llvm/lib/Transforms/Scalar/LoopLoadElimination.cpp

//===- LoopLoadElimination.cpp - Loop Load Elimination Pass ---------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implement a loop-aware load elimination pass.
//
// It uses LoopAccessAnalysis to identify loop-carried dependences with a
// distance of one between stores and loads.  These form the candidates for the
// transformation.  The source value of each store then propagated to the user
// of the corresponding load.  This makes the load dead.
//
// The pass can also version the loop and add memchecks in order to prove that
// may-aliasing stores can't change the value in memory before it's read by the
// load.
//
//===----------------------------------------------------------------------===//

#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/LoopAccessAnalysis.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/ScalarEvolutionExpander.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/Module.h"
#include "llvm/Pass.h"
#include "llvm/Support/Debug.h"
#include "llvm/Transforms/Utils/LoopVersioning.h"
#include <forward_list>

#define LLE_OPTION "loop-load-elim"
#define DEBUG_TYPE LLE_OPTION

using namespace llvm;

static cl::opt<unsigned> CheckPerElim(
    "runtime-check-per-loop-load-elim", cl::Hidden,
    cl::desc("Max number of memchecks allowed per eliminated load on average"),
    cl::init(1));

static cl::opt<unsigned> LoadElimSCEVCheckThreshold(
    "loop-load-elimination-scev-check-threshold", cl::init(8), cl::Hidden,
    cl::desc("The maximum number of SCEV checks allowed for Loop "
             "Load Elimination"));


STATISTIC(NumLoopLoadEliminted, "Number of loads eliminated by LLE");

namespace {

/// \brief Represent a store-to-forwarding candidate.
struct StoreToLoadForwardingCandidate {
  LoadInst *Load;
  StoreInst *Store;

  StoreToLoadForwardingCandidate(LoadInst *Load, StoreInst *Store)
      : Load(Load), Store(Store) {}

  /// \brief Return true if the dependence from the store to the load has a
  /// distance of one.  E.g. A[i+1] = A[i]
  bool isDependenceDistanceOfOne(PredicatedScalarEvolution &PSE) const {
    Value *LoadPtr = Load->getPointerOperand();
    Value *StorePtr = Store->getPointerOperand();
    Type *LoadPtrType = LoadPtr->getType();
    Type *LoadType = LoadPtrType->getPointerElementType();

    assert(LoadPtrType->getPointerAddressSpace() ==
               StorePtr->getType()->getPointerAddressSpace() &&
           LoadType == StorePtr->getType()->getPointerElementType() &&
           "Should be a known dependence");

    auto &DL = Load->getParent()->getModule()->getDataLayout();
    unsigned TypeByteSize = DL.getTypeAllocSize(const_cast<Type *>(LoadType));

    auto *LoadPtrSCEV = cast<SCEVAddRecExpr>(PSE.getSCEV(LoadPtr));
    auto *StorePtrSCEV = cast<SCEVAddRecExpr>(PSE.getSCEV(StorePtr));

    // We don't need to check non-wrapping here because forward/backward
    // dependence wouldn't be valid if these weren't monotonic accesses.
    auto *Dist = cast<SCEVConstant>(
        PSE.getSE()->getMinusSCEV(StorePtrSCEV, LoadPtrSCEV));
    const APInt &Val = Dist->getAPInt();
    return Val.abs() == TypeByteSize;
  }

  Value *getLoadPtr() const { return Load->getPointerOperand(); }

#ifndef NDEBUG
  friend raw_ostream &operator<<(raw_ostream &OS,
                                 const StoreToLoadForwardingCandidate &Cand) {
    OS << *Cand.Store << " -->\n";
    OS.indent(2) << *Cand.Load << "\n";
    return OS;
  }
#endif
};

/// \brief Check if the store dominates all latches, so as long as there is no
/// intervening store this value will be loaded in the next iteration.
bool doesStoreDominatesAllLatches(BasicBlock *StoreBlock, Loop *L,
                                  DominatorTree *DT) {
  SmallVector<BasicBlock *, 8> Latches;
  L->getLoopLatches(Latches);
  return std::all_of(Latches.begin(), Latches.end(),
                     [&](const BasicBlock *Latch) {
                       return DT->dominates(StoreBlock, Latch);
                     });
}

/// \brief The per-loop class that does most of the work.
class LoadEliminationForLoop {
public:
  LoadEliminationForLoop(Loop *L, LoopInfo *LI, const LoopAccessInfo &LAI,
                         DominatorTree *DT)
      : L(L), LI(LI), LAI(LAI), DT(DT), PSE(LAI.PSE) {}

  /// \brief Look through the loop-carried and loop-independent dependences in
  /// this loop and find store->load dependences.
  ///
  /// Note that no candidate is returned if LAA has failed to analyze the loop
  /// (e.g. if it's not bottom-tested, contains volatile memops, etc.)
  std::forward_list<StoreToLoadForwardingCandidate>
  findStoreToLoadDependences(const LoopAccessInfo &LAI) {
    std::forward_list<StoreToLoadForwardingCandidate> Candidates;

    const auto *Deps = LAI.getDepChecker().getDependences();
    if (!Deps)
      return Candidates;

    // Find store->load dependences (consequently true dep).  Both lexically
    // forward and backward dependences qualify.  Disqualify loads that have
    // other unknown dependences.

    SmallSet<Instruction *, 4> LoadsWithUnknownDepedence;

    for (const auto &Dep : *Deps) {
      Instruction *Source = Dep.getSource(LAI);
      Instruction *Destination = Dep.getDestination(LAI);

      if (Dep.Type == MemoryDepChecker::Dependence::Unknown) {
        if (isa<LoadInst>(Source))
          LoadsWithUnknownDepedence.insert(Source);
        if (isa<LoadInst>(Destination))
          LoadsWithUnknownDepedence.insert(Destination);
        continue;
      }

      if (Dep.isBackward())
        // Note that the designations source and destination follow the program
        // order, i.e. source is always first.  (The direction is given by the
        // DepType.)
        std::swap(Source, Destination);
      else
        assert(Dep.isForward() && "Needs to be a forward dependence");

      auto *Store = dyn_cast<StoreInst>(Source);
      if (!Store)
        continue;
      auto *Load = dyn_cast<LoadInst>(Destination);
      if (!Load)
        continue;
      Candidates.emplace_front(Load, Store);
    }

    if (!LoadsWithUnknownDepedence.empty())
      Candidates.remove_if([&](const StoreToLoadForwardingCandidate &C) {
        return LoadsWithUnknownDepedence.count(C.Load);
      });

    return Candidates;
  }

  /// \brief Return the index of the instruction according to program order.
  unsigned getInstrIndex(Instruction *Inst) {
    auto I = InstOrder.find(Inst);
    assert(I != InstOrder.end() && "No index for instruction");
    return I->second;
  }

  /// \brief If a load has multiple candidates associated (i.e. different
  /// stores), it means that it could be forwarding from multiple stores
  /// depending on control flow.  Remove these candidates.
  ///
  /// Here, we rely on LAA to include the relevant loop-independent dependences.
  /// LAA is known to omit these in the very simple case when the read and the
  /// write within an alias set always takes place using the *same* pointer.
  ///
  /// However, we know that this is not the case here, i.e. we can rely on LAA
  /// to provide us with loop-independent dependences for the cases we're
  /// interested.  Consider the case for example where a loop-independent
  /// dependece S1->S2 invalidates the forwarding S3->S2.
  ///
  ///         A[i]   = ...   (S1)
  ///         ...    = A[i]  (S2)
  ///         A[i+1] = ...   (S3)
  ///
  /// LAA will perform dependence analysis here because there are two
  /// *different* pointers involved in the same alias set (&A[i] and &A[i+1]).
  void removeDependencesFromMultipleStores(
      std::forward_list<StoreToLoadForwardingCandidate> &Candidates) {
    // If Store is nullptr it means that we have multiple stores forwarding to
    // this store.
    typedef DenseMap<LoadInst *, const StoreToLoadForwardingCandidate *>
        LoadToSingleCandT;
    LoadToSingleCandT LoadToSingleCand;

    for (const auto &Cand : Candidates) {
      bool NewElt;
      LoadToSingleCandT::iterator Iter;

      std::tie(Iter, NewElt) =
          LoadToSingleCand.insert(std::make_pair(Cand.Load, &Cand));
      if (!NewElt) {
        const StoreToLoadForwardingCandidate *&OtherCand = Iter->second;
        // Already multiple stores forward to this load.
        if (OtherCand == nullptr)
          continue;

        // Handle the very basic of case when the two stores are in the same
        // block so deciding which one forwards is easy.  The later one forwards
        // as long as they both have a dependence distance of one to the load.
        if (Cand.Store->getParent() == OtherCand->Store->getParent() &&
            Cand.isDependenceDistanceOfOne(PSE) &&
            OtherCand->isDependenceDistanceOfOne(PSE)) {
          // They are in the same block, the later one will forward to the load.
          if (getInstrIndex(OtherCand->Store) < getInstrIndex(Cand.Store))
            OtherCand = &Cand;
        } else
          OtherCand = nullptr;
      }
    }

    Candidates.remove_if([&](const StoreToLoadForwardingCandidate &Cand) {
      if (LoadToSingleCand[Cand.Load] != &Cand) {
        DEBUG(dbgs() << "Removing from candidates: \n" << Cand
                     << "  The load may have multiple stores forwarding to "
                     << "it\n");
        return true;
      }
      return false;
    });
  }

  /// \brief Given two pointers operations by their RuntimePointerChecking
  /// indices, return true if they require an alias check.
  ///
  /// We need a check if one is a pointer for a candidate load and the other is
  /// a pointer for a possibly intervening store.
  bool needsChecking(unsigned PtrIdx1, unsigned PtrIdx2,
                     const SmallSet<Value *, 4> &PtrsWrittenOnFwdingPath,
                     const std::set<Value *> &CandLoadPtrs) {
    Value *Ptr1 =
        LAI.getRuntimePointerChecking()->getPointerInfo(PtrIdx1).PointerValue;
    Value *Ptr2 =
        LAI.getRuntimePointerChecking()->getPointerInfo(PtrIdx2).PointerValue;
    return ((PtrsWrittenOnFwdingPath.count(Ptr1) && CandLoadPtrs.count(Ptr2)) ||
            (PtrsWrittenOnFwdingPath.count(Ptr2) && CandLoadPtrs.count(Ptr1)));
  }

  /// \brief Return pointers that are possibly written to on the path from a
  /// forwarding store to a load.
  ///
  /// These pointers need to be alias-checked against the forwarding candidates.
  SmallSet<Value *, 4> findPointersWrittenOnForwardingPath(
      const SmallVectorImpl<StoreToLoadForwardingCandidate> &Candidates) {
    // From FirstStore to LastLoad neither of the elimination candidate loads
    // should overlap with any of the stores.
    //
    // E.g.:
    //
    // st1 C[i]
    // ld1 B[i] <-------,
    // ld0 A[i] <----,  |              * LastLoad
    // ...           |  |
    // st2 E[i]      |  |
    // st3 B[i+1] -- | -'              * FirstStore
    // st0 A[i+1] ---'
    // st4 D[i]
    //
    // st0 forwards to ld0 if the accesses in st4 and st1 don't overlap with
    // ld0.

    LoadInst *LastLoad =
        std::max_element(Candidates.begin(), Candidates.end(),
                         [&](const StoreToLoadForwardingCandidate &A,
                             const StoreToLoadForwardingCandidate &B) {
                           return getInstrIndex(A.Load) < getInstrIndex(B.Load);
                         })
            ->Load;
    StoreInst *FirstStore =
        std::min_element(Candidates.begin(), Candidates.end(),
                         [&](const StoreToLoadForwardingCandidate &A,
                             const StoreToLoadForwardingCandidate &B) {
                           return getInstrIndex(A.Store) <
                                  getInstrIndex(B.Store);
                         })
            ->Store;

    // We're looking for stores after the first forwarding store until the end
    // of the loop, then from the beginning of the loop until the last
    // forwarded-to load.  Collect the pointer for the stores.
    SmallSet<Value *, 4> PtrsWrittenOnFwdingPath;

    auto InsertStorePtr = [&](Instruction *I) {
      if (auto *S = dyn_cast<StoreInst>(I))
        PtrsWrittenOnFwdingPath.insert(S->getPointerOperand());
    };
    const auto &MemInstrs = LAI.getDepChecker().getMemoryInstructions();
    std::for_each(MemInstrs.begin() + getInstrIndex(FirstStore) + 1,
                  MemInstrs.end(), InsertStorePtr);
    std::for_each(MemInstrs.begin(), &MemInstrs[getInstrIndex(LastLoad)],
                  InsertStorePtr);

    return PtrsWrittenOnFwdingPath;
  }

  /// \brief Determine the pointer alias checks to prove that there are no
  /// intervening stores.
  SmallVector<RuntimePointerChecking::PointerCheck, 4> collectMemchecks(
      const SmallVectorImpl<StoreToLoadForwardingCandidate> &Candidates) {

    SmallSet<Value *, 4> PtrsWrittenOnFwdingPath =
        findPointersWrittenOnForwardingPath(Candidates);

    // Collect the pointers of the candidate loads.
    // FIXME: SmallSet does not work with std::inserter.
    std::set<Value *> CandLoadPtrs;
    std::transform(Candidates.begin(), Candidates.end(),
                   std::inserter(CandLoadPtrs, CandLoadPtrs.begin()),
                   std::mem_fn(&StoreToLoadForwardingCandidate::getLoadPtr));

    const auto &AllChecks = LAI.getRuntimePointerChecking()->getChecks();
    SmallVector<RuntimePointerChecking::PointerCheck, 4> Checks;

    std::copy_if(AllChecks.begin(), AllChecks.end(), std::back_inserter(Checks),
                 [&](const RuntimePointerChecking::PointerCheck &Check) {
                   for (auto PtrIdx1 : Check.first->Members)
                     for (auto PtrIdx2 : Check.second->Members)
                       if (needsChecking(PtrIdx1, PtrIdx2,
                                         PtrsWrittenOnFwdingPath, CandLoadPtrs))
                         return true;
                   return false;
                 });

    DEBUG(dbgs() << "\nPointer Checks (count: " << Checks.size() << "):\n");
    DEBUG(LAI.getRuntimePointerChecking()->printChecks(dbgs(), Checks));

    return Checks;
  }

  /// \brief Perform the transformation for a candidate.
  void
  propagateStoredValueToLoadUsers(const StoreToLoadForwardingCandidate &Cand,
                                  SCEVExpander &SEE) {
    //
    // loop:
    //      %x = load %gep_i
    //         = ... %x
    //      store %y, %gep_i_plus_1
    //
    // =>
    //
    // ph:
    //      %x.initial = load %gep_0
    // loop:
    //      %x.storeforward = phi [%x.initial, %ph] [%y, %loop]
    //      %x = load %gep_i            <---- now dead
    //         = ... %x.storeforward
    //      store %y, %gep_i_plus_1

    Value *Ptr = Cand.Load->getPointerOperand();
    auto *PtrSCEV = cast<SCEVAddRecExpr>(PSE.getSCEV(Ptr));
    auto *PH = L->getLoopPreheader();
    Value *InitialPtr = SEE.expandCodeFor(PtrSCEV->getStart(), Ptr->getType(),
                                          PH->getTerminator());
    Value *Initial =
        new LoadInst(InitialPtr, "load_initial", PH->getTerminator());
    PHINode *PHI = PHINode::Create(Initial->getType(), 2, "store_forwarded",
                                   &L->getHeader()->front());
    PHI->addIncoming(Initial, PH);
    PHI->addIncoming(Cand.Store->getOperand(0), L->getLoopLatch());

    Cand.Load->replaceAllUsesWith(PHI);
  }

  /// \brief Top-level driver for each loop: find store->load forwarding
  /// candidates, add run-time checks and perform transformation.
  bool processLoop() {
    DEBUG(dbgs() << "\nIn \"" << L->getHeader()->getParent()->getName()
                 << "\" checking " << *L << "\n");
    // Look for store-to-load forwarding cases across the
    // backedge. E.g.:
    //
    // loop:
    //      %x = load %gep_i
    //         = ... %x
    //      store %y, %gep_i_plus_1
    //
    // =>
    //
    // ph:
    //      %x.initial = load %gep_0
    // loop:
    //      %x.storeforward = phi [%x.initial, %ph] [%y, %loop]
    //      %x = load %gep_i            <---- now dead
    //         = ... %x.storeforward
    //      store %y, %gep_i_plus_1

    // First start with store->load dependences.
    auto StoreToLoadDependences = findStoreToLoadDependences(LAI);
    if (StoreToLoadDependences.empty())
      return false;

    // Generate an index for each load and store according to the original
    // program order.  This will be used later.
    InstOrder = LAI.getDepChecker().generateInstructionOrderMap();

    // To keep things simple for now, remove those where the load is potentially
    // fed by multiple stores.
    removeDependencesFromMultipleStores(StoreToLoadDependences);
    if (StoreToLoadDependences.empty())
      return false;

    // Filter the candidates further.
    SmallVector<StoreToLoadForwardingCandidate, 4> Candidates;
    unsigned NumForwarding = 0;
    for (const StoreToLoadForwardingCandidate Cand : StoreToLoadDependences) {
      DEBUG(dbgs() << "Candidate " << Cand);
      // Only progagate value if they are of the same type.
      if (Cand.Store->getPointerOperand()->getType() !=
          Cand.Load->getPointerOperand()->getType())
        continue;

      // Make sure that the stored values is available everywhere in the loop in
      // the next iteration.
      if (!doesStoreDominatesAllLatches(Cand.Store->getParent(), L, DT))
        continue;

      // Check whether the SCEV difference is the same as the induction step,
      // thus we load the value in the next iteration.
      if (!Cand.isDependenceDistanceOfOne(PSE))
        continue;

      ++NumForwarding;
      DEBUG(dbgs()
            << NumForwarding
            << ". Valid store-to-load forwarding across the loop backedge\n");
      Candidates.push_back(Cand);
    }
    if (Candidates.empty())
      return false;

    // Check intervening may-alias stores.  These need runtime checks for alias
    // disambiguation.
    SmallVector<RuntimePointerChecking::PointerCheck, 4> Checks =
        collectMemchecks(Candidates);

    // Too many checks are likely to outweigh the benefits of forwarding.
    if (Checks.size() > Candidates.size() * CheckPerElim) {
      DEBUG(dbgs() << "Too many run-time checks needed.\n");
      return false;
    }

    if (LAI.PSE.getUnionPredicate().getComplexity() >
        LoadElimSCEVCheckThreshold) {
      DEBUG(dbgs() << "Too many SCEV run-time checks needed.\n");
      return false;
    }

    if (!Checks.empty() || !LAI.PSE.getUnionPredicate().isAlwaysTrue()) {
      if (L->getHeader()->getParent()->optForSize()) {
        DEBUG(dbgs() << "Versioning is needed but not allowed when optimizing "
                        "for size.\n");
        return false;
      }

      // Point of no-return, start the transformation.  First, version the loop
      // if necessary.

      LoopVersioning LV(LAI, L, LI, DT, PSE.getSE(), false);
      LV.setAliasChecks(std::move(Checks));
      LV.setSCEVChecks(LAI.PSE.getUnionPredicate());
      LV.versionLoop();
    }

    // Next, propagate the value stored by the store to the users of the load.
    // Also for the first iteration, generate the initial value of the load.
    SCEVExpander SEE(*PSE.getSE(), L->getHeader()->getModule()->getDataLayout(),
                     "storeforward");
    for (const auto &Cand : Candidates)
      propagateStoredValueToLoadUsers(Cand, SEE);
    NumLoopLoadEliminted += NumForwarding;

    return true;
  }

private:
  Loop *L;

  /// \brief Maps the load/store instructions to their index according to
  /// program order.
  DenseMap<Instruction *, unsigned> InstOrder;

  // Analyses used.
  LoopInfo *LI;
  const LoopAccessInfo &LAI;
  DominatorTree *DT;
  PredicatedScalarEvolution PSE;
};

/// \brief The pass.  Most of the work is delegated to the per-loop
/// LoadEliminationForLoop class.
class LoopLoadElimination : public FunctionPass {
public:
  LoopLoadElimination() : FunctionPass(ID) {
    initializeLoopLoadEliminationPass(*PassRegistry::getPassRegistry());
  }

  bool runOnFunction(Function &F) override {
    auto *LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
    auto *LAA = &getAnalysis<LoopAccessAnalysis>();
    auto *DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();

    // Build up a worklist of inner-loops to vectorize. This is necessary as the
    // act of distributing a loop creates new loops and can invalidate iterators
    // across the loops.
    SmallVector<Loop *, 8> Worklist;

    for (Loop *TopLevelLoop : *LI)
      for (Loop *L : depth_first(TopLevelLoop))
        // We only handle inner-most loops.
        if (L->empty())
          Worklist.push_back(L);

    // Now walk the identified inner loops.
    bool Changed = false;
    for (Loop *L : Worklist) {
      const LoopAccessInfo &LAI = LAA->getInfo(L, ValueToValueMap());
      // The actual work is performed by LoadEliminationForLoop.
      LoadEliminationForLoop LEL(L, LI, LAI, DT);
      Changed |= LEL.processLoop();
    }

    // Process each loop nest in the function.
    return Changed;
  }

  void getAnalysisUsage(AnalysisUsage &AU) const override {
    AU.addRequired<LoopInfoWrapperPass>();
    AU.addPreserved<LoopInfoWrapperPass>();
    AU.addRequired<LoopAccessAnalysis>();
    AU.addRequired<ScalarEvolutionWrapperPass>();
    AU.addRequired<DominatorTreeWrapperPass>();
    AU.addPreserved<DominatorTreeWrapperPass>();
  }

  static char ID;
};
}

char LoopLoadElimination::ID;
static const char LLE_name[] = "Loop Load Elimination";

INITIALIZE_PASS_BEGIN(LoopLoadElimination, LLE_OPTION, LLE_name, false, false)
INITIALIZE_PASS_DEPENDENCY(LoopInfoWrapperPass)
INITIALIZE_PASS_DEPENDENCY(LoopAccessAnalysis)
INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
INITIALIZE_PASS_DEPENDENCY(ScalarEvolutionWrapperPass)
INITIALIZE_PASS_END(LoopLoadElimination, LLE_OPTION, LLE_name, false, false)

namespace llvm {
FunctionPass *createLoopLoadEliminationPass() {
  return new LoopLoadElimination();
}
}