedStride = StrideExpr;
const SCEV *CastedBECount = MaxBTC;
ScalarEvolution *SE = PSE->getSE();
if (BETypeSizeBits >= StrideTypeSizeBits)
CastedStride = SE->getNoopOrSignExtend(StrideExpr, MaxBTC->getType());
else
CastedBECount = SE->getZeroExtendExpr(MaxBTC, StrideExpr->getType());
const SCEV *StrideMinusBETaken = SE->getMinusSCEV(CastedStride, CastedBECount);
// Since TripCount == BackEdgeTakenCount + 1, checking:
// "Stride >= TripCount" is equivalent to checking:
// Stride - MaxBTC> 0
if (SE->isKnownPositive(StrideMinusBETaken)) {
LLVM_DEBUG(
dbgs() << "LAA: Stride>=TripCount; No point in versioning as the "
"Stride==1 predicate will imply that the loop executes "
"at most once.\n");
return;
}
LLVM_DEBUG(dbgs() << "LAA: Found a strided access that we can version.\n");
// Strip back off the integer cast, and check that our result is a
// SCEVUnknown as we expect.
const SCEV *StrideBase = StrideExpr;
if (const auto *C = dyn_cast<SCEVIntegralCastExpr>(StrideBase))
StrideBase = C->getOperand();
assert(SE->isLoopInvariant(StrideBase, TheLoop) &&
"users of the map rely on the stride being loop invariant");
SymbolicStrides[Ptr] = cast<SCEVUnknown>(StrideBase);
}
LoopAccessInfo::LoopAccessInfo(Loop *L, ScalarEvolution *SE,
const TargetTransformInfo *TTI,
const TargetLibraryInfo *TLI, AAResults *AA,
DominatorTree *DT, LoopInfo *LI,
AssumptionCache *AC, bool AllowPartial)
: PSE(std::make_unique<PredicatedScalarEvolution>(*SE, *L)),
PtrRtChecking(nullptr), TheLoop(L), AllowPartial(AllowPartial) {
unsigned MaxTargetVectorWidthInBits = std::numeric_limits<unsigned>::max();
if (TTI && !TTI->enableScalableVectorization())
// Scale the vector width by 2 as rough estimate to also consider
// interleaving.
MaxTargetVectorWidthInBits =
TTI->getRegisterBitWidth(TargetTransformInfo::RGK_FixedWidthVector) * 2;
DepChecker = std::make_unique<MemoryDepChecker>(
*PSE, AC, DT, L, SymbolicStrides, MaxTargetVectorWidthInBits, LoopGuards);
PtrRtChecking =
std::make_unique<RuntimePointerChecking>(*DepChecker, SE, LoopGuards);
if (canAnalyzeLoop())
CanVecMem = analyzeLoop(AA, LI, TLI, DT);
}
void LoopAccessInfo::print(raw_ostream &OS, unsigned Depth) const {
if (CanVecMem) {
OS.indent(Depth) << "Memory dependences are safe";
const MemoryDepChecker &DC = getDepChecker();
if (!DC.isSafeForAnyVectorWidth())
OS << " with a maximum safe vector width of "
<< DC.getMaxSafeVectorWidthInBits() << " bits";
if (!DC.isSafeForAnyStoreLoadForwardDistances()) {
uint64_t SLDist = DC.getStoreLoadForwardSafeDistanceInBits();
OS << ", with a maximum safe store-load forward width of " << SLDist
<< " bits";
}
if (PtrRtChecking->Need)
OS << " with run-time checks";
OS << "\n";
}
if (HasConvergentOp)
OS.indent(Depth) << "Has convergent operation in loop\n";
if (Report)
OS.indent(Depth) << "Report: " << Report->getMsg() << "\n";
if (auto *Dependences = DepChecker->getDependences()) {
OS.indent(Depth) << "Dependences:\n";
for (const auto &Dep : *Dependences) {
Dep.print(OS, Depth + 2, DepChecker->getMemoryInstructions());
OS << "\n";
}
} else
OS.indent(Depth) << "Too many dependences, not recorded\n";
// List the pair of accesses need run-time checks to prove independence.
PtrRtChecking->print(OS, Depth);
if (PtrRtChecking->Need && !HasCompletePtrRtChecking)
OS.indent(Depth) << "Generated run-time checks are incomplete\n";
OS << "\n";
OS.indent(Depth)
<< "Non vectorizable stores to invariant address were "
<< (HasStoreStoreDependenceInvolvingLoopInvariantAddress ||
HasLoadStoreDependenceInvolvingLoopInvariantAddress
? ""
: "not ")
<< "found in loop.\n";
OS.indent(Depth) << "SCEV assumptions:\n";
PSE->getPredicate().print(OS, Depth);
OS << "\n";
OS.indent(Depth) << "Expressions re-written:\n";
PSE->print(OS, Depth);
}
const LoopAccessInfo &LoopAccessInfoManager::getInfo(Loop &L,
bool AllowPartial) {
const auto &[It, Inserted] = LoopAccessInfoMap.try_emplace(&L);
// We need to create the LoopAccessInfo if either we don't already have one,
// or if it was created with a different value of AllowPartial.
if (Inserted || It->second->hasAllowPartial() != AllowPartial)
It->second = std::make_unique<LoopAccessInfo>(&L, &SE, TTI, TLI, &AA, &DT,
&LI, AC, AllowPartial);
return *It->second;
}
void LoopAccessInfoManager::clear() {
// Collect LoopAccessInfo entries that may keep references to IR outside the
// analyzed loop or SCEVs that may have been modified or invalidated. At the
// moment, that is loops requiring memory or SCEV runtime checks, as those cache
// SCEVs, e.g. for pointer expressions.
LoopAccessInfoMap.remove_if([](const auto &Entry) {
const auto &LAI = Entry.second;
return !(LAI->getRuntimePointerChecking()->getChecks().empty() &&
LAI->getPSE().getPredicate().isAlwaysTrue());
});
}
bool LoopAccessInfoManager::invalidate(
Function &F, const PreservedAnalyses &PA,
FunctionAnalysisManager::Invalidator &Inv) {
// Check whether our analysis is preserved.
auto PAC = PA.getChecker<LoopAccessAnalysis>();
if (!PAC.preserved() && !PAC.preservedSet<AllAnalysesOn<Function>>())
// If not, give up now.
return true;
// Check whether the analyses we depend on became invalid for any reason.
// Skip checking TargetLibraryAnalysis as it is immutable and can't become
// invalid.
return Inv.invalidate<AAManager>(F, PA) ||
Inv.invalidate<ScalarEvolutionAnalysis>(F, PA) ||
Inv.invalidate<LoopAnalysis>(F, PA) ||
Inv.invalidate<DominatorTreeAnalysis>(F, PA);
}
LoopAccessInfoManager LoopAccessAnalysis::run(Function &F,
FunctionAnalysisManager &FAM) {
auto &SE = FAM.getResult<ScalarEvolutionAnalysis>(F);
auto &AA = FAM.getResult<AAManager>(F);
auto &DT = FAM.getResult<DominatorTreeAnalysis>(F);
auto &LI = FAM.getResult<LoopAnalysis>(F);
auto &TTI = FAM.getResult<TargetIRAnalysis>(F);
auto &TLI = FAM.getResult<TargetLibraryAnalysis>(F);
auto &AC = FAM.getResult<AssumptionAnalysis>(F);
return LoopAccessInfoManager(SE, AA, DT, LI, &TTI, &TLI, &AC);
}
AnalysisKey LoopAccessAnalysis::Key;