#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/BlockFrequencyInfo.h"
#include "llvm/Analysis/CodeMetrics.h"
+#include "llvm/Analysis/LoopAccessAnalysis.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/LoopIterator.h"
#include "llvm/Analysis/LoopPass.h"
#include "llvm/Transforms/Utils/BasicBlockUtils.h"
#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/VectorUtils.h"
+#include "llvm/Transforms/Utils/LoopUtils.h"
#include <algorithm>
#include <map>
#include <tuple>
STATISTIC(LoopsVectorized, "Number of loops vectorized");
STATISTIC(LoopsAnalyzed, "Number of loops analyzed for vectorization");
-static cl::opt<unsigned>
-VectorizationFactor("force-vector-width", cl::init(0), cl::Hidden,
- cl::desc("Sets the SIMD width. Zero is autoselect."));
-
-static cl::opt<unsigned>
-VectorizationInterleave("force-vector-interleave", cl::init(0), cl::Hidden,
- cl::desc("Sets the vectorization interleave count. "
- "Zero is autoselect."));
-
static cl::opt<bool>
EnableIfConversion("enable-if-conversion", cl::init(true), cl::Hidden,
cl::desc("Enable if-conversion during vectorization."));
/// We don't unroll loops with a known constant trip count below this number.
static const unsigned TinyTripCountUnrollThreshold = 128;
-/// When performing memory disambiguation checks at runtime do not make more
-/// than this number of comparisons.
-static const unsigned RuntimeMemoryCheckThreshold = 8;
-
-/// Maximum simd width.
-static const unsigned MaxVectorWidth = 64;
-
static cl::opt<unsigned> ForceTargetNumScalarRegs(
"force-target-num-scalar-regs", cl::init(0), cl::Hidden,
cl::desc("A flag that overrides the target's number of scalar registers."));
class LoopVectorizationCostModel;
class LoopVectorizeHints;
-/// Optimization analysis message produced during vectorization. Messages inform
-/// the user why vectorization did not occur.
-class VectorizationReport {
- std::string Message;
- raw_string_ostream Out;
- Instruction *Instr;
-
+/// \brief This modifies LoopAccessReport to initialize message with
+/// loop-vectorizer-specific part.
+class VectorizationReport : public LoopAccessReport {
public:
- VectorizationReport(Instruction *I = nullptr) : Out(Message), Instr(I) {
- Out << "loop not vectorized: ";
- }
-
- template <typename A> VectorizationReport &operator<<(const A &Value) {
- Out << Value;
- return *this;
- }
-
- Instruction *getInstr() { return Instr; }
-
- std::string &str() { return Out.str(); }
- operator Twine() { return Out.str(); }
-
- /// \brief Emit an analysis note with the debug location from the instruction
- /// in \p Message if available. Otherwise use the location of \p TheLoop.
- static void emitAnalysis(VectorizationReport &Message,
- const Function *TheFunction,
- const Loop *TheLoop);
+ VectorizationReport(Instruction *I = nullptr)
+ : LoopAccessReport("loop not vectorized: ", I) {}
+
+ /// \brief This allows promotion of the loop-access analysis report into the
+ /// loop-vectorizer report. It modifies the message to add the
+ /// loop-vectorizer-specific part of the message.
+ explicit VectorizationReport(const LoopAccessReport &R)
+ : LoopAccessReport(Twine("loop not vectorized: ") + R.str(),
+ R.getInstr()) {}
};
+/// A helper function for converting Scalar types to vector types.
+/// If the incoming type is void, we return void. If the VF is 1, we return
+/// the scalar type.
+static Type* ToVectorTy(Type *Scalar, unsigned VF) {
+ if (Scalar->isVoidTy() || VF == 1)
+ return Scalar;
+ return VectorType::get(Scalar, VF);
+}
+
/// InnerLoopVectorizer vectorizes loops which contain only one basic
/// block to a specified vectorization factor (VF).
/// This class performs the widening of scalars into vectors, or multiple
class InnerLoopVectorizer {
public:
InnerLoopVectorizer(Loop *OrigLoop, ScalarEvolution *SE, LoopInfo *LI,
- DominatorTree *DT, const DataLayout *DL,
- const TargetLibraryInfo *TLI, unsigned VecWidth,
+ DominatorTree *DT, const TargetLibraryInfo *TLI,
+ const TargetTransformInfo *TTI, unsigned VecWidth,
unsigned UnrollFactor)
- : OrigLoop(OrigLoop), SE(SE), LI(LI), DT(DT), DL(DL), TLI(TLI),
+ : OrigLoop(OrigLoop), SE(SE), LI(LI), DT(DT), TLI(TLI), TTI(TTI),
VF(VecWidth), UF(UnrollFactor), Builder(SE->getContext()),
Induction(nullptr), OldInduction(nullptr), WidenMap(UnrollFactor),
- Legal(nullptr) {}
+ Legal(nullptr), AddedSafetyChecks(false) {}
// Perform the actual loop widening (vectorization).
void vectorize(LoopVectorizationLegality *L) {
updateAnalysis();
}
+ // Return true if any runtime check is added.
+ bool IsSafetyChecksAdded() {
+ return AddedSafetyChecks;
+ }
+
virtual ~InnerLoopVectorizer() {}
protected:
typedef DenseMap<std::pair<BasicBlock*, BasicBlock*>,
VectorParts> EdgeMaskCache;
- /// \brief Add code that checks at runtime if the accessed arrays overlap.
- ///
- /// Returns a pair of instructions where the first element is the first
- /// instruction generated in possibly a sequence of instructions and the
- /// second value is the final comparator value or NULL if no check is needed.
- std::pair<Instruction *, Instruction *> addRuntimeCheck(Instruction *Loc);
-
/// \brief Add checks for strides that where assumed to be 1.
///
/// Returns the last check instruction and the first check instruction in the
DominatorTree *DT;
/// Alias Analysis.
AliasAnalysis *AA;
- /// Data Layout.
- const DataLayout *DL;
/// Target Library Info.
const TargetLibraryInfo *TLI;
+ /// Target Transform Info.
+ const TargetTransformInfo *TTI;
/// The vectorization SIMD factor to use. Each vector will have this many
/// vector elements.
EdgeMaskCache MaskCache;
LoopVectorizationLegality *Legal;
+
+ // Record whether runtime check is added.
+ bool AddedSafetyChecks;
};
class InnerLoopUnroller : public InnerLoopVectorizer {
public:
InnerLoopUnroller(Loop *OrigLoop, ScalarEvolution *SE, LoopInfo *LI,
- DominatorTree *DT, const DataLayout *DL,
- const TargetLibraryInfo *TLI, unsigned UnrollFactor) :
- InnerLoopVectorizer(OrigLoop, SE, LI, DT, DL, TLI, 1, UnrollFactor) { }
+ DominatorTree *DT, const TargetLibraryInfo *TLI,
+ const TargetTransformInfo *TTI, unsigned UnrollFactor)
+ : InnerLoopVectorizer(OrigLoop, SE, LI, DT, TLI, TTI, 1, UnrollFactor) {}
private:
void scalarizeInstruction(Instruction *Instr,
std::string Result;
if (L) {
raw_string_ostream OS(Result);
- const DebugLoc LoopDbgLoc = L->getStartLoc();
- if (!LoopDbgLoc.isUnknown())
- LoopDbgLoc.print(L->getHeader()->getContext(), OS);
+ if (const DebugLoc LoopDbgLoc = L->getStartLoc())
+ LoopDbgLoc.print(OS);
else
// Just print the module name.
OS << L->getHeader()->getParent()->getParent()->getModuleIdentifier();
}
}
-void VectorizationReport::emitAnalysis(VectorizationReport &Message,
- const Function *TheFunction,
- const Loop *TheLoop) {
- DebugLoc DL = TheLoop->getStartLoc();
- if (Instruction *I = Message.getInstr())
- DL = I->getDebugLoc();
- emitOptimizationRemarkAnalysis(TheFunction->getContext(), DEBUG_TYPE,
- *TheFunction, DL, Message.str());
-}
-
/// \brief Propagate known metadata from one instruction to a vector of others.
static void propagateMetadata(SmallVectorImpl<Value *> &To, const Instruction *From) {
for (Value *V : To)
propagateMetadata(I, From);
}
-namespace {
-/// This struct holds information about the memory runtime legality
-/// check that a group of pointers do not overlap.
-struct RuntimePointerCheck {
- RuntimePointerCheck() : Need(false) {}
-
- /// Reset the state of the pointer runtime information.
- void reset() {
- Need = false;
- Pointers.clear();
- Starts.clear();
- Ends.clear();
- IsWritePtr.clear();
- DependencySetId.clear();
- AliasSetId.clear();
- }
-
- /// Insert a pointer and calculate the start and end SCEVs.
- void insert(ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr,
- unsigned DepSetId, unsigned ASId, ValueToValueMap &Strides);
-
- /// This flag indicates if we need to add the runtime check.
- bool Need;
- /// Holds the pointers that we need to check.
- SmallVector<TrackingVH<Value>, 2> Pointers;
- /// Holds the pointer value at the beginning of the loop.
- SmallVector<const SCEV*, 2> Starts;
- /// Holds the pointer value at the end of the loop.
- SmallVector<const SCEV*, 2> Ends;
- /// Holds the information if this pointer is used for writing to memory.
- SmallVector<bool, 2> IsWritePtr;
- /// Holds the id of the set of pointers that could be dependent because of a
- /// shared underlying object.
- SmallVector<unsigned, 2> DependencySetId;
- /// Holds the id of the disjoint alias set to which this pointer belongs.
- SmallVector<unsigned, 2> AliasSetId;
-};
-
-/// \brief Drive the analysis of memory accesses in the loop
-///
-/// This class is responsible for analyzing the memory accesses of a loop. It
-/// collects the accesses and then its main helper the AccessAnalysis class
-/// finds and categorizes the dependences in buildDependenceSets.
-///
-/// For memory dependences that can be analyzed at compile time, it determines
-/// whether the dependence is part of cycle inhibiting vectorization. This work
-/// is delegated to the MemoryDepChecker class.
-///
-/// For memory dependences that cannot be determined at compile time, it
-/// generates run-time checks to prove independence. This is done by
-/// AccessAnalysis::canCheckPtrAtRT and the checks are maintained by the
-/// RuntimePointerCheck class.
-class LoopAccessAnalysis {
-public:
- /// \brief Collection of parameters used from the vectorizer.
- struct VectorizerParams {
- /// \brief Maximum simd width.
- unsigned MaxVectorWidth;
-
- /// \brief VF as overridden by the user.
- unsigned VectorizationFactor;
- /// \brief Interleave factor as overridden by the user.
- unsigned VectorizationInterleave;
-
- /// \\brief When performing memory disambiguation checks at runtime do not
- /// make more than this number of comparisons.
- unsigned RuntimeMemoryCheckThreshold;
-
- VectorizerParams(unsigned MaxVectorWidth,
- unsigned VectorizationFactor,
- unsigned VectorizationInterleave,
- unsigned RuntimeMemoryCheckThreshold) :
- MaxVectorWidth(MaxVectorWidth),
- VectorizationFactor(VectorizationFactor),
- VectorizationInterleave(VectorizationInterleave),
- RuntimeMemoryCheckThreshold(RuntimeMemoryCheckThreshold) {}
- };
-
- LoopAccessAnalysis(Function *F, Loop *L, ScalarEvolution *SE,
- const DataLayout *DL, const TargetLibraryInfo *TLI,
- AliasAnalysis *AA, DominatorTree *DT,
- const VectorizerParams &VectParams) :
- TheFunction(F), TheLoop(L), SE(SE), DL(DL), TLI(TLI), AA(AA), DT(DT),
- NumLoads(0), NumStores(0), MaxSafeDepDistBytes(-1U),
- VectParams(VectParams) {}
-
- /// Return true we can analyze the memory accesses in the loop and there are
- /// no memory dependence cycles. Replaces symbolic strides using Strides.
- bool canVectorizeMemory(ValueToValueMap &Strides);
-
- RuntimePointerCheck *getRuntimePointerCheck() { return &PtrRtCheck; }
-
- /// Return true if the block BB needs to be predicated in order for the loop
- /// to be vectorized.
- bool blockNeedsPredication(BasicBlock *BB);
-
- /// Returns true if the value V is uniform within the loop.
- bool isUniform(Value *V);
-
- unsigned getMaxSafeDepDistBytes() { return MaxSafeDepDistBytes; }
-
-private:
- void emitAnalysis(Report &Message) {
- emitLoopAnalysis(Message, TheFunction, TheLoop);
- }
-
- /// We need to check that all of the pointers in this list are disjoint
- /// at runtime.
- RuntimePointerCheck PtrRtCheck;
- Function *TheFunction;
- Loop *TheLoop;
- ScalarEvolution *SE;
- const DataLayout *DL;
- const TargetLibraryInfo *TLI;
- AliasAnalysis *AA;
- DominatorTree *DT;
-
- unsigned NumLoads;
- unsigned NumStores;
-
- unsigned MaxSafeDepDistBytes;
-
- /// \brief Vectorizer parameters used by the analysis.
- VectorizerParams VectParams;
-};
-} // end anonymous namespace
-
/// LoopVectorizationLegality checks if it is legal to vectorize a loop, and
/// to what vectorization factor.
/// This class does not look at the profitability of vectorization, only the
/// induction variable and the different reduction variables.
class LoopVectorizationLegality {
public:
- LoopVectorizationLegality(Loop *L, ScalarEvolution *SE, const DataLayout *DL,
- DominatorTree *DT, TargetLibraryInfo *TLI,
- AliasAnalysis *AA, Function *F,
- const TargetTransformInfo *TTI)
- : NumPredStores(0), TheLoop(L), SE(SE), DL(DL), TLI(TLI), TheFunction(F),
- TTI(TTI), Induction(nullptr), WidestIndTy(nullptr),
- LAA(F, L, SE, DL, TLI, AA, DT,
- {MaxVectorWidth, VectorizationFactor, VectorizationInterleave,
- RuntimeMemoryCheckThreshold}),
- HasFunNoNaNAttr(false) {
- }
+ LoopVectorizationLegality(Loop *L, ScalarEvolution *SE, DominatorTree *DT,
+ TargetLibraryInfo *TLI, AliasAnalysis *AA,
+ Function *F, const TargetTransformInfo *TTI,
+ LoopAccessAnalysis *LAA)
+ : NumPredStores(0), TheLoop(L), SE(SE), TLI(TLI), TheFunction(F),
+ TTI(TTI), DT(DT), LAA(LAA), LAI(nullptr), Induction(nullptr),
+ WidestIndTy(nullptr), HasFunNoNaNAttr(false) {}
/// This enum represents the kinds of reductions that we support.
enum ReductionKind {
Index = B.CreateNeg(Index);
else if (!StepValue->isOne())
Index = B.CreateMul(Index, StepValue);
- return B.CreateGEP(StartValue, Index);
+ return B.CreateGEP(nullptr, StartValue, Index);
case IK_NoInduction:
return nullptr;
bool isUniformAfterVectorization(Instruction* I) { return Uniforms.count(I); }
/// Returns the information that we collected about runtime memory check.
- RuntimePointerCheck *getRuntimePointerCheck() {
- return LAA.getRuntimePointerCheck();
+ const LoopAccessInfo::RuntimePointerCheck *getRuntimePointerCheck() const {
+ return LAI->getRuntimePointerCheck();
+ }
+
+ const LoopAccessInfo *getLAI() const {
+ return LAI;
}
/// This function returns the identity element (or neutral element) for
/// the operation K.
static Constant *getReductionIdentity(ReductionKind K, Type *Tp);
- unsigned getMaxSafeDepDistBytes() { return LAA.getMaxSafeDepDistBytes(); }
+ unsigned getMaxSafeDepDistBytes() { return LAI->getMaxSafeDepDistBytes(); }
bool hasStride(Value *V) { return StrideSet.count(V); }
bool mustCheckStrides() { return !StrideSet.empty(); }
return (MaskedOp.count(I) != 0);
}
unsigned getNumStores() const {
- return NumStores;
+ return LAI->getNumStores();
}
unsigned getNumLoads() const {
- return NumLoads;
+ return LAI->getNumLoads();
}
unsigned getNumPredStores() const {
return NumPredStores;
void collectStridedAccess(Value *LoadOrStoreInst);
/// Report an analysis message to assist the user in diagnosing loops that are
- /// not vectorized.
- void emitAnalysis(VectorizationReport &Message) {
- VectorizationReport::emitAnalysis(Message, TheFunction, TheLoop);
+ /// not vectorized. These are handled as LoopAccessReport rather than
+ /// VectorizationReport because the << operator of VectorizationReport returns
+ /// LoopAccessReport.
+ void emitAnalysis(const LoopAccessReport &Message) {
+ LoopAccessReport::emitAnalysis(Message, TheFunction, TheLoop, LV_NAME);
}
- unsigned NumLoads;
- unsigned NumStores;
unsigned NumPredStores;
/// The loop that we evaluate.
Loop *TheLoop;
/// Scev analysis.
ScalarEvolution *SE;
- /// DataLayout analysis.
- const DataLayout *DL;
/// Target Library Info.
TargetLibraryInfo *TLI;
/// Parent function
Function *TheFunction;
/// Target Transform Info
const TargetTransformInfo *TTI;
+ /// Dominator Tree.
+ DominatorTree *DT;
+ // LoopAccess analysis.
+ LoopAccessAnalysis *LAA;
+ // And the loop-accesses info corresponding to this loop. This pointer is
+ // null until canVectorizeMemory sets it up.
+ const LoopAccessInfo *LAI;
// --- vectorization state --- //
/// This set holds the variables which are known to be uniform after
/// vectorization.
SmallPtrSet<Instruction*, 4> Uniforms;
- LoopAccessAnalysis LAA;
+
/// Can we assume the absence of NaNs.
bool HasFunNoNaNAttr;
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(Loop *L, ScalarEvolution *SE, LoopInfo *LI,
LoopVectorizationLegality *Legal,
const TargetTransformInfo &TTI,
- const DataLayout *DL, const TargetLibraryInfo *TLI,
- AssumptionCache *AC, const Function *F,
- const LoopVectorizeHints *Hints)
- : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), DL(DL), TLI(TLI),
+ const TargetLibraryInfo *TLI, AssumptionCache *AC,
+ const Function *F, const LoopVectorizeHints *Hints)
+ : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), TLI(TLI),
TheFunction(F), Hints(Hints) {
CodeMetrics::collectEphemeralValues(L, AC, EphValues);
}
/// width. Vector width of one means scalar.
unsigned getInstructionCost(Instruction *I, unsigned VF);
- /// A helper function for converting Scalar types to vector types.
- /// If the incoming type is void, we return void. If the VF is 1, we return
- /// the scalar type.
- static Type* ToVectorTy(Type *Scalar, unsigned VF);
-
/// Returns whether the instruction is a load or store and will be a emitted
/// as a vector operation.
bool isConsecutiveLoadOrStore(Instruction *I);
/// Report an analysis message to assist the user in diagnosing loops that are
- /// not vectorized.
- void emitAnalysis(VectorizationReport &Message) {
- VectorizationReport::emitAnalysis(Message, TheFunction, TheLoop);
+ /// not vectorized. These are handled as LoopAccessReport rather than
+ /// VectorizationReport because the << operator of VectorizationReport returns
+ /// LoopAccessReport.
+ void emitAnalysis(const LoopAccessReport &Message) {
+ LoopAccessReport::emitAnalysis(Message, TheFunction, TheLoop, LV_NAME);
}
/// Values used only by @llvm.assume calls.
LoopVectorizationLegality *Legal;
/// Vector target information.
const TargetTransformInfo &TTI;
- /// Target data layout information.
- const DataLayout *DL;
/// Target Library Info.
const TargetLibraryInfo *TLI;
const Function *TheFunction;
bool validate(unsigned Val) {
switch (Kind) {
case HK_WIDTH:
- return isPowerOf2_32(Val) && Val <= MaxVectorWidth;
+ return isPowerOf2_32(Val) && Val <= VectorizerParams::MaxVectorWidth;
case HK_UNROLL:
return isPowerOf2_32(Val) && Val <= MaxInterleaveFactor;
case HK_FORCE:
};
LoopVectorizeHints(const Loop *L, bool DisableInterleaving)
- : Width("vectorize.width", VectorizationFactor, HK_WIDTH),
+ : Width("vectorize.width", VectorizerParams::VectorizationFactor,
+ HK_WIDTH),
Interleave("interleave.count", DisableInterleaving, HK_UNROLL),
Force("vectorize.enable", FK_Undefined, HK_FORCE),
TheLoop(L) {
getHintsFromMetadata();
// force-vector-interleave overrides DisableInterleaving.
- if (VectorizationInterleave.getNumOccurrences() > 0)
- Interleave.Value = VectorizationInterleave;
+ if (VectorizerParams::isInterleaveForced())
+ Interleave.Value = VectorizerParams::VectorizationInterleave;
DEBUG(if (DisableInterleaving && Interleave.Value == 1) dbgs()
<< "LV: Interleaving disabled by the pass manager\n");
}
ScalarEvolution *SE;
- const DataLayout *DL;
LoopInfo *LI;
TargetTransformInfo *TTI;
DominatorTree *DT;
TargetLibraryInfo *TLI;
AliasAnalysis *AA;
AssumptionCache *AC;
+ LoopAccessAnalysis *LAA;
bool DisableUnrolling;
bool AlwaysVectorize;
bool runOnFunction(Function &F) override {
SE = &getAnalysis<ScalarEvolution>();
- DataLayoutPass *DLP = getAnalysisIfAvailable<DataLayoutPass>();
- DL = DLP ? &DLP->getDataLayout() : nullptr;
LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
TTI = &getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
TLI = TLIP ? &TLIP->getTLI() : nullptr;
AA = &getAnalysis<AliasAnalysis>();
AC = &getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
+ LAA = &getAnalysis<LoopAccessAnalysis>();
// Compute some weights outside of the loop over the loops. Compute this
// using a BranchProbability to re-use its scaling math.
if (!TTI->getNumberOfRegisters(true))
return false;
- if (!DL) {
- DEBUG(dbgs() << "\nLV: Not vectorizing " << F.getName()
- << ": Missing data layout\n");
- return false;
- }
-
// Build up a worklist of inner-loops to vectorize. This is necessary as
// the act of vectorizing or partially unrolling a loop creates new loops
// and can invalidate iterators across the loops.
return Changed;
}
+ static void AddRuntimeUnrollDisableMetaData(Loop *L) {
+ SmallVector<Metadata *, 4> MDs;
+ // Reserve first location for self reference to the LoopID metadata node.
+ MDs.push_back(nullptr);
+ bool IsUnrollMetadata = false;
+ MDNode *LoopID = L->getLoopID();
+ if (LoopID) {
+ // First find existing loop unrolling disable metadata.
+ for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) {
+ MDNode *MD = dyn_cast<MDNode>(LoopID->getOperand(i));
+ if (MD) {
+ const MDString *S = dyn_cast<MDString>(MD->getOperand(0));
+ IsUnrollMetadata =
+ S && S->getString().startswith("llvm.loop.unroll.disable");
+ }
+ MDs.push_back(LoopID->getOperand(i));
+ }
+ }
+
+ if (!IsUnrollMetadata) {
+ // Add runtime unroll disable metadata.
+ LLVMContext &Context = L->getHeader()->getContext();
+ SmallVector<Metadata *, 1> DisableOperands;
+ DisableOperands.push_back(
+ MDString::get(Context, "llvm.loop.unroll.runtime.disable"));
+ MDNode *DisableNode = MDNode::get(Context, DisableOperands);
+ MDs.push_back(DisableNode);
+ MDNode *NewLoopID = MDNode::get(Context, MDs);
+ // Set operand 0 to refer to the loop id itself.
+ NewLoopID->replaceOperandWith(0, NewLoopID);
+ L->setLoopID(NewLoopID);
+ }
+ }
+
bool processLoop(Loop *L) {
assert(L->empty() && "Only process inner loops.");
}
// Check if it is legal to vectorize the loop.
- LoopVectorizationLegality LVL(L, SE, DL, DT, TLI, AA, F, TTI);
+ LoopVectorizationLegality LVL(L, SE, DT, TLI, AA, F, TTI, LAA);
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, AC, F,
- &Hints);
+ LoopVectorizationCostModel CM(L, SE, LI, &LVL, *TTI, TLI, AC, F, &Hints);
// Check the function attributes to find out if this function should be
// optimized for size.
// We decided not to vectorize, but we may want to unroll.
- InnerLoopUnroller Unroller(L, SE, LI, DT, DL, TLI, UF);
+ InnerLoopUnroller Unroller(L, SE, LI, DT, TLI, TTI, UF);
Unroller.vectorize(&LVL);
} else {
// If we decided that it is *legal* to vectorize the loop then do it.
- InnerLoopVectorizer LB(L, SE, LI, DT, DL, TLI, VF.Width, UF);
+ InnerLoopVectorizer LB(L, SE, LI, DT, TLI, TTI, VF.Width, UF);
LB.vectorize(&LVL);
++LoopsVectorized;
+ // Add metadata to disable runtime unrolling scalar loop when there's no
+ // runtime check about strides and memory. Because at this situation,
+ // scalar loop is rarely used not worthy to be unrolled.
+ if (!LB.IsSafetyChecksAdded())
+ AddRuntimeUnrollDisableMetaData(L);
+
// Report the vectorization decision.
emitOptimizationRemark(
F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(),
AU.addRequired<ScalarEvolution>();
AU.addRequired<TargetTransformInfoWrapperPass>();
AU.addRequired<AliasAnalysis>();
+ AU.addRequired<LoopAccessAnalysis>();
AU.addPreserved<LoopInfoWrapperPass>();
AU.addPreserved<DominatorTreeWrapperPass>();
AU.addPreserved<AliasAnalysis>();
// LoopVectorizationCostModel.
//===----------------------------------------------------------------------===//
-static Value *stripIntegerCast(Value *V) {
- if (CastInst *CI = dyn_cast<CastInst>(V))
- if (CI->getOperand(0)->getType()->isIntegerTy())
- return CI->getOperand(0);
- return V;
-}
-
-///\brief Replaces the symbolic stride in a pointer SCEV expression by one.
-///
-/// If \p OrigPtr is not null, use it to look up the stride value instead of
-/// \p Ptr.
-static const SCEV *replaceSymbolicStrideSCEV(ScalarEvolution *SE,
- ValueToValueMap &PtrToStride,
- Value *Ptr, Value *OrigPtr = nullptr) {
-
- const SCEV *OrigSCEV = SE->getSCEV(Ptr);
-
- // If there is an entry in the map return the SCEV of the pointer with the
- // symbolic stride replaced by one.
- ValueToValueMap::iterator SI = PtrToStride.find(OrigPtr ? OrigPtr : Ptr);
- if (SI != PtrToStride.end()) {
- Value *StrideVal = SI->second;
-
- // Strip casts.
- StrideVal = stripIntegerCast(StrideVal);
-
- // Replace symbolic stride by one.
- Value *One = ConstantInt::get(StrideVal->getType(), 1);
- ValueToValueMap RewriteMap;
- RewriteMap[StrideVal] = One;
-
- const SCEV *ByOne =
- SCEVParameterRewriter::rewrite(OrigSCEV, *SE, RewriteMap, true);
- DEBUG(dbgs() << "LV: Replacing SCEV: " << *OrigSCEV << " by: " << *ByOne
- << "\n");
- return ByOne;
- }
-
- // Otherwise, just return the SCEV of the original pointer.
- return SE->getSCEV(Ptr);
-}
-
-void RuntimePointerCheck::insert(ScalarEvolution *SE, Loop *Lp, Value *Ptr,
- bool WritePtr, unsigned DepSetId,
- unsigned ASId, ValueToValueMap &Strides) {
- // Get the stride replaced scev.
- const SCEV *Sc = replaceSymbolicStrideSCEV(SE, Strides, Ptr);
- const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(Sc);
- assert(AR && "Invalid addrec expression");
- const SCEV *Ex = SE->getBackedgeTakenCount(Lp);
- const SCEV *ScEnd = AR->evaluateAtIteration(Ex, *SE);
- Pointers.push_back(Ptr);
- Starts.push_back(AR->getStart());
- Ends.push_back(ScEnd);
- IsWritePtr.push_back(WritePtr);
- DependencySetId.push_back(DepSetId);
- AliasSetId.push_back(ASId);
-}
-
Value *InnerLoopVectorizer::getBroadcastInstrs(Value *V) {
// We need to place the broadcast of invariant variables outside the loop.
Instruction *Instr = dyn_cast<Instruction>(V);
/// \brief Find the operand of the GEP that should be checked for consecutive
/// stores. This ignores trailing indices that have no effect on the final
/// pointer.
-static unsigned getGEPInductionOperand(const DataLayout *DL,
- const GetElementPtrInst *Gep) {
+static unsigned getGEPInductionOperand(const GetElementPtrInst *Gep) {
+ const DataLayout &DL = Gep->getModule()->getDataLayout();
unsigned LastOperand = Gep->getNumOperands() - 1;
- unsigned GEPAllocSize = DL->getTypeAllocSize(
+ unsigned GEPAllocSize = DL.getTypeAllocSize(
cast<PointerType>(Gep->getType()->getScalarType())->getElementType());
// Walk backwards and try to peel off zeros.
// If it's a type with the same allocation size as the result of the GEP we
// can peel off the zero index.
- if (DL->getTypeAllocSize(*GEPTI) != GEPAllocSize)
+ if (DL.getTypeAllocSize(*GEPTI) != GEPAllocSize)
break;
--LastOperand;
}
return II.getConsecutiveDirection();
}
- unsigned InductionOperand = getGEPInductionOperand(DL, Gep);
+ unsigned InductionOperand = getGEPInductionOperand(Gep);
// Check that all of the gep indices are uniform except for our induction
// operand.
return 0;
}
-bool LoopAccessAnalysis::isUniform(Value *V) {
- return (SE->isLoopInvariant(SE->getSCEV(V), TheLoop));
-}
-
bool LoopVectorizationLegality::isUniform(Value *V) {
- return LAA.isUniform(V);
+ return LAI->isUniform(V);
}
InnerLoopVectorizer::VectorParts&
unsigned Alignment = LI ? LI->getAlignment() : SI->getAlignment();
// An alignment of 0 means target abi alignment. We need to use the scalar's
// target abi alignment in such a case.
+ const DataLayout &DL = Instr->getModule()->getDataLayout();
if (!Alignment)
- Alignment = DL->getABITypeAlignment(ScalarDataTy);
+ Alignment = DL.getABITypeAlignment(ScalarDataTy);
unsigned AddressSpace = Ptr->getType()->getPointerAddressSpace();
- unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ScalarDataTy);
- unsigned VectorElementSize = DL->getTypeStoreSize(DataTy)/VF;
+ unsigned ScalarAllocatedSize = DL.getTypeAllocSize(ScalarDataTy);
+ unsigned VectorElementSize = DL.getTypeStoreSize(DataTy) / VF;
if (SI && Legal->blockNeedsPredication(SI->getParent()) &&
!Legal->isMaskRequired(SI))
// The last index does not have to be the induction. It can be
// consecutive and be a function of the index. For example A[I+1];
unsigned NumOperands = Gep->getNumOperands();
- unsigned InductionOperand = getGEPInductionOperand(DL, Gep);
+ unsigned InductionOperand = getGEPInductionOperand(Gep);
// Create the new GEP with the new induction variable.
GetElementPtrInst *Gep2 = cast<GetElementPtrInst>(Gep->clone());
for (unsigned Part = 0; Part < UF; ++Part) {
// Calculate the pointer for the specific unroll-part.
- Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF));
+ Value *PartPtr =
+ Builder.CreateGEP(nullptr, Ptr, Builder.getInt32(Part * VF));
if (Reverse) {
// If we store to reverse consecutive memory locations then we need
StoredVal[Part] = reverseVector(StoredVal[Part]);
// If the address is consecutive but reversed, then the
// wide store needs to start at the last vector element.
- PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(-Part * VF));
- PartPtr = Builder.CreateGEP(PartPtr, Builder.getInt32(1 - VF));
+ PartPtr = Builder.CreateGEP(nullptr, Ptr, Builder.getInt32(-Part * VF));
+ PartPtr = Builder.CreateGEP(nullptr, PartPtr, Builder.getInt32(1 - VF));
Mask[Part] = reverseVector(Mask[Part]);
}
setDebugLocFromInst(Builder, LI);
for (unsigned Part = 0; Part < UF; ++Part) {
// Calculate the pointer for the specific unroll-part.
- Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF));
+ Value *PartPtr =
+ Builder.CreateGEP(nullptr, Ptr, Builder.getInt32(Part * VF));
if (Reverse) {
// If the address is consecutive but reversed, then the
// 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));
+ PartPtr = Builder.CreateGEP(nullptr, Ptr, Builder.getInt32(-Part * VF));
+ PartPtr = Builder.CreateGEP(nullptr, PartPtr, Builder.getInt32(1 - VF));
Mask[Part] = reverseVector(Mask[Part]);
}
return std::make_pair(FirstInst, TheCheck);
}
-std::pair<Instruction *, Instruction *>
-InnerLoopVectorizer::addRuntimeCheck(Instruction *Loc) {
- RuntimePointerCheck *PtrRtCheck = Legal->getRuntimePointerCheck();
-
- Instruction *tnullptr = nullptr;
- if (!PtrRtCheck->Need)
- return std::pair<Instruction *, Instruction *>(tnullptr, tnullptr);
-
- unsigned NumPointers = PtrRtCheck->Pointers.size();
- SmallVector<TrackingVH<Value> , 2> Starts;
- SmallVector<TrackingVH<Value> , 2> Ends;
-
- LLVMContext &Ctx = Loc->getContext();
- SCEVExpander Exp(*SE, "induction");
- Instruction *FirstInst = nullptr;
-
- for (unsigned i = 0; i < NumPointers; ++i) {
- Value *Ptr = PtrRtCheck->Pointers[i];
- const SCEV *Sc = SE->getSCEV(Ptr);
-
- if (SE->isLoopInvariant(Sc, OrigLoop)) {
- DEBUG(dbgs() << "LV: Adding RT check for a loop invariant ptr:" <<
- *Ptr <<"\n");
- Starts.push_back(Ptr);
- Ends.push_back(Ptr);
- } else {
- DEBUG(dbgs() << "LV: Adding RT check for range:" << *Ptr << '\n');
- unsigned AS = Ptr->getType()->getPointerAddressSpace();
-
- // Use this type for pointer arithmetic.
- Type *PtrArithTy = Type::getInt8PtrTy(Ctx, AS);
-
- Value *Start = Exp.expandCodeFor(PtrRtCheck->Starts[i], PtrArithTy, Loc);
- Value *End = Exp.expandCodeFor(PtrRtCheck->Ends[i], PtrArithTy, Loc);
- Starts.push_back(Start);
- Ends.push_back(End);
- }
- }
-
- IRBuilder<> ChkBuilder(Loc);
- // Our instructions might fold to a constant.
- Value *MemoryRuntimeCheck = nullptr;
- for (unsigned i = 0; i < NumPointers; ++i) {
- for (unsigned j = i+1; j < NumPointers; ++j) {
- // No need to check if two readonly pointers intersect.
- if (!PtrRtCheck->IsWritePtr[i] && !PtrRtCheck->IsWritePtr[j])
- continue;
-
- // Only need to check pointers between two different dependency sets.
- if (PtrRtCheck->DependencySetId[i] == PtrRtCheck->DependencySetId[j])
- continue;
- // Only need to check pointers in the same alias set.
- if (PtrRtCheck->AliasSetId[i] != PtrRtCheck->AliasSetId[j])
- continue;
-
- unsigned AS0 = Starts[i]->getType()->getPointerAddressSpace();
- unsigned AS1 = Starts[j]->getType()->getPointerAddressSpace();
-
- assert((AS0 == Ends[j]->getType()->getPointerAddressSpace()) &&
- (AS1 == Ends[i]->getType()->getPointerAddressSpace()) &&
- "Trying to bounds check pointers with different address spaces");
-
- Type *PtrArithTy0 = Type::getInt8PtrTy(Ctx, AS0);
- Type *PtrArithTy1 = Type::getInt8PtrTy(Ctx, AS1);
-
- Value *Start0 = ChkBuilder.CreateBitCast(Starts[i], PtrArithTy0, "bc");
- Value *Start1 = ChkBuilder.CreateBitCast(Starts[j], PtrArithTy1, "bc");
- Value *End0 = ChkBuilder.CreateBitCast(Ends[i], PtrArithTy1, "bc");
- Value *End1 = ChkBuilder.CreateBitCast(Ends[j], PtrArithTy0, "bc");
-
- Value *Cmp0 = ChkBuilder.CreateICmpULE(Start0, End1, "bound0");
- FirstInst = getFirstInst(FirstInst, Cmp0, Loc);
- Value *Cmp1 = ChkBuilder.CreateICmpULE(Start1, End0, "bound1");
- FirstInst = getFirstInst(FirstInst, Cmp1, Loc);
- Value *IsConflict = ChkBuilder.CreateAnd(Cmp0, Cmp1, "found.conflict");
- FirstInst = getFirstInst(FirstInst, IsConflict, Loc);
- if (MemoryRuntimeCheck) {
- IsConflict = ChkBuilder.CreateOr(MemoryRuntimeCheck, IsConflict,
- "conflict.rdx");
- FirstInst = getFirstInst(FirstInst, IsConflict, Loc);
- }
- MemoryRuntimeCheck = IsConflict;
- }
- }
-
- // We have to do this trickery because the IRBuilder might fold the check to a
- // constant expression in which case there is no Instruction anchored in a
- // the block.
- Instruction *Check = BinaryOperator::CreateAnd(MemoryRuntimeCheck,
- ConstantInt::getTrue(Ctx));
- ChkBuilder.Insert(Check, "memcheck.conflict");
- FirstInst = getFirstInst(FirstInst, Check, Loc);
- return std::make_pair(FirstInst, Check);
-}
-
void InnerLoopVectorizer::createEmptyLoop() {
/*
In this function we generate a new loop. The new loop will contain
ExitCount = SE->getAddExpr(BackedgeTakeCount,
SE->getConstant(BackedgeTakeCount->getType(), 1));
+ const DataLayout &DL = OldBasicBlock->getModule()->getDataLayout();
+
// Expand the trip count and place the new instructions in the preheader.
// Notice that the pre-header does not change, only the loop body.
- SCEVExpander Exp(*SE, "induction");
+ SCEVExpander Exp(*SE, DL, "induction");
// We need to test whether the backedge-taken count is uint##_max. Adding one
// to it will cause overflow and an incorrect loop trip count in the vector
std::tie(FirstCheckInst, StrideCheck) =
addStrideCheck(LastBypassBlock->getTerminator());
if (StrideCheck) {
+ AddedSafetyChecks = true;
// Create a new block containing the stride check.
BasicBlock *CheckBlock =
LastBypassBlock->splitBasicBlock(FirstCheckInst, "vector.stridecheck");
// faster.
Instruction *MemRuntimeCheck;
std::tie(FirstCheckInst, MemRuntimeCheck) =
- addRuntimeCheck(LastBypassBlock->getTerminator());
+ Legal->getLAI()->addRuntimeCheck(LastBypassBlock->getTerminator());
if (MemRuntimeCheck) {
+ AddedSafetyChecks = true;
// Create a new block containing the memory check.
BasicBlock *CheckBlock =
- LastBypassBlock->splitBasicBlock(MemRuntimeCheck, "vector.memcheck");
+ LastBypassBlock->splitBasicBlock(FirstCheckInst, "vector.memcheck");
if (ParentLoop)
ParentLoop->addBasicBlockToLoop(CheckBlock, *LI);
LoopBypassBlocks.push_back(CheckBlock);
}
}
-Value *createMinMaxOp(IRBuilder<> &Builder,
- LoopVectorizationLegality::MinMaxReductionKind RK,
- Value *Left,
- Value *Right) {
+static Value *createMinMaxOp(IRBuilder<> &Builder,
+ LoopVectorizationLegality::MinMaxReductionKind RK,
+ Value *Left, Value *Right) {
CmpInst::Predicate P = CmpInst::ICMP_NE;
switch (RK) {
default:
return V;
}
+/// Estimate the overhead of scalarizing a value. Insert and Extract are set if
+/// the result needs to be inserted and/or extracted from vectors.
+static unsigned getScalarizationOverhead(Type *Ty, bool Insert, bool Extract,
+ const TargetTransformInfo &TTI) {
+ if (Ty->isVoidTy())
+ return 0;
+
+ assert(Ty->isVectorTy() && "Can only scalarize vectors");
+ unsigned Cost = 0;
+
+ for (int i = 0, e = Ty->getVectorNumElements(); i < e; ++i) {
+ if (Insert)
+ Cost += TTI.getVectorInstrCost(Instruction::InsertElement, Ty, i);
+ if (Extract)
+ Cost += TTI.getVectorInstrCost(Instruction::ExtractElement, Ty, i);
+ }
+
+ return Cost;
+}
+
+// Estimate cost of a call instruction CI if it were vectorized with factor VF.
+// Return the cost of the instruction, including scalarization overhead if it's
+// needed. The flag NeedToScalarize shows if the call needs to be scalarized -
+// i.e. either vector version isn't available, or is too expensive.
+static unsigned getVectorCallCost(CallInst *CI, unsigned VF,
+ const TargetTransformInfo &TTI,
+ const TargetLibraryInfo *TLI,
+ bool &NeedToScalarize) {
+ Function *F = CI->getCalledFunction();
+ StringRef FnName = CI->getCalledFunction()->getName();
+ Type *ScalarRetTy = CI->getType();
+ SmallVector<Type *, 4> Tys, ScalarTys;
+ for (auto &ArgOp : CI->arg_operands())
+ ScalarTys.push_back(ArgOp->getType());
+
+ // Estimate cost of scalarized vector call. The source operands are assumed
+ // to be vectors, so we need to extract individual elements from there,
+ // execute VF scalar calls, and then gather the result into the vector return
+ // value.
+ unsigned ScalarCallCost = TTI.getCallInstrCost(F, ScalarRetTy, ScalarTys);
+ if (VF == 1)
+ return ScalarCallCost;
+
+ // Compute corresponding vector type for return value and arguments.
+ Type *RetTy = ToVectorTy(ScalarRetTy, VF);
+ for (unsigned i = 0, ie = ScalarTys.size(); i != ie; ++i)
+ Tys.push_back(ToVectorTy(ScalarTys[i], VF));
+
+ // Compute costs of unpacking argument values for the scalar calls and
+ // packing the return values to a vector.
+ unsigned ScalarizationCost =
+ getScalarizationOverhead(RetTy, true, false, TTI);
+ for (unsigned i = 0, ie = Tys.size(); i != ie; ++i)
+ ScalarizationCost += getScalarizationOverhead(Tys[i], false, true, TTI);
+
+ unsigned Cost = ScalarCallCost * VF + ScalarizationCost;
+
+ // If we can't emit a vector call for this function, then the currently found
+ // cost is the cost we need to return.
+ NeedToScalarize = true;
+ if (!TLI || !TLI->isFunctionVectorizable(FnName, VF) || CI->isNoBuiltin())
+ return Cost;
+
+ // If the corresponding vector cost is cheaper, return its cost.
+ unsigned VectorCallCost = TTI.getCallInstrCost(nullptr, RetTy, Tys);
+ if (VectorCallCost < Cost) {
+ NeedToScalarize = false;
+ return VectorCallCost;
+ }
+ return Cost;
+}
+
+// Estimate cost of an intrinsic call instruction CI if it were vectorized with
+// factor VF. Return the cost of the instruction, including scalarization
+// overhead if it's needed.
+static unsigned getVectorIntrinsicCost(CallInst *CI, unsigned VF,
+ const TargetTransformInfo &TTI,
+ const TargetLibraryInfo *TLI) {
+ Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
+ assert(ID && "Expected intrinsic call!");
+
+ Type *RetTy = ToVectorTy(CI->getType(), VF);
+ SmallVector<Type *, 4> Tys;
+ for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i)
+ Tys.push_back(ToVectorTy(CI->getArgOperand(i)->getType(), VF));
+
+ return TTI.getIntrinsicInstrCost(ID, RetTy, Tys);
+}
+
void InnerLoopVectorizer::vectorizeLoop() {
//===------------------------------------------------===//
//
Module *M = BB->getParent()->getParent();
CallInst *CI = cast<CallInst>(it);
+
+ StringRef FnName = CI->getCalledFunction()->getName();
+ Function *F = CI->getCalledFunction();
+ Type *RetTy = ToVectorTy(CI->getType(), VF);
+ SmallVector<Type *, 4> Tys;
+ for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i)
+ Tys.push_back(ToVectorTy(CI->getArgOperand(i)->getType(), VF));
+
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:
+ if (ID &&
+ (ID == Intrinsic::assume || ID == Intrinsic::lifetime_end ||
+ ID == Intrinsic::lifetime_start)) {
scalarizeInstruction(it);
break;
- default:
- bool HasScalarOpd = hasVectorInstrinsicScalarOpd(ID, 1);
- for (unsigned Part = 0; Part < UF; ++Part) {
- SmallVector<Value *, 4> Args;
- for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i) {
- if (HasScalarOpd && i == 1) {
- Args.push_back(CI->getArgOperand(i));
- continue;
- }
- VectorParts &Arg = getVectorValue(CI->getArgOperand(i));
- Args.push_back(Arg[Part]);
- }
- Type *Tys[] = {CI->getType()};
- if (VF > 1)
- Tys[0] = VectorType::get(CI->getType()->getScalarType(), VF);
+ }
+ // The flag shows whether we use Intrinsic or a usual Call for vectorized
+ // version of the instruction.
+ // Is it beneficial to perform intrinsic call compared to lib call?
+ bool NeedToScalarize;
+ unsigned CallCost = getVectorCallCost(CI, VF, *TTI, TLI, NeedToScalarize);
+ bool UseVectorIntrinsic =
+ ID && getVectorIntrinsicCost(CI, VF, *TTI, TLI) <= CallCost;
+ if (!UseVectorIntrinsic && NeedToScalarize) {
+ scalarizeInstruction(it);
+ break;
+ }
- Function *F = Intrinsic::getDeclaration(M, ID, Tys);
- Entry[Part] = Builder.CreateCall(F, Args);
+ for (unsigned Part = 0; Part < UF; ++Part) {
+ SmallVector<Value *, 4> Args;
+ for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i) {
+ Value *Arg = CI->getArgOperand(i);
+ // Some intrinsics have a scalar argument - don't replace it with a
+ // vector.
+ if (!UseVectorIntrinsic || !hasVectorInstrinsicScalarOpd(ID, i)) {
+ VectorParts &VectorArg = getVectorValue(CI->getArgOperand(i));
+ Arg = VectorArg[Part];
+ }
+ Args.push_back(Arg);
}
- propagateMetadata(Entry, it);
- break;
+ Function *VectorF;
+ if (UseVectorIntrinsic) {
+ // Use vector version of the intrinsic.
+ Type *TysForDecl[] = {CI->getType()};
+ if (VF > 1)
+ TysForDecl[0] = VectorType::get(CI->getType()->getScalarType(), VF);
+ VectorF = Intrinsic::getDeclaration(M, ID, TysForDecl);
+ } else {
+ // Use vector version of the library call.
+ StringRef VFnName = TLI->getVectorizedFunction(FnName, VF);
+ assert(!VFnName.empty() && "Vector function name is empty.");
+ VectorF = M->getFunction(VFnName);
+ if (!VectorF) {
+ // Generate a declaration
+ FunctionType *FTy = FunctionType::get(RetTy, Tys, false);
+ VectorF =
+ Function::Create(FTy, Function::ExternalLinkage, VFnName, M);
+ VectorF->copyAttributesFrom(F);
+ }
+ }
+ assert(VectorF && "Can't create vector function.");
+ Entry[Part] = Builder.CreateCall(VectorF, Args);
}
+
+ propagateMetadata(Entry, it);
break;
}
collectLoopUniforms();
DEBUG(dbgs() << "LV: We can vectorize this loop" <<
- (LAA.getRuntimePointerCheck()->Need ? " (with a runtime bound check)" :
+ (LAI->getRuntimePointerCheck()->Need ? " (with a runtime bound check)" :
"")
<<"!\n");
// Look for the attribute signaling the absence of NaNs.
Function &F = *Header->getParent();
+ const DataLayout &DL = F.getParent()->getDataLayout();
if (F.hasFnAttribute("no-nans-fp-math"))
- HasFunNoNaNAttr = F.getAttributes().getAttribute(
- AttributeSet::FunctionIndex,
- "no-nans-fp-math").getValueAsString() == "true";
+ HasFunNoNaNAttr =
+ F.getFnAttribute("no-nans-fp-math").getValueAsString() == "true";
// For each block in the loop.
for (Loop::block_iterator bb = TheLoop->block_begin(),
if (IK_NoInduction != IK) {
// Get the widest type.
if (!WidestIndTy)
- WidestIndTy = convertPointerToIntegerType(*DL, PhiTy);
+ WidestIndTy = convertPointerToIntegerType(DL, PhiTy);
else
- WidestIndTy = getWiderType(*DL, PhiTy, WidestIndTy);
+ WidestIndTy = getWiderType(DL, PhiTy, WidestIndTy);
// Int inductions are special because we only allow one IV.
if (IK == IK_IntInduction && StepValue->isOne()) {
return false;
}// end of PHI handling
- // We still don't handle functions. However, we can ignore dbg intrinsic
- // calls and we do handle certain intrinsic and libm functions.
+ // We handle calls that:
+ // * Are debug info intrinsics.
+ // * Have a mapping to an IR intrinsic.
+ // * Have a vector version available.
CallInst *CI = dyn_cast<CallInst>(it);
- if (CI && !getIntrinsicIDForCall(CI, TLI) && !isa<DbgInfoIntrinsic>(CI)) {
+ if (CI && !getIntrinsicIDForCall(CI, TLI) && !isa<DbgInfoIntrinsic>(CI) &&
+ !(CI->getCalledFunction() && TLI &&
+ TLI->isFunctionVectorizable(CI->getCalledFunction()->getName()))) {
emitAnalysis(VectorizationReport(it) <<
"call instruction cannot be vectorized");
- DEBUG(dbgs() << "LV: Found a call site.\n");
+ DEBUG(dbgs() << "LV: Found a non-intrinsic, non-libfunc callsite.\n");
return false;
}
///\brief Remove GEPs whose indices but the last one are loop invariant and
/// return the induction operand of the gep pointer.
-static Value *stripGetElementPtr(Value *Ptr, ScalarEvolution *SE,
- const DataLayout *DL, Loop *Lp) {
+static Value *stripGetElementPtr(Value *Ptr, ScalarEvolution *SE, Loop *Lp) {
GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(Ptr);
if (!GEP)
return Ptr;
- unsigned InductionOperand = getGEPInductionOperand(DL, GEP);
+ unsigned InductionOperand = getGEPInductionOperand(GEP);
// Check that all of the gep indices are uniform except for our induction
// operand.
///\brief Get the stride of a pointer access in a loop.
/// Looks for symbolic strides "a[i*stride]". Returns the symbolic stride as a
/// pointer to the Value, or null otherwise.
-static Value *getStrideFromPointer(Value *Ptr, ScalarEvolution *SE,
- const DataLayout *DL, Loop *Lp) {
+static Value *getStrideFromPointer(Value *Ptr, ScalarEvolution *SE, Loop *Lp) {
const PointerType *PtrTy = dyn_cast<PointerType>(Ptr->getType());
if (!PtrTy || PtrTy->isAggregateType())
return nullptr;
// The size of the pointer access.
int64_t PtrAccessSize = 1;
- Ptr = stripGetElementPtr(Ptr, SE, DL, Lp);
+ Ptr = stripGetElementPtr(Ptr, SE, Lp);
const SCEV *V = SE->getSCEV(Ptr);
if (Ptr != OrigPtr)
// Strip off the size of access multiplication if we are still analyzing the
// pointer.
if (OrigPtr == Ptr) {
- DL->getTypeAllocSize(PtrTy->getElementType());
+ const DataLayout &DL = Lp->getHeader()->getModule()->getDataLayout();
+ DL.getTypeAllocSize(PtrTy->getElementType());
if (const SCEVMulExpr *M = dyn_cast<SCEVMulExpr>(V)) {
if (M->getOperand(0)->getSCEVType() != scConstant)
return nullptr;
else
return;
- Value *Stride = getStrideFromPointer(Ptr, SE, DL, TheLoop);
+ Value *Stride = getStrideFromPointer(Ptr, SE, TheLoop);
if (!Stride)
return;
}
}
-namespace {
-/// \brief Analyses memory accesses in a loop.
-///
-/// Checks whether run time pointer checks are needed and builds sets for data
-/// dependence checking.
-class AccessAnalysis {
-public:
- /// \brief Read or write access location.
- typedef PointerIntPair<Value *, 1, bool> MemAccessInfo;
- typedef SmallPtrSet<MemAccessInfo, 8> MemAccessInfoSet;
-
- /// \brief Set of potential dependent memory accesses.
- typedef EquivalenceClasses<MemAccessInfo> DepCandidates;
-
- AccessAnalysis(const DataLayout *Dl, AliasAnalysis *AA, DepCandidates &DA) :
- DL(Dl), AST(*AA), DepCands(DA), IsRTCheckNeeded(false) {}
-
- /// \brief Register a load and whether it is only read from.
- void addLoad(AliasAnalysis::Location &Loc, bool IsReadOnly) {
- Value *Ptr = const_cast<Value*>(Loc.Ptr);
- AST.add(Ptr, AliasAnalysis::UnknownSize, Loc.AATags);
- Accesses.insert(MemAccessInfo(Ptr, false));
- if (IsReadOnly)
- ReadOnlyPtr.insert(Ptr);
- }
-
- /// \brief Register a store.
- void addStore(AliasAnalysis::Location &Loc) {
- Value *Ptr = const_cast<Value*>(Loc.Ptr);
- AST.add(Ptr, AliasAnalysis::UnknownSize, Loc.AATags);
- Accesses.insert(MemAccessInfo(Ptr, true));
- }
-
- /// \brief Check whether we can check the pointers at runtime for
- /// non-intersection.
- bool canCheckPtrAtRT(RuntimePointerCheck &RtCheck, unsigned &NumComparisons,
- ScalarEvolution *SE, Loop *TheLoop,
- ValueToValueMap &Strides,
- bool ShouldCheckStride = false);
-
- /// \brief Goes over all memory accesses, checks whether a RT check is needed
- /// and builds sets of dependent accesses.
- void buildDependenceSets() {
- processMemAccesses();
- }
-
- bool isRTCheckNeeded() { return IsRTCheckNeeded; }
-
- bool isDependencyCheckNeeded() { return !CheckDeps.empty(); }
- void resetDepChecks() { CheckDeps.clear(); }
-
- MemAccessInfoSet &getDependenciesToCheck() { return CheckDeps; }
-
-private:
- typedef SetVector<MemAccessInfo> PtrAccessSet;
-
- /// \brief Go over all memory access and check whether runtime pointer checks
- /// are needed /// and build sets of dependency check candidates.
- void processMemAccesses();
-
- /// Set of all accesses.
- PtrAccessSet Accesses;
-
- /// Set of accesses that need a further dependence check.
- MemAccessInfoSet CheckDeps;
-
- /// Set of pointers that are read only.
- SmallPtrSet<Value*, 16> ReadOnlyPtr;
-
- const DataLayout *DL;
-
- /// An alias set tracker to partition the access set by underlying object and
- //intrinsic property (such as TBAA metadata).
- AliasSetTracker AST;
-
- /// Sets of potentially dependent accesses - members of one set share an
- /// underlying pointer. The set "CheckDeps" identfies which sets really need a
- /// dependence check.
- DepCandidates &DepCands;
-
- bool IsRTCheckNeeded;
-};
-
-} // end anonymous namespace
-
-/// \brief Check whether a pointer can participate in a runtime bounds check.
-static bool hasComputableBounds(ScalarEvolution *SE, ValueToValueMap &Strides,
- Value *Ptr) {
- const SCEV *PtrScev = replaceSymbolicStrideSCEV(SE, Strides, Ptr);
- const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(PtrScev);
- if (!AR)
- return false;
-
- return AR->isAffine();
-}
-
-/// \brief Check the stride of the pointer and ensure that it does not wrap in
-/// the address space.
-static int isStridedPtr(ScalarEvolution *SE, const DataLayout *DL, Value *Ptr,
- const Loop *Lp, ValueToValueMap &StridesMap);
-
-bool AccessAnalysis::canCheckPtrAtRT(
- RuntimePointerCheck &RtCheck,
- unsigned &NumComparisons, ScalarEvolution *SE, Loop *TheLoop,
- ValueToValueMap &StridesMap, bool ShouldCheckStride) {
- // Find pointers with computable bounds. We are going to use this information
- // to place a runtime bound check.
- bool CanDoRT = true;
-
- bool IsDepCheckNeeded = isDependencyCheckNeeded();
- NumComparisons = 0;
-
- // We assign a consecutive id to access from different alias sets.
- // Accesses between different groups doesn't need to be checked.
- unsigned ASId = 1;
- for (auto &AS : AST) {
- unsigned NumReadPtrChecks = 0;
- unsigned NumWritePtrChecks = 0;
-
- // 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;
-
- for (auto A : AS) {
- Value *Ptr = A.getValue();
- bool IsWrite = Accesses.count(MemAccessInfo(Ptr, true));
- MemAccessInfo Access(Ptr, IsWrite);
-
- if (IsWrite)
- ++NumWritePtrChecks;
- else
- ++NumReadPtrChecks;
-
- if (hasComputableBounds(SE, StridesMap, Ptr) &&
- // When we run after a failing dependency check we have to make sure we
- // don't have wrapping pointers.
- (!ShouldCheckStride ||
- isStridedPtr(SE, DL, Ptr, TheLoop, StridesMap) == 1)) {
- // The id of the dependence set.
- unsigned DepId;
-
- if (IsDepCheckNeeded) {
- 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++;
-
- RtCheck.insert(SE, TheLoop, Ptr, IsWrite, DepId, ASId, StridesMap);
-
- DEBUG(dbgs() << "LV: Found a runtime check ptr:" << *Ptr << '\n');
- } else {
- CanDoRT = false;
- }
- }
-
- if (IsDepCheckNeeded && CanDoRT && RunningDepId == 2)
- NumComparisons += 0; // Only one dependence set.
- else {
- NumComparisons += (NumWritePtrChecks * (NumReadPtrChecks +
- NumWritePtrChecks - 1));
- }
-
- ++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.DependencySetId[i] == RtCheck.DependencySetId[j])
- continue;
- // Only need to check pointers in the same alias set.
- if (RtCheck.AliasSetId[i] != RtCheck.AliasSetId[j])
- continue;
-
- Value *PtrI = RtCheck.Pointers[i];
- Value *PtrJ = RtCheck.Pointers[j];
-
- unsigned ASi = PtrI->getType()->getPointerAddressSpace();
- unsigned ASj = PtrJ->getType()->getPointerAddressSpace();
- if (ASi != ASj) {
- DEBUG(dbgs() << "LV: Runtime check would require comparison between"
- " different address spaces\n");
- return false;
- }
- }
- }
-
- return CanDoRT;
-}
-
-void AccessAnalysis::processMemAccesses() {
- // 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.
-
- DEBUG(dbgs() << "LV: Processing memory accesses...\n");
- DEBUG(dbgs() << " AST: "; AST.dump());
- DEBUG(dbgs() << "LV: Accesses:\n");
- DEBUG({
- for (auto A : Accesses)
- dbgs() << "\t" << *A.getPointer() << " (" <<
- (A.getInt() ? "write" : (ReadOnlyPtr.count(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 (auto &AS : AST) {
- // Note that both the alias-set tracker and the alias sets themselves used
- // linked lists internally and so the iteration order here is deterministic
- // (matching the original instruction order within each set).
-
- bool SetHasWrite = false;
-
- // Map of pointers to last access encountered.
- typedef DenseMap<Value*, MemAccessInfo> UnderlyingObjToAccessMap;
- UnderlyingObjToAccessMap ObjToLastAccess;
-
- // Set of access to check after all writes have been processed.
- PtrAccessSet DeferredAccesses;
-
- // Iterate over each alias set twice, once to process read/write pointers,
- // and then to process read-only pointers.
- for (int SetIteration = 0; SetIteration < 2; ++SetIteration) {
- bool UseDeferred = SetIteration > 0;
- PtrAccessSet &S = UseDeferred ? DeferredAccesses : Accesses;
-
- for (auto AV : AS) {
- Value *Ptr = AV.getValue();
-
- // 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 AC : S) {
- if (AC.getPointer() != Ptr)
- continue;
-
- bool IsWrite = AC.getInt();
-
- // If we're using the deferred access set, then it contains only
- // reads.
- bool IsReadOnlyPtr = ReadOnlyPtr.count(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.count(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) {
- 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) && SetHasWrite) {
- CheckDeps.insert(Access);
- IsRTCheckNeeded = true;
- }
-
- if (IsWrite)
- SetHasWrite = true;
-
- // 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) {
- UnderlyingObjToAccessMap::iterator Prev =
- ObjToLastAccess.find(UnderlyingObj);
- if (Prev != ObjToLastAccess.end())
- DepCands.unionSets(Access, Prev->second);
-
- ObjToLastAccess[UnderlyingObj] = Access;
- }
- }
- }
- }
- }
-}
-
-namespace {
-/// \brief Checks memory dependences among accesses to the same underlying
-/// object to determine whether there vectorization is legal or not (and at
-/// which vectorization factor).
-///
-/// This class works under the assumption that we already checked that memory
-/// locations with different underlying pointers are "must-not alias".
-/// We use the ScalarEvolution framework to symbolically evalutate access
-/// functions pairs. Since we currently don't restructure the loop we can rely
-/// on the program order of memory accesses to determine their safety.
-/// At the moment we will only deem accesses as safe for:
-/// * A negative constant distance assuming program order.
-///
-/// Safe: tmp = a[i + 1]; OR a[i + 1] = x;
-/// a[i] = tmp; y = a[i];
-///
-/// The latter case is safe because later checks guarantuee that there can't
-/// be a cycle through a phi node (that is, we check that "x" and "y" is not
-/// the same variable: a header phi can only be an induction or a reduction, a
-/// reduction can't have a memory sink, an induction can't have a memory
-/// source). This is important and must not be violated (or we have to
-/// resort to checking for cycles through memory).
-///
-/// * A positive constant distance assuming program order that is bigger
-/// than the biggest memory access.
-///
-/// tmp = a[i] OR b[i] = x
-/// a[i+2] = tmp y = b[i+2];
-///
-/// Safe distance: 2 x sizeof(a[0]), and 2 x sizeof(b[0]), respectively.
-///
-/// * Zero distances and all accesses have the same size.
-///
-class MemoryDepChecker {
-public:
- typedef PointerIntPair<Value *, 1, bool> MemAccessInfo;
- typedef SmallPtrSet<MemAccessInfo, 8> MemAccessInfoSet;
-
- MemoryDepChecker(ScalarEvolution *Se, const DataLayout *Dl, const Loop *L,
- const LoopAccessAnalysis::VectorizerParams &VectParams)
- : SE(Se), DL(Dl), InnermostLoop(L), AccessIdx(0),
- ShouldRetryWithRuntimeCheck(false), VectParams(VectParams) {}
-
- /// \brief Register the location (instructions are given increasing numbers)
- /// of a write access.
- void addAccess(StoreInst *SI) {
- Value *Ptr = SI->getPointerOperand();
- Accesses[MemAccessInfo(Ptr, true)].push_back(AccessIdx);
- InstMap.push_back(SI);
- ++AccessIdx;
- }
-
- /// \brief Register the location (instructions are given increasing numbers)
- /// of a write access.
- void addAccess(LoadInst *LI) {
- Value *Ptr = LI->getPointerOperand();
- Accesses[MemAccessInfo(Ptr, false)].push_back(AccessIdx);
- InstMap.push_back(LI);
- ++AccessIdx;
- }
-
- /// \brief Check whether the dependencies between the accesses are safe.
- ///
- /// Only checks sets with elements in \p CheckDeps.
- bool areDepsSafe(AccessAnalysis::DepCandidates &AccessSets,
- MemAccessInfoSet &CheckDeps, ValueToValueMap &Strides);
-
- /// \brief The maximum number of bytes of a vector register we can vectorize
- /// the accesses safely with.
- unsigned getMaxSafeDepDistBytes() { return MaxSafeDepDistBytes; }
-
- /// \brief In same cases when the dependency check fails we can still
- /// vectorize the loop with a dynamic array access check.
- bool shouldRetryWithRuntimeCheck() { return ShouldRetryWithRuntimeCheck; }
-
-private:
- ScalarEvolution *SE;
- const DataLayout *DL;
- const Loop *InnermostLoop;
-
- /// \brief Maps access locations (ptr, read/write) to program order.
- DenseMap<MemAccessInfo, std::vector<unsigned> > Accesses;
-
- /// \brief Memory access instructions in program order.
- SmallVector<Instruction *, 16> InstMap;
-
- /// \brief The program order index to be used for the next instruction.
- unsigned AccessIdx;
-
- // We can access this many bytes in parallel safely.
- unsigned MaxSafeDepDistBytes;
-
- /// \brief If we see a non-constant dependence distance we can still try to
- /// vectorize this loop with runtime checks.
- bool ShouldRetryWithRuntimeCheck;
-
- /// \brief Vectorizer parameters used by the analysis.
- LoopAccessAnalysis::VectorizerParams VectParams;
-
- /// \brief Check whether there is a plausible dependence between the two
- /// accesses.
- ///
- /// Access \p A must happen before \p B in program order. The two indices
- /// identify the index into the program order map.
- ///
- /// This function checks whether there is a plausible dependence (or the
- /// absence of such can't be proved) between the two accesses. If there is a
- /// plausible dependence but the dependence distance is bigger than one
- /// element access it records this distance in \p MaxSafeDepDistBytes (if this
- /// distance is smaller than any other distance encountered so far).
- /// Otherwise, this function returns true signaling a possible dependence.
- bool isDependent(const MemAccessInfo &A, unsigned AIdx,
- const MemAccessInfo &B, unsigned BIdx,
- ValueToValueMap &Strides);
-
- /// \brief Check whether the data dependence could prevent store-load
- /// forwarding.
- bool couldPreventStoreLoadForward(unsigned Distance, unsigned TypeByteSize);
-};
-
-} // end anonymous namespace
-
-static bool isInBoundsGep(Value *Ptr) {
- if (GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(Ptr))
- return GEP->isInBounds();
- return false;
-}
-
-/// \brief Check whether the access through \p Ptr has a constant stride.
-static int isStridedPtr(ScalarEvolution *SE, const DataLayout *DL, Value *Ptr,
- const Loop *Lp, ValueToValueMap &StridesMap) {
- const Type *Ty = Ptr->getType();
- assert(Ty->isPointerTy() && "Unexpected non-ptr");
-
- // Make sure that the pointer does not point to aggregate types.
- const PointerType *PtrTy = cast<PointerType>(Ty);
- if (PtrTy->getElementType()->isAggregateType()) {
- DEBUG(dbgs() << "LV: Bad stride - Not a pointer to a scalar type" << *Ptr <<
- "\n");
- return 0;
- }
-
- const SCEV *PtrScev = replaceSymbolicStrideSCEV(SE, StridesMap, Ptr);
-
- const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(PtrScev);
- if (!AR) {
- DEBUG(dbgs() << "LV: Bad stride - Not an AddRecExpr pointer "
- << *Ptr << " SCEV: " << *PtrScev << "\n");
- return 0;
- }
-
- // The accesss function must stride over the innermost loop.
- if (Lp != AR->getLoop()) {
- DEBUG(dbgs() << "LV: Bad stride - Not striding over innermost loop " <<
- *Ptr << " SCEV: " << *PtrScev << "\n");
- }
-
- // The address calculation must not wrap. Otherwise, a dependence could be
- // inverted.
- // An inbounds getelementptr that is a AddRec with a unit stride
- // cannot wrap per definition. The unit stride requirement is checked later.
- // An getelementptr without an inbounds attribute and unit stride would have
- // to access the pointer value "0" which is undefined behavior in address
- // space 0, therefore we can also vectorize this case.
- bool IsInBoundsGEP = isInBoundsGep(Ptr);
- bool IsNoWrapAddRec = AR->getNoWrapFlags(SCEV::NoWrapMask);
- bool IsInAddressSpaceZero = PtrTy->getAddressSpace() == 0;
- if (!IsNoWrapAddRec && !IsInBoundsGEP && !IsInAddressSpaceZero) {
- DEBUG(dbgs() << "LV: Bad stride - Pointer may wrap in the address space "
- << *Ptr << " SCEV: " << *PtrScev << "\n");
- return 0;
- }
-
- // Check the step is constant.
- const SCEV *Step = AR->getStepRecurrence(*SE);
-
- // Calculate the pointer stride and check if it is consecutive.
- const SCEVConstant *C = dyn_cast<SCEVConstant>(Step);
- if (!C) {
- DEBUG(dbgs() << "LV: Bad stride - Not a constant strided " << *Ptr <<
- " SCEV: " << *PtrScev << "\n");
- return 0;
- }
-
- int64_t Size = DL->getTypeAllocSize(PtrTy->getElementType());
- const APInt &APStepVal = C->getValue()->getValue();
-
- // Huge step value - give up.
- if (APStepVal.getBitWidth() > 64)
- return 0;
-
- int64_t StepVal = APStepVal.getSExtValue();
-
- // Strided access.
- int64_t Stride = StepVal / Size;
- int64_t Rem = StepVal % Size;
- if (Rem)
- return 0;
-
- // If the SCEV could wrap but we have an inbounds gep with a unit stride we
- // know we can't "wrap around the address space". In case of address space
- // zero we know that this won't happen without triggering undefined behavior.
- if (!IsNoWrapAddRec && (IsInBoundsGEP || IsInAddressSpaceZero) &&
- Stride != 1 && Stride != -1)
- return 0;
-
- return Stride;
-}
-
-bool MemoryDepChecker::couldPreventStoreLoadForward(unsigned Distance,
- unsigned TypeByteSize) {
- // 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.
- const unsigned NumCyclesForStoreLoadThroughMemory = 8*TypeByteSize;
- // Maximum vector factor.
- unsigned MaxVFWithoutSLForwardIssues = VectParams.MaxVectorWidth*TypeByteSize;
- if(MaxSafeDepDistBytes < MaxVFWithoutSLForwardIssues)
- MaxVFWithoutSLForwardIssues = MaxSafeDepDistBytes;
-
- for (unsigned vf = 2*TypeByteSize; vf <= MaxVFWithoutSLForwardIssues;
- vf *= 2) {
- if (Distance % vf && Distance / vf < NumCyclesForStoreLoadThroughMemory) {
- MaxVFWithoutSLForwardIssues = (vf >>=1);
- break;
- }
- }
-
- if (MaxVFWithoutSLForwardIssues< 2*TypeByteSize) {
- DEBUG(dbgs() << "LV: Distance " << Distance <<
- " that could cause a store-load forwarding conflict\n");
- return true;
- }
-
- if (MaxVFWithoutSLForwardIssues < MaxSafeDepDistBytes &&
- MaxVFWithoutSLForwardIssues != VectParams.MaxVectorWidth*TypeByteSize)
- MaxSafeDepDistBytes = MaxVFWithoutSLForwardIssues;
- return false;
-}
-
-bool MemoryDepChecker::isDependent(const MemAccessInfo &A, unsigned AIdx,
- const MemAccessInfo &B, unsigned BIdx,
- ValueToValueMap &Strides) {
- assert (AIdx < BIdx && "Must pass arguments in program order");
-
- Value *APtr = A.getPointer();
- Value *BPtr = B.getPointer();
- bool AIsWrite = A.getInt();
- bool BIsWrite = B.getInt();
-
- // Two reads are independent.
- if (!AIsWrite && !BIsWrite)
+bool LoopVectorizationLegality::canVectorizeMemory() {
+ LAI = &LAA->getInfo(TheLoop, Strides);
+ auto &OptionalReport = LAI->getReport();
+ if (OptionalReport)
+ emitAnalysis(VectorizationReport(*OptionalReport));
+ if (!LAI->canVectorizeMemory())
return false;
- // We cannot check pointers in different address spaces.
- if (APtr->getType()->getPointerAddressSpace() !=
- BPtr->getType()->getPointerAddressSpace())
- return true;
-
- const SCEV *AScev = replaceSymbolicStrideSCEV(SE, Strides, APtr);
- const SCEV *BScev = replaceSymbolicStrideSCEV(SE, Strides, BPtr);
-
- int StrideAPtr = isStridedPtr(SE, DL, APtr, InnermostLoop, Strides);
- int StrideBPtr = isStridedPtr(SE, DL, BPtr, InnermostLoop, Strides);
-
- const SCEV *Src = AScev;
- const SCEV *Sink = BScev;
-
- // If the induction step is negative we have to invert source and sink of the
- // dependence.
- if (StrideAPtr < 0) {
- //Src = BScev;
- //Sink = AScev;
- std::swap(APtr, BPtr);
- std::swap(Src, Sink);
- std::swap(AIsWrite, BIsWrite);
- std::swap(AIdx, BIdx);
- std::swap(StrideAPtr, StrideBPtr);
- }
-
- const SCEV *Dist = SE->getMinusSCEV(Sink, Src);
-
- DEBUG(dbgs() << "LV: Src Scev: " << *Src << "Sink Scev: " << *Sink
- << "(Induction step: " << StrideAPtr << ")\n");
- DEBUG(dbgs() << "LV: Distance for " << *InstMap[AIdx] << " to "
- << *InstMap[BIdx] << ": " << *Dist << "\n");
-
- // Need consecutive accesses. We don't want to vectorize
- // "A[B[i]] += ..." and similar code or pointer arithmetic that could wrap in
- // the address space.
- if (!StrideAPtr || !StrideBPtr || StrideAPtr != StrideBPtr){
- DEBUG(dbgs() << "Non-consecutive pointer access\n");
- return true;
- }
-
- const SCEVConstant *C = dyn_cast<SCEVConstant>(Dist);
- if (!C) {
- DEBUG(dbgs() << "LV: Dependence because of non-constant distance\n");
- ShouldRetryWithRuntimeCheck = true;
- return true;
- }
-
- Type *ATy = APtr->getType()->getPointerElementType();
- Type *BTy = BPtr->getType()->getPointerElementType();
- unsigned TypeByteSize = DL->getTypeAllocSize(ATy);
-
- // Negative distances are not plausible dependencies.
- const APInt &Val = C->getValue()->getValue();
- if (Val.isNegative()) {
- bool IsTrueDataDependence = (AIsWrite && !BIsWrite);
- if (IsTrueDataDependence &&
- (couldPreventStoreLoadForward(Val.abs().getZExtValue(), TypeByteSize) ||
- ATy != BTy))
- return true;
-
- DEBUG(dbgs() << "LV: Dependence is negative: NoDep\n");
+ if (LAI->hasStoreToLoopInvariantAddress()) {
+ emitAnalysis(
+ VectorizationReport()
+ << "write to a loop invariant address could not be vectorized");
+ DEBUG(dbgs() << "LV: We don't allow storing to uniform addresses\n");
return false;
}
- // Write to the same location with the same size.
- // Could be improved to assert type sizes are the same (i32 == float, etc).
- if (Val == 0) {
- if (ATy == BTy)
- return false;
- DEBUG(dbgs() << "LV: Zero dependence difference but different types\n");
- return true;
- }
-
- assert(Val.isStrictlyPositive() && "Expect a positive value");
-
- // Positive distance bigger than max vectorization factor.
- if (ATy != BTy) {
- DEBUG(dbgs() <<
- "LV: ReadWrite-Write positive dependency with different types\n");
+ if (LAI->getNumRuntimePointerChecks() >
+ VectorizerParams::RuntimeMemoryCheckThreshold) {
+ emitAnalysis(VectorizationReport()
+ << LAI->getNumRuntimePointerChecks() << " exceeds limit of "
+ << VectorizerParams::RuntimeMemoryCheckThreshold
+ << " dependent memory operations checked at runtime");
+ DEBUG(dbgs() << "LV: Too many memory checks needed.\n");
return false;
}
-
- unsigned Distance = (unsigned) Val.getZExtValue();
-
- // Bail out early if passed-in parameters make vectorization not feasible.
- unsigned ForcedFactor = (VectParams.VectorizationFactor ?
- VectParams.VectorizationFactor : 1);
- unsigned ForcedUnroll = (VectParams.VectorizationInterleave ?
- VectParams.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
- // bigger than the currrent maximum size.
- if (Distance < 2*TypeByteSize ||
- 2*TypeByteSize > MaxSafeDepDistBytes ||
- Distance < TypeByteSize * ForcedUnroll * ForcedFactor) {
- DEBUG(dbgs() << "LV: Failure because of Positive distance "
- << Val.getSExtValue() << '\n');
- return true;
- }
-
- MaxSafeDepDistBytes = Distance < MaxSafeDepDistBytes ?
- Distance : MaxSafeDepDistBytes;
-
- bool IsTrueDataDependence = (!AIsWrite && BIsWrite);
- if (IsTrueDataDependence &&
- couldPreventStoreLoadForward(Distance, TypeByteSize))
- return true;
-
- DEBUG(dbgs() << "LV: Positive distance " << Val.getSExtValue() <<
- " with max VF = " << MaxSafeDepDistBytes / TypeByteSize << '\n');
-
- return false;
-}
-
-bool MemoryDepChecker::areDepsSafe(AccessAnalysis::DepCandidates &AccessSets,
- MemAccessInfoSet &CheckDeps,
- ValueToValueMap &Strides) {
-
- MaxSafeDepDistBytes = -1U;
- while (!CheckDeps.empty()) {
- MemAccessInfo CurAccess = *CheckDeps.begin();
-
- // Get the relevant memory access set.
- EquivalenceClasses<MemAccessInfo>::iterator I =
- AccessSets.findValue(AccessSets.getLeaderValue(CurAccess));
-
- // Check accesses within this set.
- EquivalenceClasses<MemAccessInfo>::member_iterator AI, AE;
- AI = AccessSets.member_begin(I), AE = AccessSets.member_end();
-
- // Check every access pair.
- while (AI != AE) {
- CheckDeps.erase(*AI);
- EquivalenceClasses<MemAccessInfo>::member_iterator OI = std::next(AI);
- while (OI != AE) {
- // Check every accessing instruction pair in program order.
- for (std::vector<unsigned>::iterator I1 = Accesses[*AI].begin(),
- I1E = Accesses[*AI].end(); I1 != I1E; ++I1)
- for (std::vector<unsigned>::iterator I2 = Accesses[*OI].begin(),
- I2E = Accesses[*OI].end(); I2 != I2E; ++I2) {
- if (*I1 < *I2 && isDependent(*AI, *I1, *OI, *I2, Strides))
- return false;
- if (*I2 < *I1 && isDependent(*OI, *I2, *AI, *I1, Strides))
- return false;
- }
- ++OI;
- }
- AI++;
- }
- }
return true;
}
-bool LoopAccessAnalysis::canVectorizeMemory(ValueToValueMap &Strides) {
-
- typedef SmallVector<Value*, 16> ValueVector;
- typedef SmallPtrSet<Value*, 16> ValueSet;
-
- // Holds the Load and Store *instructions*.
- ValueVector Loads;
- ValueVector Stores;
-
- // Holds all the different accesses in the loop.
- unsigned NumReads = 0;
- unsigned NumReadWrites = 0;
-
- PtrRtCheck.Pointers.clear();
- PtrRtCheck.Need = false;
-
- const bool IsAnnotatedParallel = TheLoop->isAnnotatedParallel();
- MemoryDepChecker DepChecker(SE, DL, TheLoop, VectParams);
-
- // For each block.
- for (Loop::block_iterator bb = TheLoop->block_begin(),
- be = TheLoop->block_end(); bb != be; ++bb) {
-
- // Scan the BB and collect legal loads and stores.
- for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e;
- ++it) {
-
- // If this is a load, save it. If this instruction can read from memory
- // but is not a load, then we quit. Notice that we don't handle function
- // calls that read or write.
- if (it->mayReadFromMemory()) {
- // 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.
- CallInst *Call = dyn_cast<CallInst>(it);
- if (Call && getIntrinsicIDForCall(Call, TLI))
- continue;
-
- LoadInst *Ld = dyn_cast<LoadInst>(it);
- if (!Ld || (!Ld->isSimple() && !IsAnnotatedParallel)) {
- emitAnalysis(VectorizationReport(Ld)
- << "read with atomic ordering or volatile read");
- DEBUG(dbgs() << "LV: Found a non-simple load.\n");
- return false;
- }
- NumLoads++;
- Loads.push_back(Ld);
- DepChecker.addAccess(Ld);
- continue;
- }
-
- // Save 'store' instructions. Abort if other instructions write to memory.
- if (it->mayWriteToMemory()) {
- StoreInst *St = dyn_cast<StoreInst>(it);
- if (!St) {
- emitAnalysis(VectorizationReport(it) <<
- "instruction cannot be vectorized");
- return false;
- }
- if (!St->isSimple() && !IsAnnotatedParallel) {
- emitAnalysis(VectorizationReport(St)
- << "write with atomic ordering or volatile write");
- DEBUG(dbgs() << "LV: Found a non-simple store.\n");
- return false;
- }
- NumStores++;
- Stores.push_back(St);
- DepChecker.addAccess(St);
- }
- } // Next instr.
- } // Next block.
-
- // 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()) {
- DEBUG(dbgs() << "LV: Found a read-only loop!\n");
- return true;
- }
-
- AccessAnalysis::DepCandidates DependentAccesses;
- AccessAnalysis Accesses(DL, AA, DependentAccesses);
-
- // 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.
- ValueSet Seen;
-
- ValueVector::iterator I, IE;
- for (I = Stores.begin(), IE = Stores.end(); I != IE; ++I) {
- StoreInst *ST = cast<StoreInst>(*I);
- Value* Ptr = ST->getPointerOperand();
-
- if (isUniform(Ptr)) {
- emitAnalysis(
- VectorizationReport(ST)
- << "write to a loop invariant address could not be vectorized");
- DEBUG(dbgs() << "LV: We don't allow storing to uniform addresses\n");
- return false;
- }
-
- // 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).second) {
- ++NumReadWrites;
-
- AliasAnalysis::Location Loc = AA->getLocation(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()))
- Loc.AATags.TBAA = nullptr;
-
- Accesses.addStore(Loc);
- }
- }
-
- if (IsAnnotatedParallel) {
- DEBUG(dbgs()
- << "LV: A loop annotated parallel, ignore memory dependency "
- << "checks.\n");
- return true;
- }
-
- for (I = Loads.begin(), IE = Loads.end(); I != IE; ++I) {
- LoadInst *LD = cast<LoadInst>(*I);
- 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 consecutive.
- // If the address of i is unknown (for example A[B[i]]) then we may
- // 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).second ||
- !isStridedPtr(SE, DL, Ptr, TheLoop, Strides)) {
- ++NumReads;
- IsReadOnlyPtr = true;
- }
-
- AliasAnalysis::Location Loc = AA->getLocation(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()))
- Loc.AATags.TBAA = nullptr;
-
- Accesses.addLoad(Loc, 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.
- if (NumReadWrites == 1 && NumReads == 0) {
- DEBUG(dbgs() << "LV: Found a write-only loop!\n");
- return true;
- }
-
- // Build dependence sets and check whether we need a runtime pointer bounds
- // check.
- Accesses.buildDependenceSets();
- bool NeedRTCheck = Accesses.isRTCheckNeeded();
-
- // Find pointers with computable bounds. We are going to use this information
- // to place a runtime bound check.
- unsigned NumComparisons = 0;
- bool CanDoRT = false;
- if (NeedRTCheck)
- CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE, TheLoop,
- Strides);
-
- DEBUG(dbgs() << "LV: We need to do " << NumComparisons <<
- " pointer comparisons.\n");
-
- // If we only have one set of dependences to check pointers among we don't
- // need a runtime check.
- if (NumComparisons == 0 && NeedRTCheck)
- NeedRTCheck = false;
-
- // Check that we did not collect too many pointers or found an unsizeable
- // pointer.
- if (!CanDoRT || NumComparisons > VectParams.RuntimeMemoryCheckThreshold) {
- PtrRtCheck.reset();
- CanDoRT = false;
- }
-
- if (CanDoRT) {
- DEBUG(dbgs() << "LV: We can perform a memory runtime check if needed.\n");
- }
-
- if (NeedRTCheck && !CanDoRT) {
- emitAnalysis(VectorizationReport() << "cannot identify array bounds");
- DEBUG(dbgs() << "LV: We can't vectorize because we can't find " <<
- "the array bounds.\n");
- PtrRtCheck.reset();
- return false;
- }
-
- PtrRtCheck.Need = NeedRTCheck;
-
- bool CanVecMem = true;
- if (Accesses.isDependencyCheckNeeded()) {
- DEBUG(dbgs() << "LV: Checking memory dependencies\n");
- CanVecMem = DepChecker.areDepsSafe(
- DependentAccesses, Accesses.getDependenciesToCheck(), Strides);
- MaxSafeDepDistBytes = DepChecker.getMaxSafeDepDistBytes();
-
- if (!CanVecMem && DepChecker.shouldRetryWithRuntimeCheck()) {
- DEBUG(dbgs() << "LV: Retrying with memory checks\n");
- NeedRTCheck = true;
-
- // Clear the dependency checks. We assume they are not needed.
- Accesses.resetDepChecks();
-
- PtrRtCheck.reset();
- PtrRtCheck.Need = true;
-
- CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE,
- TheLoop, Strides, true);
- // Check that we did not collect too many pointers or found an unsizeable
- // pointer.
- if (!CanDoRT || NumComparisons > VectParams.RuntimeMemoryCheckThreshold) {
- if (!CanDoRT && NumComparisons > 0)
- emitAnalysis(VectorizationReport()
- << "cannot check memory dependencies at runtime");
- else
- emitAnalysis(VectorizationReport()
- << NumComparisons << " exceeds limit of "
- << VectParams.RuntimeMemoryCheckThreshold
- << " dependent memory operations checked at runtime");
- DEBUG(dbgs() << "LV: Can't vectorize with memory checks\n");
- PtrRtCheck.reset();
- return false;
- }
-
- CanVecMem = true;
- }
- }
-
- if (!CanVecMem)
- emitAnalysis(VectorizationReport() <<
- "unsafe dependent memory operations in loop");
-
- DEBUG(dbgs() << "LV: We" << (NeedRTCheck ? "" : " don't") <<
- " need a runtime memory check.\n");
-
- return CanVecMem;
-}
-
-bool LoopVectorizationLegality::canVectorizeMemory() {
- return LAA.canVectorizeMemory(Strides);
-}
-
static bool hasMultipleUsesOf(Instruction *I,
SmallPtrSetImpl<Instruction *> &Insts) {
unsigned NumUses = 0;
}
}
-LoopVectorizationLegality::InductionKind
-LoopVectorizationLegality::isInductionVariable(PHINode *Phi,
- ConstantInt *&StepValue) {
+bool llvm::isInductionPHI(PHINode *Phi, ScalarEvolution *SE,
+ ConstantInt *&StepValue) {
Type *PhiTy = Phi->getType();
// We only handle integer and pointer inductions variables.
if (!PhiTy->isIntegerTy() && !PhiTy->isPointerTy())
- return IK_NoInduction;
+ return false;
// Check that the PHI is consecutive.
const SCEV *PhiScev = SE->getSCEV(Phi);
const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(PhiScev);
if (!AR) {
DEBUG(dbgs() << "LV: PHI is not a poly recurrence.\n");
- return IK_NoInduction;
+ return false;
}
const SCEV *Step = AR->getStepRecurrence(*SE);
// Calculate the pointer stride and check if it is consecutive.
const SCEVConstant *C = dyn_cast<SCEVConstant>(Step);
if (!C)
- return IK_NoInduction;
+ return false;
ConstantInt *CV = C->getValue();
if (PhiTy->isIntegerTy()) {
StepValue = CV;
- return IK_IntInduction;
+ return true;
}
assert(PhiTy->isPointerTy() && "The PHI must be a pointer");
// The pointer stride cannot be determined if the pointer element type is not
// sized.
if (!PointerElementType->isSized())
- return IK_NoInduction;
+ return false;
- int64_t Size = static_cast<int64_t>(DL->getTypeAllocSize(PointerElementType));
+ const DataLayout &DL = Phi->getModule()->getDataLayout();
+ int64_t Size = static_cast<int64_t>(DL.getTypeAllocSize(PointerElementType));
int64_t CVSize = CV->getSExtValue();
if (CVSize % Size)
- return IK_NoInduction;
+ return false;
StepValue = ConstantInt::getSigned(CV->getType(), CVSize / Size);
+ return true;
+}
+
+LoopVectorizationLegality::InductionKind
+LoopVectorizationLegality::isInductionVariable(PHINode *Phi,
+ ConstantInt *&StepValue) {
+ if (!isInductionPHI(Phi, SE, StepValue))
+ return IK_NoInduction;
+
+ Type *PhiTy = Phi->getType();
+ // Found an Integer induction variable.
+ if (PhiTy->isIntegerTy())
+ return IK_IntInduction;
+ // Found an Pointer induction variable.
return IK_PtrInduction;
}
return Inductions.count(PN);
}
-bool LoopAccessAnalysis::blockNeedsPredication(BasicBlock *BB) {
- assert(TheLoop->contains(BB) && "Unknown block used");
-
- // Blocks that do not dominate the latch need predication.
- BasicBlock* Latch = TheLoop->getLoopLatch();
- return !DT->dominates(BB, Latch);
-}
-
bool LoopVectorizationLegality::blockNeedsPredication(BasicBlock *BB) {
- return LAA.blockNeedsPredication(BB);
+ return LoopAccessInfo::blockNeedsPredication(BB, TheLoop, DT);
}
bool LoopVectorizationLegality::blockCanBePredicated(BasicBlock *BB,
unsigned LoopVectorizationCostModel::getWidestType() {
unsigned MaxWidth = 8;
+ const DataLayout &DL = TheFunction->getParent()->getDataLayout();
// For each block.
for (Loop::block_iterator bb = TheLoop->block_begin(),
continue;
MaxWidth = std::max(MaxWidth,
- (unsigned)DL->getTypeSizeInBits(T->getScalarType()));
+ (unsigned)DL.getTypeSizeInBits(T->getScalarType()));
}
}
return SmallUF;
}
+ // Unroll if this is a large loop (small loops are already dealt with by this
+ // point) that could benefit from interleaved unrolling.
+ bool HasReductions = (Legal->getReductionVars()->size() > 0);
+ if (TTI.enableAggressiveInterleaving(HasReductions)) {
+ DEBUG(dbgs() << "LV: Unrolling to expose ILP.\n");
+ return UF;
+ }
+
DEBUG(dbgs() << "LV: Not Unrolling.\n");
return 1;
}
// Scalarized loads/stores.
int ConsecutiveStride = Legal->isConsecutivePtr(Ptr);
bool Reverse = ConsecutiveStride < 0;
- unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ValTy);
- unsigned VectorElementSize = DL->getTypeStoreSize(VectorTy)/VF;
+ const DataLayout &DL = I->getModule()->getDataLayout();
+ unsigned ScalarAllocatedSize = DL.getTypeAllocSize(ValTy);
+ unsigned VectorElementSize = DL.getTypeStoreSize(VectorTy) / VF;
if (!ConsecutiveStride || ScalarAllocatedSize != VectorElementSize) {
bool IsComplexComputation =
isLikelyComplexAddressComputation(Ptr, Legal, SE, TheLoop);
// Wide load/stores.
unsigned Cost = TTI.getAddressComputationCost(VectorTy);
- Cost += TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS);
+ if (Legal->isMaskRequired(I))
+ Cost += TTI.getMaskedMemoryOpCost(I->getOpcode(), VectorTy, Alignment,
+ AS);
+ else
+ Cost += TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS);
if (Reverse)
Cost += TTI.getShuffleCost(TargetTransformInfo::SK_Reverse,
return TTI.getCastInstrCost(I->getOpcode(), VectorTy, SrcVecTy);
}
case Instruction::Call: {
+ bool NeedToScalarize;
CallInst *CI = cast<CallInst>(I);
- Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
- assert(ID && "Not an intrinsic call!");
- Type *RetTy = ToVectorTy(CI->getType(), VF);
- SmallVector<Type*, 4> Tys;
- for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i)
- Tys.push_back(ToVectorTy(CI->getArgOperand(i)->getType(), VF));
- return TTI.getIntrinsicInstrCost(ID, RetTy, Tys);
+ unsigned CallCost = getVectorCallCost(CI, VF, TTI, TLI, NeedToScalarize);
+ if (getIntrinsicIDForCall(CI, TLI))
+ return std::min(CallCost, getVectorIntrinsicCost(CI, VF, TTI, TLI));
+ return CallCost;
}
default: {
// We are scalarizing the instruction. Return the cost of the scalar
}// end of switch.
}
-Type* LoopVectorizationCostModel::ToVectorTy(Type *Scalar, unsigned VF) {
- if (Scalar->isVoidTy() || VF == 1)
- return Scalar;
- return VectorType::get(Scalar, VF);
-}
-
char LoopVectorize::ID = 0;
static const char lv_name[] = "Loop Vectorization";
INITIALIZE_PASS_BEGIN(LoopVectorize, LV_NAME, lv_name, false, false)
INITIALIZE_PASS_DEPENDENCY(LCSSA)
INITIALIZE_PASS_DEPENDENCY(LoopInfoWrapperPass)
INITIALIZE_PASS_DEPENDENCY(LoopSimplify)
+INITIALIZE_PASS_DEPENDENCY(LoopAccessAnalysis)
INITIALIZE_PASS_END(LoopVectorize, LV_NAME, lv_name, false, false)
namespace llvm {
projects / oota-llvm.git / blobdiff