#include "llvm/ADT/Statistic.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/Analysis/AliasAnalysis.h"
+#include "llvm/Analysis/AliasSetTracker.h"
+#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"
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."));
+ VectorizationFactor("force-vector-width", cl::init(0), cl::Hidden,
+ 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. "
- "Zero is autoselect."));
+ 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,
"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(
"enable-cond-stores-vec", cl::init(false), cl::Hidden,
cl::desc("Enable if predication of stores during vectorization."));
+static cl::opt<unsigned> MaxNestedScalarReductionUF(
+ "max-nested-scalar-reduction-unroll", cl::init(2), cl::Hidden,
+ cl::desc("The maximum unroll factor to use when unrolling a scalar "
+ "reduction in a nested loop."));
+
namespace {
// Forward declarations.
class LoopVectorizationLegality;
class LoopVectorizationCostModel;
-
-/// Optimization analysis message produced during vectorization. Messages inform
-/// the user why vectorization did not occur.
-class Report {
- std::string Message;
- raw_string_ostream Out;
- Instruction *Instr;
-
-public:
- Report(Instruction *I = nullptr) : Out(Message), Instr(I) {
- Out << "loop not vectorized: ";
- }
-
- template <typename A> Report &operator<<(const A &Value) {
- Out << Value;
- return *this;
- }
-
- Instruction *getInstr() { return Instr; }
-
- std::string &str() { return Out.str(); }
- operator Twine() { return Out.str(); }
-};
+class LoopVectorizeHints;
/// InnerLoopVectorizer vectorizes loops which contain only one basic
/// block to a specified vectorization factor (VF).
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
/// element.
virtual Value *getBroadcastInstrs(Value *V);
- /// This function adds 0, 1, 2 ... to each vector element, starting at zero.
- /// If Negate is set then negative numbers are added e.g. (0, -1, -2, ...).
- /// The sequence starts at StartIndex.
- virtual Value *getConsecutiveVector(Value* Val, int StartIdx, bool Negate);
+ /// This function adds (StartIdx, StartIdx + Step, StartIdx + 2*Step, ...)
+ /// to each vector element of Val. The sequence starts at StartIndex.
+ virtual Value *getStepVector(Value *Val, int StartIdx, Value *Step);
/// When we go over instructions in the basic block we rely on previous
/// values within the current basic block or on loop invariant values.
LoopInfo *LI;
/// Dominator Tree.
DominatorTree *DT;
+ /// Alias Analysis.
+ AliasAnalysis *AA;
/// Data Layout.
const DataLayout *DL;
/// Target Library Info.
bool IfPredicateStore = false) override;
void vectorizeMemoryInstruction(Instruction *Instr) override;
Value *getBroadcastInstrs(Value *V) override;
- Value *getConsecutiveVector(Value* Val, int StartIdx, bool Negate) override;
+ Value *getStepVector(Value *Val, int StartIdx, Value *Step) override;
Value *reverseVector(Value *Vec) override;
};
// non-speculated memory access when the condition was false, this would be
// caught by the runtime overlap checks).
if (Kind != LLVMContext::MD_tbaa &&
+ Kind != LLVMContext::MD_alias_scope &&
+ Kind != LLVMContext::MD_noalias &&
Kind != LLVMContext::MD_fpmath)
continue;
/// induction variable and the different reduction variables.
class LoopVectorizationLegality {
public:
- unsigned NumLoads;
- unsigned NumStores;
- unsigned NumPredStores;
-
LoopVectorizationLegality(Loop *L, ScalarEvolution *SE, const DataLayout *DL,
DominatorTree *DT, TargetLibraryInfo *TLI,
- Function *F)
- : NumLoads(0), NumStores(0), NumPredStores(0), TheLoop(L), SE(SE), DL(DL),
- DT(DT), TLI(TLI), TheFunction(F), Induction(nullptr),
- WidestIndTy(nullptr), HasFunNoNaNAttr(false), MaxSafeDepDistBytes(-1U) {
- }
+ AliasAnalysis *AA, Function *F,
+ const TargetTransformInfo *TTI)
+ : NumPredStores(0), TheLoop(L), SE(SE), DL(DL), TLI(TLI), TheFunction(F),
+ TTI(TTI), DT(DT), Induction(nullptr), WidestIndTy(nullptr),
+ LAI(F, L, SE, DL, TLI, AA, DT,
+ LoopAccessInfo::VectorizerParams(
+ MaxVectorWidth, VectorizationFactor, VectorizationInterleave,
+ RuntimeMemoryCheckThreshold)),
+ HasFunNoNaNAttr(false) {}
/// This enum represents the kinds of reductions that we support.
enum ReductionKind {
/// This enum represents the kinds of inductions that we support.
enum InductionKind {
- IK_NoInduction, ///< Not an induction variable.
- IK_IntInduction, ///< Integer induction variable. Step = 1.
- IK_ReverseIntInduction, ///< Reverse int induction variable. Step = -1.
- IK_PtrInduction, ///< Pointer induction var. Step = sizeof(elem).
- IK_ReversePtrInduction ///< Reverse ptr indvar. Step = - sizeof(elem).
+ IK_NoInduction, ///< Not an induction variable.
+ IK_IntInduction, ///< Integer induction variable. Step = C.
+ IK_PtrInduction ///< Pointer induction var. Step = C / sizeof(elem).
};
// This enum represents the kind of minmax reduction.
MinMaxReductionKind MinMaxKind;
};
- /// 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();
+ /// A struct for saving information about induction variables.
+ struct InductionInfo {
+ InductionInfo(Value *Start, InductionKind K, ConstantInt *Step)
+ : StartValue(Start), IK(K), StepValue(Step) {
+ assert(IK != IK_NoInduction && "Not an induction");
+ assert(StartValue && "StartValue is null");
+ assert(StepValue && !StepValue->isZero() && "StepValue is zero");
+ assert((IK != IK_PtrInduction || StartValue->getType()->isPointerTy()) &&
+ "StartValue is not a pointer for pointer induction");
+ assert((IK != IK_IntInduction || StartValue->getType()->isIntegerTy()) &&
+ "StartValue is not an integer for integer induction");
+ assert(StepValue->getType()->isIntegerTy() &&
+ "StepValue is not an integer");
+ }
+ InductionInfo()
+ : StartValue(nullptr), IK(IK_NoInduction), StepValue(nullptr) {}
+
+ /// Get the consecutive direction. Returns:
+ /// 0 - unknown or non-consecutive.
+ /// 1 - consecutive and increasing.
+ /// -1 - consecutive and decreasing.
+ int getConsecutiveDirection() const {
+ if (StepValue && (StepValue->isOne() || StepValue->isMinusOne()))
+ return StepValue->getSExtValue();
+ return 0;
}
- /// Insert a pointer and calculate the start and end SCEVs.
- void insert(ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr,
- unsigned DepSetId, 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;
- };
+ /// Compute the transformed value of Index at offset StartValue using step
+ /// StepValue.
+ /// For integer induction, returns StartValue + Index * StepValue.
+ /// For pointer induction, returns StartValue[Index * StepValue].
+ /// FIXME: The newly created binary instructions should contain nsw/nuw
+ /// flags, which can be found from the original scalar operations.
+ Value *transform(IRBuilder<> &B, Value *Index) const {
+ switch (IK) {
+ case IK_IntInduction:
+ assert(Index->getType() == StartValue->getType() &&
+ "Index type does not match StartValue type");
+ if (StepValue->isMinusOne())
+ return B.CreateSub(StartValue, Index);
+ if (!StepValue->isOne())
+ Index = B.CreateMul(Index, StepValue);
+ return B.CreateAdd(StartValue, Index);
+
+ case IK_PtrInduction:
+ if (StepValue->isMinusOne())
+ Index = B.CreateNeg(Index);
+ else if (!StepValue->isOne())
+ Index = B.CreateMul(Index, StepValue);
+ return B.CreateGEP(StartValue, Index);
+
+ case IK_NoInduction:
+ return nullptr;
+ }
+ llvm_unreachable("invalid enum");
+ }
- /// A struct for saving information about induction variables.
- struct InductionInfo {
- InductionInfo(Value *Start, InductionKind K) : StartValue(Start), IK(K) {}
- InductionInfo() : StartValue(nullptr), IK(IK_NoInduction) {}
/// Start value.
TrackingVH<Value> StartValue;
/// Induction kind.
InductionKind IK;
+ /// Step value.
+ ConstantInt *StepValue;
};
/// ReductionList contains the reduction descriptors for all
bool isUniformAfterVectorization(Instruction* I) { return Uniforms.count(I); }
/// Returns the information that we collected about runtime memory check.
- RuntimePointerCheck *getRuntimePointerCheck() { return &PtrRtCheck; }
+ LoopAccessInfo::RuntimePointerCheck *getRuntimePointerCheck() {
+ return LAI.getRuntimePointerCheck();
+ }
+
+ LoopAccessInfo *getLAI() { 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 MaxSafeDepDistBytes; }
+ unsigned getMaxSafeDepDistBytes() { return LAI.getMaxSafeDepDistBytes(); }
bool hasStride(Value *V) { return StrideSet.count(V); }
bool mustCheckStrides() { return !StrideSet.empty(); }
}
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); }
+ unsigned getNumStores() const { return LAI.getNumStores(); }
+ unsigned getNumLoads() const { return LAI.getNumLoads(); }
+ unsigned getNumPredStores() const { return NumPredStores; }
+
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
/// Return true if all of the instructions in the block can be speculatively
/// executed. \p SafePtrs is a list of addresses that are known to be legal
/// and we know that we can read from them without segfault.
- bool blockCanBePredicated(BasicBlock *BB, SmallPtrSet<Value *, 8>& SafePtrs);
+ bool blockCanBePredicated(BasicBlock *BB, SmallPtrSetImpl<Value *> &SafePtrs);
/// Returns True, if 'Phi' is the kind of reduction variable for type
/// 'Kind'. If this is a reduction variable, it adds it to ReductionList.
/// pattern corresponding to a min(X, Y) or max(X, Y).
static ReductionInstDesc isMinMaxSelectCmpPattern(Instruction *I,
ReductionInstDesc &Prev);
- /// Returns the induction kind of Phi. This function may return NoInduction
- /// if the PHI is not an induction variable.
- InductionKind isInductionVariable(PHINode *Phi);
+ /// Returns the induction kind of Phi and record the step. This function may
+ /// return NoInduction if the PHI is not an induction variable.
+ InductionKind isInductionVariable(PHINode *Phi, ConstantInt *&StepValue);
/// \brief Collect memory access with loop invariant strides.
///
/// Looks for accesses like "a[i * StrideA]" where "StrideA" is loop
/// invariant.
- void collectStridedAcccess(Value *LoadOrStoreInst);
+ void collectStridedAccess(Value *LoadOrStoreInst);
/// Report an analysis message to assist the user in diagnosing loops that are
/// not vectorized.
- void emitAnalysis(Report &Message) {
- DebugLoc DL = TheLoop->getStartLoc();
- if (Instruction *I = Message.getInstr())
- DL = I->getDebugLoc();
- emitOptimizationRemarkAnalysis(TheFunction->getContext(), DEBUG_TYPE,
- *TheFunction, DL, Message.str());
+ void emitAnalysis(VectorizationReport &Message) {
+ VectorizationReport::emitAnalysis(Message, TheFunction, TheLoop);
}
+ unsigned NumPredStores;
+
/// The loop that we evaluate.
Loop *TheLoop;
/// Scev analysis.
ScalarEvolution *SE;
/// DataLayout analysis.
const DataLayout *DL;
- /// Dominators.
- DominatorTree *DT;
/// Target Library Info.
TargetLibraryInfo *TLI;
/// Parent function
Function *TheFunction;
+ /// Target Transform Info
+ const TargetTransformInfo *TTI;
+ /// Dominator Tree.
+ DominatorTree *DT;
// --- vectorization state --- //
SmallPtrSet<Value*, 4> AllowedExit;
/// This set holds the variables which are known to be uniform after
/// vectorization.
- SmallPtrSet<Instruction*, 4> Uniforms;
- /// We need to check that all of the pointers in this list are disjoint
- /// at runtime.
- RuntimePointerCheck PtrRtCheck;
+ SmallPtrSet<Instruction *, 4> Uniforms;
+ LoopAccessInfo LAI;
/// Can we assume the absence of NaNs.
bool HasFunNoNaNAttr;
- unsigned MaxSafeDepDistBytes;
-
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
LoopVectorizationCostModel(Loop *L, ScalarEvolution *SE, LoopInfo *LI,
LoopVectorizationLegality *Legal,
const TargetTransformInfo &TTI,
- const DataLayout *DL, const TargetLibraryInfo *TLI)
- : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), DL(DL), TLI(TLI) {}
+ 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),
+ TheFunction(F), Hints(Hints) {
+ CodeMetrics::collectEphemeralValues(L, AC, EphValues);
+ }
/// Information about vectorization costs
struct VectorizationFactor {
/// This method checks every power of two up to VF. If UserVF is not ZERO
/// then this vectorization factor will be selected if vectorization is
/// possible.
- VectorizationFactor selectVectorizationFactor(bool OptForSize,
- unsigned UserVF,
- bool ForceVectorization);
+ VectorizationFactor selectVectorizationFactor(bool OptForSize);
/// \return The size (in bits) of the widest type in the code that
/// needs to be vectorized. We ignore values that remain scalar such as
/// based on register pressure and other parameters.
/// VF and LoopCost are the selected vectorization factor and the cost of the
/// selected VF.
- unsigned selectUnrollFactor(bool OptForSize, unsigned UserUF, unsigned VF,
- unsigned LoopCost);
+ unsigned selectUnrollFactor(bool OptForSize, unsigned VF, unsigned LoopCost);
/// \brief A struct that represents some properties of the register usage
/// of a loop.
/// 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);
+ }
+
+ /// Values used only by @llvm.assume calls.
+ SmallPtrSet<const Value *, 32> EphValues;
+
/// The loop that we evaluate.
Loop *TheLoop;
/// Scev analysis.
const DataLayout *DL;
/// Target Library Info.
const TargetLibraryInfo *TLI;
+ const Function *TheFunction;
+ // Loop Vectorize Hint.
+ const LoopVectorizeHints *Hints;
};
/// Utility class for getting and setting loop vectorizer hints in the form
/// of loop metadata.
+/// This class keeps a number of loop annotations locally (as member variables)
+/// and can, upon request, write them back as metadata on the loop. It will
+/// initially scan the loop for existing metadata, and will update the local
+/// values based on information in the loop.
+/// We cannot write all values to metadata, as the mere presence of some info,
+/// for example 'force', means a decision has been made. So, we need to be
+/// careful NOT to add them if the user hasn't specifically asked so.
class LoopVectorizeHints {
+ enum HintKind {
+ HK_WIDTH,
+ HK_UNROLL,
+ HK_FORCE
+ };
+
+ /// Hint - associates name and validation with the hint value.
+ struct Hint {
+ const char * Name;
+ unsigned Value; // This may have to change for non-numeric values.
+ HintKind Kind;
+
+ Hint(const char * Name, unsigned Value, HintKind Kind)
+ : Name(Name), Value(Value), Kind(Kind) { }
+
+ bool validate(unsigned Val) {
+ switch (Kind) {
+ case HK_WIDTH:
+ return isPowerOf2_32(Val) && Val <= MaxVectorWidth;
+ case HK_UNROLL:
+ return isPowerOf2_32(Val) && Val <= MaxInterleaveFactor;
+ case HK_FORCE:
+ return (Val <= 1);
+ }
+ return false;
+ }
+ };
+
+ /// Vectorization width.
+ Hint Width;
+ /// Vectorization interleave factor.
+ Hint Interleave;
+ /// Vectorization forced
+ Hint Force;
+
+ /// Return the loop metadata prefix.
+ static StringRef Prefix() { return "llvm.loop."; }
+
public:
enum ForceKind {
FK_Undefined = -1, ///< Not selected.
FK_Enabled = 1, ///< Forcing enabled.
};
- LoopVectorizeHints(const Loop *L, bool DisableUnrolling)
- : Width(VectorizationFactor),
- Unroll(DisableUnrolling),
- Force(FK_Undefined),
- LoopID(L->getLoopID()) {
- getHints(L);
- // force-vector-unroll overrides DisableUnrolling.
- if (VectorizationUnroll.getNumOccurrences() > 0)
- Unroll = VectorizationUnroll;
-
- DEBUG(if (DisableUnrolling && Unroll == 1) dbgs()
- << "LV: Unrolling disabled by the pass manager\n");
- }
+ LoopVectorizeHints(const Loop *L, bool DisableInterleaving)
+ : Width("vectorize.width", VectorizationFactor, HK_WIDTH),
+ Interleave("interleave.count", DisableInterleaving, HK_UNROLL),
+ Force("vectorize.enable", FK_Undefined, HK_FORCE),
+ TheLoop(L) {
+ // Populate values with existing loop metadata.
+ getHintsFromMetadata();
- /// Return the loop vectorizer metadata prefix.
- static StringRef Prefix() { return "llvm.loop.vectorize."; }
+ // force-vector-interleave overrides DisableInterleaving.
+ if (VectorizationInterleave.getNumOccurrences() > 0)
+ Interleave.Value = VectorizationInterleave;
- MDNode *createHint(LLVMContext &Context, StringRef Name, unsigned V) const {
- 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);
+ 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(Loop *L) {
- LLVMContext &Context = L->getHeader()->getContext();
-
- Width = 1;
-
- // Create a new loop id with one more operand for the already_vectorized
- // hint. If the loop already has a loop id then copy the existing operands.
- SmallVector<Value*, 4> Vals(1);
- if (LoopID)
- for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i)
- Vals.push_back(LoopID->getOperand(i));
-
- Vals.push_back(createHint(Context, Twine(Prefix(), "width").str(), Width));
- Vals.push_back(createHint(Context, Twine(Prefix(), "unroll").str(), 1));
-
- MDNode *NewLoopID = MDNode::get(Context, Vals);
- // Set operand 0 to refer to the loop id itself.
- NewLoopID->replaceOperandWith(0, NewLoopID);
-
- L->setLoopID(NewLoopID);
- if (LoopID)
- LoopID->replaceAllUsesWith(NewLoopID);
-
- LoopID = NewLoopID;
+ void setAlreadyVectorized() {
+ Width.Value = Interleave.Value = 1;
+ Hint Hints[] = {Width, Interleave};
+ writeHintsToMetadata(Hints);
}
+ /// Dumps all the hint information.
std::string emitRemark() const {
- Report R;
- R << "vectorization ";
- switch (Force) {
- case LoopVectorizeHints::FK_Disabled:
- R << "is explicitly disabled";
- break;
- case LoopVectorizeHints::FK_Enabled:
- R << "is explicitly enabled";
- if (Width != 0 && Unroll != 0)
- R << " with width " << Width << " and interleave count " << Unroll;
- else if (Width != 0)
- R << " with width " << Width;
- else if (Unroll != 0)
- R << " with interleave count " << Unroll;
- break;
- case LoopVectorizeHints::FK_Undefined:
- R << "was not specified";
- break;
+ VectorizationReport R;
+ if (Force.Value == LoopVectorizeHints::FK_Disabled)
+ R << "vectorization is explicitly disabled";
+ else {
+ R << "use -Rpass-analysis=loop-vectorize for more info";
+ if (Force.Value == LoopVectorizeHints::FK_Enabled) {
+ R << " (Force=true";
+ if (Width.Value != 0)
+ R << ", Vector Width=" << Width.Value;
+ if (Interleave.Value != 0)
+ R << ", Interleave Count=" << Interleave.Value;
+ R << ")";
+ }
}
+
return R.str();
}
- unsigned getWidth() const { return Width; }
- unsigned getUnroll() const { return Unroll; }
- enum ForceKind getForce() const { return Force; }
- MDNode *getLoopID() const { return LoopID; }
+ unsigned getWidth() const { return Width.Value; }
+ unsigned getInterleave() const { return Interleave.Value; }
+ enum ForceKind getForce() const { return (ForceKind)Force.Value; }
private:
- /// Find hints specified in the loop metadata.
- void getHints(const Loop *L) {
+ /// Find hints specified in the loop metadata and update local values.
+ void getHintsFromMetadata() {
+ MDNode *LoopID = TheLoop->getLoopID();
if (!LoopID)
return;
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.
if (!S)
continue;
- // Check if the hint starts with the vectorizer prefix.
- StringRef Hint = S->getString();
- if (!Hint.startswith(Prefix()))
- continue;
- // Remove the prefix.
- Hint = Hint.substr(Prefix().size(), StringRef::npos);
-
+ // Check if the hint starts with the loop metadata prefix.
+ StringRef Name = S->getString();
if (Args.size() == 1)
- getHint(Hint, Args[0]);
+ setHint(Name, Args[0]);
}
}
- // Check string hint with one operand.
- void getHint(StringRef Hint, Value *Arg) {
- const ConstantInt *C = dyn_cast<ConstantInt>(Arg);
+ /// Checks string hint with one operand and set value if valid.
+ void setHint(StringRef Name, Metadata *Arg) {
+ if (!Name.startswith(Prefix()))
+ return;
+ Name = Name.substr(Prefix().size(), StringRef::npos);
+
+ const ConstantInt *C = mdconst::dyn_extract<ConstantInt>(Arg);
if (!C) return;
unsigned Val = C->getZExtValue();
- if (Hint == "width") {
- if (isPowerOf2_32(Val) && Val <= MaxVectorWidth)
- Width = Val;
- else
- DEBUG(dbgs() << "LV: ignoring invalid width hint metadata\n");
- } else if (Hint == "unroll") {
- if (isPowerOf2_32(Val) && Val <= MaxUnrollFactor)
- Unroll = Val;
- else
- DEBUG(dbgs() << "LV: ignoring invalid unroll hint metadata\n");
- } else if (Hint == "enable") {
- if (C->getBitWidth() == 1)
- Force = Val == 1 ? LoopVectorizeHints::FK_Enabled
- : LoopVectorizeHints::FK_Disabled;
- else
- DEBUG(dbgs() << "LV: ignoring invalid enable hint metadata\n");
- } else {
- DEBUG(dbgs() << "LV: ignoring unknown hint " << Hint << '\n');
+ Hint *Hints[] = {&Width, &Interleave, &Force};
+ for (auto H : Hints) {
+ if (Name == H->Name) {
+ if (H->validate(Val))
+ H->Value = Val;
+ else
+ DEBUG(dbgs() << "LV: ignoring invalid hint '" << Name << "'\n");
+ break;
+ }
}
}
- /// Vectorization width.
- unsigned Width;
- /// Vectorization unroll factor.
- unsigned Unroll;
- /// Vectorization forced
- enum ForceKind Force;
+ /// Create a new hint from name / value pair.
+ MDNode *createHintMetadata(StringRef Name, unsigned V) const {
+ LLVMContext &Context = TheLoop->getHeader()->getContext();
+ 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, ArrayRef<Hint> HintTypes) {
+ MDString* Name = dyn_cast<MDString>(Node->getOperand(0));
+ if (!Name)
+ return false;
+
+ for (auto H : HintTypes)
+ if (Name->getString().endswith(H.Name))
+ return true;
+ return false;
+ }
+
+ /// Sets current hints into loop metadata, keeping other values intact.
+ void writeHintsToMetadata(ArrayRef<Hint> HintTypes) {
+ if (HintTypes.size() == 0)
+ return;
+
+ // Reserve the first element to LoopID (see below).
+ SmallVector<Metadata *, 4> MDs(1);
+ // If the loop already has metadata, then ignore the existing operands.
+ MDNode *LoopID = TheLoop->getLoopID();
+ if (LoopID) {
+ for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) {
+ MDNode *Node = cast<MDNode>(LoopID->getOperand(i));
+ // If node in update list, ignore old value.
+ if (!matchesHintMetadataName(Node, HintTypes))
+ MDs.push_back(Node);
+ }
+ }
+
+ // Now, add the missing hints.
+ for (auto H : HintTypes)
+ 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, MDs);
+ // Set operand 0 to refer to the loop id itself.
+ NewLoopID->replaceOperandWith(0, NewLoopID);
+
+ TheLoop->setLoopID(NewLoopID);
+ }
- MDNode *LoopID;
+ /// The loop these hints belong to.
+ const Loop *TheLoop;
};
static void emitMissedWarning(Function *F, Loop *L,
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");
DominatorTree *DT;
BlockFrequencyInfo *BFI;
TargetLibraryInfo *TLI;
+ AliasAnalysis *AA;
+ AssumptionCache *AC;
bool DisableUnrolling;
bool AlwaysVectorize;
SE = &getAnalysis<ScalarEvolution>();
DataLayoutPass *DLP = getAnalysisIfAvailable<DataLayoutPass>();
DL = DLP ? &DLP->getDataLayout() : nullptr;
- LI = &getAnalysis<LoopInfo>();
- TTI = &getAnalysis<TargetTransformInfo>();
+ LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
+ TTI = &getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
BFI = &getAnalysis<BlockFrequencyInfo>();
- TLI = getAnalysisIfAvailable<TargetLibraryInfo>();
+ auto *TLIP = getAnalysisIfAvailable<TargetLibraryInfoWrapperPass>();
+ TLI = TLIP ? &TLIP->getTLI() : nullptr;
+ AA = &getAnalysis<AliasAnalysis>();
+ AC = &getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
// 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, 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);
+ LoopVectorizationCostModel CM(L, SE, LI, &LVL, *TTI, DL, TLI, AC, F,
+ &Hints);
// Check the function attributes to find out if this function should be
// optimized for size.
// Select the optimal vectorization factor.
const LoopVectorizationCostModel::VectorizationFactor VF =
- CM.selectVectorizationFactor(OptForSize, Hints.getWidth(),
- Hints.getForce() ==
- LoopVectorizeHints::FK_Enabled);
+ CM.selectVectorizationFactor(OptForSize);
// Select the unroll factor.
const unsigned UF =
- CM.selectUnrollFactor(OptForSize, Hints.getUnroll(), VF.Width, VF.Cost);
+ CM.selectUnrollFactor(OptForSize, VF.Width, VF.Cost);
DEBUG(dbgs() << "LV: Found a vectorizable loop (" << VF.Width << ") in "
<< DebugLocStr << '\n');
}
// Mark the loop as already vectorized to avoid vectorizing again.
- Hints.setAlreadyVectorized(L);
+ Hints.setAlreadyVectorized();
DEBUG(verifyFunction(*L->getHeader()->getParent()));
return true;
}
void getAnalysisUsage(AnalysisUsage &AU) const override {
+ AU.addRequired<AssumptionCacheTracker>();
AU.addRequiredID(LoopSimplifyID);
AU.addRequiredID(LCSSAID);
AU.addRequired<BlockFrequencyInfo>();
AU.addRequired<DominatorTreeWrapperPass>();
- AU.addRequired<LoopInfo>();
+ AU.addRequired<LoopInfoWrapperPass>();
AU.addRequired<ScalarEvolution>();
- AU.addRequired<TargetTransformInfo>();
- AU.addPreserved<LoopInfo>();
+ AU.addRequired<TargetTransformInfoWrapperPass>();
+ AU.addRequired<AliasAnalysis>();
+ 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 LoopVectorizationLegality::RuntimePointerCheck::insert(
- ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr, unsigned DepSetId,
- 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);
-}
-
Value *InnerLoopVectorizer::getBroadcastInstrs(Value *V) {
// We need to place the broadcast of invariant variables outside the loop.
Instruction *Instr = dyn_cast<Instruction>(V);
return Shuf;
}
-Value *InnerLoopVectorizer::getConsecutiveVector(Value* Val, int StartIdx,
- bool Negate) {
+Value *InnerLoopVectorizer::getStepVector(Value *Val, int StartIdx,
+ Value *Step) {
assert(Val->getType()->isVectorTy() && "Must be a vector");
assert(Val->getType()->getScalarType()->isIntegerTy() &&
"Elem must be an integer");
+ assert(Step->getType() == Val->getType()->getScalarType() &&
+ "Step has wrong type");
// Create the types.
Type *ITy = Val->getType()->getScalarType();
VectorType *Ty = cast<VectorType>(Val->getType());
SmallVector<Constant*, 8> Indices;
// Create a vector of consecutive numbers from zero to VF.
- for (int i = 0; i < VLen; ++i) {
- int64_t Idx = Negate ? (-i) : i;
- Indices.push_back(ConstantInt::get(ITy, StartIdx + Idx, Negate));
- }
+ for (int i = 0; i < VLen; ++i)
+ Indices.push_back(ConstantInt::get(ITy, StartIdx + i));
// Add the consecutive indices to the vector value.
Constant *Cv = ConstantVector::get(Indices);
assert(Cv->getType() == Val->getType() && "Invalid consecutive vec");
- return Builder.CreateAdd(Val, Cv, "induction");
+ Step = Builder.CreateVectorSplat(VLen, Step);
+ assert(Step->getType() == Val->getType() && "Invalid step vec");
+ // FIXME: The newly created binary instructions should contain nsw/nuw flags,
+ // which can be found from the original scalar operations.
+ Step = Builder.CreateMul(Cv, Step);
+ return Builder.CreateAdd(Val, Step, "induction");
}
/// \brief Find the operand of the GEP that should be checked for consecutive
PHINode *Phi = dyn_cast_or_null<PHINode>(Ptr);
if (Phi && Inductions.count(Phi)) {
InductionInfo II = Inductions[Phi];
- if (IK_PtrInduction == II.IK)
- return 1;
- else if (IK_ReversePtrInduction == II.IK)
- return -1;
+ return II.getConsecutiveDirection();
}
GetElementPtrInst *Gep = dyn_cast_or_null<GetElementPtrInst>(Ptr);
return 0;
InductionInfo II = Inductions[Phi];
- if (IK_PtrInduction == II.IK)
- return 1;
- else if (IK_ReversePtrInduction == II.IK)
- return -1;
+ return II.getConsecutiveDirection();
}
unsigned InductionOperand = getGEPInductionOperand(DL, Gep);
return 0;
}
-bool LoopVectorizationLegality::isUniform(Value *V) {
- return (SE->isLoopInvariant(SE->getSCEV(V), TheLoop));
-}
+bool LoopVectorizationLegality::isUniform(Value *V) { return LAI.isUniform(V); }
InnerLoopVectorizer::VectorParts&
InnerLoopVectorizer::getVectorValue(Value *V) {
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)
Ptr = Builder.CreateExtractElement(PtrVal[0], Zero);
}
+ VectorParts Mask = createBlockInMask(Instr->getParent());
// Handle Stores:
if (SI) {
assert(!Legal->isUniform(SI->getPointerOperand()) &&
// We don't want to update the value in the map as it might be used in
// another expression. So don't use a reference type for "StoredVal".
VectorParts StoredVal = getVectorValue(SI->getValueOperand());
-
+
for (unsigned Part = 0; Part < UF; ++Part) {
// Calculate the pointer for the specific unroll-part.
Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF));
// 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));
+ Mask[Part] = reverseVector(Mask[Part]);
}
Value *VecPtr = Builder.CreateBitCast(PartPtr,
DataTy->getPointerTo(AddressSpace));
- StoreInst *NewSI =
- Builder.CreateAlignedStore(StoredVal[Part], VecPtr, Alignment);
+
+ Instruction *NewSI;
+ if (Legal->isMaskRequired(SI))
+ NewSI = Builder.CreateMaskedStore(StoredVal[Part], VecPtr, Alignment,
+ Mask[Part]);
+ 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));
+ Mask[Part] = reverseVector(Mask[Part]);
}
+ Instruction* NewLI;
Value *VecPtr = Builder.CreateBitCast(PartPtr,
DataTy->getPointerTo(AddressSpace));
- LoadInst *NewLI = Builder.CreateAlignedLoad(VecPtr, Alignment, "wide.load");
+ if (Legal->isMaskRequired(LI))
+ NewLI = Builder.CreateMaskedLoad(VecPtr, Alignment, Mask[Part],
+ UndefValue::get(DataTy),
+ "wide.masked.load");
+ else
+ NewLI = Builder.CreateAlignedLoad(VecPtr, Alignment, "wide.load");
propagateMetadata(NewLI, LI);
Entry[Part] = Reverse ? reverseVector(NewLI) : NewLI;
}
Cmp = Builder.CreateICmp(ICmpInst::ICMP_EQ, Cmp, ConstantInt::get(Cmp->getType(), 1));
CondBlock = IfBlock->splitBasicBlock(InsertPt, "cond.store");
LoopVectorBody.push_back(CondBlock);
- VectorLp->addBasicBlockToLoop(CondBlock, LI->getBase());
+ VectorLp->addBasicBlockToLoop(CondBlock, *LI);
// Update Builder with newly created basic block.
Builder.SetInsertPoint(InsertPt);
}
if (IfPredicateStore) {
BasicBlock *NewIfBlock = CondBlock->splitBasicBlock(InsertPt, "else");
LoopVectorBody.push_back(NewIfBlock);
- VectorLp->addBasicBlockToLoop(NewIfBlock, LI->getBase());
+ VectorLp->addBasicBlockToLoop(NewIfBlock, *LI);
Builder.SetInsertPoint(InsertPt);
Instruction *OldBr = IfBlock->getTerminator();
BranchInst::Create(CondBlock, NewIfBlock, Cmp, OldBr);
return std::make_pair(FirstInst, TheCheck);
}
-std::pair<Instruction *, Instruction *>
-InnerLoopVectorizer::addRuntimeCheck(Instruction *Loc) {
- LoopVectorizationLegality::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;
-
- 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
// before calling any utilities such as SCEV that require valid LoopInfo.
if (ParentLoop) {
ParentLoop->addChildLoop(Lp);
- ParentLoop->addBasicBlockToLoop(ScalarPH, LI->getBase());
- ParentLoop->addBasicBlockToLoop(VectorPH, LI->getBase());
- ParentLoop->addBasicBlockToLoop(MiddleBlock, LI->getBase());
+ ParentLoop->addBasicBlockToLoop(ScalarPH, *LI);
+ ParentLoop->addBasicBlockToLoop(VectorPH, *LI);
+ ParentLoop->addBasicBlockToLoop(MiddleBlock, *LI);
} else {
LI->addTopLevelLoop(Lp);
}
- Lp->addBasicBlockToLoop(VecBody, LI->getBase());
+ Lp->addBasicBlockToLoop(VecBody, *LI);
// Use this IR builder to create the loop instructions (Phi, Br, Cmp)
// inside the loop.
BasicBlock *CheckBlock =
LastBypassBlock->splitBasicBlock(PastOverflowCheck, "overflow.checked");
if (ParentLoop)
- ParentLoop->addBasicBlockToLoop(CheckBlock, LI->getBase());
+ ParentLoop->addBasicBlockToLoop(CheckBlock, *LI);
LoopBypassBlocks.push_back(CheckBlock);
Instruction *OldTerm = LastBypassBlock->getTerminator();
BranchInst::Create(ScalarPH, CheckBlock, CheckBCOverflow, OldTerm);
BasicBlock *CheckBlock =
LastBypassBlock->splitBasicBlock(FirstCheckInst, "vector.stridecheck");
if (ParentLoop)
- ParentLoop->addBasicBlockToLoop(CheckBlock, LI->getBase());
+ ParentLoop->addBasicBlockToLoop(CheckBlock, *LI);
LoopBypassBlocks.push_back(CheckBlock);
// Replace the branch into the memory check block with a conditional branch
// faster.
Instruction *MemRuntimeCheck;
std::tie(FirstCheckInst, MemRuntimeCheck) =
- addRuntimeCheck(LastBypassBlock->getTerminator());
+ Legal->getLAI()->addRuntimeCheck(LastBypassBlock->getTerminator());
if (MemRuntimeCheck) {
// 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->getBase());
+ ParentLoop->addBasicBlockToLoop(CheckBlock, *LI);
LoopBypassBlocks.push_back(CheckBlock);
// Replace the branch into the memory check block with a conditional branch
Value *CRD = BypassBuilder.CreateSExtOrTrunc(CountRoundDown,
II.StartValue->getType(),
"cast.crd");
- EndValue = BypassBuilder.CreateAdd(CRD, II.StartValue , "ind.end");
- break;
- }
- case LoopVectorizationLegality::IK_ReverseIntInduction: {
- // Convert the CountRoundDown variable to the PHI size.
- Value *CRD = BypassBuilder.CreateSExtOrTrunc(CountRoundDown,
- II.StartValue->getType(),
- "cast.crd");
- // Handle reverse integer induction counter.
- EndValue = BypassBuilder.CreateSub(II.StartValue, CRD, "rev.ind.end");
+ EndValue = II.transform(BypassBuilder, CRD);
+ EndValue->setName("ind.end");
break;
}
case LoopVectorizationLegality::IK_PtrInduction: {
- // For pointer induction variables, calculate the offset using
- // the end index.
- EndValue = BypassBuilder.CreateGEP(II.StartValue, CountRoundDown,
- "ptr.ind.end");
- break;
- }
- case LoopVectorizationLegality::IK_ReversePtrInduction: {
- // The value at the end of the loop for the reverse pointer is calculated
- // by creating a GEP with a negative index starting from the start value.
- Value *Zero = ConstantInt::get(CountRoundDown->getType(), 0);
- Value *NegIdx = BypassBuilder.CreateSub(Zero, CountRoundDown,
- "rev.ind.end");
- EndValue = BypassBuilder.CreateGEP(II.StartValue, NegIdx,
- "rev.ptr.ind.end");
+ EndValue = II.transform(BypassBuilder, CountRoundDown);
+ EndValue->setName("ptr.ind.end");
break;
}
}// end of case
LoopScalarBody = OldBasicBlock;
LoopVectorizeHints Hints(Lp, true);
- Hints.setAlreadyVectorized(Lp);
+ Hints.setAlreadyVectorized();
}
/// This function returns the identity element (or neutral element) for
}
// 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) {
LoopVectorizationLegality::InductionInfo II =
Legal->getInductionVars()->lookup(P);
+ // FIXME: The newly created binary instructions should contain nsw/nuw flags,
+ // which can be found from the original scalar operations.
switch (II.IK) {
case LoopVectorizationLegality::IK_NoInduction:
llvm_unreachable("Unknown induction");
Value *NormalizedIdx = Builder.CreateSub(Induction, ExtendedIdx,
"normalized.idx");
NormalizedIdx = Builder.CreateSExtOrTrunc(NormalizedIdx, PhiTy);
- Broadcasted = Builder.CreateAdd(II.StartValue, NormalizedIdx,
- "offset.idx");
+ Broadcasted = II.transform(Builder, NormalizedIdx);
+ Broadcasted->setName("offset.idx");
}
Broadcasted = getBroadcastInstrs(Broadcasted);
// After broadcasting the induction variable we need to make the vector
// consecutive by adding 0, 1, 2, etc.
for (unsigned part = 0; part < UF; ++part)
- Entry[part] = getConsecutiveVector(Broadcasted, VF * part, false);
+ Entry[part] = getStepVector(Broadcasted, VF * part, II.StepValue);
return;
}
- case LoopVectorizationLegality::IK_ReverseIntInduction:
case LoopVectorizationLegality::IK_PtrInduction:
- case LoopVectorizationLegality::IK_ReversePtrInduction:
- // Handle reverse integer and pointer inductions.
- Value *StartIdx = ExtendedIdx;
- // This is the normalized GEP that starts counting at zero.
- Value *NormalizedIdx = Builder.CreateSub(Induction, StartIdx,
- "normalized.idx");
-
- // Handle the reverse integer induction variable case.
- if (LoopVectorizationLegality::IK_ReverseIntInduction == II.IK) {
- IntegerType *DstTy = cast<IntegerType>(II.StartValue->getType());
- Value *CNI = Builder.CreateSExtOrTrunc(NormalizedIdx, DstTy,
- "resize.norm.idx");
- Value *ReverseInd = Builder.CreateSub(II.StartValue, CNI,
- "reverse.idx");
-
- // This is a new value so do not hoist it out.
- Value *Broadcasted = getBroadcastInstrs(ReverseInd);
- // After broadcasting the induction variable we need to make the
- // vector consecutive by adding ... -3, -2, -1, 0.
- for (unsigned part = 0; part < UF; ++part)
- Entry[part] = getConsecutiveVector(Broadcasted, -(int)VF * part,
- true);
- return;
- }
-
// Handle the pointer induction variable case.
assert(P->getType()->isPointerTy() && "Unexpected type.");
-
- // Is this a reverse induction ptr or a consecutive induction ptr.
- bool Reverse = (LoopVectorizationLegality::IK_ReversePtrInduction ==
- II.IK);
-
+ // This is the normalized GEP that starts counting at zero.
+ Value *NormalizedIdx =
+ Builder.CreateSub(Induction, ExtendedIdx, "normalized.idx");
// This is the vector of results. Notice that we don't generate
// vector geps because scalar geps result in better code.
for (unsigned part = 0; part < UF; ++part) {
if (VF == 1) {
- int EltIndex = (part) * (Reverse ? -1 : 1);
+ int EltIndex = part;
Constant *Idx = ConstantInt::get(Induction->getType(), EltIndex);
- Value *GlobalIdx;
- if (Reverse)
- GlobalIdx = Builder.CreateSub(Idx, NormalizedIdx, "gep.ridx");
- else
- GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx, "gep.idx");
-
- Value *SclrGep = Builder.CreateGEP(II.StartValue, GlobalIdx,
- "next.gep");
+ Value *GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx);
+ Value *SclrGep = II.transform(Builder, GlobalIdx);
+ SclrGep->setName("next.gep");
Entry[part] = SclrGep;
continue;
}
Value *VecVal = UndefValue::get(VectorType::get(P->getType(), VF));
for (unsigned int i = 0; i < VF; ++i) {
- int EltIndex = (i + part * VF) * (Reverse ? -1 : 1);
+ int EltIndex = i + part * VF;
Constant *Idx = ConstantInt::get(Induction->getType(), EltIndex);
- Value *GlobalIdx;
- if (!Reverse)
- GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx, "gep.idx");
- else
- GlobalIdx = Builder.CreateSub(Idx, NormalizedIdx, "gep.ridx");
-
- Value *SclrGep = Builder.CreateGEP(II.StartValue, GlobalIdx,
- "next.gep");
+ Value *GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx);
+ Value *SclrGep = II.transform(Builder, GlobalIdx);
+ SclrGep->setName("next.gep");
VecVal = Builder.CreateInsertElement(VecVal, SclrGep,
Builder.getInt32(i),
"insert.gep");
// Nothing to do for PHIs and BR, since we already took care of the
// loop control flow instructions.
continue;
- case Instruction::PHI:{
+ case Instruction::PHI: {
// Vectorize PHINodes.
widenPHIInstruction(it, Entry, UF, VF, PV);
continue;
for (unsigned Part = 0; Part < UF; ++Part) {
Value *V = Builder.CreateBinOp(BinOp->getOpcode(), A[Part], B[Part]);
- // Update the NSW, NUW and Exact flags. Notice: V can be an Undef.
- BinaryOperator *VecOp = dyn_cast<BinaryOperator>(V);
- if (VecOp && isa<OverflowingBinaryOperator>(BinOp)) {
- VecOp->setHasNoSignedWrap(BinOp->hasNoSignedWrap());
- VecOp->setHasNoUnsignedWrap(BinOp->hasNoUnsignedWrap());
- }
- if (VecOp && isa<PossiblyExactOperator>(VecOp))
- VecOp->setIsExact(BinOp->isExact());
-
- // Copy the fast-math flags.
- if (VecOp && isa<FPMathOperator>(V))
- VecOp->setFastMathFlags(it->getFastMathFlags());
+ if (BinaryOperator *VecOp = dyn_cast<BinaryOperator>(V))
+ VecOp->copyIRFlags(BinOp);
Entry[Part] = V;
}
Value *ScalarCast = Builder.CreateCast(CI->getOpcode(), Induction,
CI->getType());
Value *Broadcasted = getBroadcastInstrs(ScalarCast);
+ LoopVectorizationLegality::InductionInfo II =
+ Legal->getInductionVars()->lookup(OldInduction);
+ Constant *Step =
+ ConstantInt::getSigned(CI->getType(), II.StepValue->getSExtValue());
for (unsigned Part = 0; Part < UF; ++Part)
- Entry[Part] = getConsecutiveVector(Broadcasted, VF * Part, false);
+ Entry[Part] = getStepVector(Broadcasted, VF * Part, Step);
propagateMetadata(Entry, it);
break;
}
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());
}
bool LoopVectorizationLegality::canVectorizeWithIfConvert() {
if (!EnableIfConversion) {
- emitAnalysis(Report() << "if-conversion is disabled");
+ emitAnalysis(VectorizationReport() << "if-conversion is disabled");
return false;
}
// We don't support switch statements inside loops.
if (!isa<BranchInst>(BB->getTerminator())) {
- emitAnalysis(Report(BB->getTerminator())
+ emitAnalysis(VectorizationReport(BB->getTerminator())
<< "loop contains a switch statement");
return false;
}
// We must be able to predicate all blocks that need to be predicated.
if (blockNeedsPredication(BB)) {
if (!blockCanBePredicated(BB, SafePointes)) {
- emitAnalysis(Report(BB->getTerminator())
+ emitAnalysis(VectorizationReport(BB->getTerminator())
<< "control flow cannot be substituted for a select");
return false;
}
} else if (BB != Header && !canIfConvertPHINodes(BB)) {
- emitAnalysis(Report(BB->getTerminator())
+ emitAnalysis(VectorizationReport(BB->getTerminator())
<< "control flow cannot be substituted for a select");
return false;
}
// be canonicalized.
if (!TheLoop->getLoopPreheader()) {
emitAnalysis(
- Report() << "loop control flow is not understood by vectorizer");
+ VectorizationReport() <<
+ "loop control flow is not understood by vectorizer");
return false;
}
// We can only vectorize innermost loops.
- if (TheLoop->getSubLoopsVector().size()) {
- emitAnalysis(Report() << "loop is not the innermost loop");
+ if (!TheLoop->getSubLoopsVector().empty()) {
+ emitAnalysis(VectorizationReport() << "loop is not the innermost loop");
return false;
}
// We must have a single backedge.
if (TheLoop->getNumBackEdges() != 1) {
emitAnalysis(
- Report() << "loop control flow is not understood by vectorizer");
+ VectorizationReport() <<
+ "loop control flow is not understood by vectorizer");
return false;
}
// We must have a single exiting block.
if (!TheLoop->getExitingBlock()) {
emitAnalysis(
- Report() << "loop control flow is not understood by vectorizer");
+ VectorizationReport() <<
+ "loop control flow is not understood by vectorizer");
+ 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(
+ VectorizationReport() <<
+ "loop control flow is not understood by vectorizer");
return false;
}
// ScalarEvolution needs to be able to find the exit count.
const SCEV *ExitCount = SE->getBackedgeTakenCount(TheLoop);
if (ExitCount == SE->getCouldNotCompute()) {
- emitAnalysis(Report() << "could not determine number of loop iterations");
+ emitAnalysis(VectorizationReport() <<
+ "could not determine number of loop iterations");
DEBUG(dbgs() << "LV: SCEV could not compute the loop exit count.\n");
return false;
}
// Collect all of the variables that remain uniform after vectorization.
collectLoopUniforms();
- DEBUG(dbgs() << "LV: We can vectorize this loop" <<
- (PtrRtCheck.Need ? " (with a runtime bound check)" : "")
- <<"!\n");
+ DEBUG(dbgs() << "LV: We can vectorize this loop"
+ << (LAI.getRuntimePointerCheck()->Need
+ ? " (with a runtime bound check)"
+ : "") << "!\n");
// Okay! We can vectorize. At this point we don't have any other mem analysis
// which may limit our maximum vectorization factor, so just return true with
/// \brief Check that the instruction has outside loop users and is not an
/// identified reduction variable.
static bool hasOutsideLoopUser(const Loop *TheLoop, Instruction *Inst,
- SmallPtrSet<Value *, 4> &Reductions) {
+ SmallPtrSetImpl<Value *> &Reductions) {
// Reduction instructions are allowed to have exit users. All other
// instructions must not have external users.
if (!Reductions.count(Inst))
// Look for the attribute signaling the absence of NaNs.
Function &F = *Header->getParent();
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 (!PhiTy->isIntegerTy() &&
!PhiTy->isFloatingPointTy() &&
!PhiTy->isPointerTy()) {
- emitAnalysis(Report(it)
+ emitAnalysis(VectorizationReport(it)
<< "loop control flow is not understood by vectorizer");
DEBUG(dbgs() << "LV: Found an non-int non-pointer PHI.\n");
return false;
// identified reduction value with an outside user.
if (!hasOutsideLoopUser(TheLoop, it, AllowedExit))
continue;
- emitAnalysis(Report(it) << "value that could not be identified as "
- "reduction is used outside the loop");
+ emitAnalysis(VectorizationReport(it) <<
+ "value could not be identified as "
+ "an induction or reduction variable");
return false;
}
- // We only allow if-converted PHIs with more than two incoming values.
+ // We only allow if-converted PHIs with exactly two incoming values.
if (Phi->getNumIncomingValues() != 2) {
- emitAnalysis(Report(it)
+ emitAnalysis(VectorizationReport(it)
<< "control flow not understood by vectorizer");
DEBUG(dbgs() << "LV: Found an invalid PHI.\n");
return false;
// This is the value coming from the preheader.
Value *StartValue = Phi->getIncomingValueForBlock(PreHeader);
+ ConstantInt *StepValue = nullptr;
// Check if this is an induction variable.
- InductionKind IK = isInductionVariable(Phi);
+ InductionKind IK = isInductionVariable(Phi, StepValue);
if (IK_NoInduction != IK) {
// Get the widest type.
WidestIndTy = getWiderType(*DL, PhiTy, WidestIndTy);
// Int inductions are special because we only allow one IV.
- if (IK == IK_IntInduction) {
+ if (IK == IK_IntInduction && StepValue->isOne()) {
// Use the phi node with the widest type as induction. Use the last
// one if there are multiple (no good reason for doing this other
// than it is expedient).
}
DEBUG(dbgs() << "LV: Found an induction variable.\n");
- Inductions[Phi] = InductionInfo(StartValue, IK);
+ Inductions[Phi] = InductionInfo(StartValue, IK, StepValue);
// Until we explicitly handle the case of an induction variable with
// an outside loop user we have to give up vectorizing this loop.
if (hasOutsideLoopUser(TheLoop, it, AllowedExit)) {
- emitAnalysis(Report(it) << "use of induction value outside of the "
- "loop is not handled by vectorizer");
+ emitAnalysis(VectorizationReport(it) <<
+ "use of induction value outside of the "
+ "loop is not handled by vectorizer");
return false;
}
continue;
}
- emitAnalysis(Report(it) << "unvectorizable operation");
+ emitAnalysis(VectorizationReport(it) <<
+ "value that could not be identified as "
+ "reduction is used outside the loop");
DEBUG(dbgs() << "LV: Found an unidentified PHI."<< *Phi <<"\n");
return false;
}// end of PHI handling
// calls and we do handle certain intrinsic and libm functions.
CallInst *CI = dyn_cast<CallInst>(it);
if (CI && !getIntrinsicIDForCall(CI, TLI) && !isa<DbgInfoIntrinsic>(CI)) {
- emitAnalysis(Report(it) << "call instruction cannot be vectorized");
+ emitAnalysis(VectorizationReport(it) <<
+ "call instruction cannot be vectorized");
DEBUG(dbgs() << "LV: Found a call site.\n");
return false;
}
if (CI &&
hasVectorInstrinsicScalarOpd(getIntrinsicIDForCall(CI, TLI), 1)) {
if (!SE->isLoopInvariant(SE->getSCEV(CI->getOperand(1)), TheLoop)) {
- emitAnalysis(Report(it)
+ emitAnalysis(VectorizationReport(it)
<< "intrinsic instruction cannot be vectorized");
DEBUG(dbgs() << "LV: Found unvectorizable intrinsic " << *CI << "\n");
return false;
// Also, we can't vectorize extractelement instructions.
if ((!VectorType::isValidElementType(it->getType()) &&
!it->getType()->isVoidTy()) || isa<ExtractElementInst>(it)) {
- emitAnalysis(Report(it)
+ emitAnalysis(VectorizationReport(it)
<< "instruction return type cannot be vectorized");
DEBUG(dbgs() << "LV: Found unvectorizable type.\n");
return false;
if (StoreInst *ST = dyn_cast<StoreInst>(it)) {
Type *T = ST->getValueOperand()->getType();
if (!VectorType::isValidElementType(T)) {
- emitAnalysis(Report(ST) << "store instruction cannot be vectorized");
+ emitAnalysis(VectorizationReport(ST) <<
+ "store instruction cannot be vectorized");
return false;
}
if (EnableMemAccessVersioning)
- collectStridedAcccess(ST);
+ collectStridedAccess(ST);
}
if (EnableMemAccessVersioning)
if (LoadInst *LI = dyn_cast<LoadInst>(it))
- collectStridedAcccess(LI);
+ collectStridedAccess(LI);
// Reduction instructions are allowed to have exit users.
// All other instructions must not have external users.
if (hasOutsideLoopUser(TheLoop, it, AllowedExit)) {
- emitAnalysis(Report(it) << "value cannot be used outside the loop");
+ emitAnalysis(VectorizationReport(it) <<
+ "value cannot be used outside the loop");
return false;
}
if (!Induction) {
DEBUG(dbgs() << "LV: Did not find one integer induction var.\n");
if (Inductions.empty()) {
- emitAnalysis(Report()
+ emitAnalysis(VectorizationReport()
<< "loop induction variable could not be identified");
return false;
}
return Stride;
}
-void LoopVectorizationLegality::collectStridedAcccess(Value *MemAccess) {
+void LoopVectorizationLegality::collectStridedAccess(Value *MemAccess) {
Value *Ptr = nullptr;
if (LoadInst *LI = dyn_cast<LoadInst>(MemAccess))
Ptr = LI->getPointerOperand();
if (I->getType()->isPointerTy() && isConsecutivePtr(I))
Worklist.insert(Worklist.end(), I->op_begin(), I->op_end());
- while (Worklist.size()) {
+ while (!Worklist.empty()) {
Instruction *I = dyn_cast<Instruction>(Worklist.back());
Worklist.pop_back();
}
}
-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, DepCandidates &DA) :
- DL(Dl), DepCands(DA), AreAllWritesIdentified(true),
- AreAllReadsIdentified(true), IsRTCheckNeeded(false) {}
-
- /// \brief Register a load and whether it is only read from.
- void addLoad(Value *Ptr, bool IsReadOnly) {
- Accesses.insert(MemAccessInfo(Ptr, false));
- if (IsReadOnly)
- ReadOnlyPtr.insert(Ptr);
- }
-
- /// \brief Register a store.
- void addStore(Value *Ptr) {
- Accesses.insert(MemAccessInfo(Ptr, true));
- }
-
- /// \brief Check whether we can check the pointers at runtime for
- /// non-intersection.
- bool canCheckPtrAtRT(LoopVectorizationLegality::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() {
- // Process read-write pointers first.
- processMemAccesses(false);
- // Next, process read pointers.
- processMemAccesses(true);
- }
-
- bool isRTCheckNeeded() { return IsRTCheckNeeded; }
-
- bool isDependencyCheckNeeded() { return !CheckDeps.empty(); }
- void resetDepChecks() { CheckDeps.clear(); }
-
- MemAccessInfoSet &getDependenciesToCheck() { return CheckDeps; }
-
-private:
- typedef SetVector<MemAccessInfo> PtrAccessSet;
- typedef DenseMap<Value*, MemAccessInfo> UnderlyingObjToAccessMap;
-
- /// \brief Go over all memory access or only the deferred ones if
- /// \p UseDeferred is true and check whether runtime pointer checks are needed
- /// and build sets of dependency check candidates.
- void processMemAccesses(bool UseDeferred);
-
- /// Set of all accesses.
- PtrAccessSet Accesses;
-
- /// Set of access to check after all writes have been processed.
- PtrAccessSet DeferredAccesses;
-
- /// Map of pointers to last access encountered.
- UnderlyingObjToAccessMap ObjToLastAccess;
-
- /// Set of accesses that need a further dependence check.
- MemAccessInfoSet CheckDeps;
-
- /// Set of pointers that are read only.
- SmallPtrSet<Value*, 16> ReadOnlyPtr;
-
- /// Set of underlying objects already written to.
- SmallPtrSet<Value*, 16> WriteObjects;
-
- const DataLayout *DL;
-
- /// 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 AreAllWritesIdentified;
- bool AreAllReadsIdentified;
- 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(
- LoopVectorizationLegality::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.
- unsigned NumReadPtrChecks = 0;
- unsigned NumWritePtrChecks = 0;
- bool CanDoRT = true;
-
- bool IsDepCheckNeeded = isDependencyCheckNeeded();
- // 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 (PtrAccessSet::iterator AI = Accesses.begin(), AE = Accesses.end();
- AI != AE; ++AI) {
- const MemAccessInfo &Access = *AI;
- Value *Ptr = Access.getPointer();
- bool IsWrite = Access.getInt();
-
- // Just add write checks if we have both.
- if (!IsWrite && Accesses.count(MemAccessInfo(Ptr, true)))
- continue;
-
- 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, 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));
- }
-
- // 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;
-
- 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;
-}
-
-static bool isFunctionScopeIdentifiedObject(Value *Ptr) {
- return isNoAliasArgument(Ptr) || isNoAliasCall(Ptr) || isa<AllocaInst>(Ptr);
-}
-
-void AccessAnalysis::processMemAccesses(bool UseDeferred) {
- // 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.
-
- PtrAccessSet &S = UseDeferred ? DeferredAccesses : Accesses;
- for (PtrAccessSet::iterator AI = S.begin(), AE = S.end(); AI != AE; ++AI) {
- const MemAccessInfo &Access = *AI;
- Value *Ptr = Access.getPointer();
- bool IsWrite = Access.getInt();
-
- 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".
- bool IsReadOnlyPtr = ReadOnlyPtr.count(Ptr) && !IsWrite;
- if (!UseDeferred && IsReadOnlyPtr) {
- DeferredAccesses.insert(Access);
- continue;
- }
-
- bool NeedDepCheck = false;
- // Check whether there is the possibility of dependency because of
- // underlying objects being the same.
- typedef SmallVector<Value*, 16> ValueVector;
- ValueVector TempObjects;
- GetUnderlyingObjects(Ptr, TempObjects, DL);
- for (ValueVector::iterator UI = TempObjects.begin(), UE = TempObjects.end();
- UI != UE; ++UI) {
- Value *UnderlyingObj = *UI;
-
- // If this is a write then it needs to be an identified object. If this a
- // read and all writes (so far) are identified function scope objects we
- // don't need an identified underlying object but only an Argument (the
- // next write is going to invalidate this assumption if it is
- // unidentified).
- // This is a micro-optimization for the case where all writes are
- // identified and we have one argument pointer.
- // Otherwise, we do need a runtime check.
- if ((IsWrite && !isFunctionScopeIdentifiedObject(UnderlyingObj)) ||
- (!IsWrite && (!AreAllWritesIdentified ||
- !isa<Argument>(UnderlyingObj)) &&
- !isIdentifiedObject(UnderlyingObj))) {
- DEBUG(dbgs() << "LV: Found an unidentified " <<
- (IsWrite ? "write" : "read" ) << " ptr: " << *UnderlyingObj <<
- "\n");
- IsRTCheckNeeded = (IsRTCheckNeeded ||
- !isIdentifiedObject(UnderlyingObj) ||
- !AreAllReadsIdentified);
-
- if (IsWrite)
- AreAllWritesIdentified = false;
- if (!IsWrite)
- AreAllReadsIdentified = false;
- }
-
- // 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) && WriteObjects.count(UnderlyingObj))
- NeedDepCheck = true;
-
- if (IsWrite)
- WriteObjects.insert(UnderlyingObj);
-
- // Create sets of pointers connected by shared underlying objects.
- UnderlyingObjToAccessMap::iterator Prev =
- ObjToLastAccess.find(UnderlyingObj);
- if (Prev != ObjToLastAccess.end())
- DepCands.unionSets(Access, Prev->second);
-
- ObjToLastAccess[UnderlyingObj] = Access;
- }
-
- if (NeedDepCheck)
- CheckDeps.insert(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)
- : SE(Se), DL(Dl), InnermostLoop(L), AccessIdx(0),
- ShouldRetryWithRuntimeCheck(false) {}
-
- /// \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 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 = 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 != 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)
- return false;
-
- 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");
- 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");
- return false;
- }
-
- unsigned Distance = (unsigned) Val.getZExtValue();
-
- // Bail out early if passed-in parameters make vectorization not feasible.
- unsigned ForcedFactor = VectorizationFactor ? VectorizationFactor : 1;
- unsigned ForcedUnroll = VectorizationUnroll ? VectorizationUnroll : 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 LoopVectorizationLegality::canVectorizeMemory() {
-
- 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);
-
- // 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(Report(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(Report(it) << "instruction cannot be vectorized");
- return false;
- }
- if (!St->isSimple() && !IsAnnotatedParallel) {
- emitAnalysis(Report(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, 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(
- Report(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)) {
- ++NumReadWrites;
- Accesses.addStore(Ptr);
- }
- }
-
- 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) || !isStridedPtr(SE, DL, Ptr, TheLoop, Strides)) {
- ++NumReads;
- IsReadOnlyPtr = true;
- }
- Accesses.addLoad(Ptr, 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 > RuntimeMemoryCheckThreshold) {
- PtrRtCheck.reset();
- CanDoRT = false;
- }
-
- if (CanDoRT) {
- DEBUG(dbgs() << "LV: We can perform a memory runtime check if needed.\n");
- }
-
- if (NeedRTCheck && !CanDoRT) {
- emitAnalysis(Report() << "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 > RuntimeMemoryCheckThreshold) {
- if (!CanDoRT && NumComparisons > 0)
- emitAnalysis(Report()
- << "cannot check memory dependencies at runtime");
- else
- emitAnalysis(Report()
- << NumComparisons << " exceeds limit of "
- << 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(Report() << "unsafe dependent memory operations in loop");
-
- DEBUG(dbgs() << "LV: We" << (NeedRTCheck ? "" : " don't") <<
- " need a runtime memory check.\n");
-
- return CanVecMem;
+ return LAI.canVectorizeMemory(Strides);
}
static bool hasMultipleUsesOf(Instruction *I,
- SmallPtrSet<Instruction *, 8> &Insts) {
+ SmallPtrSetImpl<Instruction *> &Insts) {
unsigned NumUses = 0;
for(User::op_iterator Use = I->op_begin(), E = I->op_end(); Use != E; ++Use) {
if (Insts.count(dyn_cast<Instruction>(*Use)))
return false;
}
-static bool areAllUsesIn(Instruction *I, SmallPtrSet<Instruction *, 8> &Set) {
+static bool areAllUsesIn(Instruction *I, SmallPtrSetImpl<Instruction *> &Set) {
for(User::op_iterator Use = I->op_begin(), E = I->op_end(); Use != E; ++Use)
if (!Set.count(dyn_cast<Instruction>(*Use)))
return false;
// 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
ReductionKind Kind,
ReductionInstDesc &Prev) {
bool FP = I->getType()->isFloatingPointTy();
- bool FastMath = (FP && I->isCommutative() && I->isAssociative());
+ bool FastMath = FP && I->hasUnsafeAlgebra();
switch (I->getOpcode()) {
default:
return ReductionInstDesc(false, I);
return ReductionInstDesc(Kind == RK_IntegerXor, I);
case Instruction::FMul:
return ReductionInstDesc(Kind == RK_FloatMult && FastMath, I);
+ case Instruction::FSub:
case Instruction::FAdd:
return ReductionInstDesc(Kind == RK_FloatAdd && FastMath, I);
case Instruction::FCmp:
}
LoopVectorizationLegality::InductionKind
-LoopVectorizationLegality::isInductionVariable(PHINode *Phi) {
+LoopVectorizationLegality::isInductionVariable(PHINode *Phi,
+ ConstantInt *&StepValue) {
Type *PhiTy = Phi->getType();
// We only handle integer and pointer inductions variables.
if (!PhiTy->isIntegerTy() && !PhiTy->isPointerTy())
DEBUG(dbgs() << "LV: PHI is not a poly recurrence.\n");
return IK_NoInduction;
}
- const SCEV *Step = AR->getStepRecurrence(*SE);
-
- // Integer inductions need to have a stride of one.
- if (PhiTy->isIntegerTy()) {
- if (Step->isOne())
- return IK_IntInduction;
- if (Step->isAllOnesValue())
- return IK_ReverseIntInduction;
- return IK_NoInduction;
- }
+ 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;
+ ConstantInt *CV = C->getValue();
+ if (PhiTy->isIntegerTy()) {
+ StepValue = CV;
+ return IK_IntInduction;
+ }
+
assert(PhiTy->isPointerTy() && "The PHI must be a pointer");
- uint64_t Size = DL->getTypeAllocSize(PhiTy->getPointerElementType());
- if (C->getValue()->equalsInt(Size))
- return IK_PtrInduction;
- else if (C->getValue()->equalsInt(0 - Size))
- return IK_ReversePtrInduction;
+ Type *PointerElementType = PhiTy->getPointerElementType();
+ // The pointer stride cannot be determined if the pointer element type is not
+ // sized.
+ if (!PointerElementType->isSized())
+ return IK_NoInduction;
- return IK_NoInduction;
+ int64_t Size = static_cast<int64_t>(DL->getTypeAllocSize(PointerElementType));
+ int64_t CVSize = CV->getSExtValue();
+ if (CVSize % Size)
+ return IK_NoInduction;
+ StepValue = ConstantInt::getSigned(CV->getType(), CVSize / Size);
+ return IK_PtrInduction;
}
bool LoopVectorizationLegality::isInductionVariable(const Value *V) {
}
bool LoopVectorizationLegality::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);
+ return LoopAccessInfo::blockNeedsPredication(BB, TheLoop, DT);
}
bool LoopVectorizationLegality::blockCanBePredicated(BasicBlock *BB,
- SmallPtrSet<Value *, 8>& SafePtrs) {
+ 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;
}
}
}
LoopVectorizationCostModel::VectorizationFactor
-LoopVectorizationCostModel::selectVectorizationFactor(bool OptForSize,
- unsigned UserVF,
- bool ForceVectorization) {
+LoopVectorizationCostModel::selectVectorizationFactor(bool OptForSize) {
// Width 1 means no vectorize
VectorizationFactor Factor = { 1U, 0U };
if (OptForSize && Legal->getRuntimePointerCheck()->Need) {
+ emitAnalysis(VectorizationReport() <<
+ "runtime pointer checks needed. Enable vectorization of this "
+ "loop with '#pragma clang loop vectorize(enable)' when "
+ "compiling with -Os");
DEBUG(dbgs() << "LV: Aborting. Runtime ptr check is required in Os.\n");
return Factor;
}
- if (!EnableCondStoresVectorization && Legal->NumPredStores) {
+ if (!EnableCondStoresVectorization && Legal->getNumPredStores()) {
+ emitAnalysis(VectorizationReport() <<
+ "store that is conditionally executed prevents vectorization");
DEBUG(dbgs() << "LV: No vectorization. There are conditional stores.\n");
return Factor;
}
// 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 (OptForSize) {
// If we are unable to calculate the trip count then don't try to vectorize.
if (TC < 2) {
+ emitAnalysis
+ (VectorizationReport() <<
+ "unable to calculate the loop count due to complex control flow");
DEBUG(dbgs() << "LV: Aborting. A tail loop is required in Os.\n");
return Factor;
}
// If the trip count that we found modulo the vectorization factor is not
// zero then we require a tail.
if (VF < 2) {
+ emitAnalysis(VectorizationReport() <<
+ "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;
}
}
+ int UserVF = Hints->getWidth();
if (UserVF != 0) {
assert(isPowerOf2_32(UserVF) && "VF needs to be a power of two");
DEBUG(dbgs() << "LV: Using user VF " << UserVF << ".\n");
unsigned Width = 1;
DEBUG(dbgs() << "LV: Scalar loop costs: " << (int)ScalarCost << ".\n");
+ bool ForceVectorization = Hints->getForce() == LoopVectorizeHints::FK_Enabled;
// Ignore scalar width, because the user explicitly wants vectorization.
if (ForceVectorization && VF > 1) {
Width = 2;
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;
unsigned
LoopVectorizationCostModel::selectUnrollFactor(bool OptForSize,
- unsigned UserUF,
unsigned VF,
unsigned LoopCost) {
// -- The unroll heuristics --
// We unroll the loop in order to expose ILP and reduce the loop overhead.
// There are many micro-architectural considerations that we can't predict
- // at this level. For example frontend pressure (on decode or fetch) due to
+ // at this level. For example, frontend pressure (on decode or fetch) due to
// code size, or the number and capabilities of the execution ports.
//
// We use the following heuristics to select the unroll factor:
- // 1. If the code has reductions the we unroll in order to break the cross
+ // 1. If the code has reductions, then we unroll in order to break the cross
// iteration dependency.
- // 2. If the loop is really small then we unroll in order to reduce the loop
+ // 2. If the loop is really small, then we unroll in order to reduce the loop
// overhead.
// 3. We don't unroll if we think that we will spill registers to memory due
// to the increased register pressure.
// Use the user preference, unless 'auto' is selected.
+ int UserUF = Hints->getInterleave();
if (UserUF != 0)
return UserUF;
- // When we optimize for size we don't unroll.
+ // When we optimize for size, we don't unroll.
if (OptForSize)
return 1;
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;
// Unroll until store/load ports (estimated by max unroll factor) are
// saturated.
- unsigned StoresUF = UF / (Legal->NumStores ? Legal->NumStores : 1);
- unsigned LoadsUF = UF / (Legal->NumLoads ? Legal->NumLoads : 1);
+ unsigned NumStores = Legal->getNumStores();
+ unsigned NumLoads = Legal->getNumLoads();
+ unsigned StoresUF = UF / (NumStores ? NumStores : 1);
+ unsigned LoadsUF = UF / (NumLoads ? NumLoads : 1);
+
+ // If we have a scalar reduction (vector reductions are already dealt with
+ // by this point), we can increase the critical path length if the loop
+ // we're unrolling is inside another loop. Limit, by default to 2, so the
+ // critical path only gets increased by one reduction operation.
+ if (Legal->getReductionVars()->size() &&
+ TheLoop->getLoopDepth() > 1) {
+ unsigned F = static_cast<unsigned>(MaxNestedScalarReductionUF);
+ SmallUF = std::min(SmallUF, F);
+ StoresUF = std::min(StoresUF, F);
+ LoadsUF = std::min(LoadsUF, F);
+ }
if (EnableLoadStoreRuntimeUnroll && std::max(StoresUF, LoadsUF) > SmallUF) {
DEBUG(dbgs() << "LV: Unrolling to saturate store or load ports.\n");
// 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.
TargetTransformInfo::OK_AnyValue;
TargetTransformInfo::OperandValueKind Op2VK =
TargetTransformInfo::OK_AnyValue;
+ TargetTransformInfo::OperandValueProperties Op1VP =
+ TargetTransformInfo::OP_None;
+ TargetTransformInfo::OperandValueProperties Op2VP =
+ TargetTransformInfo::OP_None;
Value *Op2 = I->getOperand(1);
// Check for a splat of a constant or for a non uniform vector of constants.
- if (isa<ConstantInt>(Op2))
+ if (isa<ConstantInt>(Op2)) {
+ ConstantInt *CInt = cast<ConstantInt>(Op2);
+ if (CInt && CInt->getValue().isPowerOf2())
+ Op2VP = TargetTransformInfo::OP_PowerOf2;
Op2VK = TargetTransformInfo::OK_UniformConstantValue;
- else if (isa<ConstantVector>(Op2) || isa<ConstantDataVector>(Op2)) {
+ } else if (isa<ConstantVector>(Op2) || isa<ConstantDataVector>(Op2)) {
Op2VK = TargetTransformInfo::OK_NonUniformConstantValue;
- if (cast<Constant>(Op2)->getSplatValue() != nullptr)
+ Constant *SplatValue = cast<Constant>(Op2)->getSplatValue();
+ if (SplatValue) {
+ ConstantInt *CInt = dyn_cast<ConstantInt>(SplatValue);
+ if (CInt && CInt->getValue().isPowerOf2())
+ Op2VP = TargetTransformInfo::OP_PowerOf2;
Op2VK = TargetTransformInfo::OK_UniformConstantValue;
+ }
}
- return TTI.getArithmeticInstrCost(I->getOpcode(), VectorTy, Op1VK, Op2VK);
+ return TTI.getArithmeticInstrCost(I->getOpcode(), VectorTy, Op1VK, Op2VK,
+ Op1VP, Op2VP);
}
case Instruction::Select: {
SelectInst *SI = cast<SelectInst>(I);
// 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,
char LoopVectorize::ID = 0;
static const char lv_name[] = "Loop Vectorization";
INITIALIZE_PASS_BEGIN(LoopVectorize, LV_NAME, lv_name, false, false)
-INITIALIZE_AG_DEPENDENCY(TargetTransformInfo)
+INITIALIZE_PASS_DEPENDENCY(TargetTransformInfoWrapperPass)
+INITIALIZE_AG_DEPENDENCY(AliasAnalysis)
+INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_DEPENDENCY(BlockFrequencyInfo)
INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
INITIALIZE_PASS_DEPENDENCY(ScalarEvolution)
INITIALIZE_PASS_DEPENDENCY(LCSSA)
-INITIALIZE_PASS_DEPENDENCY(LoopInfo)
+INITIALIZE_PASS_DEPENDENCY(LoopInfoWrapperPass)
INITIALIZE_PASS_DEPENDENCY(LoopSimplify)
INITIALIZE_PASS_END(LoopVectorize, LV_NAME, lv_name, false, false)
ConstantInt::get(Cond[Part]->getType(), 1));
CondBlock = IfBlock->splitBasicBlock(InsertPt, "cond.store");
LoopVectorBody.push_back(CondBlock);
- VectorLp->addBasicBlockToLoop(CondBlock, LI->getBase());
+ VectorLp->addBasicBlockToLoop(CondBlock, *LI);
// Update Builder with newly created basic block.
Builder.SetInsertPoint(InsertPt);
}
if (IfPredicateStore) {
BasicBlock *NewIfBlock = CondBlock->splitBasicBlock(InsertPt, "else");
LoopVectorBody.push_back(NewIfBlock);
- VectorLp->addBasicBlockToLoop(NewIfBlock, LI->getBase());
+ VectorLp->addBasicBlockToLoop(NewIfBlock, *LI);
Builder.SetInsertPoint(InsertPt);
Instruction *OldBr = IfBlock->getTerminator();
BranchInst::Create(CondBlock, NewIfBlock, Cmp, OldBr);
return V;
}
-Value *InnerLoopUnroller::getConsecutiveVector(Value* Val, int StartIdx,
- bool Negate) {
+Value *InnerLoopUnroller::getStepVector(Value *Val, int StartIdx, Value *Step) {
// When unrolling and the VF is 1, we only need to add a simple scalar.
Type *ITy = Val->getType();
assert(!ITy->isVectorTy() && "Val must be a scalar");
- Constant *C = ConstantInt::get(ITy, StartIdx, Negate);
- return Builder.CreateAdd(Val, C, "induction");
+ Constant *C = ConstantInt::get(ITy, StartIdx);
+ return Builder.CreateAdd(Val, Builder.CreateMul(C, Step), "induction");
}
projects / oota-llvm.git / blobdiff