File Manager

003 File Manager

Current Path: /usr/src/contrib/llvm-project/llvm/lib/Target/AMDGPU

usr / src / contrib / llvm-project / llvm / lib / Target / AMDGPU /

📁 ..
📄 AMDGPU.h(11.46 KB)
📄 AMDGPU.td(36.97 KB)
📄 AMDGPUAliasAnalysis.cpp(5.58 KB)
📄 AMDGPUAliasAnalysis.h(3.32 KB)
📄 AMDGPUAlwaysInlinePass.cpp(4.83 KB)
📄 AMDGPUAnnotateKernelFeatures.cpp(11.94 KB)
📄 AMDGPUAnnotateUniformValues.cpp(6.13 KB)
📄 AMDGPUArgumentUsageInfo.cpp(7.66 KB)
📄 AMDGPUArgumentUsageInfo.h(4.81 KB)
📄 AMDGPUAsmPrinter.cpp(50.42 KB)
📄 AMDGPUAsmPrinter.h(5.13 KB)
📄 AMDGPUAtomicOptimizer.cpp(23.79 KB)
📄 AMDGPUCallLowering.cpp(28.66 KB)
📄 AMDGPUCallLowering.h(2.37 KB)
📄 AMDGPUCallingConv.td(7.33 KB)
📄 AMDGPUCodeGenPrepare.cpp(46.42 KB)
📄 AMDGPUCombine.td(2.79 KB)
📄 AMDGPUExportClustering.cpp(4.52 KB)
📄 AMDGPUExportClustering.h(533 B)
📄 AMDGPUFeatures.td(1.81 KB)
📄 AMDGPUFixFunctionBitcasts.cpp(1.87 KB)
📄 AMDGPUFrameLowering.cpp(1.98 KB)
📄 AMDGPUFrameLowering.h(1.39 KB)
📄 AMDGPUGISel.td(11.57 KB)
📄 AMDGPUGenRegisterBankInfo.def(5.83 KB)
📄 AMDGPUGlobalISelUtils.cpp(1.77 KB)
📄 AMDGPUGlobalISelUtils.h(2.07 KB)
📄 AMDGPUHSAMetadataStreamer.cpp(31.21 KB)
📄 AMDGPUHSAMetadataStreamer.h(5.46 KB)
📄 AMDGPUISelDAGToDAG.cpp(101.59 KB)
📄 AMDGPUISelLowering.cpp(168.65 KB)
📄 AMDGPUISelLowering.h(19.23 KB)
📄 AMDGPUInline.cpp(7.97 KB)
📄 AMDGPUInstrInfo.cpp(1.71 KB)
📄 AMDGPUInstrInfo.h(1.66 KB)
📄 AMDGPUInstrInfo.td(17.18 KB)
📄 AMDGPUInstructionSelector.cpp(128.53 KB)
📄 AMDGPUInstructionSelector.h(11.04 KB)
📄 AMDGPUInstructions.td(25.36 KB)
📄 AMDGPULegalizerInfo.cpp(149.32 KB)
📄 AMDGPULegalizerInfo.h(8.49 KB)
📄 AMDGPULibCalls.cpp(53.89 KB)
📄 AMDGPULibFunc.cpp(37.85 KB)
📄 AMDGPULibFunc.h(10.99 KB)
📄 AMDGPULowerIntrinsics.cpp(4.55 KB)
📄 AMDGPULowerKernelArguments.cpp(8.89 KB)
📄 AMDGPULowerKernelAttributes.cpp(7.78 KB)
📄 AMDGPUMCInstLower.cpp(14.27 KB)
📄 AMDGPUMachineCFGStructurizer.cpp(101.97 KB)
📄 AMDGPUMachineFunction.cpp(2.24 KB)
📄 AMDGPUMachineFunction.h(2.13 KB)
📄 AMDGPUMachineModuleInfo.cpp(1.34 KB)
📄 AMDGPUMachineModuleInfo.h(5.46 KB)
📄 AMDGPUMacroFusion.cpp(2.28 KB)
📄 AMDGPUMacroFusion.h(679 B)
📄 AMDGPUOpenCLEnqueuedBlockLowering.cpp(5.31 KB)
📄 AMDGPUPTNote.h(1.29 KB)
📄 AMDGPUPerfHintAnalysis.cpp(12.17 KB)
📄 AMDGPUPerfHintAnalysis.h(1.67 KB)
📄 AMDGPUPostLegalizerCombiner.cpp(12.02 KB)
📄 AMDGPUPreLegalizerCombiner.cpp(5.45 KB)
📄 AMDGPUPrintfRuntimeBinding.cpp(21.7 KB)
📄 AMDGPUPromoteAlloca.cpp(35.24 KB)
📄 AMDGPUPropagateAttributes.cpp(11.76 KB)
📄 AMDGPURegBankCombiner.cpp(5.36 KB)
📄 AMDGPURegisterBankInfo.cpp(161.67 KB)
📄 AMDGPURegisterBankInfo.h(7.41 KB)
📄 AMDGPURegisterBanks.td(921 B)
📄 AMDGPURewriteOutArguments.cpp(15.82 KB)
📄 AMDGPUSearchableTables.td(21.04 KB)
📄 AMDGPUSubtarget.cpp(29.62 KB)
📄 AMDGPUSubtarget.h(35.82 KB)
📄 AMDGPUTargetMachine.cpp(42.67 KB)
📄 AMDGPUTargetMachine.h(4.52 KB)
📄 AMDGPUTargetObjectFile.cpp(1.54 KB)
📄 AMDGPUTargetObjectFile.h(1.14 KB)
📄 AMDGPUTargetTransformInfo.cpp(39.07 KB)
📄 AMDGPUTargetTransformInfo.h(11.11 KB)
📄 AMDGPUUnifyDivergentExitNodes.cpp(13.84 KB)
📄 AMDGPUUnifyMetadata.cpp(4.46 KB)
📄 AMDILCFGStructurizer.cpp(56.32 KB)
📄 AMDKernelCodeT.h(32.84 KB)
📁 AsmParser
📄 BUFInstructions.td(110.75 KB)
📄 CaymanInstructions.td(7.93 KB)
📄 DSInstructions.td(52.37 KB)
📁 Disassembler
📄 EvergreenInstructions.td(28.24 KB)
📄 FLATInstructions.td(66.93 KB)
📄 GCNDPPCombine.cpp(19.92 KB)
📄 GCNHazardRecognizer.cpp(45.3 KB)
📄 GCNHazardRecognizer.h(3.96 KB)
📄 GCNILPSched.cpp(11.3 KB)
📄 GCNIterativeScheduler.cpp(20.62 KB)
📄 GCNIterativeScheduler.h(4.16 KB)
📄 GCNMinRegStrategy.cpp(8.47 KB)
📄 GCNNSAReassign.cpp(10.92 KB)
📄 GCNProcessors.td(4.84 KB)
📄 GCNRegBankReassign.cpp(26.68 KB)
📄 GCNRegPressure.cpp(16.27 KB)
📄 GCNRegPressure.h(9.15 KB)
📄 GCNSchedStrategy.cpp(21.67 KB)
📄 GCNSchedStrategy.h(3.77 KB)
📁 MCTargetDesc
📄 MIMGInstructions.td(39.85 KB)
📄 R600.td(1.51 KB)
📄 R600AsmPrinter.cpp(4.46 KB)
📄 R600AsmPrinter.h(1.5 KB)
📄 R600ClauseMergePass.cpp(7.38 KB)
📄 R600ControlFlowFinalizer.cpp(23.4 KB)
📄 R600Defines.h(4.25 KB)
📄 R600EmitClauseMarkers.cpp(12.1 KB)
📄 R600ExpandSpecialInstrs.cpp(10.11 KB)
📄 R600FrameLowering.cpp(1.83 KB)
📄 R600FrameLowering.h(1.25 KB)
📄 R600ISelLowering.cpp(81.88 KB)
📄 R600ISelLowering.h(4.8 KB)
📄 R600InstrFormats.td(11.58 KB)
📄 R600InstrInfo.cpp(49.47 KB)
📄 R600InstrInfo.h(13.7 KB)
📄 R600Instructions.td(55.13 KB)
📄 R600MachineFunctionInfo.cpp(551 B)
📄 R600MachineFunctionInfo.h(824 B)
📄 R600MachineScheduler.cpp(13.57 KB)
📄 R600MachineScheduler.h(2.53 KB)
📄 R600OpenCLImageTypeLoweringPass.cpp(11.75 KB)
📄 R600OptimizeVectorRegisters.cpp(13.4 KB)
📄 R600Packetizer.cpp(13.4 KB)
📄 R600Processors.td(4.42 KB)
📄 R600RegisterInfo.cpp(3.95 KB)
📄 R600RegisterInfo.h(2 KB)
📄 R600RegisterInfo.td(9.75 KB)
📄 R600Schedule.td(1.62 KB)
📄 R700Instructions.td(783 B)
📄 SIAddIMGInit.cpp(6.24 KB)
📄 SIAnnotateControlFlow.cpp(11.18 KB)
📄 SIDefines.h(20.86 KB)
📄 SIFixSGPRCopies.cpp(29.46 KB)
📄 SIFixVGPRCopies.cpp(2 KB)
📄 SIFixupVectorISel.cpp(8.75 KB)
📄 SIFoldOperands.cpp(54.56 KB)
📄 SIFormMemoryClauses.cpp(12.76 KB)
📄 SIFrameLowering.cpp(48.08 KB)
📄 SIFrameLowering.h(2.98 KB)
📄 SIISelLowering.cpp(423.43 KB)
📄 SIISelLowering.h(22.13 KB)
📄 SIInsertHardClauses.cpp(7.01 KB)
📄 SIInsertSkips.cpp(15.29 KB)
📄 SIInsertWaitcnts.cpp(58.33 KB)
📄 SIInstrFormats.td(9.44 KB)
📄 SIInstrInfo.cpp(247.15 KB)
📄 SIInstrInfo.h(41.24 KB)
📄 SIInstrInfo.td(90.7 KB)
📄 SIInstructions.td(77.7 KB)
📄 SILoadStoreOptimizer.cpp(76.21 KB)
📄 SILowerControlFlow.cpp(22.66 KB)
📄 SILowerI1Copies.cpp(27.83 KB)
📄 SILowerSGPRSpills.cpp(12.68 KB)
📄 SIMachineFunctionInfo.cpp(20.01 KB)
📄 SIMachineFunctionInfo.h(26.91 KB)
📄 SIMachineScheduler.cpp(69.44 KB)
📄 SIMachineScheduler.h(15.65 KB)
📄 SIMemoryLegalizer.cpp(45.84 KB)
📄 SIModeRegister.cpp(17.43 KB)
📄 SIOptimizeExecMasking.cpp(12.81 KB)
📄 SIOptimizeExecMaskingPreRA.cpp(11.13 KB)
📄 SIPeepholeSDWA.cpp(42.84 KB)
📄 SIPostRABundler.cpp(3.6 KB)
📄 SIPreAllocateWWMRegs.cpp(6.09 KB)
📄 SIPreEmitPeephole.cpp(10.51 KB)
📄 SIProgramInfo.h(2.04 KB)
📄 SIRegisterInfo.cpp(71.51 KB)
📄 SIRegisterInfo.h(13.04 KB)
📄 SIRegisterInfo.td(37.28 KB)
📄 SIRemoveShortExecBranches.cpp(4.96 KB)
📄 SISchedule.td(7.58 KB)
📄 SIShrinkInstructions.cpp(26.86 KB)
📄 SIWholeQuadMode.cpp(30.22 KB)
📄 SMInstructions.td(48.14 KB)
📄 SOPInstructions.td(60.51 KB)
📁 TargetInfo
📁 Utils
📄 VIInstrFormats.td(645 B)
📄 VOP1Instructions.td(35.53 KB)
📄 VOP2Instructions.td(65.04 KB)
📄 VOP3Instructions.td(53.14 KB)
📄 VOP3PInstructions.td(26.47 KB)
📄 VOPCInstructions.td(63.31 KB)
📄 VOPInstructions.td(23.76 KB)

Editing: GCNRegPressure.cpp

//===- GCNRegPressure.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
//
//===----------------------------------------------------------------------===//
///
/// \file
/// This file implements the GCNRegPressure class.
///
//===----------------------------------------------------------------------===//

#include "GCNRegPressure.h"
#include "AMDGPUSubtarget.h"
#include "SIRegisterInfo.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/CodeGen/LiveInterval.h"
#include "llvm/CodeGen/LiveIntervals.h"
#include "llvm/CodeGen/MachineInstr.h"
#include "llvm/CodeGen/MachineOperand.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/RegisterPressure.h"
#include "llvm/CodeGen/SlotIndexes.h"
#include "llvm/CodeGen/TargetRegisterInfo.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/MC/LaneBitmask.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/raw_ostream.h"
#include <algorithm>
#include <cassert>

using namespace llvm;

#define DEBUG_TYPE "machine-scheduler"

#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD
void llvm::printLivesAt(SlotIndex SI,
                        const LiveIntervals &LIS,
                        const MachineRegisterInfo &MRI) {
  dbgs() << "Live regs at " << SI << ": "
         << *LIS.getInstructionFromIndex(SI);
  unsigned Num = 0;
  for (unsigned I = 0, E = MRI.getNumVirtRegs(); I != E; ++I) {
    const unsigned Reg = Register::index2VirtReg(I);
    if (!LIS.hasInterval(Reg))
      continue;
    const auto &LI = LIS.getInterval(Reg);
    if (LI.hasSubRanges()) {
      bool firstTime = true;
      for (const auto &S : LI.subranges()) {
        if (!S.liveAt(SI)) continue;
        if (firstTime) {
          dbgs() << "  " << printReg(Reg, MRI.getTargetRegisterInfo())
                 << '\n';
          firstTime = false;
        }
        dbgs() << "  " << S << '\n';
        ++Num;
      }
    } else if (LI.liveAt(SI)) {
      dbgs() << "  " << LI << '\n';
      ++Num;
    }
  }
  if (!Num) dbgs() << "  <none>\n";
}
#endif

bool llvm::isEqual(const GCNRPTracker::LiveRegSet &S1,
                   const GCNRPTracker::LiveRegSet &S2) {
  if (S1.size() != S2.size())
    return false;

for (const auto &P : S1) {
    auto I = S2.find(P.first);
    if (I == S2.end() || I->second != P.second)
      return false;
  }
  return true;
}

///////////////////////////////////////////////////////////////////////////////
// GCNRegPressure

unsigned GCNRegPressure::getRegKind(unsigned Reg,
                                    const MachineRegisterInfo &MRI) {
  assert(Register::isVirtualRegister(Reg));
  const auto RC = MRI.getRegClass(Reg);
  auto STI = static_cast<const SIRegisterInfo*>(MRI.getTargetRegisterInfo());
  return STI->isSGPRClass(RC) ?
    (STI->getRegSizeInBits(*RC) == 32 ? SGPR32 : SGPR_TUPLE) :
    STI->hasAGPRs(RC) ?
      (STI->getRegSizeInBits(*RC) == 32 ? AGPR32 : AGPR_TUPLE) :
      (STI->getRegSizeInBits(*RC) == 32 ? VGPR32 : VGPR_TUPLE);
}

void GCNRegPressure::inc(unsigned Reg,
                         LaneBitmask PrevMask,
                         LaneBitmask NewMask,
                         const MachineRegisterInfo &MRI) {
  if (SIRegisterInfo::getNumCoveredRegs(NewMask) ==
      SIRegisterInfo::getNumCoveredRegs(PrevMask))
    return;

int Sign = 1;
  if (NewMask < PrevMask) {
    std::swap(NewMask, PrevMask);
    Sign = -1;
  }

switch (auto Kind = getRegKind(Reg, MRI)) {
  case SGPR32:
  case VGPR32:
  case AGPR32:
    Value[Kind] += Sign;
    break;

case SGPR_TUPLE:
  case VGPR_TUPLE:
  case AGPR_TUPLE:
    assert(PrevMask < NewMask);

Value[Kind == SGPR_TUPLE ? SGPR32 : Kind == AGPR_TUPLE ? AGPR32 : VGPR32] +=
      Sign * SIRegisterInfo::getNumCoveredRegs(~PrevMask & NewMask);

if (PrevMask.none()) {
      assert(NewMask.any());
      Value[Kind] += Sign * MRI.getPressureSets(Reg).getWeight();
    }
    break;

default: llvm_unreachable("Unknown register kind");
  }
}

bool GCNRegPressure::less(const GCNSubtarget &ST,
                          const GCNRegPressure& O,
                          unsigned MaxOccupancy) const {
  const auto SGPROcc = std::min(MaxOccupancy,
                                ST.getOccupancyWithNumSGPRs(getSGPRNum()));
  const auto VGPROcc = std::min(MaxOccupancy,
                                ST.getOccupancyWithNumVGPRs(getVGPRNum()));
  const auto OtherSGPROcc = std::min(MaxOccupancy,
                                ST.getOccupancyWithNumSGPRs(O.getSGPRNum()));
  const auto OtherVGPROcc = std::min(MaxOccupancy,
                                ST.getOccupancyWithNumVGPRs(O.getVGPRNum()));

const auto Occ = std::min(SGPROcc, VGPROcc);
  const auto OtherOcc = std::min(OtherSGPROcc, OtherVGPROcc);
  if (Occ != OtherOcc)
    return Occ > OtherOcc;

bool SGPRImportant = SGPROcc < VGPROcc;
  const bool OtherSGPRImportant = OtherSGPROcc < OtherVGPROcc;

// if both pressures disagree on what is more important compare vgprs
  if (SGPRImportant != OtherSGPRImportant) {
    SGPRImportant = false;
  }

// compare large regs pressure
  bool SGPRFirst = SGPRImportant;
  for (int I = 2; I > 0; --I, SGPRFirst = !SGPRFirst) {
    if (SGPRFirst) {
      auto SW = getSGPRTuplesWeight();
      auto OtherSW = O.getSGPRTuplesWeight();
      if (SW != OtherSW)
        return SW < OtherSW;
    } else {
      auto VW = getVGPRTuplesWeight();
      auto OtherVW = O.getVGPRTuplesWeight();
      if (VW != OtherVW)
        return VW < OtherVW;
    }
  }
  return SGPRImportant ? (getSGPRNum() < O.getSGPRNum()):
                         (getVGPRNum() < O.getVGPRNum());
}

#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD
void GCNRegPressure::print(raw_ostream &OS, const GCNSubtarget *ST) const {
  OS << "VGPRs: " << Value[VGPR32] << ' ';
  OS << "AGPRs: " << Value[AGPR32];
  if (ST) OS << "(O" << ST->getOccupancyWithNumVGPRs(getVGPRNum()) << ')';
  OS << ", SGPRs: " << getSGPRNum();
  if (ST) OS << "(O" << ST->getOccupancyWithNumSGPRs(getSGPRNum()) << ')';
  OS << ", LVGPR WT: " << getVGPRTuplesWeight()
     << ", LSGPR WT: " << getSGPRTuplesWeight();
  if (ST) OS << " -> Occ: " << getOccupancy(*ST);
  OS << '\n';
}
#endif

static LaneBitmask getDefRegMask(const MachineOperand &MO,
                                 const MachineRegisterInfo &MRI) {
  assert(MO.isDef() && MO.isReg() && Register::isVirtualRegister(MO.getReg()));

// We don't rely on read-undef flag because in case of tentative schedule
  // tracking it isn't set correctly yet. This works correctly however since
  // use mask has been tracked before using LIS.
  return MO.getSubReg() == 0 ?
    MRI.getMaxLaneMaskForVReg(MO.getReg()) :
    MRI.getTargetRegisterInfo()->getSubRegIndexLaneMask(MO.getSubReg());
}

static LaneBitmask getUsedRegMask(const MachineOperand &MO,
                                  const MachineRegisterInfo &MRI,
                                  const LiveIntervals &LIS) {
  assert(MO.isUse() && MO.isReg() && Register::isVirtualRegister(MO.getReg()));

if (auto SubReg = MO.getSubReg())
    return MRI.getTargetRegisterInfo()->getSubRegIndexLaneMask(SubReg);

auto MaxMask = MRI.getMaxLaneMaskForVReg(MO.getReg());
  if (SIRegisterInfo::getNumCoveredRegs(MaxMask) > 1) // cannot have subregs
    return MaxMask;

// For a tentative schedule LIS isn't updated yet but livemask should remain
  // the same on any schedule. Subreg defs can be reordered but they all must
  // dominate uses anyway.
  auto SI = LIS.getInstructionIndex(*MO.getParent()).getBaseIndex();
  return getLiveLaneMask(MO.getReg(), SI, LIS, MRI);
}

static SmallVector<RegisterMaskPair, 8>
collectVirtualRegUses(const MachineInstr &MI, const LiveIntervals &LIS,
                      const MachineRegisterInfo &MRI) {
  SmallVector<RegisterMaskPair, 8> Res;
  for (const auto &MO : MI.operands()) {
    if (!MO.isReg() || !Register::isVirtualRegister(MO.getReg()))
      continue;
    if (!MO.isUse() || !MO.readsReg())
      continue;

auto const UsedMask = getUsedRegMask(MO, MRI, LIS);

auto Reg = MO.getReg();
    auto I = std::find_if(Res.begin(), Res.end(), [Reg](const RegisterMaskPair &RM) {
      return RM.RegUnit == Reg;
    });
    if (I != Res.end())
      I->LaneMask |= UsedMask;
    else
      Res.push_back(RegisterMaskPair(Reg, UsedMask));
  }
  return Res;
}

///////////////////////////////////////////////////////////////////////////////
// GCNRPTracker

LaneBitmask llvm::getLiveLaneMask(unsigned Reg,
                                  SlotIndex SI,
                                  const LiveIntervals &LIS,
                                  const MachineRegisterInfo &MRI) {
  LaneBitmask LiveMask;
  const auto &LI = LIS.getInterval(Reg);
  if (LI.hasSubRanges()) {
    for (const auto &S : LI.subranges())
      if (S.liveAt(SI)) {
        LiveMask |= S.LaneMask;
        assert(LiveMask < MRI.getMaxLaneMaskForVReg(Reg) ||
               LiveMask == MRI.getMaxLaneMaskForVReg(Reg));
      }
  } else if (LI.liveAt(SI)) {
    LiveMask = MRI.getMaxLaneMaskForVReg(Reg);
  }
  return LiveMask;
}

GCNRPTracker::LiveRegSet llvm::getLiveRegs(SlotIndex SI,
                                           const LiveIntervals &LIS,
                                           const MachineRegisterInfo &MRI) {
  GCNRPTracker::LiveRegSet LiveRegs;
  for (unsigned I = 0, E = MRI.getNumVirtRegs(); I != E; ++I) {
    auto Reg = Register::index2VirtReg(I);
    if (!LIS.hasInterval(Reg))
      continue;
    auto LiveMask = getLiveLaneMask(Reg, SI, LIS, MRI);
    if (LiveMask.any())
      LiveRegs[Reg] = LiveMask;
  }
  return LiveRegs;
}

void GCNRPTracker::reset(const MachineInstr &MI,
                         const LiveRegSet *LiveRegsCopy,
                         bool After) {
  const MachineFunction &MF = *MI.getMF();
  MRI = &MF.getRegInfo();
  if (LiveRegsCopy) {
    if (&LiveRegs != LiveRegsCopy)
      LiveRegs = *LiveRegsCopy;
  } else {
    LiveRegs = After ? getLiveRegsAfter(MI, LIS)
                     : getLiveRegsBefore(MI, LIS);
  }

MaxPressure = CurPressure = getRegPressure(*MRI, LiveRegs);
}

void GCNUpwardRPTracker::reset(const MachineInstr &MI,
                               const LiveRegSet *LiveRegsCopy) {
  GCNRPTracker::reset(MI, LiveRegsCopy, true);
}

void GCNUpwardRPTracker::recede(const MachineInstr &MI) {
  assert(MRI && "call reset first");

LastTrackedMI = &MI;

if (MI.isDebugInstr())
    return;

auto const RegUses = collectVirtualRegUses(MI, LIS, *MRI);

// calc pressure at the MI (defs + uses)
  auto AtMIPressure = CurPressure;
  for (const auto &U : RegUses) {
    auto LiveMask = LiveRegs[U.RegUnit];
    AtMIPressure.inc(U.RegUnit, LiveMask, LiveMask | U.LaneMask, *MRI);
  }
  // update max pressure
  MaxPressure = max(AtMIPressure, MaxPressure);

for (const auto &MO : MI.operands()) {
    if (!MO.isReg() || !MO.isDef() ||
        !Register::isVirtualRegister(MO.getReg()) || MO.isDead())
      continue;

auto Reg = MO.getReg();
    auto I = LiveRegs.find(Reg);
    if (I == LiveRegs.end())
      continue;
    auto &LiveMask = I->second;
    auto PrevMask = LiveMask;
    LiveMask &= ~getDefRegMask(MO, *MRI);
    CurPressure.inc(Reg, PrevMask, LiveMask, *MRI);
    if (LiveMask.none())
      LiveRegs.erase(I);
  }
  for (const auto &U : RegUses) {
    auto &LiveMask = LiveRegs[U.RegUnit];
    auto PrevMask = LiveMask;
    LiveMask |= U.LaneMask;
    CurPressure.inc(U.RegUnit, PrevMask, LiveMask, *MRI);
  }
  assert(CurPressure == getRegPressure(*MRI, LiveRegs));
}

bool GCNDownwardRPTracker::reset(const MachineInstr &MI,
                                 const LiveRegSet *LiveRegsCopy) {
  MRI = &MI.getParent()->getParent()->getRegInfo();
  LastTrackedMI = nullptr;
  MBBEnd = MI.getParent()->end();
  NextMI = &MI;
  NextMI = skipDebugInstructionsForward(NextMI, MBBEnd);
  if (NextMI == MBBEnd)
    return false;
  GCNRPTracker::reset(*NextMI, LiveRegsCopy, false);
  return true;
}

bool GCNDownwardRPTracker::advanceBeforeNext() {
  assert(MRI && "call reset first");

NextMI = skipDebugInstructionsForward(NextMI, MBBEnd);
  if (NextMI == MBBEnd)
    return false;

SlotIndex SI = LIS.getInstructionIndex(*NextMI).getBaseIndex();
  assert(SI.isValid());

// Remove dead registers or mask bits.
  for (auto &It : LiveRegs) {
    const LiveInterval &LI = LIS.getInterval(It.first);
    if (LI.hasSubRanges()) {
      for (const auto &S : LI.subranges()) {
        if (!S.liveAt(SI)) {
          auto PrevMask = It.second;
          It.second &= ~S.LaneMask;
          CurPressure.inc(It.first, PrevMask, It.second, *MRI);
        }
      }
    } else if (!LI.liveAt(SI)) {
      auto PrevMask = It.second;
      It.second = LaneBitmask::getNone();
      CurPressure.inc(It.first, PrevMask, It.second, *MRI);
    }
    if (It.second.none())
      LiveRegs.erase(It.first);
  }

MaxPressure = max(MaxPressure, CurPressure);

return true;
}

void GCNDownwardRPTracker::advanceToNext() {
  LastTrackedMI = &*NextMI++;

// Add new registers or mask bits.
  for (const auto &MO : LastTrackedMI->operands()) {
    if (!MO.isReg() || !MO.isDef())
      continue;
    Register Reg = MO.getReg();
    if (!Register::isVirtualRegister(Reg))
      continue;
    auto &LiveMask = LiveRegs[Reg];
    auto PrevMask = LiveMask;
    LiveMask |= getDefRegMask(MO, *MRI);
    CurPressure.inc(Reg, PrevMask, LiveMask, *MRI);
  }

MaxPressure = max(MaxPressure, CurPressure);
}

bool GCNDownwardRPTracker::advance() {
  // If we have just called reset live set is actual.
  if ((NextMI == MBBEnd) || (LastTrackedMI && !advanceBeforeNext()))
    return false;
  advanceToNext();
  return true;
}

bool GCNDownwardRPTracker::advance(MachineBasicBlock::const_iterator End) {
  while (NextMI != End)
    if (!advance()) return false;
  return true;
}

bool GCNDownwardRPTracker::advance(MachineBasicBlock::const_iterator Begin,
                                   MachineBasicBlock::const_iterator End,
                                   const LiveRegSet *LiveRegsCopy) {
  reset(*Begin, LiveRegsCopy);
  return advance(End);
}

#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD
static void reportMismatch(const GCNRPTracker::LiveRegSet &LISLR,
                           const GCNRPTracker::LiveRegSet &TrackedLR,
                           const TargetRegisterInfo *TRI) {
  for (auto const &P : TrackedLR) {
    auto I = LISLR.find(P.first);
    if (I == LISLR.end()) {
      dbgs() << "  " << printReg(P.first, TRI)
             << ":L" << PrintLaneMask(P.second)
             << " isn't found in LIS reported set\n";
    }
    else if (I->second != P.second) {
      dbgs() << "  " << printReg(P.first, TRI)
        << " masks doesn't match: LIS reported "
        << PrintLaneMask(I->second)
        << ", tracked "
        << PrintLaneMask(P.second)
        << '\n';
    }
  }
  for (auto const &P : LISLR) {
    auto I = TrackedLR.find(P.first);
    if (I == TrackedLR.end()) {
      dbgs() << "  " << printReg(P.first, TRI)
             << ":L" << PrintLaneMask(P.second)
             << " isn't found in tracked set\n";
    }
  }
}

bool GCNUpwardRPTracker::isValid() const {
  const auto &SI = LIS.getInstructionIndex(*LastTrackedMI).getBaseIndex();
  const auto LISLR = llvm::getLiveRegs(SI, LIS, *MRI);
  const auto &TrackedLR = LiveRegs;

if (!isEqual(LISLR, TrackedLR)) {
    dbgs() << "\nGCNUpwardRPTracker error: Tracked and"
              " LIS reported livesets mismatch:\n";
    printLivesAt(SI, LIS, *MRI);
    reportMismatch(LISLR, TrackedLR, MRI->getTargetRegisterInfo());
    return false;
  }

auto LISPressure = getRegPressure(*MRI, LISLR);
  if (LISPressure != CurPressure) {
    dbgs() << "GCNUpwardRPTracker error: Pressure sets different\nTracked: ";
    CurPressure.print(dbgs());
    dbgs() << "LIS rpt: ";
    LISPressure.print(dbgs());
    return false;
  }
  return true;
}

void GCNRPTracker::printLiveRegs(raw_ostream &OS, const LiveRegSet& LiveRegs,
                                 const MachineRegisterInfo &MRI) {
  const TargetRegisterInfo *TRI = MRI.getTargetRegisterInfo();
  for (unsigned I = 0, E = MRI.getNumVirtRegs(); I != E; ++I) {
    unsigned Reg = Register::index2VirtReg(I);
    auto It = LiveRegs.find(Reg);
    if (It != LiveRegs.end() && It->second.any())
      OS << ' ' << printVRegOrUnit(Reg, TRI) << ':'
         << PrintLaneMask(It->second);
  }
  OS << '\n';
}
#endif

003 File Manager

Editing: GCNRegPressure.cpp

Upload File

Create Folder