llvm-project/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner...

394 lines
13 KiB
C++

//=== lib/CodeGen/GlobalISel/AMDGPUPostLegalizerCombiner.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 pass does combining of machine instructions at the generic MI level,
// after the legalizer.
//
//===----------------------------------------------------------------------===//
#include "AMDGPUTargetMachine.h"
#include "AMDGPULegalizerInfo.h"
#include "llvm/CodeGen/GlobalISel/Combiner.h"
#include "llvm/CodeGen/GlobalISel/CombinerHelper.h"
#include "llvm/CodeGen/GlobalISel/CombinerInfo.h"
#include "llvm/CodeGen/GlobalISel/GISelKnownBits.h"
#include "llvm/CodeGen/GlobalISel/MIPatternMatch.h"
#include "llvm/CodeGen/MachineDominators.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/TargetPassConfig.h"
#include "llvm/Support/Debug.h"
#include "MCTargetDesc/AMDGPUMCTargetDesc.h"
#define DEBUG_TYPE "amdgpu-postlegalizer-combiner"
using namespace llvm;
using namespace MIPatternMatch;
class AMDGPUPostLegalizerCombinerHelper {
protected:
MachineIRBuilder &B;
MachineFunction &MF;
MachineRegisterInfo &MRI;
CombinerHelper &Helper;
public:
AMDGPUPostLegalizerCombinerHelper(MachineIRBuilder &B, CombinerHelper &Helper)
: B(B), MF(B.getMF()), MRI(*B.getMRI()), Helper(Helper){};
struct FMinFMaxLegacyInfo {
Register LHS;
Register RHS;
Register True;
Register False;
CmpInst::Predicate Pred;
};
// TODO: Make sure fmin_legacy/fmax_legacy don't canonicalize
bool matchFMinFMaxLegacy(MachineInstr &MI, FMinFMaxLegacyInfo &Info);
void applySelectFCmpToFMinToFMaxLegacy(MachineInstr &MI,
const FMinFMaxLegacyInfo &Info);
bool matchUCharToFloat(MachineInstr &MI);
void applyUCharToFloat(MachineInstr &MI);
// FIXME: Should be able to have 2 separate matchdatas rather than custom
// struct boilerplate.
struct CvtF32UByteMatchInfo {
Register CvtVal;
unsigned ShiftOffset;
};
bool matchCvtF32UByteN(MachineInstr &MI, CvtF32UByteMatchInfo &MatchInfo);
void applyCvtF32UByteN(MachineInstr &MI,
const CvtF32UByteMatchInfo &MatchInfo);
};
bool AMDGPUPostLegalizerCombinerHelper::matchFMinFMaxLegacy(
MachineInstr &MI, FMinFMaxLegacyInfo &Info) {
// FIXME: Combines should have subtarget predicates, and we shouldn't need
// this here.
if (!MF.getSubtarget<GCNSubtarget>().hasFminFmaxLegacy())
return false;
// FIXME: Type predicate on pattern
if (MRI.getType(MI.getOperand(0).getReg()) != LLT::scalar(32))
return false;
Register Cond = MI.getOperand(1).getReg();
if (!MRI.hasOneNonDBGUse(Cond) ||
!mi_match(Cond, MRI,
m_GFCmp(m_Pred(Info.Pred), m_Reg(Info.LHS), m_Reg(Info.RHS))))
return false;
Info.True = MI.getOperand(2).getReg();
Info.False = MI.getOperand(3).getReg();
if (!(Info.LHS == Info.True && Info.RHS == Info.False) &&
!(Info.LHS == Info.False && Info.RHS == Info.True))
return false;
switch (Info.Pred) {
case CmpInst::FCMP_FALSE:
case CmpInst::FCMP_OEQ:
case CmpInst::FCMP_ONE:
case CmpInst::FCMP_ORD:
case CmpInst::FCMP_UNO:
case CmpInst::FCMP_UEQ:
case CmpInst::FCMP_UNE:
case CmpInst::FCMP_TRUE:
return false;
default:
return true;
}
}
void AMDGPUPostLegalizerCombinerHelper::applySelectFCmpToFMinToFMaxLegacy(
MachineInstr &MI, const FMinFMaxLegacyInfo &Info) {
B.setInstrAndDebugLoc(MI);
auto buildNewInst = [&MI, this](unsigned Opc, Register X, Register Y) {
B.buildInstr(Opc, {MI.getOperand(0)}, {X, Y}, MI.getFlags());
};
switch (Info.Pred) {
case CmpInst::FCMP_ULT:
case CmpInst::FCMP_ULE:
if (Info.LHS == Info.True)
buildNewInst(AMDGPU::G_AMDGPU_FMIN_LEGACY, Info.RHS, Info.LHS);
else
buildNewInst(AMDGPU::G_AMDGPU_FMAX_LEGACY, Info.LHS, Info.RHS);
break;
case CmpInst::FCMP_OLE:
case CmpInst::FCMP_OLT: {
// We need to permute the operands to get the correct NaN behavior. The
// selected operand is the second one based on the failing compare with NaN,
// so permute it based on the compare type the hardware uses.
if (Info.LHS == Info.True)
buildNewInst(AMDGPU::G_AMDGPU_FMIN_LEGACY, Info.LHS, Info.RHS);
else
buildNewInst(AMDGPU::G_AMDGPU_FMAX_LEGACY, Info.RHS, Info.LHS);
break;
}
case CmpInst::FCMP_UGE:
case CmpInst::FCMP_UGT: {
if (Info.LHS == Info.True)
buildNewInst(AMDGPU::G_AMDGPU_FMAX_LEGACY, Info.RHS, Info.LHS);
else
buildNewInst(AMDGPU::G_AMDGPU_FMIN_LEGACY, Info.LHS, Info.RHS);
break;
}
case CmpInst::FCMP_OGT:
case CmpInst::FCMP_OGE: {
if (Info.LHS == Info.True)
buildNewInst(AMDGPU::G_AMDGPU_FMAX_LEGACY, Info.LHS, Info.RHS);
else
buildNewInst(AMDGPU::G_AMDGPU_FMIN_LEGACY, Info.RHS, Info.LHS);
break;
}
default:
llvm_unreachable("predicate should not have matched");
}
MI.eraseFromParent();
}
bool AMDGPUPostLegalizerCombinerHelper::matchUCharToFloat(MachineInstr &MI) {
Register DstReg = MI.getOperand(0).getReg();
// TODO: We could try to match extracting the higher bytes, which would be
// easier if i8 vectors weren't promoted to i32 vectors, particularly after
// types are legalized. v4i8 -> v4f32 is probably the only case to worry
// about in practice.
LLT Ty = MRI.getType(DstReg);
if (Ty == LLT::scalar(32) || Ty == LLT::scalar(16)) {
Register SrcReg = MI.getOperand(1).getReg();
unsigned SrcSize = MRI.getType(SrcReg).getSizeInBits();
assert(SrcSize == 16 || SrcSize == 32 || SrcSize == 64);
const APInt Mask = APInt::getHighBitsSet(SrcSize, SrcSize - 8);
return Helper.getKnownBits()->maskedValueIsZero(SrcReg, Mask);
}
return false;
}
void AMDGPUPostLegalizerCombinerHelper::applyUCharToFloat(MachineInstr &MI) {
B.setInstrAndDebugLoc(MI);
const LLT S32 = LLT::scalar(32);
Register DstReg = MI.getOperand(0).getReg();
Register SrcReg = MI.getOperand(1).getReg();
LLT Ty = MRI.getType(DstReg);
LLT SrcTy = MRI.getType(SrcReg);
if (SrcTy != S32)
SrcReg = B.buildAnyExtOrTrunc(S32, SrcReg).getReg(0);
if (Ty == S32) {
B.buildInstr(AMDGPU::G_AMDGPU_CVT_F32_UBYTE0, {DstReg},
{SrcReg}, MI.getFlags());
} else {
auto Cvt0 = B.buildInstr(AMDGPU::G_AMDGPU_CVT_F32_UBYTE0, {S32},
{SrcReg}, MI.getFlags());
B.buildFPTrunc(DstReg, Cvt0, MI.getFlags());
}
MI.eraseFromParent();
}
bool AMDGPUPostLegalizerCombinerHelper::matchCvtF32UByteN(
MachineInstr &MI, CvtF32UByteMatchInfo &MatchInfo) {
Register SrcReg = MI.getOperand(1).getReg();
// Look through G_ZEXT.
mi_match(SrcReg, MRI, m_GZExt(m_Reg(SrcReg)));
Register Src0;
int64_t ShiftAmt;
bool IsShr = mi_match(SrcReg, MRI, m_GLShr(m_Reg(Src0), m_ICst(ShiftAmt)));
if (IsShr || mi_match(SrcReg, MRI, m_GShl(m_Reg(Src0), m_ICst(ShiftAmt)))) {
const unsigned Offset = MI.getOpcode() - AMDGPU::G_AMDGPU_CVT_F32_UBYTE0;
unsigned ShiftOffset = 8 * Offset;
if (IsShr)
ShiftOffset += ShiftAmt;
else
ShiftOffset -= ShiftAmt;
MatchInfo.CvtVal = Src0;
MatchInfo.ShiftOffset = ShiftOffset;
return ShiftOffset < 32 && ShiftOffset >= 8 && (ShiftOffset % 8) == 0;
}
// TODO: Simplify demanded bits.
return false;
}
void AMDGPUPostLegalizerCombinerHelper::applyCvtF32UByteN(
MachineInstr &MI, const CvtF32UByteMatchInfo &MatchInfo) {
B.setInstrAndDebugLoc(MI);
unsigned NewOpc = AMDGPU::G_AMDGPU_CVT_F32_UBYTE0 + MatchInfo.ShiftOffset / 8;
const LLT S32 = LLT::scalar(32);
Register CvtSrc = MatchInfo.CvtVal;
LLT SrcTy = MRI.getType(MatchInfo.CvtVal);
if (SrcTy != S32) {
assert(SrcTy.isScalar() && SrcTy.getSizeInBits() >= 8);
CvtSrc = B.buildAnyExt(S32, CvtSrc).getReg(0);
}
assert(MI.getOpcode() != NewOpc);
B.buildInstr(NewOpc, {MI.getOperand(0)}, {CvtSrc}, MI.getFlags());
MI.eraseFromParent();
}
class AMDGPUPostLegalizerCombinerHelperState {
protected:
CombinerHelper &Helper;
AMDGPUPostLegalizerCombinerHelper &PostLegalizerHelper;
public:
AMDGPUPostLegalizerCombinerHelperState(
CombinerHelper &Helper,
AMDGPUPostLegalizerCombinerHelper &PostLegalizerHelper)
: Helper(Helper), PostLegalizerHelper(PostLegalizerHelper) {}
};
#define AMDGPUPOSTLEGALIZERCOMBINERHELPER_GENCOMBINERHELPER_DEPS
#include "AMDGPUGenPostLegalizeGICombiner.inc"
#undef AMDGPUPOSTLEGALIZERCOMBINERHELPER_GENCOMBINERHELPER_DEPS
namespace {
#define AMDGPUPOSTLEGALIZERCOMBINERHELPER_GENCOMBINERHELPER_H
#include "AMDGPUGenPostLegalizeGICombiner.inc"
#undef AMDGPUPOSTLEGALIZERCOMBINERHELPER_GENCOMBINERHELPER_H
class AMDGPUPostLegalizerCombinerInfo final : public CombinerInfo {
GISelKnownBits *KB;
MachineDominatorTree *MDT;
public:
AMDGPUGenPostLegalizerCombinerHelperRuleConfig GeneratedRuleCfg;
AMDGPUPostLegalizerCombinerInfo(bool EnableOpt, bool OptSize, bool MinSize,
const AMDGPULegalizerInfo *LI,
GISelKnownBits *KB, MachineDominatorTree *MDT)
: CombinerInfo(/*AllowIllegalOps*/ false, /*ShouldLegalizeIllegal*/ true,
/*LegalizerInfo*/ LI, EnableOpt, OptSize, MinSize),
KB(KB), MDT(MDT) {
if (!GeneratedRuleCfg.parseCommandLineOption())
report_fatal_error("Invalid rule identifier");
}
bool combine(GISelChangeObserver &Observer, MachineInstr &MI,
MachineIRBuilder &B) const override;
};
bool AMDGPUPostLegalizerCombinerInfo::combine(GISelChangeObserver &Observer,
MachineInstr &MI,
MachineIRBuilder &B) const {
CombinerHelper Helper(Observer, B, KB, MDT, LInfo);
AMDGPUPostLegalizerCombinerHelper PostLegalizerHelper(B, Helper);
AMDGPUGenPostLegalizerCombinerHelper Generated(GeneratedRuleCfg, Helper,
PostLegalizerHelper);
if (Generated.tryCombineAll(Observer, MI, B))
return true;
switch (MI.getOpcode()) {
case TargetOpcode::G_SHL:
case TargetOpcode::G_LSHR:
case TargetOpcode::G_ASHR:
// On some subtargets, 64-bit shift is a quarter rate instruction. In the
// common case, splitting this into a move and a 32-bit shift is faster and
// the same code size.
return Helper.tryCombineShiftToUnmerge(MI, 32);
}
return false;
}
#define AMDGPUPOSTLEGALIZERCOMBINERHELPER_GENCOMBINERHELPER_CPP
#include "AMDGPUGenPostLegalizeGICombiner.inc"
#undef AMDGPUPOSTLEGALIZERCOMBINERHELPER_GENCOMBINERHELPER_CPP
// Pass boilerplate
// ================
class AMDGPUPostLegalizerCombiner : public MachineFunctionPass {
public:
static char ID;
AMDGPUPostLegalizerCombiner(bool IsOptNone = false);
StringRef getPassName() const override {
return "AMDGPUPostLegalizerCombiner";
}
bool runOnMachineFunction(MachineFunction &MF) override;
void getAnalysisUsage(AnalysisUsage &AU) const override;
private:
bool IsOptNone;
};
} // end anonymous namespace
void AMDGPUPostLegalizerCombiner::getAnalysisUsage(AnalysisUsage &AU) const {
AU.addRequired<TargetPassConfig>();
AU.setPreservesCFG();
getSelectionDAGFallbackAnalysisUsage(AU);
AU.addRequired<GISelKnownBitsAnalysis>();
AU.addPreserved<GISelKnownBitsAnalysis>();
if (!IsOptNone) {
AU.addRequired<MachineDominatorTree>();
AU.addPreserved<MachineDominatorTree>();
}
MachineFunctionPass::getAnalysisUsage(AU);
}
AMDGPUPostLegalizerCombiner::AMDGPUPostLegalizerCombiner(bool IsOptNone)
: MachineFunctionPass(ID), IsOptNone(IsOptNone) {
initializeAMDGPUPostLegalizerCombinerPass(*PassRegistry::getPassRegistry());
}
bool AMDGPUPostLegalizerCombiner::runOnMachineFunction(MachineFunction &MF) {
if (MF.getProperties().hasProperty(
MachineFunctionProperties::Property::FailedISel))
return false;
auto *TPC = &getAnalysis<TargetPassConfig>();
const Function &F = MF.getFunction();
bool EnableOpt =
MF.getTarget().getOptLevel() != CodeGenOpt::None && !skipFunction(F);
const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
const AMDGPULegalizerInfo *LI
= static_cast<const AMDGPULegalizerInfo *>(ST.getLegalizerInfo());
GISelKnownBits *KB = &getAnalysis<GISelKnownBitsAnalysis>().get(MF);
MachineDominatorTree *MDT =
IsOptNone ? nullptr : &getAnalysis<MachineDominatorTree>();
AMDGPUPostLegalizerCombinerInfo PCInfo(EnableOpt, F.hasOptSize(),
F.hasMinSize(), LI, KB, MDT);
Combiner C(PCInfo, TPC);
return C.combineMachineInstrs(MF, /*CSEInfo*/ nullptr);
}
char AMDGPUPostLegalizerCombiner::ID = 0;
INITIALIZE_PASS_BEGIN(AMDGPUPostLegalizerCombiner, DEBUG_TYPE,
"Combine AMDGPU machine instrs after legalization",
false, false)
INITIALIZE_PASS_DEPENDENCY(TargetPassConfig)
INITIALIZE_PASS_DEPENDENCY(GISelKnownBitsAnalysis)
INITIALIZE_PASS_END(AMDGPUPostLegalizerCombiner, DEBUG_TYPE,
"Combine AMDGPU machine instrs after legalization", false,
false)
namespace llvm {
FunctionPass *createAMDGPUPostLegalizeCombiner(bool IsOptNone) {
return new AMDGPUPostLegalizerCombiner(IsOptNone);
}
} // end namespace llvm