#include "llvm/ADT/StringExtras.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/AliasSetTracker.h"
+#include "llvm/Analysis/AssumptionTracker.h"
#include "llvm/Analysis/BlockFrequencyInfo.h"
+#include "llvm/Analysis/CodeMetrics.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/LoopIterator.h"
#include "llvm/Analysis/LoopPass.h"
cl::desc("Sets the SIMD width. Zero is autoselect."));
static cl::opt<unsigned>
-VectorizationUnroll("force-vector-unroll", cl::init(0), cl::Hidden,
- cl::desc("Sets the vectorization unroll count. "
+VectorizationInterleave("force-vector-interleave", cl::init(0), cl::Hidden,
+ cl::desc("Sets the vectorization interleave count. "
"Zero is autoselect."));
static cl::opt<bool>
"force-target-num-vector-regs", cl::init(0), cl::Hidden,
cl::desc("A flag that overrides the target's number of vector registers."));
-/// Maximum vectorization unroll count.
-static const unsigned MaxUnrollFactor = 16;
+/// Maximum vectorization interleave count.
+static const unsigned MaxInterleaveFactor = 16;
-static cl::opt<unsigned> ForceTargetMaxScalarUnrollFactor(
- "force-target-max-scalar-unroll", cl::init(0), cl::Hidden,
- cl::desc("A flag that overrides the target's max unroll factor for scalar "
- "loops."));
+static cl::opt<unsigned> ForceTargetMaxScalarInterleaveFactor(
+ "force-target-max-scalar-interleave", cl::init(0), cl::Hidden,
+ cl::desc("A flag that overrides the target's max interleave factor for "
+ "scalar loops."));
-static cl::opt<unsigned> ForceTargetMaxVectorUnrollFactor(
- "force-target-max-vector-unroll", cl::init(0), cl::Hidden,
- cl::desc("A flag that overrides the target's max unroll factor for "
+static cl::opt<unsigned> ForceTargetMaxVectorInterleaveFactor(
+ "force-target-max-vector-interleave", cl::init(0), cl::Hidden,
+ cl::desc("A flag that overrides the target's max interleave factor for "
"vectorized loops."));
static cl::opt<unsigned> ForceTargetInstructionCost(
LoopVectorizationLegality(Loop *L, ScalarEvolution *SE, const DataLayout *DL,
DominatorTree *DT, TargetLibraryInfo *TLI,
- AliasAnalysis *AA, Function *F)
+ AliasAnalysis *AA, Function *F,
+ const TargetTransformInfo *TTI)
: NumLoads(0), NumStores(0), NumPredStores(0), TheLoop(L), SE(SE), DL(DL),
- DT(DT), TLI(TLI), AA(AA), TheFunction(F), Induction(nullptr),
+ DT(DT), TLI(TLI), AA(AA), TheFunction(F), TTI(TTI), Induction(nullptr),
WidestIndTy(nullptr), HasFunNoNaNAttr(false), MaxSafeDepDistBytes(-1U) {
}
}
SmallPtrSet<Value *, 8>::iterator strides_end() { return StrideSet.end(); }
+ /// Returns true if the target machine supports masked store operation
+ /// for the given \p DataType and kind of access to \p Ptr.
+ bool isLegalMaskedStore(Type *DataType, Value *Ptr) {
+ return TTI->isLegalMaskedStore(DataType, isConsecutivePtr(Ptr));
+ }
+ /// Returns true if the target machine supports masked load operation
+ /// for the given \p DataType and kind of access to \p Ptr.
+ bool isLegalMaskedLoad(Type *DataType, Value *Ptr) {
+ return TTI->isLegalMaskedLoad(DataType, isConsecutivePtr(Ptr));
+ }
+ /// Returns true if vector representation of the instruction \p I
+ /// requires mask.
+ bool isMaskRequired(const Instruction* I) {
+ return (MaskedOp.count(I) != 0);
+ }
private:
/// Check if a single basic block loop is vectorizable.
/// At this point we know that this is a loop with a constant trip count
AliasAnalysis *AA;
/// Parent function
Function *TheFunction;
+ /// Target Transform Info
+ const TargetTransformInfo *TTI;
// --- vectorization state --- //
ValueToValueMap Strides;
SmallPtrSet<Value *, 8> StrideSet;
+
+ /// While vectorizing these instructions we have to generate a
+ /// call to the appropriate masked intrinsic
+ SmallPtrSet<const Instruction*, 8> MaskedOp;
};
/// LoopVectorizationCostModel - estimates the expected speedups due to
LoopVectorizationLegality *Legal,
const TargetTransformInfo &TTI,
const DataLayout *DL, const TargetLibraryInfo *TLI,
- const Function *F, const LoopVectorizeHints *Hints)
- : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), DL(DL), TLI(TLI), TheFunction(F), Hints(Hints) {}
+ AssumptionTracker *AT, const Function *F,
+ const LoopVectorizeHints *Hints)
+ : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), DL(DL), TLI(TLI),
+ TheFunction(F), Hints(Hints) {
+ CodeMetrics::collectEphemeralValues(L, AT, EphValues);
+ }
/// Information about vectorization costs
struct VectorizationFactor {
*TheFunction, DL, Message.str());
}
+ /// Values used only by @llvm.assume calls.
+ SmallPtrSet<const Value *, 32> EphValues;
+
/// The loop that we evaluate.
Loop *TheLoop;
/// Scev analysis.
case HK_WIDTH:
return isPowerOf2_32(Val) && Val <= MaxVectorWidth;
case HK_UNROLL:
- return isPowerOf2_32(Val) && Val <= MaxUnrollFactor;
+ return isPowerOf2_32(Val) && Val <= MaxInterleaveFactor;
case HK_FORCE:
return (Val <= 1);
}
/// Vectorization width.
Hint Width;
- /// Vectorization unroll factor.
- Hint Unroll;
+ /// Vectorization interleave factor.
+ Hint Interleave;
/// Vectorization forced
Hint Force;
- /// Array to help iterating through all hints.
- Hint *Hints[3]; // avoiding initialisation due to MSVC2012
/// Return the loop metadata prefix.
static StringRef Prefix() { return "llvm.loop."; }
FK_Enabled = 1, ///< Forcing enabled.
};
- LoopVectorizeHints(const Loop *L, bool DisableUnrolling)
+ LoopVectorizeHints(const Loop *L, bool DisableInterleaving)
: Width("vectorize.width", VectorizationFactor, HK_WIDTH),
- Unroll("interleave.count", DisableUnrolling, HK_UNROLL),
+ Interleave("interleave.count", DisableInterleaving, HK_UNROLL),
Force("vectorize.enable", FK_Undefined, HK_FORCE),
TheLoop(L) {
- // FIXME: Move this up initialisation when MSVC requirement is 2013+
- Hints[0] = &Width;
- Hints[1] = &Unroll;
- Hints[2] = &Force;
-
// Populate values with existing loop metadata.
getHintsFromMetadata();
- // force-vector-unroll overrides DisableUnrolling.
- if (VectorizationUnroll.getNumOccurrences() > 0)
- Unroll.Value = VectorizationUnroll;
+ // force-vector-interleave overrides DisableInterleaving.
+ if (VectorizationInterleave.getNumOccurrences() > 0)
+ Interleave.Value = VectorizationInterleave;
- DEBUG(if (DisableUnrolling && Unroll.Value == 1) dbgs()
- << "LV: Unrolling disabled by the pass manager\n");
+ DEBUG(if (DisableInterleaving && Interleave.Value == 1) dbgs()
+ << "LV: Interleaving disabled by the pass manager\n");
}
/// Mark the loop L as already vectorized by setting the width to 1.
void setAlreadyVectorized() {
- Width.Value = Unroll.Value = 1;
- // FIXME: Change all lines below for this when we can use MSVC 2013+
- //writeHintsToMetadata({ Width, Unroll });
- std::vector<Hint> hints;
- hints.reserve(2);
- hints.emplace_back(Width);
- hints.emplace_back(Unroll);
- writeHintsToMetadata(std::move(hints));
+ Width.Value = Interleave.Value = 1;
+ Hint Hints[] = {Width, Interleave};
+ writeHintsToMetadata(Hints);
}
/// Dumps all the hint information.
R << " (Force=true";
if (Width.Value != 0)
R << ", Vector Width=" << Width.Value;
- if (Unroll.Value != 0)
- R << ", Interleave Count=" << Unroll.Value;
+ if (Interleave.Value != 0)
+ R << ", Interleave Count=" << Interleave.Value;
R << ")";
}
}
}
unsigned getWidth() const { return Width.Value; }
- unsigned getUnroll() const { return Unroll.Value; }
+ unsigned getInterleave() const { return Interleave.Value; }
enum ForceKind getForce() const { return (ForceKind)Force.Value; }
private:
for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) {
const MDString *S = nullptr;
- SmallVector<Value*, 4> Args;
+ SmallVector<Metadata *, 4> Args;
// The expected hint is either a MDString or a MDNode with the first
// operand a MDString.
}
/// Checks string hint with one operand and set value if valid.
- void setHint(StringRef Name, Value *Arg) {
+ void setHint(StringRef Name, Metadata *Arg) {
if (!Name.startswith(Prefix()))
return;
Name = Name.substr(Prefix().size(), StringRef::npos);
- const ConstantInt *C = dyn_cast<ConstantInt>(Arg);
+ const ConstantInt *C = mdconst::dyn_extract<ConstantInt>(Arg);
if (!C) return;
unsigned Val = C->getZExtValue();
+ Hint *Hints[] = {&Width, &Interleave, &Force};
for (auto H : Hints) {
if (Name == H->Name) {
if (H->validate(Val))
/// Create a new hint from name / value pair.
MDNode *createHintMetadata(StringRef Name, unsigned V) const {
LLVMContext &Context = TheLoop->getHeader()->getContext();
- SmallVector<Value*, 2> Vals;
- Vals.push_back(MDString::get(Context, Name));
- Vals.push_back(ConstantInt::get(Type::getInt32Ty(Context), V));
- return MDNode::get(Context, Vals);
+ Metadata *MDs[] = {MDString::get(Context, Name),
+ ConstantAsMetadata::get(
+ ConstantInt::get(Type::getInt32Ty(Context), V))};
+ return MDNode::get(Context, MDs);
}
/// Matches metadata with hint name.
- bool matchesHintMetadataName(MDNode *Node, std::vector<Hint> &HintTypes) {
+ bool matchesHintMetadataName(MDNode *Node, ArrayRef<Hint> HintTypes) {
MDString* Name = dyn_cast<MDString>(Node->getOperand(0));
if (!Name)
return false;
for (auto H : HintTypes)
- if (Name->getName().endswith(H.Name))
+ if (Name->getString().endswith(H.Name))
return true;
return false;
}
/// Sets current hints into loop metadata, keeping other values intact.
- void writeHintsToMetadata(std::vector<Hint> HintTypes) {
+ void writeHintsToMetadata(ArrayRef<Hint> HintTypes) {
if (HintTypes.size() == 0)
return;
// Reserve the first element to LoopID (see below).
- SmallVector<Value*, 4> Vals(1);
+ SmallVector<Metadata *, 4> MDs(1);
// If the loop already has metadata, then ignore the existing operands.
MDNode *LoopID = TheLoop->getLoopID();
if (LoopID) {
MDNode *Node = cast<MDNode>(LoopID->getOperand(i));
// If node in update list, ignore old value.
if (!matchesHintMetadataName(Node, HintTypes))
- Vals.push_back(Node);
+ MDs.push_back(Node);
}
}
// Now, add the missing hints.
for (auto H : HintTypes)
- Vals.push_back(
- createHintMetadata(Twine(Prefix(), H.Name).str(), H.Value));
+ MDs.push_back(createHintMetadata(Twine(Prefix(), H.Name).str(), H.Value));
// Replace current metadata node with new one.
LLVMContext &Context = TheLoop->getHeader()->getContext();
- MDNode *NewLoopID = MDNode::get(Context, Vals);
+ MDNode *NewLoopID = MDNode::get(Context, MDs);
// Set operand 0 to refer to the loop id itself.
NewLoopID->replaceOperandWith(0, NewLoopID);
TheLoop->setLoopID(NewLoopID);
- if (LoopID)
- LoopID->replaceAllUsesWith(NewLoopID);
LoopID = NewLoopID;
}
emitLoopVectorizeWarning(
F->getContext(), *F, L->getStartLoc(),
"failed explicitly specified loop vectorization");
- else if (LH.getUnroll() != 1)
+ else if (LH.getInterleave() != 1)
emitLoopInterleaveWarning(
F->getContext(), *F, L->getStartLoc(),
"failed explicitly specified loop interleaving");
BlockFrequencyInfo *BFI;
TargetLibraryInfo *TLI;
AliasAnalysis *AA;
+ AssumptionTracker *AT;
bool DisableUnrolling;
bool AlwaysVectorize;
BFI = &getAnalysis<BlockFrequencyInfo>();
TLI = getAnalysisIfAvailable<TargetLibraryInfo>();
AA = &getAnalysis<AliasAnalysis>();
+ AT = &getAnalysis<AssumptionTracker>();
// Compute some weights outside of the loop over the loops. Compute this
// using a BranchProbability to re-use its scaling math.
: (Hints.getForce() == LoopVectorizeHints::FK_Enabled
? "enabled"
: "?")) << " width=" << Hints.getWidth()
- << " unroll=" << Hints.getUnroll() << "\n");
+ << " unroll=" << Hints.getInterleave() << "\n");
// Function containing loop
Function *F = L->getHeader()->getParent();
return false;
}
- if (Hints.getWidth() == 1 && Hints.getUnroll() == 1) {
+ if (Hints.getWidth() == 1 && Hints.getInterleave() == 1) {
DEBUG(dbgs() << "LV: Not vectorizing: Disabled/already vectorized.\n");
emitOptimizationRemarkAnalysis(
F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(),
// Check the loop for a trip count threshold:
// do not vectorize loops with a tiny trip count.
- BasicBlock *Latch = L->getLoopLatch();
- const unsigned TC = SE->getSmallConstantTripCount(L, Latch);
+ const unsigned TC = SE->getSmallConstantTripCount(L);
if (TC > 0u && TC < TinyTripCountVectorThreshold) {
DEBUG(dbgs() << "LV: Found a loop with a very small trip count. "
<< "This loop is not worth vectorizing.");
}
// Check if it is legal to vectorize the loop.
- LoopVectorizationLegality LVL(L, SE, DL, DT, TLI, AA, F);
+ LoopVectorizationLegality LVL(L, SE, DL, DT, TLI, AA, F, TTI);
if (!LVL.canVectorize()) {
DEBUG(dbgs() << "LV: Not vectorizing: Cannot prove legality.\n");
emitMissedWarning(F, L, Hints);
}
// Use the cost model.
- LoopVectorizationCostModel CM(L, SE, LI, &LVL, *TTI, DL, TLI, F, &Hints);
+ LoopVectorizationCostModel CM(L, SE, LI, &LVL, *TTI, DL, TLI, AT, F,
+ &Hints);
// Check the function attributes to find out if this function should be
// optimized for size.
}
void getAnalysisUsage(AnalysisUsage &AU) const override {
+ AU.addRequired<AssumptionTracker>();
AU.addRequiredID(LoopSimplifyID);
AU.addRequiredID(LCSSAID);
AU.addRequired<BlockFrequencyInfo>();
unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ScalarDataTy);
unsigned VectorElementSize = DL->getTypeStoreSize(DataTy)/VF;
- if (SI && Legal->blockNeedsPredication(SI->getParent()))
+ if (SI && Legal->blockNeedsPredication(SI->getParent()) &&
+ !Legal->isMaskRequired(SI))
return scalarizeInstruction(Instr, true);
if (ScalarAllocatedSize != VectorElementSize)
Value *VecPtr = Builder.CreateBitCast(PartPtr,
DataTy->getPointerTo(AddressSpace));
- StoreInst *NewSI =
- Builder.CreateAlignedStore(StoredVal[Part], VecPtr, Alignment);
+
+ Instruction *NewSI;
+ if (Legal->isMaskRequired(SI)) {
+ Type *I8PtrTy =
+ Builder.getInt8PtrTy(PartPtr->getType()->getPointerAddressSpace());
+
+ Value *I8Ptr = Builder.CreateBitCast(PartPtr, I8PtrTy);
+
+ VectorParts Cond = createBlockInMask(SI->getParent());
+ SmallVector <Value *, 8> Ops;
+ Ops.push_back(I8Ptr);
+ Ops.push_back(StoredVal[Part]);
+ Ops.push_back(Builder.getInt32(Alignment));
+ Ops.push_back(Cond[Part]);
+ NewSI = Builder.CreateMaskedStore(Ops);
+ }
+ else
+ NewSI = Builder.CreateAlignedStore(StoredVal[Part], VecPtr, Alignment);
propagateMetadata(NewSI, SI);
}
return;
if (Reverse) {
// If the address is consecutive but reversed, then the
- // wide store needs to start at the last vector element.
+ // wide load needs to start at the last vector element.
PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(-Part * VF));
PartPtr = Builder.CreateGEP(PartPtr, Builder.getInt32(1 - VF));
}
- Value *VecPtr = Builder.CreateBitCast(PartPtr,
- DataTy->getPointerTo(AddressSpace));
- LoadInst *NewLI = Builder.CreateAlignedLoad(VecPtr, Alignment, "wide.load");
+ Instruction* NewLI;
+ if (Legal->isMaskRequired(LI)) {
+ Type *I8PtrTy =
+ Builder.getInt8PtrTy(PartPtr->getType()->getPointerAddressSpace());
+
+ Value *I8Ptr = Builder.CreateBitCast(PartPtr, I8PtrTy);
+
+ VectorParts SrcMask = createBlockInMask(LI->getParent());
+ SmallVector <Value *, 8> Ops;
+ Ops.push_back(I8Ptr);
+ Ops.push_back(UndefValue::get(DataTy));
+ Ops.push_back(Builder.getInt32(Alignment));
+ Ops.push_back(SrcMask[Part]);
+ NewLI = Builder.CreateMaskedLoad(Ops);
+ }
+ else {
+ Value *VecPtr = Builder.CreateBitCast(PartPtr,
+ DataTy->getPointerTo(AddressSpace));
+ NewLI = Builder.CreateAlignedLoad(VecPtr, Alignment, "wide.load");
+ }
propagateMetadata(NewLI, LI);
Entry[Part] = Reverse ? reverseVector(NewLI) : NewLI;
}
}
// Fix the vector-loop phi.
- // We created the induction variable so we know that the
- // preheader is the first entry.
- BasicBlock *VecPreheader = Induction->getIncomingBlock(0);
// Reductions do not have to start at zero. They can start with
// any loop invariant values.
// Make sure to add the reduction stat value only to the
// first unroll part.
Value *StartVal = (part == 0) ? VectorStart : Identity;
- cast<PHINode>(VecRdxPhi[part])->addIncoming(StartVal, VecPreheader);
+ cast<PHINode>(VecRdxPhi[part])->addIncoming(StartVal,
+ LoopVectorPreHeader);
cast<PHINode>(VecRdxPhi[part])->addIncoming(Val[part],
LoopVectorBody.back());
}
LCSSAPhi->addIncoming(UndefValue::get(LCSSAPhi->getType()),
LoopMiddleBlock);
}
-}
+}
InnerLoopVectorizer::VectorParts
InnerLoopVectorizer::createEdgeMask(BasicBlock *Src, BasicBlock *Dst) {
if (BinaryOperator *VecOp = dyn_cast<BinaryOperator>(V))
VecOp->copyIRFlags(BinOp);
-
+
Entry[Part] = V;
}
Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
assert(ID && "Not an intrinsic call!");
switch (ID) {
+ case Intrinsic::assume:
case Intrinsic::lifetime_end:
case Intrinsic::lifetime_start:
scalarizeInstruction(it);
DT->addNewBlock(LoopMiddleBlock, LoopBypassBlocks[1]);
DT->addNewBlock(LoopScalarPreHeader, LoopBypassBlocks[0]);
DT->changeImmediateDominator(LoopScalarBody, LoopScalarPreHeader);
- DT->changeImmediateDominator(LoopExitBlock, LoopMiddleBlock);
+ DT->changeImmediateDominator(LoopExitBlock, LoopBypassBlocks[0]);
DEBUG(DT->verifyDomTree());
}
return false;
}
+ // We only handle bottom-tested loops, i.e. loop in which the condition is
+ // checked at the end of each iteration. With that we can assume that all
+ // instructions in the loop are executed the same number of times.
+ if (TheLoop->getExitingBlock() != TheLoop->getLoopLatch()) {
+ emitAnalysis(
+ Report() << "loop control flow is not understood by vectorizer");
+ return false;
+ }
+
// We need to have a loop header.
DEBUG(dbgs() << "LV: Found a loop: " <<
TheLoop->getHeader()->getName() << '\n');
if (IsWrite)
SetHasWrite = true;
- // Create sets of pointers connected by a shared alias set and
- // underlying object.
- typedef SmallVector<Value*, 16> ValueVector;
+ // Create sets of pointers connected by a shared alias set and
+ // underlying object.
+ typedef SmallVector<Value *, 16> ValueVector;
ValueVector TempObjects;
GetUnderlyingObjects(Ptr, TempObjects, DL);
for (Value *UnderlyingObj : TempObjects) {
// Bail out early if passed-in parameters make vectorization not feasible.
unsigned ForcedFactor = VectorizationFactor ? VectorizationFactor : 1;
- unsigned ForcedUnroll = VectorizationUnroll ? VectorizationUnroll : 1;
+ unsigned ForcedUnroll = VectorizationInterleave ? VectorizationInterleave : 1;
// The distance must be bigger than the size needed for a vectorized version
// of the operation and the size of the vectorized operation must not be
// If we did *not* see this pointer before, insert it to the read-write
// list. At this phase it is only a 'write' list.
- if (Seen.insert(Ptr)) {
+ if (Seen.insert(Ptr).second) {
++NumReadWrites;
AliasAnalysis::Location Loc = AA->getLocation(ST);
// read a few words, modify, and write a few words, and some of the
// words may be written to the same address.
bool IsReadOnlyPtr = false;
- if (Seen.insert(Ptr) || !isStridedPtr(SE, DL, Ptr, TheLoop, Strides)) {
+ if (Seen.insert(Ptr).second ||
+ !isStridedPtr(SE, DL, Ptr, TheLoop, Strides)) {
++NumReads;
IsReadOnlyPtr = true;
}
// value must only be used once, except by phi nodes and min/max
// reductions which are represented as a cmp followed by a select.
ReductionInstDesc IgnoredVal(false, nullptr);
- if (VisitedInsts.insert(UI)) {
+ if (VisitedInsts.insert(UI).second) {
if (isa<PHINode>(UI))
PHIs.push_back(UI);
else
return IK_NoInduction;
assert(PhiTy->isPointerTy() && "The PHI must be a pointer");
- uint64_t Size = DL->getTypeAllocSize(PhiTy->getPointerElementType());
+ Type *PointerElementType = PhiTy->getPointerElementType();
+ // The pointer stride cannot be determined if the pointer element type is not
+ // sized.
+ if (!PointerElementType->isSized())
+ return IK_NoInduction;
+
+ uint64_t Size = DL->getTypeAllocSize(PointerElementType);
if (C->getValue()->equalsInt(Size))
return IK_PtrInduction;
else if (C->getValue()->equalsInt(0 - Size))
bool LoopVectorizationLegality::blockCanBePredicated(BasicBlock *BB,
SmallPtrSetImpl<Value *> &SafePtrs) {
+
for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) {
+ // Check that we don't have a constant expression that can trap as operand.
+ for (Instruction::op_iterator OI = it->op_begin(), OE = it->op_end();
+ OI != OE; ++OI) {
+ if (Constant *C = dyn_cast<Constant>(*OI))
+ if (C->canTrap())
+ return false;
+ }
// We might be able to hoist the load.
if (it->mayReadFromMemory()) {
LoadInst *LI = dyn_cast<LoadInst>(it);
- if (!LI || !SafePtrs.count(LI->getPointerOperand()))
+ if (!LI)
+ return false;
+ if (!SafePtrs.count(LI->getPointerOperand())) {
+ if (isLegalMaskedLoad(LI->getType(), LI->getPointerOperand())) {
+ MaskedOp.insert(LI);
+ continue;
+ }
return false;
+ }
}
// We don't predicate stores at the moment.
StoreInst *SI = dyn_cast<StoreInst>(it);
// We only support predication of stores in basic blocks with one
// predecessor.
- if (!SI || ++NumPredStores > NumberOfStoresToPredicate ||
- !SafePtrs.count(SI->getPointerOperand()) ||
- !SI->getParent()->getSinglePredecessor())
+ if (!SI)
return false;
+
+ bool isSafePtr = (SafePtrs.count(SI->getPointerOperand()) != 0);
+ bool isSinglePredecessor = SI->getParent()->getSinglePredecessor();
+
+ if (++NumPredStores > NumberOfStoresToPredicate || !isSafePtr ||
+ !isSinglePredecessor) {
+ // Build a masked store if it is legal for the target, otherwise scalarize
+ // the block.
+ bool isLegalMaskedOp =
+ isLegalMaskedStore(SI->getValueOperand()->getType(),
+ SI->getPointerOperand());
+ if (isLegalMaskedOp) {
+ --NumPredStores;
+ MaskedOp.insert(SI);
+ continue;
+ }
+ return false;
+ }
}
if (it->mayThrow())
return false;
- // Check that we don't have a constant expression that can trap as operand.
- for (Instruction::op_iterator OI = it->op_begin(), OE = it->op_end();
- OI != OE; ++OI) {
- if (Constant *C = dyn_cast<Constant>(*OI))
- if (C->canTrap())
- return false;
- }
-
// The instructions below can trap.
switch (it->getOpcode()) {
default: continue;
case Instruction::SDiv:
case Instruction::URem:
case Instruction::SRem:
- return false;
+ return false;
}
}
}
// Find the trip count.
- unsigned TC = SE->getSmallConstantTripCount(TheLoop, TheLoop->getLoopLatch());
+ unsigned TC = SE->getSmallConstantTripCount(TheLoop);
DEBUG(dbgs() << "LV: Found trip count: " << TC << '\n');
unsigned WidestType = getWidestType();
MaxVectorSize = 1;
}
- assert(MaxVectorSize <= 32 && "Did not expect to pack so many elements"
+ assert(MaxVectorSize <= 64 && "Did not expect to pack so many elements"
" into one vector!");
unsigned VF = MaxVectorSize;
// If the trip count that we found modulo the vectorization factor is not
// zero then we require a tail.
if (VF < 2) {
- emitAnalysis(Report() << "cannot optimize for size and vectorize at the same time. Enable vectorization of this loop with '#pragma clang loop vectorize(enable)' when compiling with -Os");
+ emitAnalysis(Report() << "cannot optimize for size and vectorize at the "
+ "same time. Enable vectorization of this loop "
+ "with '#pragma clang loop vectorize(enable)' "
+ "when compiling with -Os");
DEBUG(dbgs() << "LV: Aborting. A tail loop is required in Os.\n");
return Factor;
}
for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) {
Type *T = it->getType();
+ // Ignore ephemeral values.
+ if (EphValues.count(it))
+ continue;
+
// Only examine Loads, Stores and PHINodes.
if (!isa<LoadInst>(it) && !isa<StoreInst>(it) && !isa<PHINode>(it))
continue;
// to the increased register pressure.
// Use the user preference, unless 'auto' is selected.
- int UserUF = Hints->getUnroll();
+ int UserUF = Hints->getInterleave();
if (UserUF != 0)
return UserUF;
return 1;
// Do not unroll loops with a relatively small trip count.
- unsigned TC = SE->getSmallConstantTripCount(TheLoop,
- TheLoop->getLoopLatch());
+ unsigned TC = SE->getSmallConstantTripCount(TheLoop);
if (TC > 1 && TC < TinyTripCountUnrollThreshold)
return 1;
std::max(1U, (R.MaxLocalUsers - 1)));
// Clamp the unroll factor ranges to reasonable factors.
- unsigned MaxUnrollSize = TTI.getMaximumUnrollFactor();
+ unsigned MaxInterleaveSize = TTI.getMaxInterleaveFactor();
// Check if the user has overridden the unroll max.
if (VF == 1) {
- if (ForceTargetMaxScalarUnrollFactor.getNumOccurrences() > 0)
- MaxUnrollSize = ForceTargetMaxScalarUnrollFactor;
+ if (ForceTargetMaxScalarInterleaveFactor.getNumOccurrences() > 0)
+ MaxInterleaveSize = ForceTargetMaxScalarInterleaveFactor;
} else {
- if (ForceTargetMaxVectorUnrollFactor.getNumOccurrences() > 0)
- MaxUnrollSize = ForceTargetMaxVectorUnrollFactor;
+ if (ForceTargetMaxVectorInterleaveFactor.getNumOccurrences() > 0)
+ MaxInterleaveSize = ForceTargetMaxVectorInterleaveFactor;
}
// If we did not calculate the cost for VF (because the user selected the VF)
// Clamp the calculated UF to be between the 1 and the max unroll factor
// that the target allows.
- if (UF > MaxUnrollSize)
- UF = MaxUnrollSize;
+ if (UF > MaxInterleaveSize)
+ UF = MaxInterleaveSize;
else if (UF < 1)
UF = 1;
// Ignore instructions that are never used within the loop.
if (!Ends.count(I)) continue;
+ // Ignore ephemeral values.
+ if (EphValues.count(I))
+ continue;
+
// Remove all of the instructions that end at this location.
InstrList &List = TransposeEnds[i];
for (unsigned int j=0, e = List.size(); j < e; ++j)
if (isa<DbgInfoIntrinsic>(it))
continue;
+ // Ignore ephemeral values.
+ if (EphValues.count(it))
+ continue;
+
unsigned C = getInstructionCost(it, VF);
// Check if we should override the cost.
INITIALIZE_PASS_BEGIN(LoopVectorize, LV_NAME, lv_name, false, false)
INITIALIZE_AG_DEPENDENCY(TargetTransformInfo)
INITIALIZE_AG_DEPENDENCY(AliasAnalysis)
+INITIALIZE_PASS_DEPENDENCY(AssumptionTracker)
INITIALIZE_PASS_DEPENDENCY(BlockFrequencyInfo)
INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
INITIALIZE_PASS_DEPENDENCY(ScalarEvolution)