Google Git Google Git
Sign in
llvm/llvm-project/refs/heads/users/void/array-bounds-accessed/./llvm/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
blob: cb239f83382cce413c6493570c649e50501ac16a [file] [edit]
//===- LoopIdiomRecognize.cpp - Loop idiom recognition --------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This pass implements an idiom recognizer that transforms simple loops into a
// non-loop form. In cases that this kicks in, it can be a significant
// performance win.
//
// If compiling for code size we avoid idiom recognition if the resulting
// code could be larger than the code for the original loop. One way this could
// happen is if the loop is not removable after idiom recognition due to the
// presence of non-idiom instructions. The initial implementation of the
// heuristics applies to idioms in multi-block loops.
//
//===----------------------------------------------------------------------===//
//
// TODO List:
//
// Future loop memory idioms to recognize: memcmp, etc.
//
// This could recognize common matrix multiplies and dot product idioms and
// replace them with calls to BLAS (if linked in??).
//
//===----------------------------------------------------------------------===//
#include "llvm/Transforms/Scalar/LoopIdiomRecognize.h"
#include "ScalarOptions.h"
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/MapVector.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/CmpInstAnalysis.h"
#include "llvm/Analysis/HashRecognize.h"
#include "llvm/Analysis/LoopAccessAnalysis.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/LoopPass.h"
#include "llvm/Analysis/MemoryLocation.h"
#include "llvm/Analysis/MemorySSA.h"
#include "llvm/Analysis/MemorySSAUpdater.h"
#include "llvm/Analysis/MustExecute.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/ScalarEvolutionExpressions.h"
#include "llvm/Analysis/ScalarEvolutionPatternMatch.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/Constant.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DebugLoc.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/GlobalValue.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/InstrTypes.h"
#include "llvm/IR/Instruction.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/PassManager.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/IR/ProfDataUtils.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/User.h"
#include "llvm/IR/Value.h"
#include "llvm/IR/ValueHandle.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/InstructionCost.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/BuildLibCalls.h"
#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/LoopUtils.h"
#include "llvm/Transforms/Utils/ScalarEvolutionExpander.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <utility>
using namespace llvm;
using namespace SCEVPatternMatch;
#define DEBUG_TYPE "loop-idiom"
STATISTIC(NumMemSet, "Number of memset's formed from loop stores");
STATISTIC(NumMemCpy, "Number of memcpy's formed from loop load+stores");
STATISTIC(NumMemMove, "Number of memmove's formed from loop load+stores");
STATISTIC(NumStrLen, "Number of strlen's and wcslen's formed from loop loads");
STATISTIC(
NumShiftUntilBitTest,
"Number of uncountable loops recognized as 'shift until bitttest' idiom");
STATISTIC(NumShiftUntilZero,
"Number of uncountable loops recognized as 'shift until zero' idiom");
namespace llvm {
bool DisableLIRP::All;
static cl::opt<bool, true>
DisableLIRPAll("disable-" DEBUG_TYPE "-all",
cl::desc("Options to disable Loop Idiom Recognize Pass."),
cl::location(DisableLIRP::All), cl::init(false),
cl::ReallyHidden);
bool DisableLIRP::Memset;
static cl::opt<bool, true>
DisableLIRPMemset("disable-" DEBUG_TYPE "-memset",
cl::desc("Proceed with loop idiom recognize pass, but do "
"not convert loop(s) to memset."),
cl::location(DisableLIRP::Memset), cl::init(false),
cl::ReallyHidden);
bool DisableLIRP::Memcpy;
static cl::opt<bool, true>
DisableLIRPMemcpy("disable-" DEBUG_TYPE "-memcpy",
cl::desc("Proceed with loop idiom recognize pass, but do "
"not convert loop(s) to memcpy."),
cl::location(DisableLIRP::Memcpy), cl::init(false),
cl::ReallyHidden);
bool DisableLIRP::Strlen;
static cl::opt<bool, true>
DisableLIRPStrlen("disable-loop-idiom-strlen",
cl::desc("Proceed with loop idiom recognize pass, but do "
"not convert loop(s) to strlen."),
cl::location(DisableLIRP::Strlen), cl::init(false),
cl::ReallyHidden);
bool DisableLIRP::Wcslen;
static cl::opt<bool, true>
EnableLIRPWcslen("disable-loop-idiom-wcslen",
cl::desc("Proceed with loop idiom recognize pass, "
"enable conversion of loop(s) to wcslen."),
cl::location(DisableLIRP::Wcslen), cl::init(false),
cl::ReallyHidden);
bool DisableLIRP::HashRecognize;
static cl::opt<bool, true>
DisableLIRPHashRecognize("disable-" DEBUG_TYPE "-hashrecognize",
cl::desc("Proceed with loop idiom recognize pass, "
"but do not do hash-recognize analysis."),
cl::location(DisableLIRP::HashRecognize),
cl::init(false), cl::ReallyHidden);
} // namespace llvm
namespace {
class LoopIdiomRecognize {
const ScalarOptions &Opts;
Loop *CurLoop = nullptr;
AliasAnalysis *AA;
DominatorTree *DT;
LoopInfo *LI;
ScalarEvolution *SE;
TargetLibraryInfo *TLI;
const TargetTransformInfo *TTI;
const DataLayout *DL;
OptimizationRemarkEmitter &ORE;
bool ApplyCodeSizeHeuristics;
std::unique_ptr<MemorySSAUpdater> MSSAU;
public:
explicit LoopIdiomRecognize(const ScalarOptions &Opts, AliasAnalysis *AA,
DominatorTree *DT, LoopInfo *LI,
ScalarEvolution *SE, TargetLibraryInfo *TLI,
const TargetTransformInfo *TTI, MemorySSA *MSSA,
const DataLayout *DL,
OptimizationRemarkEmitter &ORE)
: Opts(Opts), AA(AA), DT(DT), LI(LI), SE(SE), TLI(TLI), TTI(TTI), DL(DL),
ORE(ORE) {
if (MSSA)
MSSAU = std::make_unique<MemorySSAUpdater>(MSSA);
}
bool runOnLoop(Loop *L);
private:
using StoreList = SmallVector<StoreInst *, 8>;
using StoreListMap = MapVector<Value *, StoreList>;
StoreListMap StoreRefsForMemset;
StoreListMap StoreRefsForMemsetPattern;
StoreList StoreRefsForMemcpy;
bool HasMemset;
bool HasMemsetPattern;
bool HasMemcpy;
/// Return code for isLegalStore()
enum LegalStoreKind {
None = 0,
Memset,
MemsetPattern,
Memcpy,
UnorderedAtomicMemcpy,
DontUse // Dummy retval never to be used. Allows catching errors in retval
// handling.
};
/// \name Countable Loop Idiom Handling
/// @{
bool runOnCountableLoop();
bool runOnLoopBlock(BasicBlock *BB, const SCEV *BECount,
SmallVectorImpl<BasicBlock *> &ExitBlocks);
void collectStores(BasicBlock *BB);
LegalStoreKind isLegalStore(StoreInst *SI);
enum class ForMemset { No, Yes };
bool processLoopStores(SmallVectorImpl<StoreInst *> &SL, const SCEV *BECount,
ForMemset For);
template <typename MemInst>
bool processLoopMemIntrinsic(
BasicBlock *BB,
bool (LoopIdiomRecognize::*Processor)(MemInst *, const SCEV *),
const SCEV *BECount);
bool processLoopMemCpy(MemCpyInst *MCI, const SCEV *BECount);
bool processLoopMemSet(MemSetInst *MSI, const SCEV *BECount);
bool processLoopStridedStore(Value *DestPtr, const SCEV *StoreSizeSCEV,
MaybeAlign StoreAlignment, Value *StoredVal,
Instruction *TheStore,
SmallPtrSetImpl<Instruction *> &Stores,
const SCEVAddRecExpr *Ev, const SCEV *BECount,
bool IsNegStride, bool IsLoopMemset = false);
bool processLoopStoreOfLoopLoad(StoreInst *SI, const SCEV *BECount);
bool processLoopStoreOfLoopLoad(Value *DestPtr, Value *SourcePtr,
const SCEV *StoreSize, MaybeAlign StoreAlign,
MaybeAlign LoadAlign, Instruction *TheStore,
Instruction *TheLoad,
const SCEVAddRecExpr *StoreEv,
const SCEVAddRecExpr *LoadEv,
const SCEV *BECount);
bool avoidLIRForMultiBlockLoop(bool IsMemset = false,
bool IsLoopMemset = false);
bool optimizeCRCLoop(const PolynomialInfo &Info);
void optimizeCRCLoopUsingClmul(const PolynomialInfo &Info);
void optimizeCRCLoopUsingTableLookup(const PolynomialInfo &Info);
/// @}
/// \name Noncountable Loop Idiom Handling
/// @{
bool runOnNoncountableLoop();
bool recognizePopcount();
void transformLoopToPopcount(BasicBlock *PreCondBB, Instruction *CntInst,
PHINode *CntPhi, Value *Var);
bool isProfitableToInsertFFS(Intrinsic::ID IntrinID, Value *InitX,
bool ZeroCheck, size_t CanonicalSize);
bool insertFFSIfProfitable(Intrinsic::ID IntrinID, Value *InitX,
Instruction *DefX, PHINode *CntPhi,
Instruction *CntInst);
bool recognizeAndInsertFFS(); /// Find First Set: ctlz or cttz
bool recognizeShiftUntilLessThan();
void transformLoopToCountable(Intrinsic::ID IntrinID, BasicBlock *PreCondBB,
Instruction *CntInst, PHINode *CntPhi,
Value *Var, Instruction *DefX,
const DebugLoc &DL, bool ZeroCheck,
bool IsCntPhiUsedOutsideLoop,
bool InsertSub = false);
bool recognizeShiftUntilBitTest();
bool recognizeShiftUntilZero();
bool recognizeAndInsertStrLen();
/// @}
};
} // end anonymous namespace
PreservedAnalyses LoopIdiomRecognizePass::run(Loop &L, LoopAnalysisManager &AM,
LoopStandardAnalysisResults &AR,
LPMUpdater &) {
if (DisableLIRP::All)
return PreservedAnalyses::all();
const auto *DL = &L.getHeader()->getDataLayout();
// For the new PM, we also can't use OptimizationRemarkEmitter as an analysis
// pass. Function analyses need to be preserved across loop transformations
// but ORE cannot be preserved (see comment before the pass definition).
OptimizationRemarkEmitter ORE(L.getHeader()->getParent());
const ScalarOptions &Opts =
L.getHeader()->getContext().getLibraryOptions().get<ScalarOptions>();
LoopIdiomRecognize LIR(Opts, &AR.AA, &AR.DT, &AR.LI, &AR.SE, &AR.TLI, &AR.TTI,
AR.MSSA, DL, ORE);
if (!LIR.runOnLoop(&L))
return PreservedAnalyses::all();
auto PA = getLoopPassPreservedAnalyses();
if (AR.MSSA)
PA.preserve<MemorySSAAnalysis>();
return PA;
}
static void deleteDeadInstruction(Instruction *I) {
I->replaceAllUsesWith(PoisonValue::get(I->getType()));
I->eraseFromParent();
}
//===----------------------------------------------------------------------===//
//
// Implementation of LoopIdiomRecognize
//
//===----------------------------------------------------------------------===//
bool LoopIdiomRecognize::runOnLoop(Loop *L) {
CurLoop = L;
// If the loop could not be converted to canonical form, it must have an
// indirectbr in it, just give up.
if (!L->getLoopPreheader())
return false;
// Disable loop idiom recognition if the function's name is a common idiom.
StringRef Name = L->getHeader()->getParent()->getName();
if (Name == "memset" || Name == "memcpy" || Name == "strlen" ||
Name == "wcslen")
return false;
// Determine if code size heuristics need to be applied.
ApplyCodeSizeHeuristics =
L->getHeader()->getParent()->hasOptSize() && Opts.use_lir_code_size_heurs;
HasMemset = TLI->has(LibFunc_memset);
// TODO: Unconditionally enable use of the memset pattern intrinsic (or at
// least, opt-in via target hook) once we are confident it will never result
// in worse codegen than without. For now, use it only when the target
// supports memset_pattern16 libcall (or unless this is overridden by
// command line option).
HasMemsetPattern = TLI->has(LibFunc_memset_pattern16);
HasMemcpy = TLI->has(LibFunc_memcpy);
if (HasMemset || HasMemsetPattern ||
Opts.loop_idiom_force_memset_pattern_intrinsic || HasMemcpy ||
!DisableLIRP::HashRecognize)
if (SE->hasLoopInvariantBackedgeTakenCount(L))
return runOnCountableLoop();
return runOnNoncountableLoop();
}
bool LoopIdiomRecognize::runOnCountableLoop() {
const SCEV *BECount = SE->getBackedgeTakenCount(CurLoop);
assert(!isa<SCEVCouldNotCompute>(BECount) &&
"runOnCountableLoop() called on a loop without a predictable"
"backedge-taken count");
// If this loop executes exactly one time, then it should be peeled, not
// optimized by this pass.
if (BECount->isZero())
return false;
SmallVector<BasicBlock *, 8> ExitBlocks;
CurLoop->getUniqueExitBlocks(ExitBlocks);
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Scanning: F["
<< CurLoop->getHeader()->getParent()->getName()
<< "] Countable Loop %" << CurLoop->getHeader()->getName()
<< "\n");
// The following transforms hoist stores/memsets into the loop pre-header.
// Give up if the loop has instructions that may throw.
SimpleLoopSafetyInfo SafetyInfo(CurLoop);
if (SafetyInfo.anyBlockMayThrow())
return false;
bool MadeChange = false;
// Scan all the blocks in the loop that are not in subloops.
for (auto *BB : CurLoop->getBlocks()) {
// Ignore blocks in subloops.
if (LI->getLoopFor(BB) != CurLoop)
continue;
MadeChange |= runOnLoopBlock(BB, BECount, ExitBlocks);
}
// Attempt to optimize a CRC loop if one is detected by HashRecognize.
if (!DisableLIRP::HashRecognize &&
Opts.loop_idiom_crc_strategy != CRCStrategyKind::Disable)
if (auto Res = HashRecognize(*CurLoop, *SE).getResult())
MadeChange |= optimizeCRCLoop(*Res);
return MadeChange;
}
static APInt getStoreStride(const SCEVAddRecExpr *StoreEv) {
const SCEVConstant *ConstStride = cast<SCEVConstant>(StoreEv->getOperand(1));
return ConstStride->getAPInt();
}
/// getMemSetPatternValue - If a strided store of the specified value is safe to
/// turn into a memset.patternn intrinsic, return the Constant that should
/// be passed in. Otherwise, return null.
///
/// TODO this function could allow more constants than it does today (e.g.
/// those over 16 bytes) now it has transitioned to being used for the
/// memset.pattern intrinsic rather than directly the memset_pattern16
/// libcall.
static Constant *getMemSetPatternValue(Value *V, const DataLayout *DL) {
// FIXME: This could check for UndefValue because it can be merged into any
// other valid pattern.
// If the value isn't a constant, we can't promote it to being in a constant
// array. We could theoretically do a store to an alloca or something, but
// that doesn't seem worthwhile.
Constant *C = dyn_cast<Constant>(V);
if (!C || isa<ConstantExpr>(C))
return nullptr;
// Only handle simple values that are a power of two bytes in size.
uint64_t Size = DL->getTypeSizeInBits(V->getType());
if (Size == 0 || (Size & 7) || (Size & (Size - 1)))
return nullptr;
// Don't care enough about darwin/ppc to implement this.
if (DL->isBigEndian())
return nullptr;
// Convert to size in bytes.
Size /= 8;
// TODO: If CI is larger than 16-bytes, we can try slicing it in half to see
// if the top and bottom are the same (e.g. for vectors and large integers).
if (Size > 16)
return nullptr;
// For now, don't handle types that aren't int, floats, or pointers.
Type *CTy = C->getType();
if (!CTy->isIntOrPtrTy() && !CTy->isFloatingPointTy())
return nullptr;
return C;
}
LoopIdiomRecognize::LegalStoreKind
LoopIdiomRecognize::isLegalStore(StoreInst *SI) {
// Don't touch volatile stores.
if (SI->isVolatile())
return LegalStoreKind::None;
// We only want simple or unordered-atomic stores.
if (!SI->isUnordered())
return LegalStoreKind::None;
// Avoid merging nontemporal stores.
if (SI->getMetadata(LLVMContext::MD_nontemporal))
return LegalStoreKind::None;
Value *StoredVal = SI->getValueOperand();
Value *StorePtr = SI->getPointerOperand();
if (DL->hasUnstableRepresentation(StoredVal->getType()))
return LegalStoreKind::None;
// Transformations could invalidate the external-state pointers
// memcpy - LangRef specifies that a valid memcpy must preserve external
// state, so no transformations are blocked by it.
// memset - We assume that a memset of 0 has an equivalent external state
// effect as a null pointer store. This is currently not explicitly
// specified, but is true of the one exemplar we have (CHERI
// capabilities). All other memset formations are not safe.
bool MustPreserveExternalState = DL->hasExternalState(StoredVal->getType()) &&
!isa<ConstantPointerNull>(StoredVal);
// Reject stores that are so large that they overflow an unsigned.
// When storing out scalable vectors we bail out for now, since the code
// below currently only works for constant strides.
TypeSize SizeInBits = DL->getTypeSizeInBits(StoredVal->getType());
if (SizeInBits.isScalable() || (SizeInBits.getFixedValue() & 7) ||
(SizeInBits.getFixedValue() >> 32) != 0)
return LegalStoreKind::None;
// See if the pointer expression is an AddRec like {base,+,1} on the current
// loop, which indicates a strided store. If we have something else, it's a
// random store we can't handle.
const SCEV *StoreEv = SE->getSCEV(StorePtr);
const SCEVConstant *Stride;
if (!match(StoreEv, m_scev_AffineAddRec(m_SCEV(), m_SCEVConstant(Stride),
m_SpecificLoop(CurLoop))))
return LegalStoreKind::None;
// See if the store can be turned into a memset.
// If the stored value is a byte-wise value (like i32 -1), then it may be
// turned into a memset of i8 -1, assuming that all the consecutive bytes
// are stored. A store of i32 0x01020304 can never be turned into a memset,
// but it can be turned into memset_pattern if the target supports it.
Value *SplatValue = isBytewiseValue(StoredVal, *DL);
// Note: memset and memset_pattern on unordered-atomic is yet not supported
bool UnorderedAtomic = SI->isUnordered() && !SI->isSimple();
// If we're allowed to form a memset, and the stored value would be
// acceptable for memset, use it.
if (!MustPreserveExternalState && !UnorderedAtomic && HasMemset &&
SplatValue && !DisableLIRP::Memset &&
// Verify that the stored value is loop invariant. If not, we can't
// promote the memset.
CurLoop->isLoopInvariant(SplatValue)) {
// It looks like we can use SplatValue.
return LegalStoreKind::Memset;
}
if (!MustPreserveExternalState && !UnorderedAtomic &&
(HasMemsetPattern || Opts.loop_idiom_force_memset_pattern_intrinsic) &&
!DisableLIRP::Memset &&
// Don't create memset_pattern16s with address spaces.
StorePtr->getType()->getPointerAddressSpace() == 0 &&
getMemSetPatternValue(StoredVal, DL)) {
// It looks like we can use PatternValue!
return LegalStoreKind::MemsetPattern;
}
// Otherwise, see if the store can be turned into a memcpy.
if (HasMemcpy && !DisableLIRP::Memcpy) {
// Check to see if the stride matches the size of the store. If so, then we
// know that every byte is touched in the loop.
unsigned StoreSize = DL->getTypeStoreSize(SI->getValueOperand()->getType());
APInt StrideAP = Stride->getAPInt();
if (StoreSize != StrideAP && StoreSize != -StrideAP)
return LegalStoreKind::None;
// The store must be feeding a non-volatile load.
LoadInst *LI = dyn_cast<LoadInst>(SI->getValueOperand());
// Only allow non-volatile loads
if (!LI || LI->isVolatile())
return LegalStoreKind::None;
// Only allow simple or unordered-atomic loads
if (!LI->isUnordered())
return LegalStoreKind::None;
// See if the pointer expression is an AddRec like {base,+,1} on the current
// loop, which indicates a strided load. If we have something else, it's a
// random load we can't handle.
const SCEV *LoadEv = SE->getSCEV(LI->getPointerOperand());
// The store and load must share the same stride.
if (!match(LoadEv, m_scev_AffineAddRec(m_SCEV(), m_scev_Specific(Stride),
m_SpecificLoop(CurLoop))))
return LegalStoreKind::None;
// Success. This store can be converted into a memcpy.
UnorderedAtomic = UnorderedAtomic || LI->isAtomic();
return UnorderedAtomic ? LegalStoreKind::UnorderedAtomicMemcpy
: LegalStoreKind::Memcpy;
}
// This store can't be transformed into a memset/memcpy.
return LegalStoreKind::None;
}
void LoopIdiomRecognize::collectStores(BasicBlock *BB) {
StoreRefsForMemset.clear();
StoreRefsForMemsetPattern.clear();
StoreRefsForMemcpy.clear();
for (Instruction &I : *BB) {
StoreInst *SI = dyn_cast<StoreInst>(&I);
if (!SI)
continue;
// Make sure this is a strided store with a constant stride.
switch (isLegalStore(SI)) {
case LegalStoreKind::None:
// Nothing to do
break;
case LegalStoreKind::Memset: {
// Find the base pointer.
Value *Ptr = getUnderlyingObject(SI->getPointerOperand());
StoreRefsForMemset[Ptr].push_back(SI);
} break;
case LegalStoreKind::MemsetPattern: {
// Find the base pointer.
Value *Ptr = getUnderlyingObject(SI->getPointerOperand());
StoreRefsForMemsetPattern[Ptr].push_back(SI);
} break;
case LegalStoreKind::Memcpy:
case LegalStoreKind::UnorderedAtomicMemcpy:
StoreRefsForMemcpy.push_back(SI);
break;
default:
assert(false && "unhandled return value");
break;
}
}
}
/// runOnLoopBlock - Process the specified block, which lives in a counted loop
/// with the specified backedge count. This block is known to be in the current
/// loop and not in any subloops.
bool LoopIdiomRecognize::runOnLoopBlock(
BasicBlock *BB, const SCEV *BECount,
SmallVectorImpl<BasicBlock *> &ExitBlocks) {
// We can only promote stores in this block if they are unconditionally
// executed in the loop. For a block to be unconditionally executed, it has
// to dominate all the exit blocks of the loop. Verify this now.
for (BasicBlock *ExitBlock : ExitBlocks)
if (!DT->dominates(BB, ExitBlock))
return false;
bool MadeChange = false;
// Look for store instructions, which may be optimized to memset/memcpy.
collectStores(BB);
// Look for a single store or sets of stores with a common base, which can be
// optimized into a memset (memset_pattern). The latter most commonly happens
// with structs and handunrolled loops.
for (auto &SL : StoreRefsForMemset)
MadeChange |= processLoopStores(SL.second, BECount, ForMemset::Yes);
for (auto &SL : StoreRefsForMemsetPattern)
MadeChange |= processLoopStores(SL.second, BECount, ForMemset::No);
// Optimize the store into a memcpy, if it feeds an similarly strided load.
for (auto &SI : StoreRefsForMemcpy)
MadeChange |= processLoopStoreOfLoopLoad(SI, BECount);
MadeChange |= processLoopMemIntrinsic<MemCpyInst>(
BB, &LoopIdiomRecognize::processLoopMemCpy, BECount);
MadeChange |= processLoopMemIntrinsic<MemSetInst>(
BB, &LoopIdiomRecognize::processLoopMemSet, BECount);
return MadeChange;
}
/// See if this store(s) can be promoted to a memset.
bool LoopIdiomRecognize::processLoopStores(SmallVectorImpl<StoreInst *> &SL,
const SCEV *BECount, ForMemset For) {
// Try to find consecutive stores that can be transformed into memsets.
SetVector<StoreInst *> Heads, Tails;
SmallDenseMap<StoreInst *, StoreInst *> ConsecutiveChain;
// Do a quadratic search on all of the given stores and find
// all of the pairs of stores that follow each other.
SmallVector<unsigned, 16> IndexQueue;
for (unsigned i = 0, e = SL.size(); i < e; ++i) {
assert(SL[i]->isSimple() && "Expected only non-volatile stores.");
Value *FirstStoredVal = SL[i]->getValueOperand();
Value *FirstStorePtr = SL[i]->getPointerOperand();
const SCEVAddRecExpr *FirstStoreEv =
cast<SCEVAddRecExpr>(SE->getSCEV(FirstStorePtr));
APInt FirstStride = getStoreStride(FirstStoreEv);
unsigned FirstStoreSize = DL->getTypeStoreSize(SL[i]->getValueOperand()->getType());
// See if we can optimize just this store in isolation.
if (FirstStride == FirstStoreSize || -FirstStride == FirstStoreSize) {
Heads.insert(SL[i]);
continue;
}
Value *FirstSplatValue = nullptr;
Constant *FirstPatternValue = nullptr;
if (For == ForMemset::Yes)
FirstSplatValue = isBytewiseValue(FirstStoredVal, *DL);
else
FirstPatternValue = getMemSetPatternValue(FirstStoredVal, DL);
assert((FirstSplatValue || FirstPatternValue) &&
"Expected either splat value or pattern value.");
IndexQueue.clear();
// If a store has multiple consecutive store candidates, search Stores
// array according to the sequence: from i+1 to e, then from i-1 to 0.
// This is because usually pairing with immediate succeeding or preceding
// candidate create the best chance to find memset opportunity.
unsigned j = 0;
for (j = i + 1; j < e; ++j)
IndexQueue.push_back(j);
for (j = i; j > 0; --j)
IndexQueue.push_back(j - 1);
for (auto &k : IndexQueue) {
assert(SL[k]->isSimple() && "Expected only non-volatile stores.");
Value *SecondStorePtr = SL[k]->getPointerOperand();
const SCEVAddRecExpr *SecondStoreEv =
cast<SCEVAddRecExpr>(SE->getSCEV(SecondStorePtr));
APInt SecondStride = getStoreStride(SecondStoreEv);
if (FirstStride != SecondStride)
continue;
Value *SecondStoredVal = SL[k]->getValueOperand();
Value *SecondSplatValue = nullptr;
Constant *SecondPatternValue = nullptr;
if (For == ForMemset::Yes)
SecondSplatValue = isBytewiseValue(SecondStoredVal, *DL);
else
SecondPatternValue = getMemSetPatternValue(SecondStoredVal, DL);
assert((SecondSplatValue || SecondPatternValue) &&
"Expected either splat value or pattern value.");
if (isConsecutiveAccess(SL[i], SL[k], *DL, *SE, false)) {
if (For == ForMemset::Yes) {
if (isa<UndefValue>(FirstSplatValue))
FirstSplatValue = SecondSplatValue;
if (FirstSplatValue != SecondSplatValue)
continue;
} else {
if (isa<UndefValue>(FirstPatternValue))
FirstPatternValue = SecondPatternValue;
if (FirstPatternValue != SecondPatternValue)
continue;
}
Tails.insert(SL[k]);
Heads.insert(SL[i]);
ConsecutiveChain[SL[i]] = SL[k];
break;
}
}
}
// We may run into multiple chains that merge into a single chain. We mark the
// stores that we transformed so that we don't visit the same store twice.
SmallPtrSet<Value *, 16> TransformedStores;
bool Changed = false;
// For stores that start but don't end a link in the chain:
for (StoreInst *I : Heads) {
if (Tails.count(I))
continue;
// We found a store instr that starts a chain. Now follow the chain and try
// to transform it.
SmallPtrSet<Instruction *, 8> AdjacentStores;
StoreInst *HeadStore = I;
unsigned StoreSize = 0;
// Collect the chain into a list.
while (Tails.count(I) || Heads.count(I)) {
if (TransformedStores.count(I))
break;
AdjacentStores.insert(I);
StoreSize += DL->getTypeStoreSize(I->getValueOperand()->getType());
// Move to the next value in the chain.
I = ConsecutiveChain[I];
}
Value *StoredVal = HeadStore->getValueOperand();
Value *StorePtr = HeadStore->getPointerOperand();
const SCEVAddRecExpr *StoreEv = cast<SCEVAddRecExpr>(SE->getSCEV(StorePtr));
APInt Stride = getStoreStride(StoreEv);
// Check to see if the stride matches the size of the stores. If so, then
// we know that every byte is touched in the loop.
if (StoreSize != Stride && StoreSize != -Stride)
continue;
bool IsNegStride = StoreSize == -Stride;
Type *IntIdxTy = DL->getIndexType(StorePtr->getType());
const SCEV *StoreSizeSCEV = SE->getConstant(IntIdxTy, StoreSize);
if (processLoopStridedStore(StorePtr, StoreSizeSCEV,
MaybeAlign(HeadStore->getAlign()), StoredVal,
HeadStore, AdjacentStores, StoreEv, BECount,
IsNegStride)) {
TransformedStores.insert_range(AdjacentStores);
Changed = true;
}
}
return Changed;
}
/// processLoopMemIntrinsic - Template function for calling different processor
/// functions based on mem intrinsic type.
template <typename MemInst>
bool LoopIdiomRecognize::processLoopMemIntrinsic(
BasicBlock *BB,
bool (LoopIdiomRecognize::*Processor)(MemInst *, const SCEV *),
const SCEV *BECount) {
bool MadeChange = false;
for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E;) {
Instruction *Inst = &*I++;
// Look for memory instructions, which may be optimized to a larger one.
if (MemInst *MI = dyn_cast<MemInst>(Inst)) {
WeakTrackingVH InstPtr(&*I);
if (!(this->*Processor)(MI, BECount))
continue;
MadeChange = true;
// If processing the instruction invalidated our iterator, start over from
// the top of the block.
if (!InstPtr)
I = BB->begin();
}
}
return MadeChange;
}
/// processLoopMemCpy - See if this memcpy can be promoted to a large memcpy
bool LoopIdiomRecognize::processLoopMemCpy(MemCpyInst *MCI,
const SCEV *BECount) {
// We can only handle non-volatile memcpys with a constant size.
if (MCI->isVolatile() || !isa<ConstantInt>(MCI->getLength()))
return false;
// If we're not allowed to hack on memcpy, we fail. We don't mess with the
// inlined version as generating a larger inline mempcy could affect code
// size.
if (!HasMemcpy || MCI->isForceInlined() || DisableLIRP::Memcpy)
return false;
Value *Dest = MCI->getDest();
Value *Source = MCI->getSource();
if (!Dest || !Source)
return false;
// See if the load and store pointer expressions are AddRec like {base,+,1} on
// the current loop, which indicates a strided load and store. If we have
// something else, it's a random load or store we can't handle.
const SCEV *StoreEv = SE->getSCEV(Dest);
const SCEV *LoadEv = SE->getSCEV(Source);
const APInt *StoreStrideValue, *LoadStrideValue;
if (!match(StoreEv,
m_scev_AffineAddRec(m_SCEV(), m_scev_APInt(StoreStrideValue),
m_SpecificLoop(CurLoop))) ||
!match(LoadEv,
m_scev_AffineAddRec(m_SCEV(), m_scev_APInt(LoadStrideValue),
m_SpecificLoop(CurLoop))))
return false;
// Reject memcpys that are so large that they overflow an unsigned.
uint64_t SizeInBytes = cast<ConstantInt>(MCI->getLength())->getZExtValue();
if ((SizeInBytes >> 32) != 0)
return false;
// Huge stride value - give up
if (StoreStrideValue->getBitWidth() > 64 ||
LoadStrideValue->getBitWidth() > 64)
return false;
if (SizeInBytes != *StoreStrideValue && SizeInBytes != -*StoreStrideValue) {
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "SizeStrideUnequal", MCI)
<< ore::NV("Inst", "memcpy") << " in "
<< ore::NV("Function", MCI->getFunction())
<< " function will not be hoisted: "
<< ore::NV("Reason", "memcpy size is not equal to stride");
});
return false;
}
int64_t StoreStrideInt = StoreStrideValue->getSExtValue();
int64_t LoadStrideInt = LoadStrideValue->getSExtValue();
// Check if the load stride matches the store stride.
if (StoreStrideInt != LoadStrideInt)
return false;
return processLoopStoreOfLoopLoad(
Dest, Source, SE->getConstant(Dest->getType(), SizeInBytes),
MCI->getDestAlign(), MCI->getSourceAlign(), MCI, MCI,
cast<SCEVAddRecExpr>(StoreEv), cast<SCEVAddRecExpr>(LoadEv), BECount);
}
/// processLoopMemSet - See if this memset can be promoted to a large memset.
bool LoopIdiomRecognize::processLoopMemSet(MemSetInst *MSI,
const SCEV *BECount) {
// We can only handle non-volatile memsets.
if (MSI->isVolatile())
return false;
// If we're not allowed to hack on memset, we fail. We don't mess with the
// inlined version as generating a larger memset could affect code size.
if (!HasMemset || MSI->isForceInlined() || DisableLIRP::Memset)
return false;
Value *Pointer = MSI->getDest();
// See if the pointer expression is an AddRec like {base,+,1} on the current
// loop, which indicates a strided store. If we have something else, it's a
// random store we can't handle.
const SCEV *Ev = SE->getSCEV(Pointer);
const SCEV *PointerStrideSCEV;
if (!match(Ev, m_scev_AffineAddRec(m_SCEV(), m_SCEV(PointerStrideSCEV),
m_SpecificLoop(CurLoop)))) {
LLVM_DEBUG(dbgs() << " Pointer is not affine, abort\n");
return false;
}
SCEVUse MemsetSizeSCEV = SE->getSCEV(MSI->getLength());
bool IsNegStride = false;
const bool IsConstantSize = isa<ConstantInt>(MSI->getLength());
if (IsConstantSize) {
// Memset size is constant.
// Check if the pointer stride matches the memset size. If so, then
// we know that every byte is touched in the loop.
LLVM_DEBUG(dbgs() << " memset size is constant\n");
uint64_t SizeInBytes = cast<ConstantInt>(MSI->getLength())->getZExtValue();
const APInt *Stride;
if (!match(PointerStrideSCEV, m_scev_APInt(Stride)))
return false;
if (SizeInBytes != *Stride && SizeInBytes != -*Stride)
return false;
IsNegStride = SizeInBytes == -*Stride;
} else {
// Memset size is non-constant.
// Check if the pointer stride matches the memset size.
// To be conservative, the pass would not promote pointers that aren't in
// address space zero. Also, the pass only handles memset length and stride
// that are invariant for the top level loop.
LLVM_DEBUG(dbgs() << " memset size is non-constant\n");
if (Pointer->getType()->getPointerAddressSpace() != 0) {
LLVM_DEBUG(dbgs() << " pointer is not in address space zero, "
<< "abort\n");
return false;
}
if (!SE->isLoopInvariant(MemsetSizeSCEV, CurLoop)) {
LLVM_DEBUG(dbgs() << " memset size is not a loop-invariant, "
<< "abort\n");
return false;
}
// Compare positive direction PointerStrideSCEV with MemsetSizeSCEV
IsNegStride = PointerStrideSCEV->isNonConstantNegative();
SCEVUse PositiveStrideSCEV =
IsNegStride ? SCEVUse(SE->getNegativeSCEV(PointerStrideSCEV))
: SCEVUse(PointerStrideSCEV);
LLVM_DEBUG(dbgs() << " MemsetSizeSCEV: " << *MemsetSizeSCEV << "\n"
<< " PositiveStrideSCEV: " << *PositiveStrideSCEV
<< "\n");
if (PositiveStrideSCEV != MemsetSizeSCEV) {
// If an expression is covered by the loop guard, compare again and
// proceed with optimization if equal.
const SCEV *FoldedPositiveStride =
SE->applyLoopGuards(PositiveStrideSCEV, CurLoop);
const SCEV *FoldedMemsetSize =
SE->applyLoopGuards(MemsetSizeSCEV, CurLoop);
LLVM_DEBUG(dbgs() << " Try to fold SCEV based on loop guard\n"
<< " FoldedMemsetSize: " << *FoldedMemsetSize << "\n"
<< " FoldedPositiveStride: " << *FoldedPositiveStride
<< "\n");
if (FoldedPositiveStride != FoldedMemsetSize) {
LLVM_DEBUG(dbgs() << " SCEV don't match, abort\n");
return false;
}
}
}
// Verify that the memset value is loop invariant. If not, we can't promote
// the memset.
Value *SplatValue = MSI->getValue();
if (!SplatValue || !CurLoop->isLoopInvariant(SplatValue))
return false;
SmallPtrSet<Instruction *, 1> MSIs;
MSIs.insert(MSI);
return processLoopStridedStore(Pointer, SE->getSCEV(MSI->getLength()),
MSI->getDestAlign(), SplatValue, MSI, MSIs,
cast<SCEVAddRecExpr>(Ev), BECount, IsNegStride,
/*IsLoopMemset=*/true);
}
/// Return true if \p I is a (simple, loop-invariant-valued) store of the same
/// bytewise value \p SplatByte.
static bool isSameByteValueStore(Instruction &I, Value *SplatByte, Loop *L,
const DataLayout &DL) {
assert(SplatByte && "expected a bytewise splat value to match against");
auto *SI = dyn_cast<StoreInst>(&I);
if (!SI || !SI->isSimple() || !L->isLoopInvariant(SI->getValueOperand()))
return false;
return isBytewiseValue(SI->getValueOperand(), DL) == SplatByte;
}
/// mayLoopAccessLocation - Return true if the specified loop might access the
/// specified pointer location, which is a loop-strided access. The 'Access'
/// argument specifies what the verboten forms of access are (read or write).
///
/// When the access size cannot be bounded, fall back to allow stores writing
/// the same byte value \p SplatByte.
static bool mayLoopAccessLocation(Value *Ptr, ModRefInfo Access, Loop *L,
const SCEV *BECount,
const SCEV *StoreSizeSCEV, AliasAnalysis &AA,
SmallPtrSetImpl<Instruction *> &IgnoredInsts,
Value *SplatByte = nullptr,
const DataLayout *DL = nullptr) {
// Get the location that may be stored across the loop. Since the access is
// strided positively through memory, we say that the modified location starts
// at the pointer and has infinite size.
LocationSize AccessSize = LocationSize::afterPointer();
// If the loop iterates a fixed number of times, we can refine the access size
// to be exactly the size of the memset, which is (BECount+1)*StoreSize
const APInt *BECst, *ConstSize;
if (match(BECount, m_scev_APInt(BECst)) &&
match(StoreSizeSCEV, m_scev_APInt(ConstSize))) {
std::optional<uint64_t> BEInt = BECst->tryZExtValue();
std::optional<uint64_t> SizeInt = ConstSize->tryZExtValue();
// FIXME: Should this check for overflow?
if (BEInt && SizeInt)
AccessSize = LocationSize::precise((*BEInt + 1) * *SizeInt);
}
// TODO: For this to be really effective, we have to dive into the pointer
// operand in the store. Store to &A[i] of 100 will always return may alias
// with store of &A[100], we need to StoreLoc to be "A" with size of 100,
// which will then no-alias a store to &A[100].
MemoryLocation StoreLoc(Ptr, AccessSize);
// Only consult the same-byte-value fallback when the access size stayed
// infinite (non-constant trip count); with a precise size AA is accurate.
bool TrySameByteValue = !AccessSize.isPrecise() && SplatByte && DL;
for (BasicBlock *B : L->blocks())
for (Instruction &I : *B)
if (!IgnoredInsts.contains(&I) &&
isModOrRefSet(AA.getModRefInfo(&I, StoreLoc) & Access)) {
if (TrySameByteValue && isSameByteValueStore(I, SplatByte, L, *DL))
continue;
return true;
}
return false;
}
// If we have a negative stride, Start refers to the end of the memory location
// we're trying to memset. Therefore, we need to recompute the base pointer,
// which is just Start - BECount*Size.
static const SCEV *getStartForNegStride(const SCEV *Start, const SCEV *BECount,
Type *IntPtr, const SCEV *StoreSizeSCEV,
ScalarEvolution *SE) {
const SCEV *Index = SE->getTruncateOrZeroExtend(BECount, IntPtr);
if (!StoreSizeSCEV->isOne()) {
// index = back edge count * store size
Index = SE->getMulExpr(Index,
SE->getTruncateOrZeroExtend(StoreSizeSCEV, IntPtr),
SCEV::FlagNUW);
}
// base pointer = start - index * store size
return SE->getMinusSCEV(Start, Index);
}
/// Compute the number of bytes as a SCEV from the backedge taken count.
///
/// This also maps the SCEV into the provided type and tries to handle the
/// computation in a way that will fold cleanly.
static const SCEV *getNumBytes(const SCEV *BECount, Type *IntPtr,
const SCEV *StoreSizeSCEV, Loop *CurLoop,
const DataLayout *DL, ScalarEvolution *SE) {
const SCEV *TripCountSCEV =
SE->getTripCountFromExitCount(BECount, IntPtr, CurLoop);
return SE->getMulExpr(TripCountSCEV,
SE->getTruncateOrZeroExtend(StoreSizeSCEV, IntPtr),
SCEV::FlagNUW);
}
/// processLoopStridedStore - We see a strided store of some value. If we can
/// transform this into a memset or memset_pattern in the loop preheader, do so.
bool LoopIdiomRecognize::processLoopStridedStore(
Value *DestPtr, const SCEV *StoreSizeSCEV, MaybeAlign StoreAlignment,
Value *StoredVal, Instruction *TheStore,
SmallPtrSetImpl<Instruction *> &Stores, const SCEVAddRecExpr *Ev,
const SCEV *BECount, bool IsNegStride, bool IsLoopMemset) {
// We currently don't convert inline intrinsics into larger ones, to avoid
// code size increase. `processLoopMemSet` checks that the intrinsic is not
// inline before calling this function.
assert((isa<StoreInst>(TheStore) ||
!cast<MemIntrinsic>(TheStore)->isForceInlined()) &&
"inline mem intrinsics should be filtered out by callers");
Module *M = TheStore->getModule();
// The trip count of the loop and the base pointer of the addrec SCEV is
// guaranteed to be loop invariant, which means that it should dominate the
// header. This allows us to insert code for it in the preheader.
unsigned DestAS = DestPtr->getType()->getPointerAddressSpace();
BasicBlock *Preheader = CurLoop->getLoopPreheader();
IRBuilder<> Builder(Preheader->getTerminator());
SCEVExpander Expander(*SE, "loop-idiom");
SCEVExpanderCleaner ExpCleaner(Expander);
Type *DestInt8PtrTy = Builder.getPtrTy(DestAS);
Type *IntIdxTy = DL->getIndexType(DestPtr->getType());
bool Changed = false;
const SCEV *Start = Ev->getStart();
// Handle negative strided loops.
if (IsNegStride)
Start = getStartForNegStride(Start, BECount, IntIdxTy, StoreSizeSCEV, SE);
// TODO: ideally we should still be able to generate memset if SCEV expander
// is taught to generate the dependencies at the latest point.
if (!Expander.isSafeToExpand(Start))
return Changed;
// Okay, we have a strided store "p[i]" of a splattable value. We can turn
// this into a memset in the loop preheader now if we want. However, this
// would be unsafe to do if there is anything else in the loop that may read
// or write to the aliased location. Check for any overlap by generating the
// base pointer and checking the region.
Value *BasePtr =
Expander.expandCodeFor(Start, DestInt8PtrTy, Preheader->getTerminator());
// From here on out, conservatively report to the pass manager that we've
// changed the IR, even if we later clean up these added instructions. There
// may be structural differences e.g. in the order of use lists not accounted
// for in just a textual dump of the IR. This is written as a variable, even
// though statically all the places this dominates could be replaced with
// 'true', with the hope that anyone trying to be clever / "more precise" with
// the return value will read this comment, and leave them alone.
Changed = true;
Value *SplatValue = isBytewiseValue(StoredVal, *DL);
if (mayLoopAccessLocation(BasePtr, ModRefInfo::ModRef, CurLoop, BECount,
StoreSizeSCEV, *AA, Stores, SplatValue, DL))
return Changed;
if (avoidLIRForMultiBlockLoop(/*IsMemset=*/true, IsLoopMemset))
return Changed;
// Okay, everything looks good, insert the memset.
Constant *PatternValue = nullptr;
if (!SplatValue)
PatternValue = getMemSetPatternValue(StoredVal, DL);
// MemsetArg is the number of bytes for the memset libcall, and the number
// of pattern repetitions if the memset.pattern intrinsic is being used.
Value *MemsetArg;
std::optional<int64_t> BytesWritten;
if (PatternValue &&
(HasMemsetPattern || Opts.loop_idiom_force_memset_pattern_intrinsic)) {
const SCEV *TripCountS =
SE->getTripCountFromExitCount(BECount, IntIdxTy, CurLoop);
if (!Expander.isSafeToExpand(TripCountS))
return Changed;
const SCEVConstant *ConstStoreSize = dyn_cast<SCEVConstant>(StoreSizeSCEV);
if (!ConstStoreSize)
return Changed;
Value *TripCount = Expander.expandCodeFor(TripCountS, IntIdxTy,
Preheader->getTerminator());
uint64_t PatternRepsPerTrip =
(ConstStoreSize->getValue()->getZExtValue() * 8) /
DL->getTypeSizeInBits(PatternValue->getType());
// If ConstStoreSize is not equal to the width of PatternValue, then
// MemsetArg is TripCount * (ConstStoreSize/PatternValueWidth). Else
// MemSetArg is just TripCount.
MemsetArg =
PatternRepsPerTrip == 1
? TripCount
: Builder.CreateMul(TripCount,
Builder.getIntN(IntIdxTy->getIntegerBitWidth(),
PatternRepsPerTrip));
if (auto *CI = dyn_cast<ConstantInt>(TripCount))
BytesWritten =
CI->getZExtValue() * ConstStoreSize->getValue()->getZExtValue();
} else {
const SCEV *NumBytesS =
getNumBytes(BECount, IntIdxTy, StoreSizeSCEV, CurLoop, DL, SE);
// TODO: ideally we should still be able to generate memset if SCEV expander
// is taught to generate the dependencies at the latest point.
if (!Expander.isSafeToExpand(NumBytesS))
return Changed;
MemsetArg =
Expander.expandCodeFor(NumBytesS, IntIdxTy, Preheader->getTerminator());
if (auto *CI = dyn_cast<ConstantInt>(MemsetArg))
BytesWritten = CI->getZExtValue();
}
assert(MemsetArg && "MemsetArg should have been set");
AAMDNodes AATags = TheStore->getAAMetadata();
for (Instruction *Store : Stores)
AATags = AATags.merge(Store->getAAMetadata());
if (BytesWritten)
AATags = AATags.extendTo(BytesWritten.value());
else
AATags = AATags.extendTo(-1);
CallInst *NewCall;
if (SplatValue) {
NewCall = Builder.CreateMemSet(BasePtr, SplatValue, MemsetArg,
MaybeAlign(StoreAlignment),
/*isVolatile=*/false, AATags);
} else if (Opts.loop_idiom_force_memset_pattern_intrinsic ||
isLibFuncEmittable(M, TLI, LibFunc_memset_pattern16)) {
assert(isa<SCEVConstant>(StoreSizeSCEV) && "Expected constant store size");
NewCall = Builder.CreateIntrinsicWithoutFolding(
Intrinsic::experimental_memset_pattern,
{DestInt8PtrTy, PatternValue->getType(), IntIdxTy},
{BasePtr, PatternValue, MemsetArg,
ConstantInt::getFalse(M->getContext())});
if (StoreAlignment)
cast<MemSetPatternInst>(NewCall)->setDestAlignment(*StoreAlignment);
NewCall->setAAMetadata(AATags);
} else {
// Neither a memset, nor memset_pattern16
return Changed;
}
NewCall->setDebugLoc(TheStore->getDebugLoc());
if (MSSAU) {
MemoryAccess *NewMemAcc = MSSAU->createMemoryAccessInBB(
NewCall, nullptr, NewCall->getParent(), MemorySSA::BeforeTerminator);
MSSAU->insertDef(cast<MemoryDef>(NewMemAcc), true);
}
LLVM_DEBUG(dbgs() << " Formed memset: " << *NewCall << "\n"
<< " from store to: " << *Ev << " at: " << *TheStore
<< "\n");
ORE.emit([&]() {
OptimizationRemark R(DEBUG_TYPE, "ProcessLoopStridedStore",
NewCall->getDebugLoc(), Preheader);
R << "Transformed loop-strided store in "
<< ore::NV("Function", TheStore->getFunction())
<< " function into a call to "
<< ore::NV("NewFunction", NewCall->getCalledFunction())
<< "() intrinsic";
if (!Stores.empty())
R << ore::setExtraArgs();
for (auto *I : Stores) {
R << ore::NV("FromBlock", I->getParent()->getName())
<< ore::NV("ToBlock", Preheader->getName());
}
return R;
});
// Okay, the memset has been formed. Zap the original store and anything that
// feeds into it.
for (auto *I : Stores) {
if (MSSAU)
MSSAU->removeMemoryAccess(I, true);
deleteDeadInstruction(I);
}
if (MSSAU && VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA();
++NumMemSet;
ExpCleaner.markResultUsed();
return true;
}
/// If the stored value is a strided load in the same loop with the same stride
/// this may be transformable into a memcpy. This kicks in for stuff like
/// for (i) A[i] = B[i];
bool LoopIdiomRecognize::processLoopStoreOfLoopLoad(StoreInst *SI,
const SCEV *BECount) {
assert(SI->isUnordered() && "Expected only non-volatile non-ordered stores.");
Value *StorePtr = SI->getPointerOperand();
const SCEVAddRecExpr *StoreEv = cast<SCEVAddRecExpr>(SE->getSCEV(StorePtr));
unsigned StoreSize = DL->getTypeStoreSize(SI->getValueOperand()->getType());
// The store must be feeding a non-volatile load.
LoadInst *LI = cast<LoadInst>(SI->getValueOperand());
assert(LI->isUnordered() && "Expected only non-volatile non-ordered loads.");
// See if the pointer expression is an AddRec like {base,+,1} on the current
// loop, which indicates a strided load. If we have something else, it's a
// random load we can't handle.
Value *LoadPtr = LI->getPointerOperand();
const SCEVAddRecExpr *LoadEv = cast<SCEVAddRecExpr>(SE->getSCEV(LoadPtr));
const SCEV *StoreSizeSCEV = SE->getConstant(StorePtr->getType(), StoreSize);
return processLoopStoreOfLoopLoad(StorePtr, LoadPtr, StoreSizeSCEV,
SI->getAlign(), LI->getAlign(), SI, LI,
StoreEv, LoadEv, BECount);
}
namespace {
class MemmoveVerifier {
public:
explicit MemmoveVerifier(const SCEV &LoadStart, const SCEV &StoreStart,
ScalarEvolution &SE)
: DL(SE.getDataLayout()),
Off(dyn_cast<SCEVConstant>(SE.getMinusSCEV(&StoreStart, &LoadStart))),
BasePtr(dyn_cast<SCEVUnknown>(SE.getPointerBase(&StoreStart))),
IsSameObject(Off != nullptr) {}
bool loadAndStoreMayFormMemmove(unsigned StoreSize, bool IsNegStride,
const Instruction &TheLoad,
bool IsMemCpy) const {
// The store must be at a constant offset from the load, and there must be
// an underlying pointer.
if (!Off || !BasePtr)
return false;
const APInt &OffVal = Off->getAPInt();
// If null is defined then the base pointer can't be null
auto *NullBase = dyn_cast<ConstantPointerNull>(BasePtr->getValue());
if (NullBase && NullPointerIsDefined(
TheLoad.getParent()->getParent(),
NullBase->getPointerType()->getPointerAddressSpace()))
return false;
int64_t LoadSize;
if (IsMemCpy) {
// memcpy is equivalent to a sequence of byte loads and stores
LoadSize = 1;
} else {
LoadSize = DL.getTypeSizeInBits(TheLoad.getType()).getFixedValue() / 8;
if (LoadSize != StoreSize)
return false;
}
// Ensure that LoadBasePtr is after StoreBasePtr or before StoreBasePtr
// for negative stride. LoadBasePtr shouldn't overlap with StoreBasePtr.
if (IsNegStride ? OffVal.slt(LoadSize) : OffVal.sgt(-LoadSize))
return false;
return true;
}
private:
const DataLayout &DL;
const SCEVConstant *Off;
const SCEVUnknown *BasePtr;
public:
const bool IsSameObject;
};
} // namespace
bool LoopIdiomRecognize::processLoopStoreOfLoopLoad(
Value *DestPtr, Value *SourcePtr, const SCEV *StoreSizeSCEV,
MaybeAlign StoreAlign, MaybeAlign LoadAlign, Instruction *TheStore,
Instruction *TheLoad, const SCEVAddRecExpr *StoreEv,
const SCEVAddRecExpr *LoadEv, const SCEV *BECount) {
// We currently don't convert inline intrinsics into larger ones, to avoid
// code size increase. `processLoopMemCpy` checks that the intrinsic is not
// inline before calling this function.
assert((isa<StoreInst>(TheStore) ||
!cast<MemIntrinsic>(TheStore)->isForceInlined()) &&
"inline mem intrinsics should be filtered out by callers");
// The trip count of the loop and the base pointer of the addrec SCEV is
// guaranteed to be loop invariant, which means that it should dominate the
// header. This allows us to insert code for it in the preheader.
BasicBlock *Preheader = CurLoop->getLoopPreheader();
IRBuilder<> Builder(Preheader->getTerminator());
SCEVExpander Expander(*SE, "loop-idiom");
SCEVExpanderCleaner ExpCleaner(Expander);
bool Changed = false;
const SCEV *StrStart = StoreEv->getStart();
unsigned StrAS = DestPtr->getType()->getPointerAddressSpace();
Type *IntIdxTy = Builder.getIntNTy(DL->getIndexSizeInBits(StrAS));
APInt Stride = getStoreStride(StoreEv);
const SCEVConstant *ConstStoreSize = dyn_cast<SCEVConstant>(StoreSizeSCEV);
// TODO: Deal with non-constant size; Currently expect constant store size
assert(ConstStoreSize && "store size is expected to be a constant");
int64_t StoreSize = ConstStoreSize->getValue()->getZExtValue();
bool IsNegStride = StoreSize == -Stride;
// Handle negative strided loops.
if (IsNegStride)
StrStart =
getStartForNegStride(StrStart, BECount, IntIdxTy, StoreSizeSCEV, SE);
// Okay, we have a strided store "p[i]" of a loaded value. We can turn
// this into a memcpy in the loop preheader now if we want. However, this
// would be unsafe to do if there is anything else in the loop that may read
// or write the memory region we're storing to. This includes the load that
// feeds the stores. Check for an alias by generating the base address and
// checking everything.
Value *StoreBasePtr = Expander.expandCodeFor(
StrStart, Builder.getPtrTy(StrAS), Preheader->getTerminator());
// From here on out, conservatively report to the pass manager that we've
// changed the IR, even if we later clean up these added instructions. There
// may be structural differences e.g. in the order of use lists not accounted
// for in just a textual dump of the IR. This is written as a variable, even
// though statically all the places this dominates could be replaced with
// 'true', with the hope that anyone trying to be clever / "more precise" with
// the return value will read this comment, and leave them alone.
Changed = true;
SmallPtrSet<Instruction *, 2> IgnoredInsts;
IgnoredInsts.insert(TheStore);
bool IsMemCpy = isa<MemCpyInst>(TheStore);
const StringRef InstRemark = IsMemCpy ? "memcpy" : "load and store";
bool LoopAccessStore =
mayLoopAccessLocation(StoreBasePtr, ModRefInfo::ModRef, CurLoop, BECount,
StoreSizeSCEV, *AA, IgnoredInsts);
if (LoopAccessStore) {
// For memmove case it's not enough to guarantee that loop doesn't access
// TheStore and TheLoad. Additionally we need to make sure that TheStore is
// the only user of TheLoad.
if (!TheLoad->hasOneUse())
return Changed;
IgnoredInsts.insert(TheLoad);
if (mayLoopAccessLocation(StoreBasePtr, ModRefInfo::ModRef, CurLoop,
BECount, StoreSizeSCEV, *AA, IgnoredInsts)) {
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "LoopMayAccessStore",
TheStore)
<< ore::NV("Inst", InstRemark) << " in "
<< ore::NV("Function", TheStore->getFunction())
<< " function will not be hoisted: "
<< ore::NV("Reason", "The loop may access store location");
});
return Changed;
}
IgnoredInsts.erase(TheLoad);
}
const SCEV *LdStart = LoadEv->getStart();
unsigned LdAS = SourcePtr->getType()->getPointerAddressSpace();
// Handle negative strided loops.
if (IsNegStride)
LdStart =
getStartForNegStride(LdStart, BECount, IntIdxTy, StoreSizeSCEV, SE);
// For a memcpy, we have to make sure that the input array is not being
// mutated by the loop.
Value *LoadBasePtr = Expander.expandCodeFor(LdStart, Builder.getPtrTy(LdAS),
Preheader->getTerminator());
// If the store is a memcpy instruction, we must check if it will write to
// the load memory locations. So remove it from the ignored stores.
MemmoveVerifier Verifier(*LdStart, *StrStart, *SE);
if (IsMemCpy && !Verifier.IsSameObject)
IgnoredInsts.erase(TheStore);
if (mayLoopAccessLocation(LoadBasePtr, ModRefInfo::Mod, CurLoop, BECount,
StoreSizeSCEV, *AA, IgnoredInsts)) {
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "LoopMayAccessLoad", TheLoad)
<< ore::NV("Inst", InstRemark) << " in "
<< ore::NV("Function", TheStore->getFunction())
<< " function will not be hoisted: "
<< ore::NV("Reason", "The loop may access load location");
});
return Changed;
}
bool IsAtomic = TheStore->isAtomic() || TheLoad->isAtomic();
bool UseMemMove = IsMemCpy ? Verifier.IsSameObject : LoopAccessStore;
if (IsAtomic) {
// For now don't support unordered atomic memmove.
if (UseMemMove)
return Changed;
// We cannot allow unaligned ops for unordered load/store, so reject
// anything where the alignment isn't at least the element size.
assert((StoreAlign && LoadAlign) &&
"Expect unordered load/store to have align.");
if (*StoreAlign < StoreSize || *LoadAlign < StoreSize)
return Changed;
// If the element.atomic memcpy is not lowered into explicit
// loads/stores later, then it will be lowered into an element-size
// specific lib call. If the lib call doesn't exist for our store size, then
// we shouldn't generate the memcpy.
if (StoreSize > TTI->getAtomicMemIntrinsicMaxElementSize())
return Changed;
}
if (UseMemMove)
if (!Verifier.loadAndStoreMayFormMemmove(StoreSize, IsNegStride, *TheLoad,
IsMemCpy))
return Changed;
if (avoidLIRForMultiBlockLoop())
return Changed;
// Okay, everything is safe, we can transform this!
const SCEV *NumBytesS =
getNumBytes(BECount, IntIdxTy, StoreSizeSCEV, CurLoop, DL, SE);
Value *NumBytes =
Expander.expandCodeFor(NumBytesS, IntIdxTy, Preheader->getTerminator());
AAMDNodes AATags = TheLoad->getAAMetadata();
AAMDNodes StoreAATags = TheStore->getAAMetadata();
AATags = AATags.merge(StoreAATags);
if (auto CI = dyn_cast<ConstantInt>(NumBytes))
AATags = AATags.extendTo(CI->getZExtValue());
else
AATags = AATags.extendTo(-1);
CallInst *NewCall = nullptr;
// Check whether to generate an unordered atomic memcpy:
// If the load or store are atomic, then they must necessarily be unordered
// by previous checks.
if (!IsAtomic) {
if (UseMemMove)
NewCall = Builder.CreateMemMove(StoreBasePtr, StoreAlign, LoadBasePtr,
LoadAlign, NumBytes,
/*isVolatile=*/false, AATags);
else
NewCall =
Builder.CreateMemCpy(StoreBasePtr, StoreAlign, LoadBasePtr, LoadAlign,
NumBytes, /*isVolatile=*/false, AATags);
} else {
// Create the call.
// Note that unordered atomic loads/stores are *required* by the spec to
// have an alignment but non-atomic loads/stores may not.
NewCall = Builder.CreateElementUnorderedAtomicMemCpy(
StoreBasePtr, *StoreAlign, LoadBasePtr, *LoadAlign, NumBytes, StoreSize,
AATags);
}
NewCall->setDebugLoc(TheStore->getDebugLoc());
if (MSSAU) {
MemoryAccess *NewMemAcc = MSSAU->createMemoryAccessInBB(
NewCall, nullptr, NewCall->getParent(), MemorySSA::BeforeTerminator);
MSSAU->insertDef(cast<MemoryDef>(NewMemAcc), true);
}
LLVM_DEBUG(dbgs() << " Formed new call: " << *NewCall << "\n"
<< " from load ptr=" << *LoadEv << " at: " << *TheLoad
<< "\n"
<< " from store ptr=" << *StoreEv << " at: " << *TheStore
<< "\n");
ORE.emit([&]() {
return OptimizationRemark(DEBUG_TYPE, "ProcessLoopStoreOfLoopLoad",
NewCall->getDebugLoc(), Preheader)
<< "Formed a call to "
<< ore::NV("NewFunction", NewCall->getCalledFunction())
<< "() intrinsic from " << ore::NV("Inst", InstRemark)
<< " instruction in " << ore::NV("Function", TheStore->getFunction())
<< " function"
<< ore::setExtraArgs()
<< ore::NV("FromBlock", TheStore->getParent()->getName())
<< ore::NV("ToBlock", Preheader->getName());
});
// Okay, a new call to memcpy/memmove has been formed. Zap the original store
// and anything that feeds into it.
if (MSSAU)
MSSAU->removeMemoryAccess(TheStore, true);
deleteDeadInstruction(TheStore);
if (MSSAU && VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA();
if (UseMemMove)
++NumMemMove;
else
++NumMemCpy;
ExpCleaner.markResultUsed();
return true;
}
// When compiling for codesize we avoid idiom recognition for a multi-block loop
// unless it is a loop_memset idiom or a memset/memcpy idiom in a nested loop.
//
bool LoopIdiomRecognize::avoidLIRForMultiBlockLoop(bool IsMemset,
bool IsLoopMemset) {
if (ApplyCodeSizeHeuristics && CurLoop->getNumBlocks() > 1) {
if (CurLoop->isOutermost() && (!IsMemset || !IsLoopMemset)) {
LLVM_DEBUG(dbgs() << " " << CurLoop->getHeader()->getParent()->getName()
<< " : LIR " << (IsMemset ? "Memset" : "Memcpy")
<< " avoided: multi-block top-level loop\n");
return true;
}
}
return false;
}
bool LoopIdiomRecognize::optimizeCRCLoop(const PolynomialInfo &Info) {
// FIXME: Hexagon has a special HexagonLoopIdiom that optimizes CRC using
// carry-less multiplication instructions, which is more efficient than our
// Sarwate table-lookup optimization. Hence, until we're able to emit
// target-specific instructions for Hexagon, subsuming HexagonLoopIdiom,
// disable the optimization for Hexagon.
Module &M = *CurLoop->getHeader()->getModule();
Triple TT(M.getTargetTriple());
if (TT.getArch() == Triple::hexagon)
return false;
LLVMContext &Ctx = Info.LHS->getContext();
Type *CRCTy = Info.LHS->getType();
unsigned CRCBW = CRCTy->getIntegerBitWidth();
// CRC computation is mostly serial, so latency works best for comparison.
TargetTransformInfo::TargetCostKind CostKind =
TargetTransformInfo::TCK_Latency;
InstructionCost XorCost =
TTI->getArithmeticInstrCost(Instruction::Xor, CRCTy, CostKind);
InstructionCost ShiftCost =
TTI->getArithmeticInstrCost(Instruction::LShr, CRCTy, CostKind);
InstructionCost AndCost =
TTI->getArithmeticInstrCost(Instruction::And, CRCTy, CostKind);
InstructionCost SelectCost =
TTI->getCmpSelInstrCost(Instruction::Select, CRCTy, Type::getInt1Ty(Ctx),
CmpInst::BAD_ICMP_PREDICATE, CostKind);
InstructionCost LoadCost =
TTI->getMemoryOpCost(Instruction::Load, CRCTy, DL->getABITypeAlign(CRCTy),
DL->getDefaultGlobalsAddressSpace(), CostKind);
auto ClmulCost = [&](unsigned BW) {
auto *Ty = IntegerType::get(Ctx, BW);
IntrinsicCostAttributes Attrs(Intrinsic::clmul, Ty, {Ty, Ty});
return TTI->getIntrinsicInstrCost(Attrs, CostKind);
};
// Estimate the cost of the original, unoptimized loop.
InstructionCost OrigLoopCost =
(2 * ShiftCost + 2 * XorCost + AndCost + SelectCost) * Info.TripCount;
// Estimate the cost of the Sarwate lookup table optimization strategy.
// As mentioned previously, a byte-multiple trip count is required.
InstructionCost TableStrategyCost =
Info.TripCount % 8 != 0
? InstructionCost::getInvalid()
: (LoadCost + XorCost + 2 * ShiftCost) * (Info.TripCount / 8);
// Estimate the cost of the carry-less multiplication optimization strategy.
InstructionCost ClmulStrategyCost = ClmulCost(2 * Info.TripCount) +
ClmulCost(CRCBW + Info.TripCount) +
2 * XorCost + 2 * ShiftCost + AndCost;
ORE.emit([&]() {
return OptimizationRemarkAnalysis(DEBUG_TYPE, "CRCLoopCosts",
CurLoop->getStartLoc(),
CurLoop->getHeader())
<< "CRC loop costs: original="
<< ore::NV("OrigLoopCost", OrigLoopCost)
<< ", table=" << ore::NV("TableStrategyCost", TableStrategyCost)
<< ", clmul=" << ore::NV("ClmulStrategyCost", ClmulStrategyCost);
});
auto ReportMissed = [&](StringRef Reason) {
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "CRCLoopMissed",
CurLoop->getStartLoc(),
CurLoop->getHeader())
<< "CRC loop not optimized: " << Reason;
});
};
auto ReportOptimized = [&](StringRef Strategy, StringRef Reason) {
ORE.emit([&]() {
return OptimizationRemark(DEBUG_TYPE, "CRCLoopOptimized",
CurLoop->getStartLoc(), CurLoop->getHeader())
<< "CRC loop optimized using " << ore::NV("Strategy", Strategy)
<< ": " << Reason;
});
};
switch (Opts.loop_idiom_crc_strategy) {
default:
ReportMissed("disabled by user");
return false;
case CRCStrategyKind::Table:
// The table strategy is not possible in its current form without a byte-
// multiple trip count.
if (Info.TripCount % 8 == 0) {
optimizeCRCLoopUsingTableLookup(Info);
ReportOptimized("table", "forced by user");
return true;
}
ReportMissed("table strategy forced, but not possible");
return false;
case CRCStrategyKind::Clmul:
optimizeCRCLoopUsingClmul(Info);
ReportOptimized("clmul", "forced by user");
return true;
case CRCStrategyKind::Auto:
// When using the auto strategy, bail if we are optimizing for size since
// there's usually not a clear size benefit.
// TODO: The clmul optimization is around the same size in many cases, so it
// could be worth it to take advantage of that fact, especially if it would
// be much faster than the original loop.
if (ApplyCodeSizeHeuristics) {
ReportMissed("optimizing for size");
return false;
}
// Only apply an optimization if there's a clear benefit to doing so.
if (std::min(TableStrategyCost, ClmulStrategyCost) >= OrigLoopCost) {
ReportMissed("no profitable strategy");
return false;
}
if (TableStrategyCost <= ClmulStrategyCost) {
optimizeCRCLoopUsingTableLookup(Info);
ReportOptimized("table", "most profitable strategy");
} else {
optimizeCRCLoopUsingClmul(Info);
ReportOptimized("clmul", "most profitable strategy");
}
return true;
}
}
// The algorithm used in this optimization is a Polynomial (GF(2)) Barrett
// Reduction based on Intel's "Fast CRC Computation for Generic Polynomials
// Using PCLMULQDQ Instruction" white paper (December 2009).
void LoopIdiomRecognize::optimizeCRCLoopUsingClmul(const PolynomialInfo &Info) {
// TODO: If clmul exists on the target but not for the required width, it
// might be possible to split into multiple iterations of reduction.
Type *CRCTy = Info.LHS->getType();
LLVMContext &Ctx = CRCTy->getContext();
unsigned CRCBW = CRCTy->getIntegerBitWidth();
// The loop's TripCount determines how many bits of the data are processed,
// regardless of whether the actual data bit width matches (if auxiliary data
// is even used at all).
unsigned TC = Info.TripCount;
// Based on the clmul inputs, the first clmul needs 2*TC bits, and the second
// needs CRCBW+TC bits. However, only the low TC bits of the first clmul are
// used in little-endian, so a clmul in TC bits suffices in that case.
IntegerType *ClmulMuTy =
IntegerType::get(Ctx, Info.IsBigEndian ? 2 * TC : TC);
IntegerType *ClmulGPTy = IntegerType::get(Ctx, CRCBW + TC);
// First, generate the constants required for GF(2) Barrett reduction.
auto [Mu, FullGenPoly] = HashRecognize::genBarrettConstants(Info);
Value *MuConst =
ConstantInt::get(Ctx, Mu.zextOrTrunc(ClmulMuTy->getBitWidth()));
Value *GenPolyConst =
ConstantInt::get(Ctx, FullGenPoly.zext(ClmulGPTy->getBitWidth()));
IRBuilder<> Builder(CurLoop->getLoopPreheader()->getTerminator());
// If a shift needs to occur in the setup for the first clmul with MuConst, it
// will be by abs(TC - CRCBW). To ensure that the shift can work without
// losing information or creating poison, give it CRCBW + TC bits.
bool SetupShiftNeeded = Info.IsBigEndian && TC != CRCBW;
auto *SetupTy = IntegerType::get(Ctx, SetupShiftNeeded ? CRCBW + TC : TC);
// Based on the Intel white paper, in our case, we have
// R(x) = (LHS*x^TC) xor (LHSAux ? getTCBits(LHSAux)*x^CRCBW : 0)
// since the CRC loop multiplies LHS by x each iteration, and the x^CRCBW term
// of getTCBits(LHSAux) is XORed in for the significant bit check.
// Rather than compute the full R(x), we can split it in two: a quotient for
// step 1 (floor(R(x)/x^CRCBW)) and a remainder for step 3 (R(x) mod x^CRCBW).
//
// ClmulMuInput is an evolving variable that will eventually become the part
// used in step 1, which can be simplified to
// (LHS*x^(TC-CRCBW)) xor (LHSAux ? getTCBits(LHSAux) : 0).
// Thanks to restrictions imposed by HashRecognize for big-endian CRC loops,
// getTCBits(LHSAux) = LHSAux*x^(TC-CRCBW), so this can be further simplified
// to (LHS xor (LHSAux ? LHSAux : 0))*x^(TC-CRCBW).
Value *ClmulMuInput =
Builder.CreateZExtOrTrunc(Info.LHS, SetupTy, "crc.cast");
// If auxiliary data is present, XOR it in with the CRC.
if (Value *Data = Info.LHSAux) {
// This is usually a zext, but DataBW may exceed CRCBW+TC if both CRCBW and
// TC are small enough.
Data = Builder.CreateZExtOrTrunc(Data, SetupTy, "data.cast");
ClmulMuInput = Builder.CreateXor(ClmulMuInput, Data, "xor.crc.data");
}
// Align the current CRC with TripCount (multiply or divide by x^(TC-CRCBW)).
if (SetupShiftNeeded) {
ClmulMuInput =
TC > CRCBW
? Builder.CreateShl(ClmulMuInput, TC - CRCBW, "crc.align.tc")
: Builder.CreateLShr(ClmulMuInput, CRCBW - TC, "crc.align.tc");
}
// Zero out any bits above (TC-1) for calculation since the original loop
// doesn't use them in the significant bit checks.
if (SetupTy->getBitWidth() > TC) {
auto *Mask =
ConstantInt::get(Ctx, APInt::getLowBitsSet(SetupTy->getBitWidth(), TC));
ClmulMuInput = Builder.CreateAnd(ClmulMuInput, Mask, "crc.tcbits");
}
// Step 1: T1(x) = floor(R(x)/x^CRCBW) * mu
// Input is TC bits and mu is TC+1 bits, so result will be 2*TC bits.
ClmulMuInput =
Builder.CreateZExtOrTrunc(ClmulMuInput, ClmulMuTy, "tcbits.cast");
Value *ClmulMu = Builder.CreateBinaryIntrinsic(
Intrinsic::clmul, ClmulMuInput, MuConst, /*FMFSource=*/{}, "clmul.mu");
// Calculate floor(T1(x)/x^TC) for step 2.
Value *ClmulGPInput =
Info.IsBigEndian ? Builder.CreateLShr(ClmulMu, TC, "quot.lshr") : ClmulMu;
// Step 2: T2(x) = floor(T1(x)/x^TC) * P(x)
// Input is TC bits and P(x) is CRCBW+1 bits, so result will be CRCBW+TC bits.
ClmulGPInput =
Builder.CreateZExtOrTrunc(ClmulGPInput, ClmulGPTy, "quot.cast");
Value *ClmulGP = Builder.CreateBinaryIntrinsic(Intrinsic::clmul, ClmulGPInput,
GenPolyConst,
/*FMFSource=*/{}, "clmul.gp");
// Calculate the least significant part of R(x) for step 3 as specified above.
// R(x) mod x^CRCBW = LHS*x^TC mod x^CRCBW, though the (mod x^CRCBW) is
// handled later on when truncating back to CRCBW for ComputedValue.
Value *CRCNext = Builder.CreateZExt(Info.LHS, ClmulGPTy, "crc.recast");
if (Info.IsBigEndian)
CRCNext = Builder.CreateShl(CRCNext, TC, "crc.shl");
// Step 3: C(x) = (R(x) xor T2(x)) mod x^CRCBW
CRCNext = Builder.CreateXor(CRCNext, ClmulGP, "xor.crc.mult");
if (!Info.IsBigEndian)
CRCNext = Builder.CreateLShr(CRCNext, TC, "crc.lshr");
// Bring the result back down the the CRC bit width.
CRCNext = Builder.CreateTrunc(CRCNext, CRCTy, "crc.next");
// Replace the result of the loop with the new computed CRC value.
Info.ComputedValue->replaceUsesOutsideBlock(CRCNext, CurLoop->getLoopLatch());
// Finally, clean up the loop as much as possible so it can be trivially
// deleted.
{
for (PHINode &PN : make_early_inc_range(CurLoop->getHeader()->phis())) {
PN.replaceAllUsesWith(PoisonValue::get(PN.getType()));
RecursivelyDeleteDeadPHINode(&PN);
}
// Replace the exit condition with constant true/false to always cause a
// branch to the exit block.
deleteDeadInstruction(CurLoop->getLatchCmpInst());
auto *BrInst = cast<CondBrInst>(CurLoop->getLoopLatch()->getTerminator());
BrInst->setCondition(ConstantInt::getBool(
Ctx, BrInst->getSuccessor(0) == CurLoop->getExitBlock()));
SE->forgetLoop(CurLoop);
}
}
void LoopIdiomRecognize::optimizeCRCLoopUsingTableLookup(
const PolynomialInfo &Info) {
assert(Info.TripCount % 8 == 0 && "A byte-multiple trip count is required");
// First, create a new GlobalVariable corresponding to the
// Sarwate-lookup-table.
Type *CRCTy = Info.LHS->getType();
unsigned CRCBW = CRCTy->getIntegerBitWidth();
std::array<Constant *, 256> CRCConstants;
transform(HashRecognize::genSarwateTable(Info.RHS, Info.IsBigEndian),
CRCConstants.begin(),
[CRCTy](const APInt &E) { return ConstantInt::get(CRCTy, E); });
Constant *ConstArray =
ConstantArray::get(ArrayType::get(CRCTy, 256), CRCConstants);
GlobalVariable *GV = new GlobalVariable(
*CurLoop->getHeader()->getModule(), ConstArray->getType(), true,
GlobalValue::PrivateLinkage, ConstArray, ".crctable");
PHINode *IV = CurLoop->getCanonicalInductionVariable();
SmallVector<PHINode *, 2> Cleanup;
// Next, mark all PHIs for removal except IV.
{
for (PHINode &PN : CurLoop->getHeader()->phis()) {
if (&PN == IV)
continue;
PN.replaceAllUsesWith(PoisonValue::get(PN.getType()));
Cleanup.push_back(&PN);
}
}
// Next, fix up the trip count.
{
unsigned NewBTC = (Info.TripCount / 8) - 1;
BasicBlock *LoopBlk = CurLoop->getLoopLatch();
CondBrInst *BrInst = cast<CondBrInst>(LoopBlk->getTerminator());
CmpPredicate ExitPred = BrInst->getSuccessor(0) == LoopBlk
? ICmpInst::Predicate::ICMP_NE
: ICmpInst::Predicate::ICMP_EQ;
Instruction *ExitCond = CurLoop->getLatchCmpInst();
Value *ExitLimit = ConstantInt::get(IV->getType(), NewBTC);
IRBuilder<> Builder(ExitCond);
Value *NewExitCond =
Builder.CreateICmp(ExitPred, IV, ExitLimit, "exit.cond");
ExitCond->replaceAllUsesWith(NewExitCond);
deleteDeadInstruction(ExitCond);
}
// Finally, fill the loop with the Sarwate-table-lookup logic, and replace all
// uses of ComputedValue.
//
// Little-endian:
// crc = (crc >> 8) ^ tbl[(iv'th byte of data) ^ (bottom byte of crc)]
// Big-Endian:
// crc = (crc << 8) ^ tbl[(iv'th byte of data) ^ (top byte of crc)]
{
auto LoByte = [](IRBuilderBase &Builder, Value *Op, const Twine &Name) {
return Builder.CreateZExtOrTrunc(
Op, IntegerType::getInt8Ty(Op->getContext()), Name);
};
auto HiIdx = [LoByte, CRCBW](IRBuilderBase &Builder, Value *Op,
const Twine &Name) {
// Shift the top bits of Op to the bottom byte by using the CRC bitwidth
// as a reference.
if (CRCBW != 8) {
Op = CRCBW > 8 ? Builder.CreateLShr(Op, CRCBW - 8, Name)
: Builder.CreateShl(Op, 8 - CRCBW, Name);
}
return LoByte(Builder, Op, Name + ".lo.byte");
};
IRBuilder<> Builder(CurLoop->getHeader()->getFirstNonPHIIt());
// Create the CRC PHI, and initialize its incoming value to the initial
// value of CRC.
PHINode *CRCPhi = Builder.CreatePHI(CRCTy, 2, "crc");
CRCPhi->addIncoming(Info.LHS, CurLoop->getLoopPreheader());
// CRC is now an evolving variable, initialized to the PHI.
Value *CRC = CRCPhi;
// TableIndexer = ((top|bottom) byte of CRC). It is XOR'ed with (iv'th byte
// of LHSAux), if LHSAux is non-nullptr.
Value *Indexer = CRC;
if (Value *Data = Info.LHSAux) {
Type *DataTy = Data->getType();
// To index into the (iv'th byte of LHSAux), we multiply iv by 8, and we
// shift right by that amount, and take the lo-byte (in the little-endian
// case), or shift left by that amount, and take the hi-idx (in the
// big-endian case).
Value *IVBits = Builder.CreateZExtOrTrunc(
Builder.CreateShl(IV, 3, "iv.bits"), DataTy, "iv.indexer");
Value *DataIndexer =
Info.IsBigEndian ? Builder.CreateShl(Data, IVBits, "data.indexer")
: Builder.CreateLShr(Data, IVBits, "data.indexer");
Indexer = Builder.CreateXor(
DataIndexer,
Builder.CreateZExtOrTrunc(Indexer, DataTy, "crc.indexer.cast"),
"crc.data.indexer");
}
Indexer = Info.IsBigEndian ? HiIdx(Builder, Indexer, "indexer.hi")
: LoByte(Builder, Indexer, "indexer.lo");
// Always index into a GEP using the index type.
Indexer = Builder.CreateZExt(
Indexer, SE->getDataLayout().getIndexType(GV->getType()),
"indexer.ext");
// CRCTableLd = CRCTable[(iv'th byte of data) ^ (top|bottom) byte of CRC].
Value *CRCTableGEP =
Builder.CreateInBoundsGEP(CRCTy, GV, Indexer, "tbl.ptradd");
Instruction *CRCTableLd = Builder.CreateLoad(CRCTy, CRCTableGEP, "tbl.ld");
// Update MemorySSA since we just created a new load instruction.
if (MSSAU) {
auto *NewMemAcc = MSSAU->createMemoryAccessInBB(
CRCTableLd, /*Definition=*/nullptr, CRCTableLd->getParent(),
MemorySSA::Beginning);
MSSAU->insertUse(cast<MemoryUse>(NewMemAcc), /*RenameUses=*/true);
}
// CRCNext = (CRC (<<|>>) 8) ^ CRCTableLd, or simply CRCTableLd in case of
// CRC-8.
Value *CRCNext = CRCTableLd;
if (CRCBW > 8) {
Value *CRCShift = Info.IsBigEndian
? Builder.CreateShl(CRC, 8, "crc.be.shift")
: Builder.CreateLShr(CRC, 8, "crc.le.shift");
CRCNext = Builder.CreateXor(CRCShift, CRCTableLd, "crc.next");
}
// Connect the back-edge for the loop, and RAUW the ComputedValue.
CRCPhi->addIncoming(CRCNext, CurLoop->getLoopLatch());
Info.ComputedValue->replaceUsesOutsideBlock(CRCNext,
CurLoop->getLoopLatch());
}
// Cleanup.
{
for (PHINode *PN : Cleanup)
RecursivelyDeleteDeadPHINode(PN);
SE->forgetLoop(CurLoop);
if (MSSAU && VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA();
}
}
bool LoopIdiomRecognize::runOnNoncountableLoop() {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Scanning: F["
<< CurLoop->getHeader()->getParent()->getName()
<< "] Noncountable Loop %"
<< CurLoop->getHeader()->getName() << "\n");
return recognizePopcount() || recognizeAndInsertFFS() ||
recognizeShiftUntilBitTest() || recognizeShiftUntilZero() ||
recognizeShiftUntilLessThan() || recognizeAndInsertStrLen();
}
/// Check if the given conditional branch is based on the comparison between
/// a variable and zero, and if the variable is non-zero or zero (JmpOnZero is
/// true), the control yields to the loop entry. If the branch matches the
/// behavior, the variable involved in the comparison is returned. This function
/// will be called to see if the precondition and postcondition of the loop are
/// in desirable form.
static Value *matchCondition(CondBrInst *BI, BasicBlock *LoopEntry,
bool JmpOnZero = false) {
ICmpInst *Cond = dyn_cast<ICmpInst>(BI->getCondition());
if (!Cond)
return nullptr;
auto *CmpZero = dyn_cast<ConstantInt>(Cond->getOperand(1));
if (!CmpZero || !CmpZero->isZero())
return nullptr;
BasicBlock *TrueSucc = BI->getSuccessor(0);
BasicBlock *FalseSucc = BI->getSuccessor(1);
if (JmpOnZero)
std::swap(TrueSucc, FalseSucc);
ICmpInst::Predicate Pred = Cond->getPredicate();
if ((Pred == ICmpInst::ICMP_NE && TrueSucc == LoopEntry) ||
(Pred == ICmpInst::ICMP_EQ && FalseSucc == LoopEntry))
return Cond->getOperand(0);
return nullptr;
}
namespace {
class StrlenVerifier {
public:
explicit StrlenVerifier(const Loop *CurLoop, ScalarEvolution *SE,
const TargetLibraryInfo *TLI)
: CurLoop(CurLoop), SE(SE), TLI(TLI) {}
bool isValidStrlenIdiom() {
// Give up if the loop has multiple blocks, multiple backedges, or
// multiple exit blocks
if (CurLoop->getNumBackEdges() != 1 || CurLoop->getNumBlocks() != 1 ||
!CurLoop->getUniqueExitBlock())
return false;
// It should have a preheader and a branch instruction.
BasicBlock *Preheader = CurLoop->getLoopPreheader();
if (!Preheader ||
!isa<UncondBrInst, CondBrInst>(Preheader->getTerminator()))
return false;
// The loop exit must be conditioned on an icmp with 0 the null terminator.
// The icmp operand has to be a load on some SSA reg that increments
// by 1 in the loop.
BasicBlock *LoopBody = *CurLoop->block_begin();
// Skip if the body is too big as it most likely is not a strlen idiom.
if (!LoopBody || LoopBody->size() >= 15)
return false;
CondBrInst *LoopTerm = dyn_cast<CondBrInst>(LoopBody->getTerminator());
if (!LoopTerm)
return false;
Value *LoopCond = matchCondition(LoopTerm, LoopBody);
if (!LoopCond)
return false;
LoadInst *LoopLoad = dyn_cast<LoadInst>(LoopCond);
if (!LoopLoad || LoopLoad->getPointerAddressSpace() != 0)
return false;
OperandType = LoopLoad->getType();
if (!OperandType || !OperandType->isIntegerTy())
return false;
// See if the pointer expression is an AddRec with constant step a of form
// ({n,+,a}) where a is the width of the char type.
Value *IncPtr = LoopLoad->getPointerOperand();
const SCEV *LoadEv = SE->getSCEV(IncPtr);
const APInt *Step;
if (!match(LoadEv,
m_scev_AffineAddRec(m_SCEV(LoadBaseEv), m_scev_APInt(Step))))
return false;
LLVM_DEBUG(dbgs() << "pointer load scev: " << *LoadEv << "\n");
uint64_t StepSize = Step->getZExtValue();
// Verify that StepSize is consistent with platform char width.
OpWidth = OperandType->getIntegerBitWidth();
unsigned WcharSize = TLI->getWCharSize(*LoopLoad->getModule());
if (OpWidth != 8 && OpWidth != 16 && OpWidth != 32)
return false;
if (StepSize != OpWidth / 8)
return false;
if (OpWidth >= 16) {
if (OpWidth != WcharSize * 8)
return false;
// Don't call `wcslen` if the base pointer isn't known to be aligned.
const DataLayout &DL = LoopLoad->getDataLayout();
Align OperandAlign = DL.getABITypeAlign(OperandType);
if (LoopLoad->getAlign() < OperandAlign) {
// For null pointers the load base number of trailing zeros may be
// larger than what `Align` allows. So take log2 of operand alignment
// instead and compare trailing zeros instead of alignments.
if (SE->getMinTrailingZeros(LoadBaseEv) < Log2(OperandAlign))
return false;
}
}
// Scan every instruction in the loop to ensure there are no side effects.
for (Instruction &I : *LoopBody)
if (I.mayHaveSideEffects())
return false;
BasicBlock *LoopExitBB = CurLoop->getExitBlock();
if (!LoopExitBB)
return false;
for (PHINode &PN : LoopExitBB->phis()) {
if (!SE->isSCEVable(PN.getType()))
return false;
const SCEV *Ev = SE->getSCEV(&PN);
if (!Ev)
return false;
LLVM_DEBUG(dbgs() << "loop exit phi scev: " << *Ev << "\n");
// Since we verified that the loop trip count will be a valid strlen
// idiom, we can expand all lcssa phi with {n,+,1} as (n + strlen) and use
// SCEVExpander materialize the loop output.
const SCEVAddRecExpr *AddRecEv = dyn_cast<SCEVAddRecExpr>(Ev);
if (!AddRecEv || !AddRecEv->isAffine())
return false;
// We only want RecAddExpr with recurrence step that is constant. This
// is good enough for all the idioms we want to recognize. Later we expand
// and materialize the recurrence as {base,+,a} -> (base + a * strlen)
if (!isa<SCEVConstant>(AddRecEv->getStepRecurrence(*SE)))
return false;
}
return true;
}
public:
const Loop *CurLoop;
ScalarEvolution *SE;
const TargetLibraryInfo *TLI;
unsigned OpWidth;
ConstantInt *StepSizeCI;
const SCEV *LoadBaseEv;
Type *OperandType;
};
} // namespace
/// The Strlen Idiom we are trying to detect has the following structure
///
/// preheader:
/// ...
/// br label %body, ...
///
/// body:
/// ... ; %0 is incremented by a gep
/// %1 = load i8, ptr %0, align 1
/// %2 = icmp eq i8 %1, 0
/// br i1 %2, label %exit, label %body
///
/// exit:
/// %lcssa = phi [%0, %body], ...
///
/// We expect the strlen idiom to have a load of a character type that
/// is compared against '\0', and such load pointer operand must have scev
/// expression of the form {%str,+,c} where c is a ConstantInt of the
/// appropiate character width for the idiom, and %str is the base of the string
/// And, that all lcssa phis have the form {...,+,n} where n is a constant,
///
/// When transforming the output of the strlen idiom, the lccsa phi are
/// expanded using SCEVExpander as {base scev,+,a} -> (base scev + a * strlen)
/// and all subsequent uses are replaced. For example,
///
/// \code{.c}
/// const char* base = str;
/// while (*str != '\0')
/// ++str;
/// size_t result = str - base;
/// \endcode
///
/// will be transformed as follows: The idiom will be replaced by a strlen
/// computation to compute the address of the null terminator of the string.
///
/// \code{.c}
/// const char* base = str;
/// const char* end = base + strlen(str);
/// size_t result = end - base;
/// \endcode
///
/// In the case we index by an induction variable, as long as the induction
/// variable has a constant int increment, we can replace all such indvars
/// with the closed form computation of strlen
///
/// \code{.c}
/// size_t i = 0;
/// while (str[i] != '\0')
/// ++i;
/// size_t result = i;
/// \endcode
///
/// Will be replaced by
///
/// \code{.c}
/// size_t i = 0 + strlen(str);
/// size_t result = i;
/// \endcode
///
bool LoopIdiomRecognize::recognizeAndInsertStrLen() {
if (DisableLIRP::All)
return false;
StrlenVerifier Verifier(CurLoop, SE, TLI);
if (!Verifier.isValidStrlenIdiom())
return false;
BasicBlock *Preheader = CurLoop->getLoopPreheader();
BasicBlock *LoopBody = *CurLoop->block_begin();
BasicBlock *LoopExitBB = CurLoop->getExitBlock();
CondBrInst *LoopTerm = cast<CondBrInst>(LoopBody->getTerminator());
assert(Preheader && LoopBody && LoopExitBB &&
"Should be verified to be valid by StrlenVerifier");
if (Verifier.OpWidth == 8) {
if (DisableLIRP::Strlen)
return false;
if (!isLibFuncEmittable(Preheader->getModule(), TLI, LibFunc_strlen))
return false;
} else {
if (DisableLIRP::Wcslen)
return false;
if (!isLibFuncEmittable(Preheader->getModule(), TLI, LibFunc_wcslen))
return false;
}
IRBuilder<> Builder(Preheader->getTerminator());
Builder.SetCurrentDebugLocation(CurLoop->getStartLoc());
SCEVExpander Expander(*SE, "strlen_idiom");
Value *MaterialzedBase = Expander.expandCodeFor(
Verifier.LoadBaseEv, Verifier.LoadBaseEv->getType(),
Builder.GetInsertPoint());
Value *StrLenFunc = nullptr;
if (Verifier.OpWidth == 8) {
StrLenFunc = emitStrLen(MaterialzedBase, Builder, *DL, TLI);
} else {
StrLenFunc = emitWcsLen(MaterialzedBase, Builder, *DL, TLI);
}
assert(StrLenFunc && "Failed to emit strlen function.");
const SCEV *StrlenEv = SE->getSCEV(StrLenFunc);
SmallVector<PHINode *, 4> Cleanup;
for (PHINode &PN : LoopExitBB->phis()) {
// We can now materialize the loop output as all phi have scev {base,+,a}.
// We expand the phi as:
// %strlen = call i64 @strlen(%str)
// %phi.new = base expression + step * %strlen
const SCEV *Ev = SE->getSCEV(&PN);
const SCEVAddRecExpr *AddRecEv = dyn_cast<SCEVAddRecExpr>(Ev);
const SCEVConstant *Step =
dyn_cast<SCEVConstant>(AddRecEv->getStepRecurrence(*SE));
const SCEV *Base = AddRecEv->getStart();
// It is safe to truncate to base since if base is narrower than size_t
// the equivalent user code will have to truncate anyways.
const SCEV *NewEv = SE->getAddExpr(
Base, SE->getMulExpr(Step, SE->getTruncateOrSignExtend(
StrlenEv, Base->getType())));
Value *MaterializedPHI = Expander.expandCodeFor(NewEv, NewEv->getType(),
Builder.GetInsertPoint());
Expander.clear();
PN.replaceAllUsesWith(MaterializedPHI);
Cleanup.push_back(&PN);
}
// All LCSSA Loop Phi are dead, the left over dead loop body can be cleaned
// up by later passes
for (PHINode *PN : Cleanup)
RecursivelyDeleteDeadPHINode(PN);
// LoopDeletion only delete invariant loops with known trip-count. We can
// update the condition so it will reliablely delete the invariant loop
assert((LoopTerm->getSuccessor(0) == LoopBody ||
LoopTerm->getSuccessor(1) == LoopBody) &&
"loop body must have a successor that is it self");
ConstantInt *NewLoopCond = LoopTerm->getSuccessor(0) == LoopBody
? Builder.getFalse()
: Builder.getTrue();
LoopTerm->setCondition(NewLoopCond);
SE->forgetLoop(CurLoop);
++NumStrLen;
LLVM_DEBUG(dbgs() << " Formed strlen idiom: " << *StrLenFunc << "\n");
ORE.emit([&]() {
return OptimizationRemark(DEBUG_TYPE, "recognizeAndInsertStrLen",
CurLoop->getStartLoc(), Preheader)
<< "Transformed " << StrLenFunc->getName() << " loop idiom";
});
return true;
}
/// Check if the given conditional branch is based on an unsigned less-than
/// comparison between a variable and a constant, and if the comparison is false
/// the control yields to the loop entry. If the branch matches the behaviour,
/// the variable involved in the comparison is returned.
static Value *matchShiftULTCondition(CondBrInst *BI, BasicBlock *LoopEntry,
APInt &Threshold) {
ICmpInst *Cond = dyn_cast<ICmpInst>(BI->getCondition());
if (!Cond)
return nullptr;
ConstantInt *CmpConst = dyn_cast<ConstantInt>(Cond->getOperand(1));
if (!CmpConst)
return nullptr;
BasicBlock *FalseSucc = BI->getSuccessor(1);
ICmpInst::Predicate Pred = Cond->getPredicate();
if (Pred == ICmpInst::ICMP_ULT && FalseSucc == LoopEntry) {
Threshold = CmpConst->getValue();
return Cond->getOperand(0);
}
return nullptr;
}
// Check if the recurrence variable `VarX` is in the right form to create
// the idiom. Returns the value coerced to a PHINode if so.
static PHINode *getRecurrenceVar(Value *VarX, Instruction *DefX,
BasicBlock *LoopEntry) {
auto *PhiX = dyn_cast<PHINode>(VarX);
if (PhiX && PhiX->getParent() == LoopEntry &&
(PhiX->getOperand(0) == DefX || PhiX->getOperand(1) == DefX))
return PhiX;
return nullptr;
}
/// Return true if the idiom is detected in the loop.
///
/// Additionally:
/// 1) \p CntInst is set to the instruction Counting Leading Zeros (CTLZ)
/// or nullptr if there is no such.
/// 2) \p CntPhi is set to the corresponding phi node
/// or nullptr if there is no such.
/// 3) \p InitX is set to the value whose CTLZ could be used.
/// 4) \p DefX is set to the instruction calculating Loop exit condition.
/// 5) \p Threshold is set to the constant involved in the unsigned less-than
/// comparison.
///
/// The core idiom we are trying to detect is:
/// \code
/// if (x0 < 2)
/// goto loop-exit // the precondition of the loop
/// cnt0 = init-val
/// do {
/// x = phi (x0, x.next); //PhiX
/// cnt = phi (cnt0, cnt.next)
///
/// cnt.next = cnt + 1;
/// ...
/// x.next = x >> 1; // DefX
/// } while (x >= 4)
/// loop-exit:
/// \endcode
static bool detectShiftUntilLessThanIdiom(Loop *CurLoop, const DataLayout &DL,
Intrinsic::ID &IntrinID,
Value *&InitX, Instruction *&CntInst,
PHINode *&CntPhi, Instruction *&DefX,
APInt &Threshold) {
BasicBlock *LoopEntry;
DefX = nullptr;
CntInst = nullptr;
CntPhi = nullptr;
LoopEntry = *(CurLoop->block_begin());
// step 1: Check if the loop-back branch is in desirable form.
auto *EntryBI = dyn_cast<CondBrInst>(LoopEntry->getTerminator());
if (!EntryBI)
return false;
if (Value *T = matchShiftULTCondition(EntryBI, LoopEntry, Threshold))
DefX = dyn_cast<Instruction>(T);
else
return false;
// step 2: Check the recurrence of variable X
if (!DefX || !isa<PHINode>(DefX))
return false;
PHINode *VarPhi = cast<PHINode>(DefX);
int Idx = VarPhi->getBasicBlockIndex(LoopEntry);
if (Idx == -1)
return false;
DefX = dyn_cast<Instruction>(VarPhi->getIncomingValue(Idx));
if (!DefX || DefX->getNumOperands() == 0 || DefX->getOperand(0) != VarPhi)
return false;
// step 3: detect instructions corresponding to "x.next = x >> 1"
if (DefX->getOpcode() != Instruction::LShr)
return false;
IntrinID = Intrinsic::ctlz;
ConstantInt *Shft = dyn_cast<ConstantInt>(DefX->getOperand(1));
if (!Shft || !Shft->isOne())
return false;
InitX = VarPhi->getIncomingValueForBlock(CurLoop->getLoopPreheader());
// step 4: Find the instruction which count the CTLZ: cnt.next = cnt + 1
// or cnt.next = cnt + -1.
// TODO: We can skip the step. If loop trip count is known (CTLZ),
// then all uses of "cnt.next" could be optimized to the trip count
// plus "cnt0". Currently it is not optimized.
// This step could be used to detect POPCNT instruction:
// cnt.next = cnt + (x.next & 1)
for (Instruction &Inst :
llvm::make_range(LoopEntry->getFirstNonPHIIt(), LoopEntry->end())) {
if (Inst.getOpcode() != Instruction::Add)
continue;
ConstantInt *Inc = dyn_cast<ConstantInt>(Inst.getOperand(1));
if (!Inc || (!Inc->isOne() && !Inc->isMinusOne()))
continue;
PHINode *Phi = getRecurrenceVar(Inst.getOperand(0), &Inst, LoopEntry);
if (!Phi)
continue;
CntInst = &Inst;
CntPhi = Phi;
break;
}
if (!CntInst)
return false;
return true;
}
/// Return true iff the idiom is detected in the loop.
///
/// Additionally:
/// 1) \p CntInst is set to the instruction counting the population bit.
/// 2) \p CntPhi is set to the corresponding phi node.
/// 3) \p Var is set to the value whose population bits are being counted.
///
/// The core idiom we are trying to detect is:
/// \code
/// if (x0 != 0)
/// goto loop-exit // the precondition of the loop
/// cnt0 = init-val;
/// do {
/// x1 = phi (x0, x2);
/// cnt1 = phi(cnt0, cnt2);
///
/// cnt2 = cnt1 + 1;
/// ...
/// x2 = x1 & (x1 - 1);
/// ...
/// } while(x != 0);
///
/// loop-exit:
/// \endcode
static bool detectPopcountIdiom(Loop *CurLoop, BasicBlock *PreCondBB,
Instruction *&CntInst, PHINode *&CntPhi,
Value *&Var) {
// step 1: Check to see if the look-back branch match this pattern:
// "if (a!=0) goto loop-entry".
BasicBlock *LoopEntry;
Instruction *DefX2, *CountInst;
Value *VarX1, *VarX0;
PHINode *PhiX, *CountPhi;
DefX2 = CountInst = nullptr;
VarX1 = VarX0 = nullptr;
PhiX = CountPhi = nullptr;
LoopEntry = *(CurLoop->block_begin());
// step 1: Check if the loop-back branch is in desirable form.
{
auto *LoopTerm = dyn_cast<CondBrInst>(LoopEntry->getTerminator());
if (!LoopTerm)
return false;
DefX2 = dyn_cast_or_null<Instruction>(matchCondition(LoopTerm, LoopEntry));
}
// step 2: detect instructions corresponding to "x2 = x1 & (x1 - 1)"
{
if (!DefX2 || DefX2->getOpcode() != Instruction::And)
return false;
BinaryOperator *SubOneOp;
if ((SubOneOp = dyn_cast<BinaryOperator>(DefX2->getOperand(0))))
VarX1 = DefX2->getOperand(1);
else {
VarX1 = DefX2->getOperand(0);
SubOneOp = dyn_cast<BinaryOperator>(DefX2->getOperand(1));
}
if (!SubOneOp || SubOneOp->getOperand(0) != VarX1)
return false;
ConstantInt *Dec = dyn_cast<ConstantInt>(SubOneOp->getOperand(1));
if (!Dec ||
!((SubOneOp->getOpcode() == Instruction::Sub && Dec->isOne()) ||
(SubOneOp->getOpcode() == Instruction::Add &&
Dec->isMinusOne()))) {
return false;
}
}
// step 3: Check the recurrence of variable X
PhiX = getRecurrenceVar(VarX1, DefX2, LoopEntry);
if (!PhiX)
return false;
// step 4: Find the instruction which count the population: cnt2 = cnt1 + 1
{
CountInst = nullptr;
for (Instruction &Inst :
llvm::make_range(LoopEntry->getFirstNonPHIIt(), LoopEntry->end())) {
if (Inst.getOpcode() != Instruction::Add)
continue;
ConstantInt *Inc = dyn_cast<ConstantInt>(Inst.getOperand(1));
if (!Inc || !Inc->isOne())
continue;
PHINode *Phi = getRecurrenceVar(Inst.getOperand(0), &Inst, LoopEntry);
if (!Phi)
continue;
// Check if the result of the instruction is live of the loop.
bool LiveOutLoop = false;
for (User *U : Inst.users()) {
if ((cast<Instruction>(U))->getParent() != LoopEntry) {
LiveOutLoop = true;
break;
}
}
if (LiveOutLoop) {
CountInst = &Inst;
CountPhi = Phi;
break;
}
}
if (!CountInst)
return false;
}
// step 5: check if the precondition is in this form:
// "if (x != 0) goto loop-head ; else goto somewhere-we-don't-care;"
{
auto *PreCondBr = dyn_cast<CondBrInst>(PreCondBB->getTerminator());
if (!PreCondBr)
return false;
Value *T = matchCondition(PreCondBr, CurLoop->getLoopPreheader());
if (T != PhiX->getOperand(0) && T != PhiX->getOperand(1))
return false;
CntInst = CountInst;
CntPhi = CountPhi;
Var = T;
}
return true;
}
/// Return true if the idiom is detected in the loop.
///
/// Additionally:
/// 1) \p CntInst is set to the instruction Counting Leading Zeros (CTLZ)
/// or nullptr if there is no such.
/// 2) \p CntPhi is set to the corresponding phi node
/// or nullptr if there is no such.
/// 3) \p Var is set to the value whose CTLZ could be used.
/// 4) \p DefX is set to the instruction calculating Loop exit condition.
///
/// The core idiom we are trying to detect is:
/// \code
/// if (x0 == 0)
/// goto loop-exit // the precondition of the loop
/// cnt0 = init-val;
/// do {
/// x = phi (x0, x.next); //PhiX
/// cnt = phi(cnt0, cnt.next);
///
/// cnt.next = cnt + 1;
/// ...
/// x.next = x >> 1; // DefX
/// ...
/// } while(x.next != 0);
///
/// loop-exit:
/// \endcode
static bool detectShiftUntilZeroIdiom(Loop *CurLoop, const DataLayout &DL,
Intrinsic::ID &IntrinID, Value *&InitX,
Instruction *&CntInst, PHINode *&CntPhi,
Instruction *&DefX) {
BasicBlock *LoopEntry;
Value *VarX = nullptr;
DefX = nullptr;
CntInst = nullptr;
CntPhi = nullptr;
LoopEntry = *(CurLoop->block_begin());
// step 1: Check if the loop-back branch is in desirable form.
auto *LoopTerm = dyn_cast<CondBrInst>(LoopEntry->getTerminator());
if (!LoopTerm)
return false;
DefX = dyn_cast_or_null<Instruction>(matchCondition(LoopTerm, LoopEntry));
// step 2: detect instructions corresponding to "x.next = x >> 1 or x << 1"
if (!DefX || !DefX->isShift())
return false;
IntrinID = DefX->getOpcode() == Instruction::Shl ? Intrinsic::cttz :
Intrinsic::ctlz;
ConstantInt *Shft = dyn_cast<ConstantInt>(DefX->getOperand(1));
if (!Shft || !Shft->isOne())
return false;
VarX = DefX->getOperand(0);
// step 3: Check the recurrence of variable X
PHINode *PhiX = getRecurrenceVar(VarX, DefX, LoopEntry);
if (!PhiX)
return false;
InitX = PhiX->getIncomingValueForBlock(CurLoop->getLoopPreheader());
// Make sure the initial value can't be negative otherwise the ashr in the
// loop might never reach zero which would make the loop infinite.
if (DefX->getOpcode() == Instruction::AShr && !isKnownNonNegative(InitX, DL))
return false;
// step 4: Find the instruction which count the CTLZ: cnt.next = cnt + 1
// or cnt.next = cnt + -1.
// TODO: We can skip the step. If loop trip count is known (CTLZ),
// then all uses of "cnt.next" could be optimized to the trip count
// plus "cnt0". Currently it is not optimized.
// This step could be used to detect POPCNT instruction:
// cnt.next = cnt + (x.next & 1)
for (Instruction &Inst :
llvm::make_range(LoopEntry->getFirstNonPHIIt(), LoopEntry->end())) {
if (Inst.getOpcode() != Instruction::Add)
continue;
ConstantInt *Inc = dyn_cast<ConstantInt>(Inst.getOperand(1));
if (!Inc || (!Inc->isOne() && !Inc->isMinusOne()))
continue;
PHINode *Phi = getRecurrenceVar(Inst.getOperand(0), &Inst, LoopEntry);
if (!Phi)
continue;
CntInst = &Inst;
CntPhi = Phi;
break;
}
if (!CntInst)
return false;
return true;
}
// Check if CTLZ / CTTZ intrinsic is profitable. Assume it is always
// profitable if we delete the loop.
bool LoopIdiomRecognize::isProfitableToInsertFFS(Intrinsic::ID IntrinID,
Value *InitX, bool ZeroCheck,
size_t CanonicalSize) {
const Value *Args[] = {InitX,
ConstantInt::getBool(InitX->getContext(), ZeroCheck)};
uint32_t HeaderSize = CurLoop->getHeader()->size();
IntrinsicCostAttributes Attrs(IntrinID, InitX->getType(), Args);
InstructionCost Cost = TTI->getIntrinsicInstrCost(
Attrs, TargetTransformInfo::TCK_SizeAndLatency);
if (HeaderSize != CanonicalSize && Cost > TargetTransformInfo::TCC_Basic)
return false;
return true;
}
/// Convert CTLZ / CTTZ idiom loop into countable loop.
/// If CTLZ / CTTZ inserted as a new trip count returns true; otherwise,
/// returns false.
bool LoopIdiomRecognize::insertFFSIfProfitable(Intrinsic::ID IntrinID,
Value *InitX, Instruction *DefX,
PHINode *CntPhi,
Instruction *CntInst) {
bool IsCntPhiUsedOutsideLoop = false;
for (User *U : CntPhi->users())
if (!CurLoop->contains(cast<Instruction>(U))) {
IsCntPhiUsedOutsideLoop = true;
break;
}
bool IsCntInstUsedOutsideLoop = false;
for (User *U : CntInst->users())
if (!CurLoop->contains(cast<Instruction>(U))) {
IsCntInstUsedOutsideLoop = true;
break;
}
// If both CntInst and CntPhi are used outside the loop the profitability
// is questionable.
if (IsCntInstUsedOutsideLoop && IsCntPhiUsedOutsideLoop)
return false;
// For some CPUs result of CTLZ(X) intrinsic is undefined
// when X is 0. If we can not guarantee X != 0, we need to check this
// when expand.
bool ZeroCheck = false;
// It is safe to assume Preheader exist as it was checked in
// parent function RunOnLoop.
BasicBlock *PH = CurLoop->getLoopPreheader();
// If we are using the count instruction outside the loop, make sure we
// have a zero check as a precondition. Without the check the loop would run
// one iteration for before any check of the input value. This means 0 and 1
// would have identical behavior in the original loop and thus
if (!IsCntPhiUsedOutsideLoop) {
auto *PreCondBB = PH->getSinglePredecessor();
auto *PreCondBI =
PreCondBB ? dyn_cast<CondBrInst>(PreCondBB->getTerminator()) : nullptr;
if (!(PreCondBI && matchCondition(PreCondBI, PH) == InitX) &&
!isKnownNonZero(
InitX, SimplifyQuery(*DL, DT, /*AC=*/nullptr, PH->getTerminator())))
return false;
ZeroCheck = true;
}
// FFS idiom loop has only 6 instructions:
// %n.addr.0 = phi [ %n, %entry ], [ %shr, %while.cond ]
// %i.0 = phi [ %i0, %entry ], [ %inc, %while.cond ]
// %shr = ashr %n.addr.0, 1
// %tobool = icmp eq %shr, 0
// %inc = add nsw %i.0, 1
// br i1 %tobool
size_t IdiomCanonicalSize = 6;
if (!isProfitableToInsertFFS(IntrinID, InitX, ZeroCheck, IdiomCanonicalSize))
return false;
transformLoopToCountable(IntrinID, PH, CntInst, CntPhi, InitX, DefX,
DefX->getDebugLoc(), ZeroCheck,
IsCntPhiUsedOutsideLoop);
return true;
}
/// Recognize CTLZ or CTTZ idiom in a non-countable loop and convert the loop
/// to countable (with CTLZ / CTTZ trip count). If CTLZ / CTTZ inserted as a new
/// trip count returns true; otherwise, returns false.
bool LoopIdiomRecognize::recognizeAndInsertFFS() {
// Give up if the loop has multiple blocks or multiple backedges.
if (CurLoop->getNumBackEdges() != 1 || CurLoop->getNumBlocks() != 1)
return false;
Intrinsic::ID IntrinID;
Value *InitX;
Instruction *DefX = nullptr;
PHINode *CntPhi = nullptr;
Instruction *CntInst = nullptr;
if (!detectShiftUntilZeroIdiom(CurLoop, *DL, IntrinID, InitX, CntInst, CntPhi,
DefX))
return false;
return insertFFSIfProfitable(IntrinID, InitX, DefX, CntPhi, CntInst);
}
bool LoopIdiomRecognize::recognizeShiftUntilLessThan() {
// Give up if the loop has multiple blocks or multiple backedges.
if (CurLoop->getNumBackEdges() != 1 || CurLoop->getNumBlocks() != 1)
return false;
Intrinsic::ID IntrinID;
Value *InitX;
Instruction *DefX = nullptr;
PHINode *CntPhi = nullptr;
Instruction *CntInst = nullptr;
APInt LoopThreshold;
if (!detectShiftUntilLessThanIdiom(CurLoop, *DL, IntrinID, InitX, CntInst,
CntPhi, DefX, LoopThreshold))
return false;
if (LoopThreshold == 2) {
// Treat as regular FFS.
return insertFFSIfProfitable(IntrinID, InitX, DefX, CntPhi, CntInst);
}
// Look for Floor Log2 Idiom.
if (LoopThreshold != 4)
return false;
// Abort if CntPhi is used outside of the loop.
for (User *U : CntPhi->users())
if (!CurLoop->contains(cast<Instruction>(U)))
return false;
// It is safe to assume Preheader exist as it was checked in
// parent function RunOnLoop.
BasicBlock *PH = CurLoop->getLoopPreheader();
auto *PreCondBB = PH->getSinglePredecessor();
if (!PreCondBB)
return false;
auto *PreCondBI = dyn_cast<CondBrInst>(PreCondBB->getTerminator());
if (!PreCondBI)
return false;
APInt PreLoopThreshold;
if (matchShiftULTCondition(PreCondBI, PH, PreLoopThreshold) != InitX ||
PreLoopThreshold != 2)
return false;
bool ZeroCheck = true;
// the loop has only 6 instructions:
// %n.addr.0 = phi [ %n, %entry ], [ %shr, %while.cond ]
// %i.0 = phi [ %i0, %entry ], [ %inc, %while.cond ]
// %shr = ashr %n.addr.0, 1
// %tobool = icmp ult %n.addr.0, C
// %inc = add nsw %i.0, 1
// br i1 %tobool
size_t IdiomCanonicalSize = 6;
if (!isProfitableToInsertFFS(IntrinID, InitX, ZeroCheck, IdiomCanonicalSize))
return false;
// log2(x) = w − 1 − clz(x)
transformLoopToCountable(IntrinID, PH, CntInst, CntPhi, InitX, DefX,
DefX->getDebugLoc(), ZeroCheck,
/*IsCntPhiUsedOutsideLoop=*/false,
/*InsertSub=*/true);
return true;
}
/// Recognizes a population count idiom in a non-countable loop.
///
/// If detected, transforms the relevant code to issue the popcount intrinsic
/// function call, and returns true; otherwise, returns false.
bool LoopIdiomRecognize::recognizePopcount() {
if (TTI->getPopcntSupport(32) != TargetTransformInfo::PSK_FastHardware)
return false;
// Counting population are usually conducted by few arithmetic instructions.
// Such instructions can be easily "absorbed" by vacant slots in a
// non-compact loop. Therefore, recognizing popcount idiom only makes sense
// in a compact loop.
// Give up if the loop has multiple blocks or multiple backedges.
if (CurLoop->getNumBackEdges() != 1 || CurLoop->getNumBlocks() != 1)
return false;
BasicBlock *LoopBody = *(CurLoop->block_begin());
if (LoopBody->size() >= 20) {
// The loop is too big, bail out.
return false;
}
// It should have a preheader containing nothing but an unconditional branch.
BasicBlock *PH = CurLoop->getLoopPreheader();
if (!PH || &PH->front() != PH->getTerminator())
return false;
auto *EntryBI = dyn_cast<UncondBrInst>(PH->getTerminator());
if (!EntryBI)
return false;
// It should have a precondition block where the generated popcount intrinsic
// function can be inserted.
auto *PreCondBB = PH->getSinglePredecessor();
if (!PreCondBB)
return false;
auto *PreCondBI = dyn_cast<CondBrInst>(PreCondBB->getTerminator());
if (!PreCondBI)
return false;
Instruction *CntInst;
PHINode *CntPhi;
Value *Val;
if (!detectPopcountIdiom(CurLoop, PreCondBB, CntInst, CntPhi, Val))
return false;
transformLoopToPopcount(PreCondBB, CntInst, CntPhi, Val);
return true;
}
static Value *createPopcntIntrinsic(IRBuilder<> &IRBuilder, Value *Val,
const DebugLoc &DL) {
Value *Ops[] = {Val};
Type *Tys[] = {Val->getType()};
IRBuilder.SetCurrentDebugLocation(DL);
return IRBuilder.CreateIntrinsic(Intrinsic::ctpop, Tys, Ops);
}
static Value *createFFSIntrinsic(IRBuilder<> &IRBuilder, Value *Val,
const DebugLoc &DL, bool ZeroCheck,
Intrinsic::ID IID) {
Value *Ops[] = {Val, IRBuilder.getInt1(ZeroCheck)};
Type *Tys[] = {Val->getType()};
IRBuilder.SetCurrentDebugLocation(DL);
return IRBuilder.CreateIntrinsic(IID, Tys, Ops);
}
/// Transform the following loop (Using CTLZ, CTTZ is similar):
/// loop:
/// CntPhi = PHI [Cnt0, CntInst]
/// PhiX = PHI [InitX, DefX]
/// CntInst = CntPhi + 1
/// DefX = PhiX >> 1
/// LOOP_BODY
/// Br: loop if (DefX != 0)
/// Use(CntPhi) or Use(CntInst)
///
/// Into:
/// If CntPhi used outside the loop:
/// CountPrev = BitWidth(InitX) - CTLZ(InitX >> 1)
/// Count = CountPrev + 1
/// else
/// Count = BitWidth(InitX) - CTLZ(InitX)
/// loop:
/// CntPhi = PHI [Cnt0, CntInst]
/// PhiX = PHI [InitX, DefX]
/// PhiCount = PHI [Count, Dec]
/// CntInst = CntPhi + 1
/// DefX = PhiX >> 1
/// Dec = PhiCount - 1
/// LOOP_BODY
/// Br: loop if (Dec != 0)
/// Use(CountPrev + Cnt0) // Use(CntPhi)
/// or
/// Use(Count + Cnt0) // Use(CntInst)
///
/// If LOOP_BODY is empty the loop will be deleted.
/// If CntInst and DefX are not used in LOOP_BODY they will be removed.
void LoopIdiomRecognize::transformLoopToCountable(
Intrinsic::ID IntrinID, BasicBlock *Preheader, Instruction *CntInst,
PHINode *CntPhi, Value *InitX, Instruction *DefX, const DebugLoc &DL,
bool ZeroCheck, bool IsCntPhiUsedOutsideLoop, bool InsertSub) {
// Step 1: Insert the CTLZ/CTTZ instruction at the end of the preheader block
IRBuilder<> Builder(Preheader->getTerminator());
Builder.SetCurrentDebugLocation(DL);
// If there are no uses of CntPhi crate:
// Count = BitWidth - CTLZ(InitX);
// NewCount = Count;
// If there are uses of CntPhi create:
// NewCount = BitWidth - CTLZ(InitX >> 1);
// Count = NewCount + 1;
Value *InitXNext;
if (IsCntPhiUsedOutsideLoop) {
if (DefX->getOpcode() == Instruction::AShr)
InitXNext = Builder.CreateAShr(InitX, 1);
else if (DefX->getOpcode() == Instruction::LShr)
InitXNext = Builder.CreateLShr(InitX, 1);
else if (DefX->getOpcode() == Instruction::Shl) // cttz
InitXNext = Builder.CreateShl(InitX, 1);
else
llvm_unreachable("Unexpected opcode!");
} else
InitXNext = InitX;
Value *Count =
createFFSIntrinsic(Builder, InitXNext, DL, ZeroCheck, IntrinID);
Type *CountTy = Count->getType();
Count = Builder.CreateSub(
ConstantInt::get(CountTy, CountTy->getIntegerBitWidth()), Count);
if (InsertSub)
Count = Builder.CreateSub(Count, ConstantInt::get(CountTy, 1));
Value *NewCount = Count;
if (IsCntPhiUsedOutsideLoop)
Count = Builder.CreateAdd(Count, ConstantInt::get(CountTy, 1));
NewCount = Builder.CreateZExtOrTrunc(NewCount, CntInst->getType());
Value *CntInitVal = CntPhi->getIncomingValueForBlock(Preheader);
if (cast<ConstantInt>(CntInst->getOperand(1))->isOne()) {
// If the counter was being incremented in the loop, add NewCount to the
// counter's initial value, but only if the initial value is not zero.
ConstantInt *InitConst = dyn_cast<ConstantInt>(CntInitVal);
if (!InitConst || !InitConst->isZero())
NewCount = Builder.CreateAdd(NewCount, CntInitVal);
} else {
// If the count was being decremented in the loop, subtract NewCount from
// the counter's initial value.
NewCount = Builder.CreateSub(CntInitVal, NewCount);
}
// Step 2: Insert new IV and loop condition:
// loop:
// ...
// PhiCount = PHI [Count, Dec]
// ...
// Dec = PhiCount - 1
// ...
// Br: loop if (Dec != 0)
BasicBlock *Body = *(CurLoop->block_begin());
auto *LbBr = cast<CondBrInst>(Body->getTerminator());
ICmpInst *LbCond = cast<ICmpInst>(LbBr->getCondition());
PHINode *TcPhi = PHINode::Create(CountTy, 2, "tcphi");
TcPhi->insertBefore(Body->begin());
Builder.SetInsertPoint(LbCond);
Instruction *TcDec = cast<Instruction>(Builder.CreateSub(
TcPhi, ConstantInt::get(CountTy, 1), "tcdec", false, true));
TcPhi->addIncoming(Count, Preheader);
TcPhi->addIncoming(TcDec, Body);
CmpInst::Predicate Pred =
(LbBr->getSuccessor(0) == Body) ? CmpInst::ICMP_NE : CmpInst::ICMP_EQ;
LbCond->setPredicate(Pred);
LbCond->setOperand(0, TcDec);
LbCond->setOperand(1, ConstantInt::get(CountTy, 0));
// Step 3: All the references to the original counter outside
// the loop are replaced with the NewCount
if (IsCntPhiUsedOutsideLoop)
CntPhi->replaceUsesOutsideBlock(NewCount, Body);
else
CntInst->replaceUsesOutsideBlock(NewCount, Body);
// step 4: Forget the "non-computable" trip-count SCEV associated with the
// loop. The loop would otherwise not be deleted even if it becomes empty.
SE->forgetLoop(CurLoop);
}
void LoopIdiomRecognize::transformLoopToPopcount(BasicBlock *PreCondBB,
Instruction *CntInst,
PHINode *CntPhi, Value *Var) {
BasicBlock *PreHead = CurLoop->getLoopPreheader();
auto *PreCondBr = cast<CondBrInst>(PreCondBB->getTerminator());
const DebugLoc &DL = CntInst->getDebugLoc();
// Assuming before transformation, the loop is following:
// if (x) // the precondition
// do { cnt++; x &= x - 1; } while(x);
// Step 1: Insert the ctpop instruction at the end of the precondition block
IRBuilder<> Builder(PreCondBr);
Value *PopCnt, *NewCount, *TripCnt;
{
PopCnt = createPopcntIntrinsic(Builder, Var, DL);
NewCount =
Builder.CreateZExtOrTrunc(PopCnt, cast<IntegerType>(CntPhi->getType()));
if (NewCount != PopCnt)
(cast<Instruction>(NewCount))->setDebugLoc(DL);
// TripCnt is exactly the number of iterations the loop has
TripCnt = PopCnt;
// If the population counter's initial value is not zero, insert Add Inst.
Value *CntInitVal = CntPhi->getIncomingValueForBlock(PreHead);
ConstantInt *InitConst = dyn_cast<ConstantInt>(CntInitVal);
if (!InitConst || !InitConst->isZero()) {
NewCount = Builder.CreateAdd(NewCount, CntInitVal);
(cast<Instruction>(NewCount))->setDebugLoc(DL);
}
}
// Step 2: Replace the precondition from "if (x == 0) goto loop-exit" to
// "if (TripCnt == 0) loop-exit". Without this change, the intrinsic
// function would be partial dead code, and downstream passes will drag
// it back from the precondition block to the preheader.
{
ICmpInst *PreCond = cast<ICmpInst>(PreCondBr->getCondition());
Value *Opnd0 = TripCnt;
Value *Opnd1 = ConstantInt::get(TripCnt->getType(), 0);
if (PreCond->getOperand(0) != Var)
std::swap(Opnd0, Opnd1);
ICmpInst *NewPreCond = cast<ICmpInst>(
Builder.CreateICmp(PreCond->getPredicate(), Opnd0, Opnd1));
PreCondBr->setCondition(NewPreCond);
RecursivelyDeleteTriviallyDeadInstructions(PreCond, TLI);
}
// Step 3: Note that the population count is exactly the trip count of the
// loop in question, which enable us to convert the loop from noncountable
// loop into a countable one. The benefit is twofold:
//
// - If the loop only counts population, the entire loop becomes dead after
// the transformation. It is a lot easier to prove a countable loop dead
// than to prove a noncountable one. (In some C dialects, an infinite loop
// isn't dead even if it computes nothing useful. In general, DCE needs
// to prove a noncountable loop finite before safely delete it.)
//
// - If the loop also performs something else, it remains alive.
// Since it is transformed to countable form, it can be aggressively
// optimized by some optimizations which are in general not applicable
// to a noncountable loop.
//
// After this step, this loop (conceptually) would look like following:
// newcnt = __builtin_ctpop(x);
// t = newcnt;
// if (x)
// do { cnt++; x &= x-1; t--) } while (t > 0);
BasicBlock *Body = *(CurLoop->block_begin());
{
auto *LbBr = cast<CondBrInst>(Body->getTerminator());
ICmpInst *LbCond = cast<ICmpInst>(LbBr->getCondition());
Type *Ty = TripCnt->getType();
PHINode *TcPhi = PHINode::Create(Ty, 2, "tcphi");
TcPhi->insertBefore(Body->begin());
Builder.SetInsertPoint(LbCond);
Instruction *TcDec = cast<Instruction>(
Builder.CreateSub(TcPhi, ConstantInt::get(Ty, 1), "tcdec",
/*HasNUW=*/true, /*HasNSW=*/false));
TcPhi->addIncoming(TripCnt, PreHead);
TcPhi->addIncoming(TcDec, Body);
CmpInst::Predicate Pred =
(LbBr->getSuccessor(0) == Body) ? CmpInst::ICMP_UGT : CmpInst::ICMP_EQ;
LbCond->setPredicate(Pred);
LbCond->setOperand(0, TcDec);
LbCond->setOperand(1, ConstantInt::get(Ty, 0));
}
// Step 4: All the references to the original population counter outside
// the loop are replaced with the NewCount -- the value returned from
// __builtin_ctpop().
CntInst->replaceUsesOutsideBlock(NewCount, Body);
// step 5: Forget the "non-computable" trip-count SCEV associated with the
// loop. The loop would otherwise not be deleted even if it becomes empty.
SE->forgetLoop(CurLoop);
}
/// Match loop-invariant value.
template <typename SubPattern_t> struct match_LoopInvariant {
SubPattern_t SubPattern;
const Loop *L;
match_LoopInvariant(const SubPattern_t &SP, const Loop *L)
: SubPattern(SP), L(L) {}
template <typename ITy> bool match(ITy *V) const {
return L->isLoopInvariant(V) && SubPattern.match(V);
}
};
/// Matches if the value is loop-invariant.
template <typename Ty>
inline match_LoopInvariant<Ty> m_LoopInvariant(const Ty &M, const Loop *L) {
return match_LoopInvariant<Ty>(M, L);
}
/// Return true if the idiom is detected in the loop.
///
/// The core idiom we are trying to detect is:
/// \code
/// entry:
/// <...>
/// %bitmask = shl i32 1, %bitpos
/// br label %loop
///
/// loop:
/// %x.curr = phi i32 [ %x, %entry ], [ %x.next, %loop ]
/// %x.curr.bitmasked = and i32 %x.curr, %bitmask
/// %x.curr.isbitunset = icmp eq i32 %x.curr.bitmasked, 0
/// %x.next = shl i32 %x.curr, 1
/// <...>
/// br i1 %x.curr.isbitunset, label %loop, label %end
///
/// end:
/// %x.curr.res = phi i32 [ %x.curr, %loop ] <...>
/// %x.next.res = phi i32 [ %x.next, %loop ] <...>
/// <...>
/// \endcode
static bool detectShiftUntilBitTestIdiom(Loop *CurLoop, Value *&BaseX,
Value *&BitMask, Value *&BitPos,
Value *&CurrX, Instruction *&NextX) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE
" Performing shift-until-bittest idiom detection.\n");
// Give up if the loop has multiple blocks or multiple backedges.
if (CurLoop->getNumBlocks() != 1 || CurLoop->getNumBackEdges() != 1) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad block/backedge count.\n");
return false;
}
BasicBlock *LoopHeaderBB = CurLoop->getHeader();
BasicBlock *LoopPreheaderBB = CurLoop->getLoopPreheader();
assert(LoopPreheaderBB && "There is always a loop preheader.");
using namespace PatternMatch;
// Step 1: Check if the loop backedge is in desirable form.
CmpPredicate Pred;
Value *CmpLHS, *CmpRHS;
BasicBlock *TrueBB, *FalseBB;
if (!match(LoopHeaderBB->getTerminator(),
m_Br(m_ICmp(Pred, m_Value(CmpLHS), m_Value(CmpRHS)),
m_BasicBlock(TrueBB), m_BasicBlock(FalseBB)))) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad backedge structure.\n");
return false;
}
// Step 2: Check if the backedge's condition is in desirable form.
auto MatchVariableBitMask = [&]() {
return ICmpInst::isEquality(Pred) && match(CmpRHS, m_Zero()) &&
match(CmpLHS,
m_c_And(m_Value(CurrX),
m_CombineAnd(
m_Value(BitMask),
m_LoopInvariant(m_Shl(m_One(), m_Value(BitPos)),
CurLoop))));
};
auto MatchDecomposableConstantBitMask = [&]() {
auto Res = llvm::decomposeBitTestICmp(
CmpLHS, CmpRHS, Pred, /*LookThroughTrunc=*/true,
/*AllowNonZeroC=*/false, /*DecomposeAnd=*/true);
if (Res && Res->Mask.isPowerOf2()) {
assert(ICmpInst::isEquality(Res->Pred));
Pred = Res->Pred;
CurrX = Res->X;
BitMask = ConstantInt::get(CurrX->getType(), Res->Mask);
BitPos = ConstantInt::get(CurrX->getType(), Res->Mask.logBase2());
return true;
}
return false;
};
if (!MatchVariableBitMask() && !MatchDecomposableConstantBitMask()) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad backedge comparison.\n");
return false;
}
// Step 3: Check if the recurrence is in desirable form.
auto *CurrXPN = dyn_cast<PHINode>(CurrX);
if (!CurrXPN || CurrXPN->getParent() != LoopHeaderBB) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Not an expected PHI node.\n");
return false;
}
BaseX = CurrXPN->getIncomingValueForBlock(LoopPreheaderBB);
NextX =
dyn_cast<Instruction>(CurrXPN->getIncomingValueForBlock(LoopHeaderBB));
assert(CurLoop->isLoopInvariant(BaseX) &&
"Expected BaseX to be available in the preheader!");
if (!NextX || (!match(NextX, m_Shl(m_Specific(CurrX), m_One())) &&
!match(NextX, m_LShr(m_Specific(CurrX), m_One())))) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad recurrence.\n");
return false;
}
// Step 4: Check if the backedge's destinations are in desirable form.
assert(ICmpInst::isEquality(Pred) &&
"Should only get equality predicates here.");
// cmp-br is commutative, so canonicalize to a single variant.
if (Pred != ICmpInst::Predicate::ICMP_EQ) {
Pred = ICmpInst::getInversePredicate(Pred);
std::swap(TrueBB, FalseBB);
}
// We expect to exit loop when comparison yields false,
// so when it yields true we should branch back to loop header.
if (TrueBB != LoopHeaderBB) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad backedge flow.\n");
return false;
}
// Okay, idiom checks out.
return true;
}
static bool isRightShiftBitTestKnownToTerminate(Value *X, Value *BitPos,
BasicBlock *Preheader,
const DataLayout &DL,
DominatorTree *DT) {
using namespace PatternMatch;
auto *Pos = dyn_cast<ConstantInt>(BitPos);
unsigned Bitwidth = X->getType()->getScalarSizeInBits();
if (!Pos || Bitwidth == 1 || Pos->getValue().uge(Bitwidth))
return false;
SimplifyQuery SQ(DL, DT, /*AC=*/nullptr, Preheader->getTerminator());
if (!isKnownNonZero(X, SQ))
return false;
if (Pos->isZero())
return true;
// The guard must rule out set bits below BitPos; those bits can never
// reach the tested position when X shifts right.
BasicBlock *GuardBB = Preheader->getSinglePredecessor();
if (!GuardBB)
return false;
auto *Guard = dyn_cast<CondBrInst>(GuardBB->getTerminator());
if (!Guard)
return false;
CmpPredicate Pred;
Value *LHS;
if (!match(Guard->getCondition(), m_ICmp(Pred, m_Value(LHS), m_Zero())) ||
!ICmpInst::isEquality(Pred))
return false;
const APInt *Mask;
if (!match(LHS, m_And(m_Specific(X), m_APInt(Mask))))
return false;
BasicBlock *ZeroSuccessor =
Guard->getSuccessor(Pred == ICmpInst::ICMP_EQ ? 0 : 1);
if (ZeroSuccessor != Preheader)
return false;
return Mask->countr_one() >= Pos->getZExtValue();
}
/// Look for the following loop:
/// \code
/// entry:
/// <...>
/// %bitmask = shl i32 1, %bitpos
/// br label %loop
///
/// loop:
/// %x.curr = phi i32 [ %x, %entry ], [ %x.next, %loop ]
/// %x.curr.bitmasked = and i32 %x.curr, %bitmask
/// %x.curr.isbitunset = icmp eq i32 %x.curr.bitmasked, 0
/// %x.next = shl i32 %x.curr, 1
/// <...>
/// br i1 %x.curr.isbitunset, label %loop, label %end
///
/// end:
/// %x.curr.res = phi i32 [ %x.curr, %loop ] <...>
/// %x.next.res = phi i32 [ %x.next, %loop ] <...>
/// <...>
/// \endcode
///
/// And transform it into:
/// \code
/// entry:
/// %bitmask = shl i32 1, %bitpos
/// %lowbitmask = add i32 %bitmask, -1
/// %mask = or i32 %lowbitmask, %bitmask
/// %x.masked = and i32 %x, %mask
/// %x.masked.numleadingzeros = call i32 @llvm.ctlz.i32(i32 %x.masked,
/// i1 true)
/// %x.masked.numactivebits = sub i32 32, %x.masked.numleadingzeros
/// %x.masked.leadingonepos = add i32 %x.masked.numactivebits, -1
/// %backedgetakencount = sub i32 %bitpos, %x.masked.leadingonepos
/// %tripcount = add i32 %backedgetakencount, 1
/// %x.curr = shl i32 %x, %backedgetakencount
/// %x.next = shl i32 %x, %tripcount
/// br label %loop
///
/// loop:
/// %loop.iv = phi i32 [ 0, %entry ], [ %loop.iv.next, %loop ]
/// %loop.iv.next = add nuw i32 %loop.iv, 1
/// %loop.ivcheck = icmp eq i32 %loop.iv.next, %tripcount
/// <...>
/// br i1 %loop.ivcheck, label %end, label %loop
///
/// end:
/// %x.curr.res = phi i32 [ %x.curr, %loop ] <...>
/// %x.next.res = phi i32 [ %x.next, %loop ] <...>
/// <...>
/// \endcode
/// A logical right-shift recurrence uses cttz of X shifted by BitPos instead,
/// provided the preheader proves that the loop will reach a set bit.
bool LoopIdiomRecognize::recognizeShiftUntilBitTest() {
bool MadeChange = false;
Value *X, *BitMask, *BitPos, *XCurr;
Instruction *XNext;
if (!detectShiftUntilBitTestIdiom(CurLoop, X, BitMask, BitPos, XCurr,
XNext)) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE
" shift-until-bittest idiom detection failed.\n");
return MadeChange;
}
LLVM_DEBUG(dbgs() << DEBUG_TYPE " shift-until-bittest idiom detected!\n");
// Ok, it is the idiom we were looking for, we *could* transform this loop,
// but is it profitable to transform?
BasicBlock *LoopHeaderBB = CurLoop->getHeader();
BasicBlock *LoopPreheaderBB = CurLoop->getLoopPreheader();
assert(LoopPreheaderBB && "There is always a loop preheader.");
BasicBlock *SuccessorBB = CurLoop->getExitBlock();
assert(SuccessorBB && "There is only a single successor.");
IRBuilder<> Builder(LoopPreheaderBB->getTerminator());
Builder.SetCurrentDebugLocation(cast<Instruction>(XCurr)->getDebugLoc());
bool IsRightShift = XNext->getOpcode() == Instruction::LShr;
Intrinsic::ID IntrID = IsRightShift ? Intrinsic::cttz : Intrinsic::ctlz;
Type *Ty = X->getType();
unsigned Bitwidth = Ty->getScalarSizeInBits();
// A right shift must be guaranteed to reach a set bit at BitPos.
if (IsRightShift && !isRightShiftBitTestKnownToTerminate(
X, BitPos, LoopPreheaderBB, *DL, DT)) {
return false;
}
TargetTransformInfo::TargetCostKind CostKind =
TargetTransformInfo::TCK_SizeAndLatency;
// The rewrite is considered to be unprofitable iff and only iff the
// intrinsic/shift we'll use are not cheap. Note that we are okay with *just*
// making the loop countable, even if nothing else changes.
IntrinsicCostAttributes Attrs(
IntrID, Ty, {PoisonValue::get(Ty), /*is_zero_poison=*/Builder.getTrue()});
InstructionCost Cost = TTI->getIntrinsicInstrCost(Attrs, CostKind);
if (Cost > TargetTransformInfo::TCC_Basic) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE
" Intrinsic is too costly, not beneficial\n");
return MadeChange;
}
if (TTI->getArithmeticInstrCost(XNext->getOpcode(), Ty, CostKind) >
TargetTransformInfo::TCC_Basic) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Shift is too costly, not beneficial\n");
return MadeChange;
}
// Ok, transform appears worthwhile.
MadeChange = true;
if (!isGuaranteedNotToBeUndefOrPoison(BitPos)) {
// BitMask may be computed from BitPos, Freeze BitPos so we can increase
// it's use count.
std::optional<BasicBlock::iterator> InsertPt = std::nullopt;
if (auto *BitPosI = dyn_cast<Instruction>(BitPos))
InsertPt = BitPosI->getInsertionPointAfterDef();
else
InsertPt = DT->getRoot()->getFirstNonPHIOrAlloca();
if (!InsertPt)
return false;
FreezeInst *BitPosFrozen =
new FreezeInst(BitPos, BitPos->getName() + ".fr", *InsertPt);
BitPos->replaceUsesWithIf(BitPosFrozen, [BitPosFrozen](Use &U) {
return U.getUser() != BitPosFrozen;
});
BitPos = BitPosFrozen;
}
// Step 1: Compute the loop trip count.
Value *LoopBackedgeTakenCount;
if (IsRightShift) {
Value *ShiftedX = Builder.CreateLShr(X, BitPos);
LoopBackedgeTakenCount = Builder.CreateIntrinsic(
IntrID, Ty, {ShiftedX, /*is_zero_poison=*/Builder.getTrue()});
} else {
Value *LowBitMask =
Builder.CreateAdd(BitMask, Constant::getAllOnesValue(Ty),
BitPos->getName() + ".lowbitmask");
Value *Mask =
Builder.CreateOr(LowBitMask, BitMask, BitPos->getName() + ".mask");
Value *XMasked = Builder.CreateAnd(X, Mask, X->getName() + ".masked");
Value *XMaskedNumLeadingZeros = Builder.CreateIntrinsic(
IntrID, Ty, {XMasked, /*is_zero_poison=*/Builder.getTrue()},
/*FMFSource=*/nullptr, XMasked->getName() + ".numleadingzeros");
Value *XMaskedNumActiveBits = Builder.CreateSub(
ConstantInt::get(Ty, Ty->getScalarSizeInBits()), XMaskedNumLeadingZeros,
XMasked->getName() + ".numactivebits",
/*HasNUW=*/true, /*HasNSW=*/Bitwidth != 2);
Value *XMaskedLeadingOnePos = Builder.CreateAdd(
XMaskedNumActiveBits, Constant::getAllOnesValue(Ty),
XMasked->getName() + ".leadingonepos", /*HasNUW=*/false,
/*HasNSW=*/Bitwidth > 2);
LoopBackedgeTakenCount = Builder.CreateSub(
BitPos, XMaskedLeadingOnePos,
CurLoop->getName() + ".backedgetakencount", /*HasNUW=*/true,
/*HasNSW=*/true);
}
// We know loop's backedge-taken count, but what's loop's trip count?
// Note that while NUW is always safe, while NSW is only for bitwidths != 2.
Value *LoopTripCount =
Builder.CreateAdd(LoopBackedgeTakenCount, ConstantInt::get(Ty, 1),
CurLoop->getName() + ".tripcount", /*HasNUW=*/true,
/*HasNSW=*/Bitwidth != 2);
// Step 2: Compute the recurrence's final value without a loop.
// NewX is always safe to compute, because `LoopBackedgeTakenCount`
// will always be smaller than `bitwidth(X)`, i.e. we never get poison.
Value *NewX = IsRightShift ? Builder.CreateLShr(X, LoopBackedgeTakenCount)
: Builder.CreateShl(X, LoopBackedgeTakenCount);
NewX->takeName(XCurr);
if (auto *I = dyn_cast<Instruction>(NewX))
I->copyIRFlags(XNext, /*IncludeWrapFlags=*/true);
Value *NewXNext;
// Shifting X by LoopTripCount is poison if the count equals Bitwidth.
// For right shifts, shift NewX by one instead. For left shifts, this can
// happen when BitPos is the highest bit and X is one, so use a single
// additional shift unless the larger shift is known to be safe.
if (IsRightShift) {
NewXNext = Builder.CreateLShr(NewX, ConstantInt::get(Ty, 1));
} else if (XNext->hasNoSignedWrap() || XNext->hasNoUnsignedWrap() ||
PatternMatch::match(BitPos,
PatternMatch::m_SpecificInt_ICMP(
ICmpInst::ICMP_NE,
APInt(Ty->getScalarSizeInBits(),
Ty->getScalarSizeInBits() - 1)))) {
NewXNext = Builder.CreateShl(X, LoopTripCount);
} else {
// Otherwise, just additionally shift by one. It's the smallest solution,
// alternatively, we could check that NewX is INT_MIN (or BitPos is )
// and select 0 instead.
NewXNext = Builder.CreateShl(NewX, ConstantInt::get(Ty, 1));
}
NewXNext->takeName(XNext);
if (auto *I = dyn_cast<Instruction>(NewXNext))
I->copyIRFlags(XNext, /*IncludeWrapFlags=*/true);
// Step 3: Adjust the successor basic block to receive the computed
// recurrence's final value instead of the recurrence itself.
XCurr->replaceUsesOutsideBlock(NewX, LoopHeaderBB);
XNext->replaceUsesOutsideBlock(NewXNext, LoopHeaderBB);
// Step 4: Rewrite the loop into a countable form, with canonical IV.
// The new canonical induction variable.
Builder.SetInsertPoint(LoopHeaderBB->begin());
auto *IV = Builder.CreatePHI(Ty, 2, CurLoop->getName() + ".iv");
// The induction itself.
// Note that while NUW is always safe, while NSW is only for bitwidths != 2.
Builder.SetInsertPoint(LoopHeaderBB->getTerminator());
auto *IVNext =
Builder.CreateAdd(IV, ConstantInt::get(Ty, 1), IV->getName() + ".next",
/*HasNUW=*/true, /*HasNSW=*/Bitwidth != 2);
// The loop trip count check.
auto *IVCheck = Builder.CreateICmpEQ(IVNext, LoopTripCount,
CurLoop->getName() + ".ivcheck");
SmallVector<uint32_t> BranchWeights;
const bool HasBranchWeights =
extractBranchWeights(*LoopHeaderBB->getTerminator(), BranchWeights);
auto *BI = Builder.CreateCondBr(IVCheck, SuccessorBB, LoopHeaderBB);
if (HasBranchWeights) {
if (SuccessorBB == LoopHeaderBB->getTerminator()->getSuccessor(1))
std::swap(BranchWeights[0], BranchWeights[1]);
// We're not changing the loop profile, so we can reuse the original loop's
// profile.
setBranchWeights(*BI, BranchWeights,
/*IsExpected=*/false);
}
LoopHeaderBB->getTerminator()->eraseFromParent();
// Populate the IV PHI.
IV->addIncoming(ConstantInt::get(Ty, 0), LoopPreheaderBB);
IV->addIncoming(IVNext, LoopHeaderBB);
// Step 5: Forget the "non-computable" trip-count SCEV associated with the
// loop. The loop would otherwise not be deleted even if it becomes empty.
SE->forgetLoop(CurLoop);
// Other passes will take care of actually deleting the loop if possible.
LLVM_DEBUG(dbgs() << DEBUG_TYPE " shift-until-bittest idiom optimized!\n");
++NumShiftUntilBitTest;
return MadeChange;
}
/// Return true if the idiom is detected in the loop.
///
/// The core idiom we are trying to detect is:
/// \code
/// entry:
/// <...>
/// %start = <...>
/// %extraoffset = <...>
/// <...>
/// br label %for.cond
///
/// loop:
/// %iv = phi i8 [ %start, %entry ], [ %iv.next, %for.cond ]
/// %nbits = add nsw i8 %iv, %extraoffset
/// %val.shifted = {{l,a}shr,shl} i8 %val, %nbits
/// %val.shifted.iszero = icmp eq i8 %val.shifted, 0
/// %iv.next = add i8 %iv, 1
/// <...>
/// br i1 %val.shifted.iszero, label %end, label %loop
///
/// end:
/// %iv.res = phi i8 [ %iv, %loop ] <...>
/// %nbits.res = phi i8 [ %nbits, %loop ] <...>
/// %val.shifted.res = phi i8 [ %val.shifted, %loop ] <...>
/// %val.shifted.iszero.res = phi i1 [ %val.shifted.iszero, %loop ] <...>
/// %iv.next.res = phi i8 [ %iv.next, %loop ] <...>
/// <...>
/// \endcode
static bool detectShiftUntilZeroIdiom(Loop *CurLoop, ScalarEvolution *SE,
Instruction *&ValShiftedIsZero,
Intrinsic::ID &IntrinID, Instruction *&IV,
Value *&Start, Value *&Val,
const SCEV *&ExtraOffsetExpr,
bool &InvertedCond) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE
" Performing shift-until-zero idiom detection.\n");
// Give up if the loop has multiple blocks or multiple backedges.
if (CurLoop->getNumBlocks() != 1 || CurLoop->getNumBackEdges() != 1) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad block/backedge count.\n");
return false;
}
Instruction *ValShifted, *NBits, *IVNext;
Value *ExtraOffset;
BasicBlock *LoopHeaderBB = CurLoop->getHeader();
BasicBlock *LoopPreheaderBB = CurLoop->getLoopPreheader();
assert(LoopPreheaderBB && "There is always a loop preheader.");
using namespace PatternMatch;
// Step 1: Check if the loop backedge, condition is in desirable form.
CmpPredicate Pred;
BasicBlock *TrueBB, *FalseBB;
if (!match(LoopHeaderBB->getTerminator(),
m_Br(m_Instruction(ValShiftedIsZero), m_BasicBlock(TrueBB),
m_BasicBlock(FalseBB))) ||
!match(ValShiftedIsZero,
m_ICmp(Pred, m_Instruction(ValShifted), m_Zero())) ||
!ICmpInst::isEquality(Pred)) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad backedge structure.\n");
return false;
}
// Step 2: Check if the comparison's operand is in desirable form.
// FIXME: Val could be a one-input PHI node, which we should look past.
if (!match(ValShifted, m_Shift(m_LoopInvariant(m_Value(Val), CurLoop),
m_Instruction(NBits)))) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad comparisons value computation.\n");
return false;
}
IntrinID = ValShifted->getOpcode() == Instruction::Shl ? Intrinsic::cttz
: Intrinsic::ctlz;
// Step 3: Check if the shift amount is in desirable form.
if (match(NBits, m_c_Add(m_Instruction(IV),
m_LoopInvariant(m_Value(ExtraOffset), CurLoop))) &&
(NBits->hasNoSignedWrap() || NBits->hasNoUnsignedWrap()))
ExtraOffsetExpr = SE->getNegativeSCEV(SE->getSCEV(ExtraOffset));
else if (match(NBits,
m_Sub(m_Instruction(IV),
m_LoopInvariant(m_Value(ExtraOffset), CurLoop))) &&
NBits->hasNoSignedWrap())
ExtraOffsetExpr = SE->getSCEV(ExtraOffset);
else {
IV = NBits;
ExtraOffsetExpr = SE->getZero(NBits->getType());
}
// Step 4: Check if the recurrence is in desirable form.
auto *IVPN = dyn_cast<PHINode>(IV);
if (!IVPN || IVPN->getParent() != LoopHeaderBB) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Not an expected PHI node.\n");
return false;
}
Start = IVPN->getIncomingValueForBlock(LoopPreheaderBB);
IVNext = dyn_cast<Instruction>(IVPN->getIncomingValueForBlock(LoopHeaderBB));
if (!IVNext || !match(IVNext, m_Add(m_Specific(IVPN), m_One()))) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad recurrence.\n");
return false;
}
// Step 4: Check if the backedge's destinations are in desirable form.
assert(ICmpInst::isEquality(Pred) &&
"Should only get equality predicates here.");
// cmp-br is commutative, so canonicalize to a single variant.
InvertedCond = Pred != ICmpInst::Predicate::ICMP_EQ;
if (InvertedCond) {
Pred = ICmpInst::getInversePredicate(Pred);
std::swap(TrueBB, FalseBB);
}
// We expect to exit loop when comparison yields true,
// so when it yields false we should branch back to loop header.
if (FalseBB != LoopHeaderBB) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Bad backedge flow.\n");
return false;
}
// The new, countable, loop will certainly only run a known number of
// iterations, It won't be infinite. But the old loop might be infinite
// under certain conditions. For logical shifts, the value will become zero
// after at most bitwidth(%Val) loop iterations. However, for arithmetic
// right-shift, iff the sign bit was set, the value will never become zero,
// and the loop may never finish.
if (ValShifted->getOpcode() == Instruction::AShr &&
!isMustProgress(CurLoop) && !SE->isKnownNonNegative(SE->getSCEV(Val))) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE " Can not prove the loop is finite.\n");
return false;
}
// Okay, idiom checks out.
return true;
}
/// Look for the following loop:
/// \code
/// entry:
/// <...>
/// %start = <...>
/// %extraoffset = <...>
/// <...>
/// br label %loop
///
/// loop:
/// %iv = phi i8 [ %start, %entry ], [ %iv.next, %loop ]
/// %nbits = add nsw i8 %iv, %extraoffset
/// %val.shifted = {{l,a}shr,shl} i8 %val, %nbits
/// %val.shifted.iszero = icmp eq i8 %val.shifted, 0
/// %iv.next = add i8 %iv, 1
/// <...>
/// br i1 %val.shifted.iszero, label %end, label %loop
///
/// end:
/// %iv.res = phi i8 [ %iv, %loop ] <...>
/// %nbits.res = phi i8 [ %nbits, %loop ] <...>
/// %val.shifted.res = phi i8 [ %val.shifted, %loop ] <...>
/// %val.shifted.iszero.res = phi i1 [ %val.shifted.iszero, %loop ] <...>
/// %iv.next.res = phi i8 [ %iv.next, %loop ] <...>
/// <...>
/// \endcode
///
/// And transform it into:
/// \code
/// entry:
/// <...>
/// %start = <...>
/// %extraoffset = <...>
/// <...>
/// %val.numleadingzeros = call i8 @llvm.ct{l,t}z.i8(i8 %val, i1 0)
/// %val.numactivebits = sub i8 8, %val.numleadingzeros
/// %extraoffset.neg = sub i8 0, %extraoffset
/// %tmp = add i8 %val.numactivebits, %extraoffset.neg
/// %iv.final = call i8 @llvm.smax.i8(i8 %tmp, i8 %start)
/// %loop.tripcount = sub i8 %iv.final, %start
/// br label %loop
///
/// loop:
/// %loop.iv = phi i8 [ 0, %entry ], [ %loop.iv.next, %loop ]
/// %loop.iv.next = add i8 %loop.iv, 1
/// %loop.ivcheck = icmp eq i8 %loop.iv.next, %loop.tripcount
/// %iv = add i8 %loop.iv, %start
/// <...>
/// br i1 %loop.ivcheck, label %end, label %loop
///
/// end:
/// %iv.res = phi i8 [ %iv.final, %loop ] <...>
/// <...>
/// \endcode
bool LoopIdiomRecognize::recognizeShiftUntilZero() {
bool MadeChange = false;
Instruction *ValShiftedIsZero;
Intrinsic::ID IntrID;
Instruction *IV;
Value *Start, *Val;
const SCEV *ExtraOffsetExpr;
bool InvertedCond;
if (!detectShiftUntilZeroIdiom(CurLoop, SE, ValShiftedIsZero, IntrID, IV,
Start, Val, ExtraOffsetExpr, InvertedCond)) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE
" shift-until-zero idiom detection failed.\n");
return MadeChange;
}
LLVM_DEBUG(dbgs() << DEBUG_TYPE " shift-until-zero idiom detected!\n");
// Ok, it is the idiom we were looking for, we *could* transform this loop,
// but is it profitable to transform?
BasicBlock *LoopHeaderBB = CurLoop->getHeader();
BasicBlock *LoopPreheaderBB = CurLoop->getLoopPreheader();
assert(LoopPreheaderBB && "There is always a loop preheader.");
BasicBlock *SuccessorBB = CurLoop->getExitBlock();
assert(SuccessorBB && "There is only a single successor.");
IRBuilder<> Builder(LoopPreheaderBB->getTerminator());
Builder.SetCurrentDebugLocation(IV->getDebugLoc());
Type *Ty = Val->getType();
unsigned Bitwidth = Ty->getScalarSizeInBits();
TargetTransformInfo::TargetCostKind CostKind =
TargetTransformInfo::TCK_SizeAndLatency;
// The rewrite is considered to be unprofitable iff and only iff the
// intrinsic we'll use are not cheap. Note that we are okay with *just*
// making the loop countable, even if nothing else changes.
IntrinsicCostAttributes Attrs(
IntrID, Ty, {PoisonValue::get(Ty), /*is_zero_poison=*/Builder.getFalse()});
InstructionCost Cost = TTI->getIntrinsicInstrCost(Attrs, CostKind);
if (Cost > TargetTransformInfo::TCC_Basic) {
LLVM_DEBUG(dbgs() << DEBUG_TYPE
" Intrinsic is too costly, not beneficial\n");
return MadeChange;
}
// Ok, transform appears worthwhile.
MadeChange = true;
bool OffsetIsZero = ExtraOffsetExpr->isZero();
// Step 1: Compute the loop's final IV value / trip count.
Value *ValNumLeadingZeros = Builder.CreateIntrinsic(
IntrID, Ty, {Val, /*is_zero_poison=*/Builder.getFalse()},
/*FMFSource=*/nullptr, Val->getName() + ".numleadingzeros");
Value *ValNumActiveBits = Builder.CreateSub(
ConstantInt::get(Ty, Ty->getScalarSizeInBits()), ValNumLeadingZeros,
Val->getName() + ".numactivebits", /*HasNUW=*/true,
/*HasNSW=*/Bitwidth != 2);
SCEVExpander Expander(*SE, "loop-idiom");
Expander.setInsertPoint(&*Builder.GetInsertPoint());
Value *ExtraOffset = Expander.expandCodeFor(ExtraOffsetExpr);
Value *ValNumActiveBitsOffset = Builder.CreateAdd(
ValNumActiveBits, ExtraOffset, ValNumActiveBits->getName() + ".offset",
/*HasNUW=*/OffsetIsZero, /*HasNSW=*/true);
Value *IVFinal = Builder.CreateIntrinsic(Intrinsic::smax, {Ty},
{ValNumActiveBitsOffset, Start},
/*FMFSource=*/nullptr, "iv.final");
auto *LoopBackedgeTakenCount = cast<Instruction>(Builder.CreateSub(
IVFinal, Start, CurLoop->getName() + ".backedgetakencount",
/*HasNUW=*/OffsetIsZero, /*HasNSW=*/true));
// FIXME: or when the offset was `add nuw`
// We know loop's backedge-taken count, but what's loop's trip count?
Value *LoopTripCount =
Builder.CreateAdd(LoopBackedgeTakenCount, ConstantInt::get(Ty, 1),
CurLoop->getName() + ".tripcount", /*HasNUW=*/true,
/*HasNSW=*/Bitwidth != 2);
// Step 2: Adjust the successor basic block to receive the original
// induction variable's final value instead of the orig. IV itself.
IV->replaceUsesOutsideBlock(IVFinal, LoopHeaderBB);
// Step 3: Rewrite the loop into a countable form, with canonical IV.
// The new canonical induction variable.
Builder.SetInsertPoint(LoopHeaderBB->begin());
auto *CIV = Builder.CreatePHI(Ty, 2, CurLoop->getName() + ".iv");
// The induction itself.
Builder.SetInsertPoint(LoopHeaderBB->getFirstNonPHIIt());
auto *CIVNext =
Builder.CreateAdd(CIV, ConstantInt::get(Ty, 1), CIV->getName() + ".next",
/*HasNUW=*/true, /*HasNSW=*/Bitwidth != 2);
// The loop trip count check.
auto *CIVCheck = Builder.CreateICmpEQ(CIVNext, LoopTripCount,
CurLoop->getName() + ".ivcheck");
auto *NewIVCheck = CIVCheck;
if (InvertedCond) {
NewIVCheck = Builder.CreateNot(CIVCheck);
NewIVCheck->takeName(ValShiftedIsZero);
}
// The original IV, but rebased to be an offset to the CIV.
auto *IVDePHId = Builder.CreateAdd(CIV, Start, "", /*HasNUW=*/false,
/*HasNSW=*/true); // FIXME: what about NUW?
IVDePHId->takeName(IV);
// The loop terminator.
Builder.SetInsertPoint(LoopHeaderBB->getTerminator());
SmallVector<uint32_t> BranchWeights;
const bool HasBranchWeights =
extractBranchWeights(*LoopHeaderBB->getTerminator(), BranchWeights);
auto *BI = Builder.CreateCondBr(CIVCheck, SuccessorBB, LoopHeaderBB);
if (HasBranchWeights) {
if (InvertedCond)
std::swap(BranchWeights[0], BranchWeights[1]);
// We're not changing the loop profile, so we can reuse the original loop's
// profile.
setBranchWeights(*BI, BranchWeights, /*IsExpected=*/false);
}
LoopHeaderBB->getTerminator()->eraseFromParent();
// Populate the IV PHI.
CIV->addIncoming(ConstantInt::get(Ty, 0), LoopPreheaderBB);
CIV->addIncoming(CIVNext, LoopHeaderBB);
// Step 4: Forget the "non-computable" trip-count SCEV associated with the
// loop. The loop would otherwise not be deleted even if it becomes empty.
SE->forgetLoop(CurLoop);
// Step 5: Try to cleanup the loop's body somewhat.
IV->replaceAllUsesWith(IVDePHId);
IV->eraseFromParent();
ValShiftedIsZero->replaceAllUsesWith(NewIVCheck);
ValShiftedIsZero->eraseFromParent();
// Other passes will take care of actually deleting the loop if possible.
LLVM_DEBUG(dbgs() << DEBUG_TYPE " shift-until-zero idiom optimized!\n");
++NumShiftUntilZero;
return MadeChange;
}