[llvm] AMDGPU: Add NextUseAnalysis Pass (PR #178873)

Matt Arsenault via llvm-commits llvm-commits at lists.llvm.org
Fri Apr 10 10:15:07 PDT 2026


================
@@ -0,0 +1,2614 @@
+//===---------------------- AMDGPUNextUseAnalysis.cpp ---------------------===//
+//
+// 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 file implements the AMDGPUNextUseAnalysis pass, a machine-level analysis
+// that computes the distance from each instruction to the "nearest" next use of
+// every live virtual register. These distances guide register spilling
+// decisions by identifying which live values are furthest from their next use
+// and are therefore the best candidates to spill.
+//
+// The analysis is based on the Braun & Hack CC'09 paper "Register Spilling and
+// Live-Range Splitting for SSA-Form Programs."
+//
+// Key concepts:
+//
+//   NextUseDistance     A loop-depth-weighted instruction count representing
+//                       how far away a register's next use is. Distances
+//                       through deeper loops are scaled by fromLoopDepth() so
+//                       that uses inside hot loops appear closer.
+//
+//   Inter-block         Pre-computed shortest weighted distances between all
+//   distances           pairs of basic blocks, used to efficiently answer
+//                       cross-block next-use queries. Each intermediate block
+//                       is weighted by fromLoopDepth() applied once per loop
+//                       boundary crossing relative to the destination.
+//
+// Configuration flags (see Config struct in the header):
+//
+//   CountPhis           Count PHI instructions toward distance and block size.
+//   ForwardOnly         Restrict inter-block distances to forward-reachable
+//                       paths.
+//   PreciseUseModeling  Model PHI uses at their incoming edge block and filter
+//                       uses with intermediate redefinitions.
+//   PromoteToPreheader  Route loop-entry and inner-loop uses to the preheader.
+//
+// This file contains:
+//
+//   - Command-line options for configuration and debug output
+//   - LiveRegUse / JSON helpers
+//   - AMDGPUNextUseAnalysisImpl (the main analysis implementation)
+//       - Instruction ID assignment and block size computation
+//       - CFG path pre-computation (reachability, loop depth, back-edges)
+//       - Inter-block distance computation
+//       - Per-register next-use distance queries and caching
+//   - AMDGPUNextUseAnalysis (public facade, pimpl)
+//   - Legacy and new pass manager wrappers
+//
+//===----------------------------------------------------------------------===//
+
+#include "AMDGPUNextUseAnalysis.h"
+#include "AMDGPU.h"
+#include "GCNRegPressure.h"
+#include "GCNSubtarget.h"
+
+#include "llvm/ADT/DenseMap.h"
+#include "llvm/ADT/PostOrderIterator.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/CodeGen/MachineBasicBlock.h"
+#include "llvm/CodeGen/MachineFunction.h"
+#include "llvm/CodeGen/MachineInstr.h"
+#include "llvm/CodeGen/MachineLoopInfo.h"
+#include "llvm/IR/ModuleSlotTracker.h"
+#include "llvm/InitializePasses.h"
+#include "llvm/Support/FileSystem.h"
+#include "llvm/Support/JSON.h"
+#include "llvm/Support/Timer.h"
+#include "llvm/Support/ToolOutputFile.h"
+#include "llvm/Support/raw_ostream.h"
+
+#include <algorithm>
+#include <limits>
+#include <string>
+
+using namespace llvm;
+
+#define DEBUG_TYPE "amdgpu-next-use-analysis"
+
+//==============================================================================
+// Options etc
+//==============================================================================
+namespace {
+
+cl::opt<bool> DumpNextUseDistance("amdgpu-next-use-analysis-dump-distance",
+                                  cl::init(false), cl::Hidden);
+
+cl::opt<bool>
+    DistanceCacheEnabled("amdgpu-next-use-analysis-distance-cache",
+                         cl::init(true), cl::Hidden,
+                         cl::desc("Enable live-reg-use distance cache"));
+
+cl::opt<std::string>
+    DumpNextUseDistanceAsJson("amdgpu-next-use-analysis-dump-distance-as-json",
+                              cl::Hidden);
+cl::opt<bool>
+    DumpNextUseDistanceVerbose("amdgpu-next-use-analysis-dump-distance-verbose",
+                               cl::init(false), cl::Hidden);
+
+// 'graphics' and 'compute' modes arose due to initial competing implementations
+// of next-use analysis that emphasized different types of workloads. This
+// implementation is a compromise that combines aspects of both. Over time, the
+// hope is we will be able to remove some of these differences and settle on a
+// more unified implementation.
+cl::opt<std::string>
+    ConfigPresetOpt("amdgpu-next-use-analysis-config", cl::Hidden,
+                    cl::init("graphics"),
+                    cl::desc("Config preset: 'graphics' or 'compute'"));
+
+cl::opt<bool> ConfigCountPhisOpt(
+    "amdgpu-next-use-analysis-count-phis", cl::Hidden,
+    cl::desc("Count PHI instructions toward distance and block size"));
+cl::opt<bool> ConfigForwardOnlyOpt(
+    "amdgpu-next-use-analysis-forward-only", cl::Hidden,
+    cl::desc("Restrict inter-block distances to forward-reachable paths"));
+cl::opt<bool> ConfigPreciseUseModelingOpt(
+    "amdgpu-next-use-analysis-precise-use-modeling", cl::Hidden,
+    cl::desc("Model PHI uses via incoming edge block with loop-aware "
+             "reachability filtering"));
+cl::opt<bool> ConfigPromoteToPreheaderOpt(
+    "amdgpu-next-use-analysis-use-preheader-model", cl::Hidden,
+    cl::desc("Promote loop-entry and inner-loop uses to the loop preheader"));
+} // namespace
+
+//==============================================================================
+// LiveRegUse - Represents a live register use with its distance. Used for
+// tracking and sorting register uses by distance.
+//==============================================================================
+namespace {
+using UseDistancePair = AMDGPUNextUseAnalysis::UseDistancePair;
+struct LiveRegUse : public UseDistancePair {
+  // 'nullptr' indicates an unset/invalid state.
+  LiveRegUse() : UseDistancePair(nullptr, 0) {}
+  LiveRegUse(const MachineOperand *Use, NextUseDistance Dist)
+      : UseDistancePair(Use, Dist) {}
+  LiveRegUse(const UseDistancePair &P) : UseDistancePair(P) {}
+
+  bool isUnset() const { return Use == nullptr; }
+
+  Register getReg() const { return Use->getReg(); }
+  unsigned getSubReg() const { return Use->getSubReg(); }
+  LaneBitmask getLaneMask(const SIRegisterInfo *TRI) const {
+    return TRI->getSubRegIndexLaneMask(Use->getSubReg());
+  }
+
+  bool isCloserThan(const LiveRegUse &X) const {
+    if (Dist < X.Dist)
+      return true;
+
+    if (Dist > X.Dist)
+      return false;
+
+    if (Use == X.Use)
+      return false;
+
+    // Ugh. When !CountPhis, PHIs and the first non-PHI instruction have id
+    // 0. In this case, consider PHIs as less than the first non-PHI
+    // instruction.
+    const MachineInstr *ThisMI = Use->getParent();
+    const MachineInstr *XMI = X.Use->getParent();
+    const MachineBasicBlock *ThisMBB = ThisMI->getParent();
+    if (ThisMBB == XMI->getParent()) {
+      if (ThisMI->isPHI() && !XMI->isPHI() &&
+          XMI == &(*ThisMBB->getFirstNonPHI()))
+        return true;
+    }
+
+    // Ensure deterministic results
+    return X.getReg() < getReg();
+  }
+
+  void print(raw_ostream &OS) const {
+    if (isUnset()) {
+      OS << "<unset>";
+      return;
+    }
+    Dist.print(OS);
+    OS << " [" << printReg(getReg());
+    if (getSubReg())
+      OS << ":" << getSubReg();
+    OS << "]";
+  }
+
+  LLVM_DUMP_METHOD void dump() const {
+    print(dbgs());
+    dbgs() << '\n';
+  }
+};
+
+inline bool updateClosest(LiveRegUse &Closest, const LiveRegUse &X) {
+  if (!Closest.Use || X.isCloserThan(Closest)) {
+    Closest = X;
+    return true;
+  }
+  return false;
+}
+
+inline bool updateFurthest(LiveRegUse &Furthest, const LiveRegUse &X) {
+  if (!Furthest.Use || Furthest.isCloserThan(X)) {
+    Furthest = X;
+    return true;
+  }
+  return false;
+}
+} // namespace
+
+//==============================================================================
+// JSON helpers
+//==============================================================================
+namespace {
+template <typename Lambda>
+void printStringAttr(json::OStream &J, const char *Name, Lambda L) {
+  J.attributeBegin(Name);
+  raw_ostream &OS = J.rawValueBegin();
+  OS << '"';
+  L(OS);
+  OS << '"';
+  J.rawValueEnd();
+  J.attributeEnd();
+}
+void printStringAttr(json::OStream &J, const char *Name, Printable P) {
+  printStringAttr(J, Name, [&](raw_ostream &OS) { OS << P; });
+}
+
+void printStringAttr(json::OStream &J, const char *Name, const MachineInstr &MI,
+                     ModuleSlotTracker &MST) {
+  printStringAttr(J, Name, [&](raw_ostream &OS) {
+    MI.print(OS, MST,
+             /* IsStandalone    */ false,
+             /* SkipOpers       */ false,
+             /* SkipDebugLoc    */ false,
+             /* AddNewLine ---> */ false,
+             /* TargetInstrInfo */ nullptr);
+  });
+}
+
+void printMBBNameAttr(json::OStream &J, const char *Name,
+                      const MachineBasicBlock &MBB, ModuleSlotTracker &MST) {
+  printStringAttr(J, Name, [&](raw_ostream &OS) {
+    MBB.printName(OS, MachineBasicBlock::PrintNameIr, &MST);
+  });
+}
+
+template <typename NameLambda, typename ValueT>
+void printAttr(json::OStream &J, NameLambda NL, ValueT V) {
+  std::string Name;
+  raw_string_ostream NameOS(Name);
+  NL(NameOS);
+  J.attribute(NameOS.str(), V);
+}
+
+template <typename ValueT>
+void printAttr(json::OStream &J, const Printable &P, ValueT V) {
+  printAttr(J, [&](raw_ostream &OS) { OS << P; }, V);
+}
+
+} // namespace
+
+//==============================================================================
+// AMDGPUNextUseAnalysisImpl
+//==============================================================================
+class llvm::AMDGPUNextUseAnalysisImpl {
+public:
+  struct CacheableNextUseDistance {
+    bool IsInstrRelative;
+    NextUseDistance Distance;
+  };
+  static constexpr bool InstrRelative = true;
+  static constexpr bool InstrInvariant = false;
+
+private:
+  const MachineFunction *MF = nullptr;
+  const SIRegisterInfo *TRI = nullptr;
+  const SIInstrInfo *TII = nullptr;
+  const MachineLoopInfo *MLI = nullptr;
+  const MachineRegisterInfo *MRI = nullptr;
+
+  using InstrIdTy = unsigned;
+  using InstrToIdMap = DenseMap<const MachineInstr *, InstrIdTy>;
+  InstrToIdMap InstrToId;
+  AMDGPUNextUseAnalysis::Config Cfg;
+
+  void initializeTables() {
+    for (const MachineBasicBlock &BB : *MF)
+      calcInstrIds(&BB, InstrToId);
+    initializeCfgPaths();
+    initializeInterBlockDistances();
+  }
+
+  void clearTables() {
+    InstrToId.clear();
+    RegUseMap.clear();
+    Paths.clear();
+
+    resetDistanceCache();
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // Instruction Ids
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  unsigned sizeOf(const MachineInstr &MI) const {
+    // When !Cfg.CountPhis, PHIs do not contribute to distances/sizes since they
+    // generally don't result in the generation of a machine instruction.
+    // FIXME: Consider using MI.isPseudo()
+    return Cfg.CountPhis ? 1 : !MI.isPHI();
+  }
+
+  void calcInstrIds(const MachineBasicBlock *BB,
+                    InstrToIdMap &MutableInstrToId) const {
+    InstrIdTy Id = 0;
+    for (auto &MI : BB->instrs()) {
+      MutableInstrToId[&MI] = Id;
+      Id += sizeOf(MI);
+    }
+  }
+
+  /// Returns MI's instruction Id. It renumbers (part of) the BB if MI is not
+  /// found in the map.
+  InstrIdTy getInstrId(const MachineInstr *MI) const {
+    auto It = InstrToId.find(MI);
+    if (It != InstrToId.end())
+      return It->second;
+
+    // Renumber the MBB.
+    // TODO: Renumber from MI onwards.
+    auto &MutableInstrToId = const_cast<InstrToIdMap &>(InstrToId);
+    calcInstrIds(MI->getParent(), MutableInstrToId);
+    return InstrToId.find(MI)->second;
+  }
+
+  // Length of the segment from MI (inclusive) to the first instruction of the
+  // basic block.
+  InstrIdTy getHeadLen(const MachineInstr *MI) const {
+    const MachineBasicBlock *MBB = MI->getParent();
+    return getInstrId(MI) + getInstrId(&MBB->instr_front()) + 1;
+  }
+
+  // Length of the segment from MI (exclusive) to the last instruction of the
+  // basic block.
+  InstrIdTy getTailLen(const MachineInstr *MI) const {
+    const MachineBasicBlock *MBB = MI->getParent();
+    return getInstrId(&MBB->instr_back()) - getInstrId(MI);
+  }
+
+  // Length of the segment from 'From' to 'To' (exclusive). Both instructions
+  // must be in the same basic block.
+  InstrIdTy getDistance(const MachineInstr *From,
+                        const MachineInstr *To) const {
+    assert(From->getParent() == To->getParent());
+    return getInstrId(To) - getInstrId(From);
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // RegUses - cache of uses by register
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  DenseMap<Register, SmallVector<const MachineOperand *>> RegUseMap;
+
+  const SmallVector<const MachineOperand *> &
+  getRegisterUses(Register Reg) const {
+    auto I = RegUseMap.find(Reg);
+    if (I != RegUseMap.end())
+      return I->second;
+
+    auto *NonConstThis = const_cast<AMDGPUNextUseAnalysisImpl *>(this);
+    SmallVector<const MachineOperand *> &Uses = NonConstThis->RegUseMap[Reg];
+    for (const MachineOperand &UseMO : MRI->use_nodbg_operands(Reg)) {
+      if (!UseMO.isUndef())
+        Uses.push_back(&UseMO);
+    }
+    return Uses;
+  }
+
+  bool hasAtLeastOneUse(Register Reg) const {
+    return !getRegisterUses(Reg).empty();
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // Paths
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  class Path
+      : public std::pair<const MachineBasicBlock *, const MachineBasicBlock *> {
+  public:
+    using Base =
+        std::pair<const MachineBasicBlock *, const MachineBasicBlock *>;
+    using Base::pair;
+    Path(const Base &Pair) : Base(Pair) {}
+
+    const MachineBasicBlock *src() const { return first; }
+    const MachineBasicBlock *dst() const { return second; }
+
+    using DenseMapInfo = llvm::DenseMapInfo<Base>;
+  };
+
+  enum EdgeKind { Back = -1, None = 0, Forward = 1 };
+  struct PathInfo {
+    EdgeKind EK;
+    bool Reachable;
+    int ForwardReachable;
+    unsigned RelativeLoopDepth;
+    std::optional<NextUseDistance> ShortestDistance;
+    std::optional<NextUseDistance> ShortestUnweightedDistance;
+    InstrIdTy Size;
+
+    PathInfo()
+        : EK(None), Reachable(false), ForwardReachable(-1),
+          RelativeLoopDepth(0), Size(0) {}
+
+    bool isBackedge() const { return EK == EdgeKind::Back; }
+
+    bool isForwardReachableSet() const { return 0 <= ForwardReachable; }
+    bool isForwardReachableUnset() const { return ForwardReachable < 0; }
+    bool isForwardReachable() const { return ForwardReachable == 1; }
+    bool isNotForwardReachable() const { return ForwardReachable == 0; }
+
+    void print(raw_ostream &OS) const {
+      const char *EKStr = EK == Back ? "back" : EK == Forward ? "fwd" : "none";
+      OS << "{ek=" << EKStr << " reach=" << Reachable
+         << " fwd-reach=" << ForwardReachable
+         << " loop-depth=" << RelativeLoopDepth << " size=" << Size;
+      if (ShortestDistance) {
+        OS << " shdist=";
+        ShortestDistance->print(OS);
+      }
+      if (ShortestUnweightedDistance) {
+        OS << " shudist=";
+        ShortestUnweightedDistance->print(OS);
+      }
+      OS << "}";
+    }
+
+    LLVM_DUMP_METHOD void dump() const {
+      print(dbgs());
+      dbgs() << '\n';
+    }
+  };
+
+  //----------------------------------------------------------------------------
+  // Path Storage - 'Paths' is lazily populated and some members are lazily
+  // computed. All mutations should go through one of the 'initializePathInfo*'
+  // flavors below.
+  //----------------------------------------------------------------------------
+  DenseMap<Path, PathInfo, Path::DenseMapInfo> Paths;
+
+  const PathInfo *maybePathInfoFor(const MachineBasicBlock *From,
+                                   const MachineBasicBlock *To) const {
+    auto I = Paths.find({From, To});
+    return I == Paths.end() ? nullptr : &I->second;
+  }
+
+  PathInfo &getOrInitPathInfo(const MachineBasicBlock *From,
+                              const MachineBasicBlock *To) const {
+    auto *NonConstThis = const_cast<AMDGPUNextUseAnalysisImpl *>(this);
+    auto &MutablePaths = NonConstThis->Paths;
+
+    Path P(From, To);
+    auto [I, Inserted] = MutablePaths.try_emplace(P);
+    if (!Inserted)
+      return I->second;
+
+    bool Reachable = calcIsReachable(P.src(), P.dst());
+
+    // Iterator may have been invalidated by calcIsReachable, so get a fresh
+    // reference to the slot.
+    return NonConstThis->initializePathInfo(MutablePaths.at(P), P,
+                                            EdgeKind::None, Reachable);
+  }
+
+  const PathInfo &pathInfoFor(const MachineBasicBlock *From,
+                              const MachineBasicBlock *To) const {
+    return getOrInitPathInfo(From, To);
+  }
+
+  //----------------------------------------------------------------------------
+  // initializePathInfo* - various flavors of PathInfo initialization. They
+  // (should) always funnel to the first flavor below.
+  //----------------------------------------------------------------------------
+  PathInfo &initializePathInfo(PathInfo &Slot, Path P, EdgeKind EK,
+                               bool Reachable) {
+    Slot.EK = EK;
+    Slot.Reachable = Reachable;
+    Slot.ForwardReachable = EK != EdgeKind::None ? (0 < EK) : -1;
+    Slot.RelativeLoopDepth =
+        Slot.Reachable ? calcRelativeLoopDepth(P.src(), P.dst()) : 0;
+    Slot.Size = P.src() == P.dst() ? calcSize(P.src()) : 0;
+    if (EK != EdgeKind::None)
+      Slot.ShortestUnweightedDistance = 0;
+    return Slot;
+  }
+
+  PathInfo &initializePathInfo(Path P, EdgeKind EK, bool Reachable) const {
+    auto *NonConstThis = const_cast<AMDGPUNextUseAnalysisImpl *>(this);
+    auto &MutablePaths = NonConstThis->Paths;
+    return NonConstThis->initializePathInfo(MutablePaths[P], P, EK, Reachable);
+  }
+
+  std::pair<PathInfo *, bool> maybeInitializePathInfo(Path P, EdgeKind EK,
+                                                      bool Reachable) const {
+    auto *NonConstThis = const_cast<AMDGPUNextUseAnalysisImpl *>(this);
+    auto &MutablePaths = NonConstThis->Paths;
+    auto [I, Inserted] = MutablePaths.try_emplace(P);
+    if (Inserted)
+      NonConstThis->initializePathInfo(I->second, P, EK, Reachable);
+    return {&I->second, Inserted};
+  }
+
+  bool initializePathInfoForwardReachable(const MachineBasicBlock *From,
+                                          const MachineBasicBlock *To,
+                                          bool Value) const {
+    PathInfo &Slot = getOrInitPathInfo(From, To);
+    assert(Slot.isForwardReachableUnset());
+    Slot.ForwardReachable = Value;
+    return Value;
+  }
+
+  NextUseDistance
+  initializePathInfoShortestDistance(const MachineBasicBlock *From,
+                                     const MachineBasicBlock *To,
+                                     NextUseDistance Value) const {
+    PathInfo &Slot = getOrInitPathInfo(From, To);
+    assert(!Slot.ShortestDistance.has_value());
+    Slot.ShortestDistance = Value;
+    return Value;
+  }
+
+  NextUseDistance
+  initializePathInfoShortestUnweightedDistance(const MachineBasicBlock *From,
+                                               const MachineBasicBlock *To,
+                                               NextUseDistance Value) const {
+    PathInfo &Slot = getOrInitPathInfo(From, To);
+    assert(!Slot.ShortestUnweightedDistance.has_value());
+    Slot.ShortestUnweightedDistance = Value;
+    return Value;
+  }
+
+  //----------------------------------------------------------------------------
+  // initialize*Paths
+  //----------------------------------------------------------------------------
+private:
+  void initializePaths(const SmallVector<Path> &ReachablePaths,
+                       const SmallVector<Path> &UnreachablePaths) const {
+    for (bool R : {true, false}) {
+      const auto &ToInit = R ? ReachablePaths : UnreachablePaths;
+      for (const Path &P : ToInit)
+        initializePathInfo(P, EdgeKind::None, R);
+    }
+  }
+
+  void
+  initializeForwardOnlyPaths(const SmallVector<Path> &ReachablePaths,
+                             const SmallVector<Path> &UnreachablePaths) const {
+    for (bool R : {true, false}) {
+      const auto &ToInit = R ? ReachablePaths : UnreachablePaths;
+      for (const Path &P : ToInit) {
+        PathInfo &Slot = getOrInitPathInfo(P.src(), P.dst());
+        assert(Slot.isForwardReachableUnset() || Slot.ForwardReachable == R);
+        Slot.ForwardReachable = R;
+      }
+    }
+  }
+
+  // Follow the control flow graph starting at the entry block until all blocks
+  // have been visited. Along the way, initialize the PathInfo for each edge
+  // traversed.
+  void initializeCfgPaths() {
+    Paths.clear();
+
+    enum VisitState { Undiscovered, Visiting, Finished };
+    DenseMap<const MachineBasicBlock *, VisitState> State;
+
+    SmallVector<const MachineBasicBlock *> Work{&MF->front()};
+    State[&MF->front()] = Undiscovered;
+
+    while (!Work.empty()) {
+      const MachineBasicBlock *Src = Work.back();
+      VisitState &SrcState = State[Src];
+
+      // A block may already be 'Finished' if it is reachable from multiple
+      // predecessors causing it to be pushed more than once while still
+      // 'Undiscovered'.
+      if (SrcState == Visiting || SrcState == Finished) {
+        Work.pop_back();
+        SrcState = Finished;
+        continue;
+      }
+
+      SrcState = Visiting;
+      for (const MachineBasicBlock *Dst : Src->successors()) {
+        const VisitState DstState = State.lookup(Dst);
+
+        EdgeKind EK;
+        if (DstState == Undiscovered) {
+          EK = EdgeKind::Forward;
+          Work.push_back(Dst);
+        } else if (DstState == Visiting) {
+          EK = EdgeKind::Back;
+        } else {
+          EK = EdgeKind::Forward;
+        }
+
+        Path P(Src, Dst);
+        assert(!Paths.contains(P));
+        initializePathInfo(P, EK, /*Reachable*/ true);
+      }
+    }
+
+    LLVM_DEBUG(dumpPaths());
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // Loop helpers
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  static bool isStandAloneLoop(const MachineLoop *Loop) {
+    return Loop->getSubLoops().empty() && Loop->isOutermost();
+  }
+
+  static MachineLoop *findChildLoop(MachineLoop *const Parent,
+                                    MachineLoop *Descendant) {
+    for (MachineLoop *L = Descendant; L != Parent; L = L->getParentLoop()) {
+      if (L->getParentLoop() == Parent)
+        return L;
+    }
+    return nullptr;
+  }
+
+  // If loops 'A' and 'B' share a common parent loop, return that loop and the
+  // depth of 'A' relative to it. Otherwise return nullptr and the loop depth of
+  // 'A'.
+  static std::pair<MachineLoop *, unsigned>
+  findCommonParent(MachineLoop *A, const MachineLoop *const B) {
+    unsigned Depth = 0;
+    for (; A != nullptr; A = A->getParentLoop(), ++Depth) {
+      if (A->contains(B))
+        break;
+    }
+    return {A, Depth};
+  }
+
+  static const MachineBasicBlock *
+  getOutermostPreheader(const MachineLoop *Loop) {
+    return Loop ? Loop->getOutermostLoop()->getLoopPreheader() : nullptr;
+  }
+
+  static MachineBasicBlock *findChildPreheader(MachineLoop *const Parent,
+                                               MachineLoop *Descendant) {
+    MachineLoop *ChildLoop = findChildLoop(Parent, Descendant);
+    return ChildLoop ? ChildLoop->getLoopPreheader() : nullptr;
+  }
+
+  static const MachineBasicBlock *mbbForPhiOp(const MachineInstr *MI,
+                                              const MachineOperand *MO) {
+    return MI->isPHI() ? MI->getOperand(MO->getOperandNo() + 1).getMBB()
+                       : nullptr;
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // Calculate features
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  InstrIdTy calcSize(const MachineBasicBlock *BB) const {
+    InstrIdTy Size = BB->size();
+    if (!Cfg.CountPhis)
+      Size -= std::distance(BB->begin(), BB->getFirstNonPHI());
+    return Size;
+  }
+
+  NextUseDistance calcWeightedSize(const MachineBasicBlock *From,
+                                   const MachineBasicBlock *To) const {
+    return NextUseDistance::fromSize(getSize(From),
+                                     getRelativeLoopDepth(From, To));
+  }
+
+  // Return the loop depth of 'From' relative to 'To'.
+  unsigned calcRelativeLoopDepth(const MachineBasicBlock *From,
+                                 const MachineBasicBlock *To) const {
+    MachineLoop *LoopFrom = MLI->getLoopFor(From);
+    MachineLoop *LoopTo = MLI->getLoopFor(To);
+
+    if (!LoopFrom)
+      return 0;
+
+    if (!LoopTo)
+      return LoopFrom->getLoopDepth();
+
+    if (LoopFrom->contains(LoopTo)) // covers LoopFrom == LoopTo
+      return 0;
+
+    if (LoopTo->contains(LoopFrom))
+      return LoopFrom->getLoopDepth() - LoopTo->getLoopDepth();
+
+    // Loops are siblings of some sort.
+    return findCommonParent(LoopFrom, LoopTo).second;
+  }
+
+  // Attempt to find a path from 'From' to 'To' using a depth first search. If
+  // 'ForwardOnly' is true, do not follow backedges. As a performance
+  // improvement, this may initialize reachable intermediate paths or paths we
+  // determine are unreachable.
+  bool calcIsReachable(const MachineBasicBlock *From,
+                       const MachineBasicBlock *To,
+                       bool ForwardOnly = false) const {
+    if (From == To && !MLI->getLoopFor(From))
+      return false;
+
+    if (!ForwardOnly && interBlockDistanceExists(From, To))
+      return true;
+
+    enum { VisitOp, PopOp };
+    using MBBOpPair = std::pair<const MachineBasicBlock *, int>;
+    SmallVector<MBBOpPair> Work{{From, VisitOp}};
+    DenseSet<const MachineBasicBlock *> Visited{From};
+
+    SmallVector<Path> IntermediatePath;
+    SmallVector<Path> Unreachable;
+
+    // Should be run at every function exit point.
+    auto Finally = [&](bool Reachable) {
+      // This is an optimization. For intermediate paths we found while
+      // calculating reachability for 'From' --> 'To', remember their
+      // reachability.
+      if (!Reachable) {
+        IntermediatePath.clear();
+        for (const MachineBasicBlock *MBB : Visited) {
+          if (MBB != From)
+            Unreachable.emplace_back(MBB, To);
+        }
+      }
+
+      if (ForwardOnly)
+        initializeForwardOnlyPaths(IntermediatePath, Unreachable);
+      else
+        initializePaths(IntermediatePath, Unreachable);
+
+      return Reachable;
+    };
+
+    while (!Work.empty()) {
+      auto [Current, Op] = Work.pop_back_val();
+
+      // Backtracking
+      if (Op == PopOp) {
+        IntermediatePath.pop_back();
+        if (ForwardOnly)
+          Unreachable.emplace_back(Current, To);
+        continue;
+      }
+
+      if (Current->succ_empty())
+        continue;
+
+      if (Current != From) {
+        IntermediatePath.emplace_back(Current, To);
+        Work.emplace_back(Current, PopOp);
+      }
+
+      for (const MachineBasicBlock *Succ : Current->successors()) {
+        if (ForwardOnly && isBackedge(Current, Succ))
+          continue;
+
+        if (Succ == To)
+          return Finally(true);
+
+        if (auto CachedReachable = isMaybeReachable(Succ, To, ForwardOnly)) {
+          if (CachedReachable.value())
+            return Finally(true);
+          Visited.insert(Succ);
+          continue;
+        }
+
+        if (Visited.insert(Succ).second)
+          Work.emplace_back(Succ, VisitOp);
+      }
+    }
+
+    return Finally(false);
+  }
+
+  //----------------------------------------------------------------------------
+  // Inter-block distance - the weighted and unweighted cost (i.e. "distance")
+  // to travel from one MachineBasicBlock to another.
+  //
+  // Values are pre-computed and stored in 'InterBlockDistances' using a
+  // backwards data-flow algorithm similar to the one described in 4.1 of a
+  // "Register Spilling and Live-Range Splitting for SSA-Form Programs" by
+  // Matthias Braun and Sebastian Hack, CC'09. This replaced a prior
+  // implementation based on Dijkstra's shortest path algorithm.
+  //----------------------------------------------------------------------------
+private:
+  struct InterBlockDistance {
+    NextUseDistance Weighted;
+    NextUseDistance Unweighted;
+    InterBlockDistance() : Weighted(-1), Unweighted(-1) {}
+    InterBlockDistance(NextUseDistance W, NextUseDistance UW)
+        : Weighted(W), Unweighted(UW) {}
+    bool operator==(const InterBlockDistance &Other) const {
+      return Weighted == Other.Weighted && Unweighted == Other.Unweighted;
+    }
+    bool operator!=(const InterBlockDistance &Other) const {
+      return !(*this == Other);
+    }
+
+    void print(raw_ostream &OS) const {
+      OS << "{W=";
+      Weighted.print(OS);
+      OS << " U=";
+      Unweighted.print(OS);
+      OS << "}";
+    }
+
+    LLVM_DUMP_METHOD void dump() const {
+      print(dbgs());
+      dbgs() << '\n';
+    }
+  };
+  using InterBlockDistanceMap =
+      DenseMap<unsigned, DenseMap<unsigned, InterBlockDistance>>;
+  InterBlockDistanceMap InterBlockDistances;
+
+  void initializeInterBlockDistances() {
+    InterBlockDistanceMap Distances;
+
+    bool Changed;
+    do {
+      Changed = false;
+      for (const MachineBasicBlock *MBB : post_order(MF)) {
+        unsigned MBBNum = MBB->getNumber();
+
+        // Save previous state for convergence check
+        InterBlockDistanceMap::mapped_type Prev = std::move(Distances[MBBNum]);
+        InterBlockDistanceMap::mapped_type Curr;
+        Curr.reserve(Prev.size());
+
+        // Direct successors are distance 0 by definition: no instructions are
+        // executed between exiting MBB and entering Succ.
+        for (const MachineBasicBlock *Succ : MBB->successors())
+          Curr[Succ->getNumber()] = InterBlockDistance(0, 0);
+
+        // Propagate further destinations through each successor.
+        for (const MachineBasicBlock *Succ : MBB->successors()) {
+          unsigned SuccNum = Succ->getNumber();
+          const unsigned UnweightedSize{getSize(Succ)};
+
+          for (const auto &[DestBlockNum, DestDist] : Distances[SuccNum]) {
+            // MBB -> MBB is considered unreachable (getInterBlockDistance
+            // asserts From != To).
+            if (DestBlockNum == MBBNum)
+              continue;
+
+            const MachineBasicBlock *DestMBB =
+                MF->getBlockNumbered(DestBlockNum);
+
+            const NextUseDistance UnweightedDist{UnweightedSize +
+                                                 DestDist.Unweighted};
+
+            unsigned SuccToDestLoopDepth = calcRelativeLoopDepth(Succ, DestMBB);
+
+            const NextUseDistance WeightedDist =
+                DestDist.Weighted +
+                NextUseDistance::fromSize(UnweightedSize, SuccToDestLoopDepth);
+
+            // Insert or update distances (take minimum)
+            auto [I, First] =
+                Curr.try_emplace(DestBlockNum, WeightedDist, UnweightedDist);
+            if (!First) {
+              InterBlockDistance &Slot = I->second;
+              Slot.Weighted = min(Slot.Weighted, WeightedDist);
+              Slot.Unweighted = min(Slot.Unweighted, UnweightedDist);
+            }
+          }
+        }
+        Changed |= (Prev != Curr);
+        Distances[MBBNum] = std::move(Curr);
+      }
+    } while (Changed);
+
+    InterBlockDistances = std::move(Distances);
+    LLVM_DEBUG(dumpInterBlockDistances());
+  }
+
+  const InterBlockDistance *
+  getInterBlockDistanceMapValue(const MachineBasicBlock *From,
+                                const MachineBasicBlock *To) const {
+    auto I = InterBlockDistances.find(From->getNumber());
+    if (I == InterBlockDistances.end())
+      return nullptr;
+    const InterBlockDistanceMap::mapped_type &FromSlot = I->second;
+    auto J = FromSlot.find(To->getNumber());
+    return J == FromSlot.end() ? nullptr : &J->second;
+  }
+
+  bool interBlockDistanceExists(const MachineBasicBlock *From,
+                                const MachineBasicBlock *To) const {
+    return getInterBlockDistanceMapValue(From, To);
+  }
+
+  NextUseDistance getInterBlockDistance(const MachineBasicBlock *From,
+                                        const MachineBasicBlock *To,
+                                        bool Unweighted) const {
+
+    assert(From != To && "The basic blocks should be different.");
+    if (!From || !To)
+      return NextUseDistance::unreachable();
+
+    if (Cfg.ForwardOnly && !isForwardReachable(From, To))
+      return NextUseDistance::unreachable();
+
+    const InterBlockDistance *BD = getInterBlockDistanceMapValue(From, To);
+    if (!BD)
+      return NextUseDistance::unreachable();
+
+    return Unweighted ? BD->Unweighted : BD->Weighted;
+  }
+
+  NextUseDistance
+  getWeightedInterBlockDistance(const MachineBasicBlock *From,
+                                const MachineBasicBlock *To) const {
+    return getInterBlockDistance(From, To, false);
+  }
+
+  NextUseDistance
+  getUnweightedInterBlockDistance(const MachineBasicBlock *From,
+                                  const MachineBasicBlock *To) const {
+    return getInterBlockDistance(From, To, true);
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // Feature getters. Use cached results if available. If not calculate.
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  InstrIdTy getSize(const MachineBasicBlock *BB) const {
+    return pathInfoFor(BB, BB).Size;
+  }
+
+  bool isReachable(const MachineBasicBlock *From,
+                   const MachineBasicBlock *To) const {
+    return pathInfoFor(From, To).Reachable;
+  }
+
+  bool isReachableOrSame(const MachineBasicBlock *From,
+                         const MachineBasicBlock *To) const {
+    return From == To || pathInfoFor(From, To).Reachable;
+  }
+
+  bool isForwardReachable(const MachineBasicBlock *From,
+                          const MachineBasicBlock *To) const {
+    const PathInfo &PI = pathInfoFor(From, To);
+    if (PI.isForwardReachableSet())
+      return PI.isForwardReachable();
+
+    return initializePathInfoForwardReachable(
+        From, To,
+        PI.Reachable && calcIsReachable(From, To, /*ForwardOnly*/ true));
+  }
+
+  // Return true/false if we know that 'To' is reachable or not from
+  // 'From'. Otherwise return 'std::nullopt'.
+  std::optional<bool> isMaybeReachable(const MachineBasicBlock *From,
+                                       const MachineBasicBlock *To,
+                                       bool ForwardOnly) const {
+    const PathInfo *PI = maybePathInfoFor(From, To);
+    if (!PI)
+      return std::nullopt;
+
+    if (ForwardOnly) {
+      if (PI->isForwardReachable())
+        return true;
+
+      if (PI->isNotForwardReachable())
+        return false;
+      return std::nullopt;
+    }
+    return PI->Reachable;
+  }
+
+  bool isBackedge(const MachineBasicBlock *From,
+                  const MachineBasicBlock *To) const {
+    return pathInfoFor(From, To).isBackedge();
+  }
+
+  // Can be used as a substitute for DT->dominates(A, B) if A and B are in the
+  // same basic block.
+  bool instrsAreInOrder(const MachineInstr *A, const MachineInstr *B) const {
+    assert(A->getParent() == B->getParent() &&
+           "instructions must be in the same basic block!");
+    if (A == B || getInstrId(A) < getInstrId(B))
+      return true;
+    if (!A->isPHI())
+      return false;
+    if (!B->isPHI())
+      return true;
+    for (auto &PHI : A->getParent()->phis()) {
+      if (&PHI == A)
+        return true;
+      if (&PHI == B)
+        return false;
+    }
+    return false;
+  }
+
+  unsigned getRelativeLoopDepth(const MachineBasicBlock *From,
+                                const MachineBasicBlock *To) const {
+    return pathInfoFor(From, To).RelativeLoopDepth;
+  }
+
+  NextUseDistance getShortestPath(const MachineBasicBlock *From,
+                                  const MachineBasicBlock *To) const {
+    std::optional<NextUseDistance> MaybeD =
+        pathInfoFor(From, To).ShortestDistance;
+    if (MaybeD.has_value())
+      return MaybeD.value();
+
+    NextUseDistance Dist = getWeightedInterBlockDistance(From, To);
+    return initializePathInfoShortestDistance(From, To, Dist);
+  }
+
+  NextUseDistance getShortestUnweightedPath(const MachineBasicBlock *From,
+                                            const MachineBasicBlock *To) const {
+    std::optional<NextUseDistance> MaybeD =
+        pathInfoFor(From, To).ShortestUnweightedDistance;
+    if (MaybeD.has_value())
+      return MaybeD.value();
+
+    return initializePathInfoShortestUnweightedDistance(
+        From, To, getUnweightedInterBlockDistance(From, To));
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  /// MBBDistPair - Represents the distance to a machine basic block.
+  /// Used for returning both the distance and the target block together.
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  struct MBBDistPair {
+    NextUseDistance Distance;
+    const MachineBasicBlock *MBB;
+    MBBDistPair() : Distance(NextUseDistance::unreachable()), MBB(nullptr) {}
+    MBBDistPair(NextUseDistance D, const MachineBasicBlock *B)
+        : Distance(D), MBB(B) {}
+
+    MBBDistPair operator+(NextUseDistance D) { return {Distance + D, MBB}; }
+    MBBDistPair &operator+=(NextUseDistance D) {
+      Distance += D;
+      return *this;
+    }
+
+    void print(raw_ostream &OS) const {
+      OS << "{";
+      Distance.print(OS);
+      if (MBB)
+        OS << " " << printMBBReference(*MBB);
+      else
+        OS << " <null>";
+      OS << "}";
+    }
+
+    LLVM_DUMP_METHOD void dump() const {
+      print(dbgs());
+      dbgs() << '\n';
+    }
+  };
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // CFG Helpers
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  // Return the shortest distance to a latch
+  MBBDistPair calcShortestDistanceToLatch(const MachineBasicBlock *CurMBB,
+                                          const MachineLoop *CurLoop) const {
+    SmallVector<MachineBasicBlock *, 2> Latches;
+    CurLoop->getLoopLatches(Latches);
+    MBBDistPair LD;
+
+    for (MachineBasicBlock *LMBB : Latches) {
+      if (LMBB == CurMBB)
+        return {0, CurMBB};
+
+      NextUseDistance Dst = getShortestPath(CurMBB, LMBB);
+      if (Dst < LD.Distance) {
+        LD.Distance = Dst;
+        LD.MBB = LMBB;
+      }
+    }
+    return LD;
+  }
+
+  // Return the shortest unweighted distance to a latch
+  MBBDistPair
+  calcShortestUnweightedDistanceToLatch(const MachineBasicBlock *CurMBB,
+                                        const MachineLoop *CurLoop) const {
+    SmallVector<MachineBasicBlock *, 2> Latches;
+    CurLoop->getLoopLatches(Latches);
+    MBBDistPair LD;
+
+    for (MachineBasicBlock *LMBB : Latches) {
+      if (LMBB == CurMBB)
+        return {0, CurMBB};
+
+      NextUseDistance Dst = getShortestUnweightedPath(CurMBB, LMBB);
+      if (Dst < LD.Distance) {
+        LD.Distance = Dst;
+        LD.MBB = LMBB;
+      }
+    }
+    return LD;
+  }
+
+  // Return the shortest distance to an exit
+  MBBDistPair calcShortestDistanceToExit(const MachineBasicBlock *CurMBB,
+                                         const MachineLoop *CurLoop) const {
+    SmallVector<std::pair<MachineBasicBlock *, MachineBasicBlock *>> ExitEdges;
+    CurLoop->getExitEdges(ExitEdges);
+    MBBDistPair LD;
+
+    for (auto [Exit, Dest] : ExitEdges) {
+      if (Exit == CurMBB)
+        return {0, CurMBB};
+
+      NextUseDistance Dst = getShortestPath(CurMBB, Exit);
+      if (Dst < LD.Distance) {
+        LD.Distance = Dst;
+        LD.MBB = Exit;
+      }
+    }
+    return LD;
+  }
+
+  // Return the shortest distance through a loop (header to latch) that goes
+  // through CurMBB.
+  MBBDistPair
+  calcShortestDistanceThroughInnermostLoop(const MachineBasicBlock *CurMBB,
+                                           MachineLoop *CurLoop) const {
+    assert(MLI->getLoopFor(CurMBB) == CurLoop);
+
+    // This is a hot spot. Check it before doing anything else.
+    if (CurLoop->getNumBlocks() == 1)
+      return {getSize(CurMBB), CurMBB};
+
+    MachineBasicBlock *LoopHeader = CurLoop->getHeader();
+    MBBDistPair LD{0, nullptr};
+
+    LD += getSize(LoopHeader);
+
+    if (CurMBB != LoopHeader)
+      LD += getShortestPath(LoopHeader, CurMBB);
+
+    if (CurLoop->isLoopExiting(CurMBB))
+      LD.MBB = CurMBB;
+    else
+      LD = calcShortestDistanceToExit(CurMBB, CurLoop) + LD.Distance;
+
+    if (CurMBB != LoopHeader && CurMBB != LD.MBB)
+      LD += getSize(CurMBB);
+
+    if (LD.MBB != LoopHeader)
+      LD += getSize(LD.MBB);
+
+    return LD;
+  }
+
+  // Return the shortest distance through a loop (header to latch) that goes
+  // through CurMBB.
+  MBBDistPair calcShortestDistanceThroughLoop(const MachineBasicBlock *CurMBB,
+                                              MachineLoop *OuterLoop) const {
+    MachineLoop *CurLoop = MLI->getLoopFor(CurMBB);
+    MBBDistPair CurLD =
+        calcShortestDistanceThroughInnermostLoop(CurMBB, CurLoop);
+
+    MachineBasicBlock *CurHdr = CurLoop->getHeader();
+    for (;;) {
+      if (OuterLoop == CurLoop)
+        return CurLD;
+
+      MachineLoop *ParentLoop = CurLoop->getParentLoop();
+      MachineBasicBlock *ParentHdr = ParentLoop->getHeader();
+
+      MBBDistPair LD{0, nullptr};
+      LD += getSize(ParentHdr);
+      LD += getShortestPath(ParentHdr, CurHdr);
+      LD += CurLD.Distance.applyLoopWeight();
+      LD = calcShortestDistanceToExit(CurLD.MBB, ParentLoop) + LD.Distance;
+      LD += getSize(LD.MBB);
+      CurLD = LD;
+      CurLoop = ParentLoop;
+      CurHdr = ParentHdr;
+    }
+    llvm_unreachable("CurMBB not contained in OuterLoop");
+  }
+
+  // Similar to calcShortestDistanceThroughLoop with LoopWeight applied to the
+  // returned distance.
+  MBBDistPair
+  calcWeightedDistanceThroughLoopViaMBB(const MachineBasicBlock *CurMBB,
+                                        MachineLoop *CurLoop) const {
+    MBBDistPair LD = calcShortestDistanceThroughLoop(CurMBB, CurLoop);
+    LD.Distance = LD.Distance.applyLoopWeight();
+    return LD;
+  }
+
+  // Return the weighted, shortest distance through a loop (header to latch).
+  // If ParentLoop is provided, use it to adjust the loop depth.
+  MBBDistPair calcWeightedDistanceThroughLoop(
+      MachineLoop *CurLoop, const MachineLoop *ParentLoop = nullptr) const {
+    const MachineBasicBlock *Hdr = CurLoop->getHeader();
+    if (CurLoop->getNumBlocks() != 1)
+      return calcWeightedDistanceThroughLoopViaMBB(Hdr, CurLoop);
+
+    unsigned LoopDepth = MLI->getLoopDepth(Hdr);
+    if (ParentLoop)
+      LoopDepth -= ParentLoop->getLoopDepth();
+
+    return {NextUseDistance::fromSize(getSize(Hdr), LoopDepth),
+            CurLoop->getLoopLatch()};
+  }
+
+  // Calculate total distance from exit point to use instruction
+  NextUseDistance appendDistanceToUse(const MBBDistPair &Exit,
+                                      const MachineInstr *UseMI,
+                                      const MachineBasicBlock *UseMBB) const {
+    return Exit.Distance + getShortestPath(Exit.MBB, UseMBB) +
+           getHeadLen(UseMI);
+  }
+
+  // Return the weighted, shortest distance through the CurLoop which is a
+  // sub-loop of UseLoop.
+  MBBDistPair calcDistanceThroughSubLoopUse(MachineLoop *CurLoop,
+                                            MachineLoop *UseLoop) const {
+    // All the sub-loops of the UseLoop will be executed before the use.
+    // Hence, we should take this into consideration in distance calculation.
+    MachineLoop *UseLoopSubLoop = findChildLoop(UseLoop, CurLoop);
+    assert(UseLoopSubLoop && "CurLoop should be nested in UseLoop");
+    return calcWeightedDistanceThroughLoop(UseLoopSubLoop, UseLoop);
+  }
+
+  // Similar to calcDistanceThroughSubLoopUse, adding the distance to 'UseMI'.
+  NextUseDistance calcDistanceThroughSubLoopToUseMI(
+      const MachineBasicBlock *CurMBB, MachineLoop *CurLoop,
+      const MachineInstr *UseMI, const MachineBasicBlock *UseMBB,
+      MachineLoop *UseLoop) const {
+    return appendDistanceToUse(calcDistanceThroughSubLoopUse(CurLoop, UseLoop),
+                               UseMI, UseMBB);
+  }
+
+  // Return the weighted distance through a loop to an outside use loop.
+  // Differentiates between uses inside or outside of the current loop nest.
+  MBBDistPair calcDistanceThroughLoopToOutsideLoopUse(
+      const MachineBasicBlock *CurMBB, MachineLoop *CurLoop,
+      const MachineBasicBlock *UseMBB, MachineLoop *UseLoop) const {
+    assert(!CurLoop->contains(UseLoop));
+
+    if (isStandAloneLoop(CurLoop))
+      return calcWeightedDistanceThroughLoopViaMBB(CurMBB, CurLoop);
+
+    MachineLoop *OutermostLoop = CurLoop->getOutermostLoop();
+    if (!OutermostLoop->contains(UseLoop)) {
+      // We should take into consideration the whole loop nest in the
+      // calculation of the distance because we will reach the use after
+      // executing the whole loop nest.
+
+      // ... But make sure that we pick a route that goes through CurMBB
+      return calcWeightedDistanceThroughLoopViaMBB(CurMBB, OutermostLoop);
+    }
+
+    // At this point we know that CurLoop and UseLoop are independent and they
+    // are in the same loop nest.
+
+    if (MLI->getLoopDepth(CurMBB) <= MLI->getLoopDepth(UseMBB))
+      return calcWeightedDistanceThroughLoop(CurLoop);
+
+    assert(CurLoop != OutermostLoop && "The loop cannot be the outermost.");
+    const unsigned UseLoopDepth = MLI->getLoopDepth(UseMBB);
+    for (;;) {
+      if (CurLoop->getLoopDepth() == UseLoopDepth)
+        break;
+      CurLoop = CurLoop->getParentLoop();
+      if (CurLoop == OutermostLoop)
+        break;
+    }
+    return calcWeightedDistanceThroughLoop(CurLoop);
+  }
+
+  // Similar to calcDistanceThroughLoopToOutsideLoopUse but adds the distance to
+  // an instruction in the loop.
+  NextUseDistance calcDistanceThroughLoopToOutsideLoopUseMI(
+      const MachineBasicBlock *CurMBB, MachineLoop *CurLoop,
+      const MachineInstr *UseMI, const MachineBasicBlock *UseMBB,
+      MachineLoop *UseLoop) const {
+    return appendDistanceToUse(calcDistanceThroughLoopToOutsideLoopUse(
+                                   CurMBB, CurLoop, UseMBB, UseLoop),
+                               UseMI, UseMBB);
+  }
+
+  // Return true if 'MO' is covered by 'LaneMask'
+  bool machineOperandCoveredBy(const MachineOperand &MO,
+                               LaneBitmask LaneMask) const {
+    LaneBitmask Mask = TRI->getSubRegIndexLaneMask(MO.getSubReg());
+    return (Mask & LaneMask) == Mask;
+  }
+
+  // Returns true iff uses of LiveReg/LiveLaneMask in PHI UseMI are coming from
+  // a backedge when starting at CurMI.
+  bool isIncomingValFromBackedge(Register LiveReg, LaneBitmask LiveLaneMask,
+                                 const MachineInstr *CurMI,
+                                 const MachineInstr *UseMI) const {
+    if (!UseMI->isPHI())
+      return false;
+
+    MachineLoop *CurLoop = MLI->getLoopFor(CurMI->getParent());
+    MachineLoop *UseLoop = MLI->getLoopFor(UseMI->getParent());
+
+    // Not a backedge if ...
+    // A: not in a loop at all
+    // B: or CurMI is in a loop outside of UseLoop
+    // C: or UseMI is not in the UseLoop header
+    if (/*A:*/ !UseLoop ||
+        /*B:*/ (CurLoop && !UseLoop->contains(CurLoop)) ||
+        /*C:*/ UseMI->getParent() != UseLoop->getHeader())
+      return false;
+
+    SmallVector<MachineBasicBlock *, 2> Latches;
+    UseLoop->getLoopLatches(Latches);
+
+    auto Ops = UseMI->operands();
+    for (auto It = std::next(Ops.begin()), ItE = Ops.end(); It != ItE;
+         It = std::next(It, 2)) {
+      auto &RegMO = *It;
+      auto &MBBMO = *std::next(It);
+      assert(RegMO.isReg() && "Expected register operand of PHI");
+      assert(MBBMO.isMBB() && "Expected MBB operand of PHI");
+      if (RegMO.getReg() == LiveReg &&
+          machineOperandCoveredBy(RegMO, LiveLaneMask)) {
+        MachineBasicBlock *IncomingBB = MBBMO.getMBB();
+        if (llvm::is_contained(Latches, IncomingBB))
+          return true;
+      }
+    }
+    return false;
+  }
+
+  // Return the distance from 'CurMI' through a parent loop backedge PHI Use
+  // ('UseMI').
+  CacheableNextUseDistance calcDistanceViaEnclosingBackedge(
+      const MachineInstr *CurMI, const MachineBasicBlock *CurMBB,
+      MachineLoop *CurLoop, const MachineInstr *UseMI,
+      const MachineBasicBlock *UseMBB, MachineLoop *UseLoop) const {
+    assert(UseLoop && "There is no backedge.");
+    assert(CurLoop && (UseLoop != CurLoop) && UseLoop->contains(CurLoop) &&
+           "Unexpected loop configuration");
+
+    InstrIdTy UseHeadLen = getHeadLen(UseMI);
+    MBBDistPair InnerLoopLD = calcDistanceThroughSubLoopUse(CurLoop, UseLoop);
+    MBBDistPair LD = calcShortestDistanceToLatch(InnerLoopLD.MBB, UseLoop);
+    return {InstrInvariant,
+            InnerLoopLD.Distance + LD.Distance + getSize(LD.MBB) + UseHeadLen};
+  }
+
+  // Optimized version of calcBackedgeDistance when we already know that CurMI
+  // and UseMI are in the same basic block
+  NextUseDistance calcBackedgeDistance(const MachineInstr *CurMI,
+                                       const MachineBasicBlock *CurMBB,
+                                       MachineLoop *CurLoop,
+                                       const MachineInstr *UseMI) const {
+    // use is in the next loop iteration
+    InstrIdTy CurTailLen = getTailLen(CurMI);
+    InstrIdTy UseHeadLen = getHeadLen(UseMI);
+    MBBDistPair LD = calcShortestUnweightedDistanceToLatch(CurMBB, CurLoop);
+    const MachineBasicBlock *HdrMBB = CurLoop->getHeader();
+    NextUseDistance Hdr = CurMBB == HdrMBB ? 0 : getSize(HdrMBB);
+    NextUseDistance Dst =
+        CurMBB == HdrMBB ? 0 : getShortestUnweightedPath(HdrMBB, CurMBB);
+
+    return CurTailLen + LD.Distance + getSize(LD.MBB) + Hdr + Dst + UseHeadLen;
+  }
+
+  //----------------------------------------------------------------------------
+  // Calculate inter-instruction distances
+  //----------------------------------------------------------------------------
+private:
+  // Calculate the shortest weighted path from MachineInstruction 'FromMI' to
+  // 'ToMI'. It is weighted distance in that paths that exit loops are made to
+  // look much further away.
+  NextUseDistance calcShortestDistance(const MachineInstr *FromMI,
+                                       const MachineInstr *ToMI) const {
+    const MachineBasicBlock *FromMBB = FromMI->getParent();
+    const MachineBasicBlock *ToMBB = ToMI->getParent();
+
+    if (FromMBB == ToMBB) {
+      NextUseDistance RV = getDistance(FromMI, ToMI);
+      assert(RV >= 0 && "unexpected negative distance from getDistance");
+      return RV;
+    }
+
+    InstrIdTy FromTailLen = getTailLen(FromMI);
+    InstrIdTy ToHeadLen = getHeadLen(ToMI);
+    NextUseDistance Dst = getShortestPath(FromMBB, ToMBB);
+    assert(Dst.isReachable() &&
+           "calcShortestDistance called for instructions in non-reachable"
+           " basic blocks!");
+    NextUseDistance RV = FromTailLen + Dst + ToHeadLen;
+    assert(RV >= 0 && "unexpected negative distance");
+    return RV;
+  }
+
+  // Calculate the shortest unweighted path from MachineInstruction 'FromMI' to
+  // 'ToMI'. In contrast with 'calcShortestDistance', distances are based solely
+  // on basic block instruction counts and traversing a loop exit does not
+  // affect the value.
+  NextUseDistance
+  calcShortestUnweightedDistance(const MachineInstr *FromMI,
+                                 const MachineInstr *ToMI) const {
+    const MachineBasicBlock *FromMBB = FromMI->getParent();
+    const MachineBasicBlock *ToMBB = ToMI->getParent();
+
+    if (FromMBB == ToMBB)
+      return getDistance(FromMI, ToMI);
+
+    InstrIdTy FromTailLen = getTailLen(FromMI);
+    InstrIdTy ToHeadLen = getHeadLen(ToMI);
+    NextUseDistance Dst = getShortestUnweightedPath(FromMBB, ToMBB);
+    assert(Dst.isReachable() &&
+           "calcShortestUnweightedDistance called for instructions in"
+           " non-reachable basic blocks!");
+    return FromTailLen + Dst + ToHeadLen;
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // calcDistanceToUse* - various flavors of calculating the distance from an
+  // instruction 'CurMI' to the use of a live [sub]register.
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  // Return the distance from 'CurMI' to a live [sub]register use ('UseMI').
+  //
+  // Cfg flags controlling behavior:
+  //   PreciseUseModeling — rewrite PHI uses to their incoming edge block;
+  //                        also selects unweighted cross-block distance
+  //   PromoteToPreheader — route loop-entry / inner-loop uses to the preheader
+  CacheableNextUseDistance
+  calcDistanceToUse(Register LiveReg, LaneBitmask LiveLaneMask,
+                    const MachineInstr &CurMI,
+                    const MachineOperand *UseMO) const {
+    const MachineInstr *UseMI = UseMO->getParent();
+    const MachineBasicBlock *CurMBB = CurMI.getParent();
+    const MachineBasicBlock *UseMBB = UseMI->getParent();
+    MachineLoop *CurLoop = MLI->getLoopFor(CurMBB);
+    MachineLoop *UseLoop = MLI->getLoopFor(UseMBB);
+
+    if (Cfg.PreciseUseModeling) {
+      // Map PHI use to the end of its incoming edge block.
+      if (const MachineBasicBlock *PhiUseEdge = mbbForPhiOp(UseMI, UseMO)) {
+        UseMI = &PhiUseEdge->back();
+        UseMBB = PhiUseEdge;
+        UseLoop = MLI->getLoopFor(PhiUseEdge);
+      }
+    }
+
+    enum class LoopConfig {
+      NoCur,
+      Same,
+      CurContainsUse,
+      UseContainsCur,
+      Siblings,
+      Unrelated
+    };
+    auto [LpCfg, PreHdr, CommonParent] = [&]()
+        -> std::tuple<LoopConfig, const MachineBasicBlock *, MachineLoop *> {
+      if (!CurLoop) {
+        return {LoopConfig::NoCur, getOutermostPreheader(UseLoop), nullptr};
+      }
+      if (CurLoop->contains(UseLoop)) {
+        return {CurMBB == UseMBB ? LoopConfig::Same
+                                 : LoopConfig::CurContainsUse,
+                findChildPreheader(CurLoop, UseLoop), nullptr};
+      }
+
+      if (MachineLoop *P = findCommonParent(UseLoop, CurLoop).first) {
+        if (P != UseLoop)
+          return {LoopConfig::Siblings, findChildPreheader(P, UseLoop), P};
+        return {LoopConfig::UseContainsCur, nullptr, nullptr};
+      }
+      return {LoopConfig::Unrelated, getOutermostPreheader(UseLoop), nullptr};
+    }();
+
+    //--------------------------------------------------------------------------
+    // Don't PromoteToPreheader
+    //--------------------------------------------------------------------------
+    if (!Cfg.PromoteToPreheader) {
+      switch (LpCfg) {
+      case LoopConfig::NoCur:
+      case LoopConfig::Same:
+      case LoopConfig::CurContainsUse:
+        return {InstrRelative, calcShortestDistance(&CurMI, UseMI)};
+
+      case LoopConfig::UseContainsCur: {
+        if (isIncomingValFromBackedge(LiveReg, LiveLaneMask, &CurMI, UseMI)) {
+          return calcDistanceViaEnclosingBackedge(&CurMI, CurMBB, CurLoop,
+                                                  UseMI, UseMBB, UseLoop);
+        }
+
+        return {InstrInvariant, calcDistanceThroughSubLoopToUseMI(
+                                    CurMBB, CurLoop, UseMI, UseMBB, UseLoop)};
+      }
+      case LoopConfig::Siblings:
+      case LoopConfig::Unrelated:
+        return {InstrInvariant, calcDistanceThroughLoopToOutsideLoopUseMI(
+                                    CurMBB, CurLoop, UseMI, UseMBB, UseLoop)};
+      }
+      llvm_unreachable("unexpected loop configuration!");
+    }
+
+    //--------------------------------------------------------------------------
+    // PromoteToPreheader
+    //--------------------------------------------------------------------------
+    if (PreHdr) {
+      UseMI = &PreHdr->back();
+      UseMBB = PreHdr;
+      UseLoop = CommonParent;
+    }
+
+    switch (LpCfg) {
+    case LoopConfig::NoCur:
+      return {InstrRelative, calcShortestUnweightedDistance(&CurMI, UseMI) -
+                                 (sizeOf(*UseMI) ? 0 : 1)};
+
+    case LoopConfig::Same:
+    case LoopConfig::CurContainsUse:
+      if (CurMBB == UseMBB && !instrsAreInOrder(&CurMI, UseMI))
+        return {InstrRelative,
+                calcBackedgeDistance(&CurMI, CurMBB, CurLoop, UseMI)};
+
+      return {InstrRelative, calcShortestUnweightedDistance(&CurMI, UseMI)};
+
+    case LoopConfig::UseContainsCur:
+    case LoopConfig::Siblings:
+      return {InstrInvariant, calcDistanceThroughSubLoopToUseMI(
+                                  CurMBB, CurLoop, UseMI, UseMBB, UseLoop)};
+
+    case LoopConfig::Unrelated:
+      return {InstrInvariant, calcDistanceThroughLoopToOutsideLoopUseMI(
+                                  CurMBB, CurLoop, UseMI, UseMBB, UseLoop)};
+    }
+    llvm_unreachable("unexpected loop configuration!");
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // getUses helpers (compute mode)
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  // Returns true if Use is reachable from MI. Handles backedges and intervening
+  // defs.
+  bool isUseReachablePrecise(const MachineInstr &MI,
+                             const MachineBasicBlock *MBB,
+                             const MachineOperand *UseMO,
+                             const MachineInstr *UseMI,
+                             const MachineBasicBlock *UseMBB) const {
+
+    // Filter out uses that are clearly unreachable
+    if (MBB != UseMBB && !isReachable(MBB, UseMBB))
+      return false;
+
+    // PHI uses are considered part of the incoming BB. Check for reachability
+    // at the edge.
+    if (const MachineBasicBlock *EdgeMBB = mbbForPhiOp(UseMI, UseMO)) {
+      if (!isReachableOrSame(MBB, EdgeMBB))
+        return false;
+    }
+
+    // Filter out uses with an intermediate def.
+    const MachineInstr *DefMI = MRI->getUniqueVRegDef(UseMO->getReg());
+    const MachineBasicBlock *DefMBB = DefMI->getParent();
+    if (MBB == UseMBB) {
+      if (UseMI->isPHI() && MBB == DefMBB)
+        return true;
+
+      if (instrsAreInOrder(&MI, UseMI))
+        return true;
+
+      // A Def in the loop means that the value at MI will not survive through
+      // to this use.
+      MachineLoop *UseLoop = MLI->getLoopFor(UseMBB);
+      return UseLoop && !UseLoop->contains(DefMBB);
+    }
+
+    if (MBB == DefMBB)
+      return instrsAreInOrder(DefMI, &MI);
+
+    MachineLoop *Loop = MLI->getLoopFor(MBB);
+    if (!Loop)
+      return true;
+
+    MachineLoop *TopLoop = Loop->getOutermostLoop();
+    return !TopLoop->contains(DefMBB) || !isReachable(MBB, DefMBB) ||
+           !isForwardReachable(UseMBB, MBB);
+  }
+
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  // Debug/Developer Helpers
+  //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+private:
+  /// Goes over all MBB pairs in \p MF, calculates the shortest path between
+  /// them.
+  void populatePathTable() {
+    for (const MachineBasicBlock &MBB1 : *MF) {
+      for (const MachineBasicBlock &MBB2 : *MF) {
+        if (&MBB1 == &MBB2)
+          continue;
+        getShortestPath(&MBB1, &MBB2);
+      }
+    }
+  }
+
+  void printPaths(raw_ostream &OS) const {
+    OS << "\n---------------- Paths --------------- {\n";
+    for (const auto &[P, PI] : Paths) {
+      OS << "  " << printMBBReference(*P.src()) << " -> "
+         << printMBBReference(*P.dst()) << ": ";
+      PI.print(OS);
+      OS << '\n';
+    }
+    OS << "}\n";
+  }
+
+  LLVM_DUMP_METHOD void dumpPaths() const { printPaths(dbgs()); }
+
+  // Legacy alias kept for existing call sites.
+  void dumpShortestPaths() const {
+    for (const auto &P : Paths) {
+      const MachineBasicBlock *From = P.first.src();
+      const MachineBasicBlock *To = P.first.dst();
+      std::optional<NextUseDistance> Dist = P.second.ShortestDistance;
+      dbgs() << "From: " << printMBBReference(*From)
+             << "-> To:" << printMBBReference(*To) << " = "
+             << Dist.value_or(-1).fmt() << "\n";
+    }
+  }
+
+  void printInterBlockDistances(raw_ostream &OS) const {
+    using MBBPair = std::pair<unsigned, unsigned>;
+    using Elem = std::pair<NextUseDistance, MBBPair>;
+    std::vector<Elem> SortedDistances;
+
+    for (const auto &[FromNum, Dsts] : InterBlockDistances) {
+      for (const auto &[ToNum, Dist] : Dsts) {
+        SortedDistances.emplace_back(Dist.Weighted, MBBPair(FromNum, ToNum));
+      }
+    }
+    std::sort(SortedDistances.begin(), SortedDistances.end(),
+              [](const auto &A, const auto &B) {
+                if (A.first != B.first)
+                  return A.first < B.first;
+
+                if (A.second.first != B.second.first)
+                  return A.second.first < B.second.first;
+
+                return A.second.second < B.second.second;
+              });
+
+    OS << "\n--------- InterBlockDistances -------- {\n";
+    for (const Elem &E : SortedDistances) {
+
+      OS << "  bb." << E.second.first << " -> bb." << E.second.second << ": ";
+      E.first.print(OS);
+      OS << '\n';
+    }
+    OS << "}\n";
+  }
+
+  LLVM_DUMP_METHOD void dumpInterBlockDistances() const {
+    printInterBlockDistances(dbgs());
+  }
+
+  void printAllDistances() {
+    auto getRegNextUseDistance =
+        [this](Register DefReg) -> std::optional<NextUseDistance> {
+      const MachineInstr &DefMI = *MRI->def_instr_begin(DefReg);
+
+      SmallVector<const MachineOperand *> Uses;
+      for (MachineOperand &UseMO : MRI->use_nodbg_operands(DefReg))
+        Uses.push_back(&UseMO);
+
+      return getShortestDistance(DefReg, DefMI, Uses);
+    };
+
+    for (const MachineBasicBlock &MBB : *MF) {
+      for (const MachineInstr &MI : MBB) {
+        for (const MachineOperand &MO : MI.operands()) {
+          if (!MO.isReg() || MO.isUse())
+            continue;
+
+          Register Reg = MO.getReg();
+          if (Reg.isPhysical() || TRI->isAGPR(*MRI, Reg))
----------------
arsenm wrote:

Shouldn't skip AGPR 

https://github.com/llvm/llvm-project/pull/178873


More information about the llvm-commits mailing list