1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-23 04:52:54 +02:00
llvm-mirror/utils/TableGen/SubtargetEmitter.cpp

1829 lines
66 KiB
C++
Raw Normal View History

//===- SubtargetEmitter.cpp - Generate subtarget enumerations -------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This tablegen backend emits subtarget enumerations.
//
//===----------------------------------------------------------------------===//
#include "CodeGenTarget.h"
#include "CodeGenSchedule.h"
#include "PredicateExpander.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/MC/MCInstrItineraries.h"
#include "llvm/MC/MCSchedule.h"
#include "llvm/MC/SubtargetFeature.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/Format.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/TableGen/Error.h"
#include "llvm/TableGen/Record.h"
#include "llvm/TableGen/TableGenBackend.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <iterator>
#include <map>
#include <string>
#include <vector>
using namespace llvm;
#define DEBUG_TYPE "subtarget-emitter"
namespace {
class SubtargetEmitter {
// Each processor has a SchedClassDesc table with an entry for each SchedClass.
// The SchedClassDesc table indexes into a global write resource table, write
// latency table, and read advance table.
struct SchedClassTables {
std::vector<std::vector<MCSchedClassDesc>> ProcSchedClasses;
std::vector<MCWriteProcResEntry> WriteProcResources;
std::vector<MCWriteLatencyEntry> WriteLatencies;
std::vector<std::string> WriterNames;
std::vector<MCReadAdvanceEntry> ReadAdvanceEntries;
// Reserve an invalid entry at index 0
SchedClassTables() {
ProcSchedClasses.resize(1);
WriteProcResources.resize(1);
WriteLatencies.resize(1);
WriterNames.push_back("InvalidWrite");
ReadAdvanceEntries.resize(1);
}
};
struct LessWriteProcResources {
bool operator()(const MCWriteProcResEntry &LHS,
const MCWriteProcResEntry &RHS) {
return LHS.ProcResourceIdx < RHS.ProcResourceIdx;
}
};
const CodeGenTarget &TGT;
RecordKeeper &Records;
CodeGenSchedModels &SchedModels;
std::string Target;
void Enumeration(raw_ostream &OS);
unsigned FeatureKeyValues(raw_ostream &OS);
unsigned CPUKeyValues(raw_ostream &OS);
void FormItineraryStageString(const std::string &Names,
Record *ItinData, std::string &ItinString,
unsigned &NStages);
void FormItineraryOperandCycleString(Record *ItinData, std::string &ItinString,
unsigned &NOperandCycles);
void FormItineraryBypassString(const std::string &Names,
Record *ItinData,
std::string &ItinString, unsigned NOperandCycles);
void EmitStageAndOperandCycleData(raw_ostream &OS,
std::vector<std::vector<InstrItinerary>>
&ProcItinLists);
void EmitItineraries(raw_ostream &OS,
std::vector<std::vector<InstrItinerary>>
&ProcItinLists);
unsigned EmitRegisterFileTables(const CodeGenProcModel &ProcModel,
raw_ostream &OS);
void EmitExtraProcessorInfo(const CodeGenProcModel &ProcModel,
raw_ostream &OS);
void EmitProcessorProp(raw_ostream &OS, const Record *R, StringRef Name,
char Separator);
void EmitProcessorResourceSubUnits(const CodeGenProcModel &ProcModel,
raw_ostream &OS);
void EmitProcessorResources(const CodeGenProcModel &ProcModel,
raw_ostream &OS);
Record *FindWriteResources(const CodeGenSchedRW &SchedWrite,
const CodeGenProcModel &ProcModel);
Record *FindReadAdvance(const CodeGenSchedRW &SchedRead,
const CodeGenProcModel &ProcModel);
void ExpandProcResources(RecVec &PRVec, std::vector<int64_t> &Cycles,
const CodeGenProcModel &ProcModel);
void GenSchedClassTables(const CodeGenProcModel &ProcModel,
SchedClassTables &SchedTables);
void EmitSchedClassTables(SchedClassTables &SchedTables, raw_ostream &OS);
void EmitProcessorModels(raw_ostream &OS);
void EmitProcessorLookup(raw_ostream &OS);
void EmitSchedModelHelpers(const std::string &ClassName, raw_ostream &OS);
void emitSchedModelHelpersImpl(raw_ostream &OS,
bool OnlyExpandMCInstPredicates = false);
void emitGenMCSubtargetInfo(raw_ostream &OS);
void EmitSchedModel(raw_ostream &OS);
void EmitHwModeCheck(const std::string &ClassName, raw_ostream &OS);
void ParseFeaturesFunction(raw_ostream &OS, unsigned NumFeatures,
unsigned NumProcs);
public:
SubtargetEmitter(RecordKeeper &R, CodeGenTarget &TGT)
: TGT(TGT), Records(R), SchedModels(TGT.getSchedModels()),
Target(TGT.getName()) {}
void run(raw_ostream &o);
};
} // end anonymous namespace
//
// Enumeration - Emit the specified class as an enumeration.
//
void SubtargetEmitter::Enumeration(raw_ostream &OS) {
2005-10-28 17:20:43 +02:00
// Get all records of class and sort
std::vector<Record*> DefList =
Records.getAllDerivedDefinitions("SubtargetFeature");
llvm::sort(DefList.begin(), DefList.end(), LessRecord());
unsigned N = DefList.size();
if (N == 0)
return;
if (N > MAX_SUBTARGET_FEATURES)
PrintFatalError("Too many subtarget features! Bump MAX_SUBTARGET_FEATURES.");
OS << "namespace " << Target << " {\n";
// Open enumeration.
OS << "enum {\n";
2011-04-01 03:56:55 +02:00
// For each record
for (unsigned i = 0; i < N; ++i) {
// Next record
Record *Def = DefList[i];
// Get and emit name
OS << " " << Def->getName() << " = " << i << ",\n";
}
// Close enumeration and namespace
OS << "};\n";
OS << "} // end namespace " << Target << "\n";
}
//
// FeatureKeyValues - Emit data of all the subtarget features. Used by the
// command line.
//
unsigned SubtargetEmitter::FeatureKeyValues(raw_ostream &OS) {
2005-10-28 17:20:43 +02:00
// Gather and sort all the features
std::vector<Record*> FeatureList =
Records.getAllDerivedDefinitions("SubtargetFeature");
if (FeatureList.empty())
return 0;
llvm::sort(FeatureList.begin(), FeatureList.end(), LessRecordFieldName());
2005-10-28 17:20:43 +02:00
// Begin feature table
OS << "// Sorted (by key) array of values for CPU features.\n"
<< "extern const llvm::SubtargetFeatureKV " << Target
<< "FeatureKV[] = {\n";
2011-04-01 03:56:55 +02:00
2005-10-28 17:20:43 +02:00
// For each feature
unsigned NumFeatures = 0;
for (unsigned i = 0, N = FeatureList.size(); i < N; ++i) {
// Next feature
Record *Feature = FeatureList[i];
StringRef Name = Feature->getName();
StringRef CommandLineName = Feature->getValueAsString("Name");
StringRef Desc = Feature->getValueAsString("Desc");
2011-04-01 03:56:55 +02:00
if (CommandLineName.empty()) continue;
2011-04-01 03:56:55 +02:00
// Emit as { "feature", "description", { featureEnum }, { i1 , i2 , ... , in } }
OS << " { "
<< "\"" << CommandLineName << "\", "
<< "\"" << Desc << "\", "
<< "{ " << Target << "::" << Name << " }, ";
RecVec ImpliesList = Feature->getValueAsListOfDefs("Implies");
2011-04-01 03:56:55 +02:00
OS << "{";
for (unsigned j = 0, M = ImpliesList.size(); j < M;) {
OS << " " << Target << "::" << ImpliesList[j]->getName();
if (++j < M) OS << ",";
}
OS << " } },\n";
++NumFeatures;
}
2011-04-01 03:56:55 +02:00
2005-10-28 17:20:43 +02:00
// End feature table
OS << "};\n";
return NumFeatures;
}
//
// CPUKeyValues - Emit data of all the subtarget processors. Used by command
// line.
//
unsigned SubtargetEmitter::CPUKeyValues(raw_ostream &OS) {
2005-10-28 17:20:43 +02:00
// Gather and sort processor information
std::vector<Record*> ProcessorList =
Records.getAllDerivedDefinitions("Processor");
llvm::sort(ProcessorList.begin(), ProcessorList.end(), LessRecordFieldName());
2005-10-28 17:20:43 +02:00
// Begin processor table
OS << "// Sorted (by key) array of values for CPU subtype.\n"
<< "extern const llvm::SubtargetFeatureKV " << Target
<< "SubTypeKV[] = {\n";
2011-04-01 03:56:55 +02:00
2005-10-28 17:20:43 +02:00
// For each processor
for (Record *Processor : ProcessorList) {
StringRef Name = Processor->getValueAsString("Name");
RecVec FeatureList = Processor->getValueAsListOfDefs("Features");
2011-04-01 03:56:55 +02:00
// Emit as { "cpu", "description", { f1 , f2 , ... fn } },
OS << " { "
<< "\"" << Name << "\", "
<< "\"Select the " << Name << " processor\", ";
2011-04-01 03:56:55 +02:00
OS << "{";
for (unsigned j = 0, M = FeatureList.size(); j < M;) {
OS << " " << Target << "::" << FeatureList[j]->getName();
if (++j < M) OS << ",";
}
// The { } is for the "implies" section of this data structure.
OS << " }, { } },\n";
}
2011-04-01 03:56:55 +02:00
2005-10-28 17:20:43 +02:00
// End processor table
OS << "};\n";
return ProcessorList.size();
}
//
// FormItineraryStageString - Compose a string containing the stage
// data initialization for the specified itinerary. N is the number
// of stages.
//
void SubtargetEmitter::FormItineraryStageString(const std::string &Name,
Record *ItinData,
std::string &ItinString,
unsigned &NStages) {
// Get states list
RecVec StageList = ItinData->getValueAsListOfDefs("Stages");
2005-10-28 17:20:43 +02:00
// For each stage
unsigned N = NStages = StageList.size();
for (unsigned i = 0; i < N;) {
// Next stage
const Record *Stage = StageList[i];
2011-04-01 03:56:55 +02:00
// Form string as ,{ cycles, u1 | u2 | ... | un, timeinc, kind }
int Cycles = Stage->getValueAsInt("Cycles");
ItinString += " { " + itostr(Cycles) + ", ";
2011-04-01 03:56:55 +02:00
// Get unit list
RecVec UnitList = Stage->getValueAsListOfDefs("Units");
2011-04-01 03:56:55 +02:00
2005-10-28 17:20:43 +02:00
// For each unit
for (unsigned j = 0, M = UnitList.size(); j < M;) {
// Add name and bitwise or
ItinString += Name + "FU::" + UnitList[j]->getName().str();
if (++j < M) ItinString += " | ";
}
2011-04-01 03:56:55 +02:00
int TimeInc = Stage->getValueAsInt("TimeInc");
ItinString += ", " + itostr(TimeInc);
int Kind = Stage->getValueAsInt("Kind");
ItinString += ", (llvm::InstrStage::ReservationKinds)" + itostr(Kind);
2005-10-28 17:20:43 +02:00
// Close off stage
ItinString += " }";
if (++i < N) ItinString += ", ";
}
}
//
// FormItineraryOperandCycleString - Compose a string containing the
// operand cycle initialization for the specified itinerary. N is the
// number of operands that has cycles specified.
//
void SubtargetEmitter::FormItineraryOperandCycleString(Record *ItinData,
std::string &ItinString, unsigned &NOperandCycles) {
// Get operand cycle list
std::vector<int64_t> OperandCycleList =
ItinData->getValueAsListOfInts("OperandCycles");
// For each operand cycle
unsigned N = NOperandCycles = OperandCycleList.size();
for (unsigned i = 0; i < N;) {
// Next operand cycle
const int OCycle = OperandCycleList[i];
2011-04-01 03:56:55 +02:00
ItinString += " " + itostr(OCycle);
if (++i < N) ItinString += ", ";
}
}
void SubtargetEmitter::FormItineraryBypassString(const std::string &Name,
Record *ItinData,
std::string &ItinString,
unsigned NOperandCycles) {
RecVec BypassList = ItinData->getValueAsListOfDefs("Bypasses");
unsigned N = BypassList.size();
unsigned i = 0;
for (; i < N;) {
ItinString += Name + "Bypass::" + BypassList[i]->getName().str();
if (++i < NOperandCycles) ItinString += ", ";
}
for (; i < NOperandCycles;) {
ItinString += " 0";
if (++i < NOperandCycles) ItinString += ", ";
}
}
//
// EmitStageAndOperandCycleData - Generate unique itinerary stages and operand
// cycle tables. Create a list of InstrItinerary objects (ProcItinLists) indexed
// by CodeGenSchedClass::Index.
//
void SubtargetEmitter::
EmitStageAndOperandCycleData(raw_ostream &OS,
std::vector<std::vector<InstrItinerary>>
&ProcItinLists) {
// Multiple processor models may share an itinerary record. Emit it once.
SmallPtrSet<Record*, 8> ItinsDefSet;
// Emit functional units for all the itineraries.
for (const CodeGenProcModel &ProcModel : SchedModels.procModels()) {
if (!ItinsDefSet.insert(ProcModel.ItinsDef).second)
continue;
RecVec FUs = ProcModel.ItinsDef->getValueAsListOfDefs("FU");
if (FUs.empty())
continue;
StringRef Name = ProcModel.ItinsDef->getName();
OS << "\n// Functional units for \"" << Name << "\"\n"
<< "namespace " << Name << "FU {\n";
for (unsigned j = 0, FUN = FUs.size(); j < FUN; ++j)
OS << " const unsigned " << FUs[j]->getName()
<< " = 1 << " << j << ";\n";
OS << "} // end namespace " << Name << "FU\n";
RecVec BPs = ProcModel.ItinsDef->getValueAsListOfDefs("BP");
if (!BPs.empty()) {
OS << "\n// Pipeline forwarding paths for itineraries \"" << Name
<< "\"\n" << "namespace " << Name << "Bypass {\n";
OS << " const unsigned NoBypass = 0;\n";
for (unsigned j = 0, BPN = BPs.size(); j < BPN; ++j)
OS << " const unsigned " << BPs[j]->getName()
<< " = 1 << " << j << ";\n";
OS << "} // end namespace " << Name << "Bypass\n";
}
}
2005-10-28 17:20:43 +02:00
// Begin stages table
std::string StageTable = "\nextern const llvm::InstrStage " + Target +
"Stages[] = {\n";
StageTable += " { 0, 0, 0, llvm::InstrStage::Required }, // No itinerary\n";
2011-04-01 03:56:55 +02:00
// Begin operand cycle table
std::string OperandCycleTable = "extern const unsigned " + Target +
"OperandCycles[] = {\n";
OperandCycleTable += " 0, // No itinerary\n";
// Begin pipeline bypass table
std::string BypassTable = "extern const unsigned " + Target +
2012-07-07 05:59:48 +02:00
"ForwardingPaths[] = {\n";
BypassTable += " 0, // No itinerary\n";
2011-04-01 03:56:55 +02:00
// For each Itinerary across all processors, add a unique entry to the stages,
2017-05-08 17:33:08 +02:00
// operand cycles, and pipeline bypass tables. Then add the new Itinerary
// object with computed offsets to the ProcItinLists result.
unsigned StageCount = 1, OperandCycleCount = 1;
std::map<std::string, unsigned> ItinStageMap, ItinOperandMap;
for (const CodeGenProcModel &ProcModel : SchedModels.procModels()) {
// Add process itinerary to the list.
ProcItinLists.resize(ProcItinLists.size()+1);
// If this processor defines no itineraries, then leave the itinerary list
// empty.
std::vector<InstrItinerary> &ItinList = ProcItinLists.back();
if (!ProcModel.hasItineraries())
continue;
StringRef Name = ProcModel.ItinsDef->getName();
2011-04-01 03:56:55 +02:00
ItinList.resize(SchedModels.numInstrSchedClasses());
assert(ProcModel.ItinDefList.size() == ItinList.size() && "bad Itins");
for (unsigned SchedClassIdx = 0, SchedClassEnd = ItinList.size();
SchedClassIdx < SchedClassEnd; ++SchedClassIdx) {
// Next itinerary data
Record *ItinData = ProcModel.ItinDefList[SchedClassIdx];
2011-04-01 03:56:55 +02:00
2005-10-28 17:20:43 +02:00
// Get string and stage count
std::string ItinStageString;
unsigned NStages = 0;
if (ItinData)
FormItineraryStageString(Name, ItinData, ItinStageString, NStages);
// Get string and operand cycle count
std::string ItinOperandCycleString;
unsigned NOperandCycles = 0;
std::string ItinBypassString;
if (ItinData) {
FormItineraryOperandCycleString(ItinData, ItinOperandCycleString,
NOperandCycles);
FormItineraryBypassString(Name, ItinData, ItinBypassString,
NOperandCycles);
}
// Check to see if stage already exists and create if it doesn't
uint16_t FindStage = 0;
if (NStages > 0) {
FindStage = ItinStageMap[ItinStageString];
if (FindStage == 0) {
// Emit as { cycles, u1 | u2 | ... | un, timeinc }, // indices
StageTable += ItinStageString + ", // " + itostr(StageCount);
if (NStages > 1)
StageTable += "-" + itostr(StageCount + NStages - 1);
StageTable += "\n";
// Record Itin class number.
ItinStageMap[ItinStageString] = FindStage = StageCount;
StageCount += NStages;
}
}
2011-04-01 03:56:55 +02:00
// Check to see if operand cycle already exists and create if it doesn't
uint16_t FindOperandCycle = 0;
if (NOperandCycles > 0) {
std::string ItinOperandString = ItinOperandCycleString+ItinBypassString;
FindOperandCycle = ItinOperandMap[ItinOperandString];
if (FindOperandCycle == 0) {
// Emit as cycle, // index
OperandCycleTable += ItinOperandCycleString + ", // ";
std::string OperandIdxComment = itostr(OperandCycleCount);
if (NOperandCycles > 1)
OperandIdxComment += "-"
+ itostr(OperandCycleCount + NOperandCycles - 1);
OperandCycleTable += OperandIdxComment + "\n";
// Record Itin class number.
2011-04-01 03:56:55 +02:00
ItinOperandMap[ItinOperandCycleString] =
FindOperandCycle = OperandCycleCount;
// Emit as bypass, // index
BypassTable += ItinBypassString + ", // " + OperandIdxComment + "\n";
OperandCycleCount += NOperandCycles;
}
}
2011-04-01 03:56:55 +02:00
// Set up itinerary as location and location + stage count
int16_t NumUOps = ItinData ? ItinData->getValueAsInt("NumMicroOps") : 0;
InstrItinerary Intinerary = {
NumUOps,
FindStage,
uint16_t(FindStage + NStages),
FindOperandCycle,
uint16_t(FindOperandCycle + NOperandCycles),
};
2005-10-28 17:20:43 +02:00
// Inject - empty slots will be 0, 0
ItinList[SchedClassIdx] = Intinerary;
}
}
// Closing stage
StageTable += " { 0, 0, 0, llvm::InstrStage::Required } // End stages\n";
StageTable += "};\n";
// Closing operand cycles
OperandCycleTable += " 0 // End operand cycles\n";
OperandCycleTable += "};\n";
BypassTable += " 0 // End bypass tables\n";
BypassTable += "};\n";
// Emit tables.
OS << StageTable;
OS << OperandCycleTable;
OS << BypassTable;
}
//
// EmitProcessorData - Generate data for processor itineraries that were
// computed during EmitStageAndOperandCycleData(). ProcItinLists lists all
// Itineraries for each processor. The Itinerary lists are indexed on
// CodeGenSchedClass::Index.
//
void SubtargetEmitter::
EmitItineraries(raw_ostream &OS,
std::vector<std::vector<InstrItinerary>> &ProcItinLists) {
// Multiple processor models may share an itinerary record. Emit it once.
SmallPtrSet<Record*, 8> ItinsDefSet;
// For each processor's machine model
std::vector<std::vector<InstrItinerary>>::iterator
ProcItinListsIter = ProcItinLists.begin();
for (CodeGenSchedModels::ProcIter PI = SchedModels.procModelBegin(),
PE = SchedModels.procModelEnd(); PI != PE; ++PI, ++ProcItinListsIter) {
Record *ItinsDef = PI->ItinsDef;
if (!ItinsDefSet.insert(ItinsDef).second)
continue;
// Get the itinerary list for the processor.
assert(ProcItinListsIter != ProcItinLists.end() && "bad iterator");
std::vector<InstrItinerary> &ItinList = *ProcItinListsIter;
// Empty itineraries aren't referenced anywhere in the tablegen output
// so don't emit them.
if (ItinList.empty())
continue;
OS << "\n";
OS << "static const llvm::InstrItinerary ";
2011-04-01 03:56:55 +02:00
// Begin processor itinerary table
OS << ItinsDef->getName() << "[] = {\n";
// For each itinerary class in CodeGenSchedClass::Index order.
for (unsigned j = 0, M = ItinList.size(); j < M; ++j) {
InstrItinerary &Intinerary = ItinList[j];
// Emit Itinerary in the form of
// { firstStage, lastStage, firstCycle, lastCycle } // index
OS << " { " <<
Intinerary.NumMicroOps << ", " <<
Intinerary.FirstStage << ", " <<
Intinerary.LastStage << ", " <<
Intinerary.FirstOperandCycle << ", " <<
Intinerary.LastOperandCycle << " }" <<
", // " << j << " " << SchedModels.getSchedClass(j).Name << "\n";
}
// End processor itinerary table
OS << " { 0, uint16_t(~0U), uint16_t(~0U), uint16_t(~0U), uint16_t(~0U) }"
"// end marker\n";
OS << "};\n";
}
}
// Emit either the value defined in the TableGen Record, or the default
// value defined in the C++ header. The Record is null if the processor does not
// define a model.
void SubtargetEmitter::EmitProcessorProp(raw_ostream &OS, const Record *R,
StringRef Name, char Separator) {
OS << " ";
int V = R ? R->getValueAsInt(Name) : -1;
if (V >= 0)
OS << V << Separator << " // " << Name;
else
OS << "MCSchedModel::Default" << Name << Separator;
OS << '\n';
}
void SubtargetEmitter::EmitProcessorResourceSubUnits(
const CodeGenProcModel &ProcModel, raw_ostream &OS) {
OS << "\nstatic const unsigned " << ProcModel.ModelName
<< "ProcResourceSubUnits[] = {\n"
<< " 0, // Invalid\n";
for (unsigned i = 0, e = ProcModel.ProcResourceDefs.size(); i < e; ++i) {
Record *PRDef = ProcModel.ProcResourceDefs[i];
if (!PRDef->isSubClassOf("ProcResGroup"))
continue;
RecVec ResUnits = PRDef->getValueAsListOfDefs("Resources");
for (Record *RUDef : ResUnits) {
Record *const RU =
SchedModels.findProcResUnits(RUDef, ProcModel, PRDef->getLoc());
for (unsigned J = 0; J < RU->getValueAsInt("NumUnits"); ++J) {
OS << " " << ProcModel.getProcResourceIdx(RU) << ", ";
}
}
OS << " // " << PRDef->getName() << "\n";
}
OS << "};\n";
}
static void EmitRetireControlUnitInfo(const CodeGenProcModel &ProcModel,
raw_ostream &OS) {
int64_t ReorderBufferSize = 0, MaxRetirePerCycle = 0;
if (Record *RCU = ProcModel.RetireControlUnit) {
ReorderBufferSize =
std::max(ReorderBufferSize, RCU->getValueAsInt("ReorderBufferSize"));
MaxRetirePerCycle =
std::max(MaxRetirePerCycle, RCU->getValueAsInt("MaxRetirePerCycle"));
}
OS << ReorderBufferSize << ", // ReorderBufferSize\n ";
OS << MaxRetirePerCycle << ", // MaxRetirePerCycle\n ";
}
static void EmitRegisterFileInfo(const CodeGenProcModel &ProcModel,
unsigned NumRegisterFiles,
unsigned NumCostEntries, raw_ostream &OS) {
if (NumRegisterFiles)
OS << ProcModel.ModelName << "RegisterFiles,\n " << (1 + NumRegisterFiles);
else
OS << "nullptr,\n 0";
OS << ", // Number of register files.\n ";
if (NumCostEntries)
OS << ProcModel.ModelName << "RegisterCosts,\n ";
else
OS << "nullptr,\n ";
OS << NumCostEntries << ", // Number of register cost entries.\n";
}
unsigned
SubtargetEmitter::EmitRegisterFileTables(const CodeGenProcModel &ProcModel,
raw_ostream &OS) {
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca This patch allows the description of register files in processor scheduling models. This addresses PR36662. A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td. Targets can optionally describe register files for their processors using that class. In particular, class RegisterFile allows to specify: - The total number of physical registers. - Which target registers are accessible through the register file. - The cost of allocating a register at register renaming stage. Example (from this patch - see file X86/X86ScheduleBtVer2.td) def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]> Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar (btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM register definitions only cost 1 physical register. The syntax allows to specify an empty set of register classes. An empty set of register classes means: this register file models all the registers specified by the Target. For each register class, users can specify an optional register cost. By default, register costs default to 1. A value of 0 for the number of physical registers means: "this register file has an unbounded number of physical registers". This patch is structured in two parts. * Part 1 - MC/Tablegen * A first part adds the tablegen definition of RegisterFile, and teaches the SubtargetEmitter how to emit information related to register files. Information about register files is accessible through an instance of MCExtraProcessorInfo. The idea behind this design is to logically partition the processor description which is only used by external tools (like llvm-mca) from the processor information used by the llvm machine schedulers. I think that this design would make easier for targets to get rid of the extra processor information if they don't want it. * Part 2 - llvm-mca related * The second part of this patch is related to changes to llvm-mca. The main differences are: 1) class RegisterFile now needs to take into account the "cost of a register" when allocating physical registers at register renaming stage. 2) Point 1. triggered a minor refactoring which lef to the removal of the "maximum 32 register files" restriction. 3) The BackendStatistics view has been updated so that we can print out extra details related to each register file implemented by the processor. The effect of point 3. is also visible in tests register-files-[1..5].s. Differential Revision: https://reviews.llvm.org/D44980 llvm-svn: 329067
2018-04-03 15:36:24 +02:00
if (llvm::all_of(ProcModel.RegisterFiles, [](const CodeGenRegisterFile &RF) {
return RF.hasDefaultCosts();
}))
return 0;
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca This patch allows the description of register files in processor scheduling models. This addresses PR36662. A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td. Targets can optionally describe register files for their processors using that class. In particular, class RegisterFile allows to specify: - The total number of physical registers. - Which target registers are accessible through the register file. - The cost of allocating a register at register renaming stage. Example (from this patch - see file X86/X86ScheduleBtVer2.td) def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]> Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar (btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM register definitions only cost 1 physical register. The syntax allows to specify an empty set of register classes. An empty set of register classes means: this register file models all the registers specified by the Target. For each register class, users can specify an optional register cost. By default, register costs default to 1. A value of 0 for the number of physical registers means: "this register file has an unbounded number of physical registers". This patch is structured in two parts. * Part 1 - MC/Tablegen * A first part adds the tablegen definition of RegisterFile, and teaches the SubtargetEmitter how to emit information related to register files. Information about register files is accessible through an instance of MCExtraProcessorInfo. The idea behind this design is to logically partition the processor description which is only used by external tools (like llvm-mca) from the processor information used by the llvm machine schedulers. I think that this design would make easier for targets to get rid of the extra processor information if they don't want it. * Part 2 - llvm-mca related * The second part of this patch is related to changes to llvm-mca. The main differences are: 1) class RegisterFile now needs to take into account the "cost of a register" when allocating physical registers at register renaming stage. 2) Point 1. triggered a minor refactoring which lef to the removal of the "maximum 32 register files" restriction. 3) The BackendStatistics view has been updated so that we can print out extra details related to each register file implemented by the processor. The effect of point 3. is also visible in tests register-files-[1..5].s. Differential Revision: https://reviews.llvm.org/D44980 llvm-svn: 329067
2018-04-03 15:36:24 +02:00
// Print the RegisterCost table first.
OS << "\n// {RegisterClassID, Register Cost}\n";
OS << "static const llvm::MCRegisterCostEntry " << ProcModel.ModelName
<< "RegisterCosts"
<< "[] = {\n";
for (const CodeGenRegisterFile &RF : ProcModel.RegisterFiles) {
// Skip register files with a default cost table.
if (RF.hasDefaultCosts())
continue;
// Add entries to the cost table.
for (const CodeGenRegisterCost &RC : RF.Costs) {
OS << " { ";
Record *Rec = RC.RCDef;
if (Rec->getValue("Namespace"))
OS << Rec->getValueAsString("Namespace") << "::";
OS << Rec->getName() << "RegClassID, " << RC.Cost << "},\n";
}
}
OS << "};\n";
// Now generate a table with register file info.
OS << "\n // {Name, #PhysRegs, #CostEntries, IndexToCostTbl}\n";
OS << "static const llvm::MCRegisterFileDesc " << ProcModel.ModelName
<< "RegisterFiles"
<< "[] = {\n"
<< " { \"InvalidRegisterFile\", 0, 0, 0 },\n";
unsigned CostTblIndex = 0;
for (const CodeGenRegisterFile &RD : ProcModel.RegisterFiles) {
OS << " { ";
OS << '"' << RD.Name << '"' << ", " << RD.NumPhysRegs << ", ";
unsigned NumCostEntries = RD.Costs.size();
OS << NumCostEntries << ", " << CostTblIndex << "},\n";
CostTblIndex += NumCostEntries;
}
OS << "};\n";
return CostTblIndex;
}
static bool EmitPfmIssueCountersTable(const CodeGenProcModel &ProcModel,
raw_ostream &OS) {
unsigned NumCounterDefs = 1 + ProcModel.ProcResourceDefs.size();
std::vector<const Record *> CounterDefs(NumCounterDefs);
bool HasCounters = false;
for (const Record *CounterDef : ProcModel.PfmIssueCounterDefs) {
const Record *&CD = CounterDefs[ProcModel.getProcResourceIdx(
CounterDef->getValueAsDef("Resource"))];
if (CD) {
PrintFatalError(CounterDef->getLoc(),
"multiple issue counters for " +
CounterDef->getValueAsDef("Resource")->getName());
}
CD = CounterDef;
HasCounters = true;
}
if (!HasCounters) {
return false;
}
OS << "\nstatic const char* " << ProcModel.ModelName
<< "PfmIssueCounters[] = {\n";
for (unsigned i = 0; i != NumCounterDefs; ++i) {
const Record *CounterDef = CounterDefs[i];
if (CounterDef) {
const auto PfmCounters = CounterDef->getValueAsListOfStrings("Counters");
if (PfmCounters.empty())
PrintFatalError(CounterDef->getLoc(), "empty counter list");
OS << " \"" << PfmCounters[0];
for (unsigned p = 1, e = PfmCounters.size(); p != e; ++p)
OS << ",\" \"" << PfmCounters[p];
OS << "\", // #" << i << " = ";
OS << CounterDef->getValueAsDef("Resource")->getName() << "\n";
} else {
OS << " nullptr, // #" << i << "\n";
}
}
OS << "};\n";
return true;
}
static void EmitPfmCounters(const CodeGenProcModel &ProcModel,
const bool HasPfmIssueCounters, raw_ostream &OS) {
OS << " {\n";
// Emit the cycle counter.
if (ProcModel.PfmCycleCounterDef)
OS << " \"" << ProcModel.PfmCycleCounterDef->getValueAsString("Counter")
<< "\", // Cycle counter.\n";
else
OS << " nullptr, // No cycle counter.\n";
// Emit a reference to issue counters table.
if (HasPfmIssueCounters)
OS << " " << ProcModel.ModelName << "PfmIssueCounters\n";
else
OS << " nullptr // No issue counters.\n";
OS << " }\n";
}
void SubtargetEmitter::EmitExtraProcessorInfo(const CodeGenProcModel &ProcModel,
raw_ostream &OS) {
// Generate a table of register file descriptors (one entry per each user
// defined register file), and a table of register costs.
unsigned NumCostEntries = EmitRegisterFileTables(ProcModel, OS);
// Generate a table of ProcRes counter names.
const bool HasPfmIssueCounters = EmitPfmIssueCountersTable(ProcModel, OS);
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca This patch allows the description of register files in processor scheduling models. This addresses PR36662. A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td. Targets can optionally describe register files for their processors using that class. In particular, class RegisterFile allows to specify: - The total number of physical registers. - Which target registers are accessible through the register file. - The cost of allocating a register at register renaming stage. Example (from this patch - see file X86/X86ScheduleBtVer2.td) def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]> Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar (btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM register definitions only cost 1 physical register. The syntax allows to specify an empty set of register classes. An empty set of register classes means: this register file models all the registers specified by the Target. For each register class, users can specify an optional register cost. By default, register costs default to 1. A value of 0 for the number of physical registers means: "this register file has an unbounded number of physical registers". This patch is structured in two parts. * Part 1 - MC/Tablegen * A first part adds the tablegen definition of RegisterFile, and teaches the SubtargetEmitter how to emit information related to register files. Information about register files is accessible through an instance of MCExtraProcessorInfo. The idea behind this design is to logically partition the processor description which is only used by external tools (like llvm-mca) from the processor information used by the llvm machine schedulers. I think that this design would make easier for targets to get rid of the extra processor information if they don't want it. * Part 2 - llvm-mca related * The second part of this patch is related to changes to llvm-mca. The main differences are: 1) class RegisterFile now needs to take into account the "cost of a register" when allocating physical registers at register renaming stage. 2) Point 1. triggered a minor refactoring which lef to the removal of the "maximum 32 register files" restriction. 3) The BackendStatistics view has been updated so that we can print out extra details related to each register file implemented by the processor. The effect of point 3. is also visible in tests register-files-[1..5].s. Differential Revision: https://reviews.llvm.org/D44980 llvm-svn: 329067
2018-04-03 15:36:24 +02:00
// Now generate a table for the extra processor info.
OS << "\nstatic const llvm::MCExtraProcessorInfo " << ProcModel.ModelName
<< "ExtraInfo = {\n ";
// Add information related to the retire control unit.
EmitRetireControlUnitInfo(ProcModel, OS);
// Add information related to the register files (i.e. where to find register
// file descriptors and register costs).
EmitRegisterFileInfo(ProcModel, ProcModel.RegisterFiles.size(),
NumCostEntries, OS);
EmitPfmCounters(ProcModel, HasPfmIssueCounters, OS);
OS << "};\n";
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca This patch allows the description of register files in processor scheduling models. This addresses PR36662. A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td. Targets can optionally describe register files for their processors using that class. In particular, class RegisterFile allows to specify: - The total number of physical registers. - Which target registers are accessible through the register file. - The cost of allocating a register at register renaming stage. Example (from this patch - see file X86/X86ScheduleBtVer2.td) def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]> Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar (btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM register definitions only cost 1 physical register. The syntax allows to specify an empty set of register classes. An empty set of register classes means: this register file models all the registers specified by the Target. For each register class, users can specify an optional register cost. By default, register costs default to 1. A value of 0 for the number of physical registers means: "this register file has an unbounded number of physical registers". This patch is structured in two parts. * Part 1 - MC/Tablegen * A first part adds the tablegen definition of RegisterFile, and teaches the SubtargetEmitter how to emit information related to register files. Information about register files is accessible through an instance of MCExtraProcessorInfo. The idea behind this design is to logically partition the processor description which is only used by external tools (like llvm-mca) from the processor information used by the llvm machine schedulers. I think that this design would make easier for targets to get rid of the extra processor information if they don't want it. * Part 2 - llvm-mca related * The second part of this patch is related to changes to llvm-mca. The main differences are: 1) class RegisterFile now needs to take into account the "cost of a register" when allocating physical registers at register renaming stage. 2) Point 1. triggered a minor refactoring which lef to the removal of the "maximum 32 register files" restriction. 3) The BackendStatistics view has been updated so that we can print out extra details related to each register file implemented by the processor. The effect of point 3. is also visible in tests register-files-[1..5].s. Differential Revision: https://reviews.llvm.org/D44980 llvm-svn: 329067
2018-04-03 15:36:24 +02:00
}
void SubtargetEmitter::EmitProcessorResources(const CodeGenProcModel &ProcModel,
raw_ostream &OS) {
EmitProcessorResourceSubUnits(ProcModel, OS);
OS << "\n// {Name, NumUnits, SuperIdx, IsBuffered, SubUnitsIdxBegin}\n";
OS << "static const llvm::MCProcResourceDesc " << ProcModel.ModelName
<< "ProcResources"
<< "[] = {\n"
<< " {\"InvalidUnit\", 0, 0, 0, 0},\n";
unsigned SubUnitsOffset = 1;
for (unsigned i = 0, e = ProcModel.ProcResourceDefs.size(); i < e; ++i) {
Record *PRDef = ProcModel.ProcResourceDefs[i];
Record *SuperDef = nullptr;
unsigned SuperIdx = 0;
unsigned NumUnits = 0;
const unsigned SubUnitsBeginOffset = SubUnitsOffset;
int BufferSize = PRDef->getValueAsInt("BufferSize");
if (PRDef->isSubClassOf("ProcResGroup")) {
RecVec ResUnits = PRDef->getValueAsListOfDefs("Resources");
for (Record *RU : ResUnits) {
NumUnits += RU->getValueAsInt("NumUnits");
SubUnitsOffset += RU->getValueAsInt("NumUnits");
}
}
else {
// Find the SuperIdx
if (PRDef->getValueInit("Super")->isComplete()) {
SuperDef =
SchedModels.findProcResUnits(PRDef->getValueAsDef("Super"),
ProcModel, PRDef->getLoc());
SuperIdx = ProcModel.getProcResourceIdx(SuperDef);
}
NumUnits = PRDef->getValueAsInt("NumUnits");
}
// Emit the ProcResourceDesc
OS << " {\"" << PRDef->getName() << "\", ";
if (PRDef->getName().size() < 15)
OS.indent(15 - PRDef->getName().size());
OS << NumUnits << ", " << SuperIdx << ", " << BufferSize << ", ";
if (SubUnitsBeginOffset != SubUnitsOffset) {
OS << ProcModel.ModelName << "ProcResourceSubUnits + "
<< SubUnitsBeginOffset;
} else {
OS << "nullptr";
}
OS << "}, // #" << i+1;
if (SuperDef)
OS << ", Super=" << SuperDef->getName();
OS << "\n";
}
OS << "};\n";
}
// Find the WriteRes Record that defines processor resources for this
// SchedWrite.
Record *SubtargetEmitter::FindWriteResources(
const CodeGenSchedRW &SchedWrite, const CodeGenProcModel &ProcModel) {
// Check if the SchedWrite is already subtarget-specific and directly
// specifies a set of processor resources.
if (SchedWrite.TheDef->isSubClassOf("SchedWriteRes"))
return SchedWrite.TheDef;
Record *AliasDef = nullptr;
for (Record *A : SchedWrite.Aliases) {
const CodeGenSchedRW &AliasRW =
SchedModels.getSchedRW(A->getValueAsDef("AliasRW"));
if (AliasRW.TheDef->getValueInit("SchedModel")->isComplete()) {
Record *ModelDef = AliasRW.TheDef->getValueAsDef("SchedModel");
if (&SchedModels.getProcModel(ModelDef) != &ProcModel)
continue;
}
if (AliasDef)
PrintFatalError(AliasRW.TheDef->getLoc(), "Multiple aliases "
"defined for processor " + ProcModel.ModelName +
" Ensure only one SchedAlias exists per RW.");
AliasDef = AliasRW.TheDef;
}
if (AliasDef && AliasDef->isSubClassOf("SchedWriteRes"))
return AliasDef;
// Check this processor's list of write resources.
Record *ResDef = nullptr;
for (Record *WR : ProcModel.WriteResDefs) {
if (!WR->isSubClassOf("WriteRes"))
continue;
if (AliasDef == WR->getValueAsDef("WriteType")
|| SchedWrite.TheDef == WR->getValueAsDef("WriteType")) {
if (ResDef) {
PrintFatalError(WR->getLoc(), "Resources are defined for both "
"SchedWrite and its alias on processor " +
ProcModel.ModelName);
}
ResDef = WR;
}
}
// TODO: If ProcModel has a base model (previous generation processor),
// then call FindWriteResources recursively with that model here.
if (!ResDef) {
PrintFatalError(ProcModel.ModelDef->getLoc(),
Twine("Processor does not define resources for ") +
SchedWrite.TheDef->getName());
}
return ResDef;
}
/// Find the ReadAdvance record for the given SchedRead on this processor or
/// return NULL.
Record *SubtargetEmitter::FindReadAdvance(const CodeGenSchedRW &SchedRead,
const CodeGenProcModel &ProcModel) {
// Check for SchedReads that directly specify a ReadAdvance.
if (SchedRead.TheDef->isSubClassOf("SchedReadAdvance"))
return SchedRead.TheDef;
// Check this processor's list of aliases for SchedRead.
Record *AliasDef = nullptr;
for (Record *A : SchedRead.Aliases) {
const CodeGenSchedRW &AliasRW =
SchedModels.getSchedRW(A->getValueAsDef("AliasRW"));
if (AliasRW.TheDef->getValueInit("SchedModel")->isComplete()) {
Record *ModelDef = AliasRW.TheDef->getValueAsDef("SchedModel");
if (&SchedModels.getProcModel(ModelDef) != &ProcModel)
continue;
}
if (AliasDef)
PrintFatalError(AliasRW.TheDef->getLoc(), "Multiple aliases "
"defined for processor " + ProcModel.ModelName +
" Ensure only one SchedAlias exists per RW.");
AliasDef = AliasRW.TheDef;
}
if (AliasDef && AliasDef->isSubClassOf("SchedReadAdvance"))
return AliasDef;
// Check this processor's ReadAdvanceList.
Record *ResDef = nullptr;
for (Record *RA : ProcModel.ReadAdvanceDefs) {
if (!RA->isSubClassOf("ReadAdvance"))
continue;
if (AliasDef == RA->getValueAsDef("ReadType")
|| SchedRead.TheDef == RA->getValueAsDef("ReadType")) {
if (ResDef) {
PrintFatalError(RA->getLoc(), "Resources are defined for both "
"SchedRead and its alias on processor " +
ProcModel.ModelName);
}
ResDef = RA;
}
}
// TODO: If ProcModel has a base model (previous generation processor),
// then call FindReadAdvance recursively with that model here.
if (!ResDef && SchedRead.TheDef->getName() != "ReadDefault") {
PrintFatalError(ProcModel.ModelDef->getLoc(),
Twine("Processor does not define resources for ") +
SchedRead.TheDef->getName());
}
return ResDef;
}
// Expand an explicit list of processor resources into a full list of implied
// resource groups and super resources that cover them.
void SubtargetEmitter::ExpandProcResources(RecVec &PRVec,
std::vector<int64_t> &Cycles,
const CodeGenProcModel &PM) {
// Default to 1 resource cycle.
Cycles.resize(PRVec.size(), 1);
for (unsigned i = 0, e = PRVec.size(); i != e; ++i) {
Record *PRDef = PRVec[i];
RecVec SubResources;
if (PRDef->isSubClassOf("ProcResGroup"))
SubResources = PRDef->getValueAsListOfDefs("Resources");
else {
SubResources.push_back(PRDef);
PRDef = SchedModels.findProcResUnits(PRDef, PM, PRDef->getLoc());
for (Record *SubDef = PRDef;
SubDef->getValueInit("Super")->isComplete();) {
if (SubDef->isSubClassOf("ProcResGroup")) {
// Disallow this for simplicitly.
PrintFatalError(SubDef->getLoc(), "Processor resource group "
" cannot be a super resources.");
}
Record *SuperDef =
SchedModels.findProcResUnits(SubDef->getValueAsDef("Super"), PM,
SubDef->getLoc());
PRVec.push_back(SuperDef);
Cycles.push_back(Cycles[i]);
SubDef = SuperDef;
}
}
for (Record *PR : PM.ProcResourceDefs) {
if (PR == PRDef || !PR->isSubClassOf("ProcResGroup"))
continue;
RecVec SuperResources = PR->getValueAsListOfDefs("Resources");
RecIter SubI = SubResources.begin(), SubE = SubResources.end();
for( ; SubI != SubE; ++SubI) {
if (!is_contained(SuperResources, *SubI)) {
break;
}
}
if (SubI == SubE) {
PRVec.push_back(PR);
Cycles.push_back(Cycles[i]);
}
}
}
}
// Generate the SchedClass table for this processor and update global
// tables. Must be called for each processor in order.
void SubtargetEmitter::GenSchedClassTables(const CodeGenProcModel &ProcModel,
SchedClassTables &SchedTables) {
SchedTables.ProcSchedClasses.resize(SchedTables.ProcSchedClasses.size() + 1);
if (!ProcModel.hasInstrSchedModel())
return;
std::vector<MCSchedClassDesc> &SCTab = SchedTables.ProcSchedClasses.back();
LLVM_DEBUG(dbgs() << "\n+++ SCHED CLASSES (GenSchedClassTables) +++\n");
for (const CodeGenSchedClass &SC : SchedModels.schedClasses()) {
LLVM_DEBUG(SC.dump(&SchedModels));
SCTab.resize(SCTab.size() + 1);
MCSchedClassDesc &SCDesc = SCTab.back();
// SCDesc.Name is guarded by NDEBUG
SCDesc.NumMicroOps = 0;
SCDesc.BeginGroup = false;
SCDesc.EndGroup = false;
SCDesc.WriteProcResIdx = 0;
SCDesc.WriteLatencyIdx = 0;
SCDesc.ReadAdvanceIdx = 0;
// A Variant SchedClass has no resources of its own.
bool HasVariants = false;
for (const CodeGenSchedTransition &CGT :
make_range(SC.Transitions.begin(), SC.Transitions.end())) {
if (CGT.ProcIndices[0] == 0 ||
is_contained(CGT.ProcIndices, ProcModel.Index)) {
HasVariants = true;
break;
}
}
if (HasVariants) {
SCDesc.NumMicroOps = MCSchedClassDesc::VariantNumMicroOps;
continue;
}
// Determine if the SchedClass is actually reachable on this processor. If
// not don't try to locate the processor resources, it will fail.
// If ProcIndices contains 0, this class applies to all processors.
assert(!SC.ProcIndices.empty() && "expect at least one procidx");
if (SC.ProcIndices[0] != 0) {
if (!is_contained(SC.ProcIndices, ProcModel.Index))
continue;
}
IdxVec Writes = SC.Writes;
IdxVec Reads = SC.Reads;
if (!SC.InstRWs.empty()) {
// This class has a default ReadWrite list which can be overridden by
// InstRW definitions.
Record *RWDef = nullptr;
for (Record *RW : SC.InstRWs) {
Record *RWModelDef = RW->getValueAsDef("SchedModel");
if (&ProcModel == &SchedModels.getProcModel(RWModelDef)) {
RWDef = RW;
break;
}
}
if (RWDef) {
Writes.clear();
Reads.clear();
SchedModels.findRWs(RWDef->getValueAsListOfDefs("OperandReadWrites"),
Writes, Reads);
}
}
if (Writes.empty()) {
// Check this processor's itinerary class resources.
for (Record *I : ProcModel.ItinRWDefs) {
RecVec Matched = I->getValueAsListOfDefs("MatchedItinClasses");
if (is_contained(Matched, SC.ItinClassDef)) {
SchedModels.findRWs(I->getValueAsListOfDefs("OperandReadWrites"),
Writes, Reads);
break;
}
}
if (Writes.empty()) {
LLVM_DEBUG(dbgs() << ProcModel.ModelName
<< " does not have resources for class " << SC.Name
<< '\n');
}
}
// Sum resources across all operand writes.
std::vector<MCWriteProcResEntry> WriteProcResources;
std::vector<MCWriteLatencyEntry> WriteLatencies;
std::vector<std::string> WriterNames;
std::vector<MCReadAdvanceEntry> ReadAdvanceEntries;
for (unsigned W : Writes) {
IdxVec WriteSeq;
SchedModels.expandRWSeqForProc(W, WriteSeq, /*IsRead=*/false,
ProcModel);
// For each operand, create a latency entry.
MCWriteLatencyEntry WLEntry;
WLEntry.Cycles = 0;
unsigned WriteID = WriteSeq.back();
WriterNames.push_back(SchedModels.getSchedWrite(WriteID).Name);
// If this Write is not referenced by a ReadAdvance, don't distinguish it
// from other WriteLatency entries.
if (!SchedModels.hasReadOfWrite(
SchedModels.getSchedWrite(WriteID).TheDef)) {
WriteID = 0;
}
WLEntry.WriteResourceID = WriteID;
for (unsigned WS : WriteSeq) {
Record *WriteRes =
FindWriteResources(SchedModels.getSchedWrite(WS), ProcModel);
// Mark the parent class as invalid for unsupported write types.
if (WriteRes->getValueAsBit("Unsupported")) {
SCDesc.NumMicroOps = MCSchedClassDesc::InvalidNumMicroOps;
break;
}
WLEntry.Cycles += WriteRes->getValueAsInt("Latency");
SCDesc.NumMicroOps += WriteRes->getValueAsInt("NumMicroOps");
SCDesc.BeginGroup |= WriteRes->getValueAsBit("BeginGroup");
SCDesc.EndGroup |= WriteRes->getValueAsBit("EndGroup");
SCDesc.BeginGroup |= WriteRes->getValueAsBit("SingleIssue");
SCDesc.EndGroup |= WriteRes->getValueAsBit("SingleIssue");
// Create an entry for each ProcResource listed in WriteRes.
RecVec PRVec = WriteRes->getValueAsListOfDefs("ProcResources");
std::vector<int64_t> Cycles =
WriteRes->getValueAsListOfInts("ResourceCycles");
ExpandProcResources(PRVec, Cycles, ProcModel);
for (unsigned PRIdx = 0, PREnd = PRVec.size();
PRIdx != PREnd; ++PRIdx) {
MCWriteProcResEntry WPREntry;
WPREntry.ProcResourceIdx = ProcModel.getProcResourceIdx(PRVec[PRIdx]);
assert(WPREntry.ProcResourceIdx && "Bad ProcResourceIdx");
WPREntry.Cycles = Cycles[PRIdx];
// If this resource is already used in this sequence, add the current
// entry's cycles so that the same resource appears to be used
// serially, rather than multiple parallel uses. This is important for
// in-order machine where the resource consumption is a hazard.
unsigned WPRIdx = 0, WPREnd = WriteProcResources.size();
for( ; WPRIdx != WPREnd; ++WPRIdx) {
if (WriteProcResources[WPRIdx].ProcResourceIdx
== WPREntry.ProcResourceIdx) {
WriteProcResources[WPRIdx].Cycles += WPREntry.Cycles;
break;
}
}
if (WPRIdx == WPREnd)
WriteProcResources.push_back(WPREntry);
}
}
WriteLatencies.push_back(WLEntry);
}
// Create an entry for each operand Read in this SchedClass.
// Entries must be sorted first by UseIdx then by WriteResourceID.
for (unsigned UseIdx = 0, EndIdx = Reads.size();
UseIdx != EndIdx; ++UseIdx) {
Record *ReadAdvance =
FindReadAdvance(SchedModels.getSchedRead(Reads[UseIdx]), ProcModel);
if (!ReadAdvance)
continue;
// Mark the parent class as invalid for unsupported write types.
if (ReadAdvance->getValueAsBit("Unsupported")) {
SCDesc.NumMicroOps = MCSchedClassDesc::InvalidNumMicroOps;
break;
}
RecVec ValidWrites = ReadAdvance->getValueAsListOfDefs("ValidWrites");
IdxVec WriteIDs;
if (ValidWrites.empty())
WriteIDs.push_back(0);
else {
for (Record *VW : ValidWrites) {
WriteIDs.push_back(SchedModels.getSchedRWIdx(VW, /*IsRead=*/false));
}
}
llvm::sort(WriteIDs.begin(), WriteIDs.end());
for(unsigned W : WriteIDs) {
MCReadAdvanceEntry RAEntry;
RAEntry.UseIdx = UseIdx;
RAEntry.WriteResourceID = W;
RAEntry.Cycles = ReadAdvance->getValueAsInt("Cycles");
ReadAdvanceEntries.push_back(RAEntry);
}
}
if (SCDesc.NumMicroOps == MCSchedClassDesc::InvalidNumMicroOps) {
WriteProcResources.clear();
WriteLatencies.clear();
ReadAdvanceEntries.clear();
}
// Add the information for this SchedClass to the global tables using basic
// compression.
//
// WritePrecRes entries are sorted by ProcResIdx.
llvm::sort(WriteProcResources.begin(), WriteProcResources.end(),
LessWriteProcResources());
SCDesc.NumWriteProcResEntries = WriteProcResources.size();
std::vector<MCWriteProcResEntry>::iterator WPRPos =
std::search(SchedTables.WriteProcResources.begin(),
SchedTables.WriteProcResources.end(),
WriteProcResources.begin(), WriteProcResources.end());
if (WPRPos != SchedTables.WriteProcResources.end())
SCDesc.WriteProcResIdx = WPRPos - SchedTables.WriteProcResources.begin();
else {
SCDesc.WriteProcResIdx = SchedTables.WriteProcResources.size();
SchedTables.WriteProcResources.insert(WPRPos, WriteProcResources.begin(),
WriteProcResources.end());
}
// Latency entries must remain in operand order.
SCDesc.NumWriteLatencyEntries = WriteLatencies.size();
std::vector<MCWriteLatencyEntry>::iterator WLPos =
std::search(SchedTables.WriteLatencies.begin(),
SchedTables.WriteLatencies.end(),
WriteLatencies.begin(), WriteLatencies.end());
if (WLPos != SchedTables.WriteLatencies.end()) {
unsigned idx = WLPos - SchedTables.WriteLatencies.begin();
SCDesc.WriteLatencyIdx = idx;
for (unsigned i = 0, e = WriteLatencies.size(); i < e; ++i)
if (SchedTables.WriterNames[idx + i].find(WriterNames[i]) ==
std::string::npos) {
SchedTables.WriterNames[idx + i] += std::string("_") + WriterNames[i];
}
}
else {
SCDesc.WriteLatencyIdx = SchedTables.WriteLatencies.size();
SchedTables.WriteLatencies.insert(SchedTables.WriteLatencies.end(),
WriteLatencies.begin(),
WriteLatencies.end());
SchedTables.WriterNames.insert(SchedTables.WriterNames.end(),
WriterNames.begin(), WriterNames.end());
}
// ReadAdvanceEntries must remain in operand order.
SCDesc.NumReadAdvanceEntries = ReadAdvanceEntries.size();
std::vector<MCReadAdvanceEntry>::iterator RAPos =
std::search(SchedTables.ReadAdvanceEntries.begin(),
SchedTables.ReadAdvanceEntries.end(),
ReadAdvanceEntries.begin(), ReadAdvanceEntries.end());
if (RAPos != SchedTables.ReadAdvanceEntries.end())
SCDesc.ReadAdvanceIdx = RAPos - SchedTables.ReadAdvanceEntries.begin();
else {
SCDesc.ReadAdvanceIdx = SchedTables.ReadAdvanceEntries.size();
SchedTables.ReadAdvanceEntries.insert(RAPos, ReadAdvanceEntries.begin(),
ReadAdvanceEntries.end());
}
}
}
// Emit SchedClass tables for all processors and associated global tables.
void SubtargetEmitter::EmitSchedClassTables(SchedClassTables &SchedTables,
raw_ostream &OS) {
// Emit global WriteProcResTable.
OS << "\n// {ProcResourceIdx, Cycles}\n"
<< "extern const llvm::MCWriteProcResEntry "
<< Target << "WriteProcResTable[] = {\n"
<< " { 0, 0}, // Invalid\n";
for (unsigned WPRIdx = 1, WPREnd = SchedTables.WriteProcResources.size();
WPRIdx != WPREnd; ++WPRIdx) {
MCWriteProcResEntry &WPREntry = SchedTables.WriteProcResources[WPRIdx];
OS << " {" << format("%2d", WPREntry.ProcResourceIdx) << ", "
<< format("%2d", WPREntry.Cycles) << "}";
if (WPRIdx + 1 < WPREnd)
OS << ',';
OS << " // #" << WPRIdx << '\n';
}
OS << "}; // " << Target << "WriteProcResTable\n";
// Emit global WriteLatencyTable.
OS << "\n// {Cycles, WriteResourceID}\n"
<< "extern const llvm::MCWriteLatencyEntry "
<< Target << "WriteLatencyTable[] = {\n"
<< " { 0, 0}, // Invalid\n";
for (unsigned WLIdx = 1, WLEnd = SchedTables.WriteLatencies.size();
WLIdx != WLEnd; ++WLIdx) {
MCWriteLatencyEntry &WLEntry = SchedTables.WriteLatencies[WLIdx];
OS << " {" << format("%2d", WLEntry.Cycles) << ", "
<< format("%2d", WLEntry.WriteResourceID) << "}";
if (WLIdx + 1 < WLEnd)
OS << ',';
OS << " // #" << WLIdx << " " << SchedTables.WriterNames[WLIdx] << '\n';
}
OS << "}; // " << Target << "WriteLatencyTable\n";
// Emit global ReadAdvanceTable.
OS << "\n// {UseIdx, WriteResourceID, Cycles}\n"
<< "extern const llvm::MCReadAdvanceEntry "
<< Target << "ReadAdvanceTable[] = {\n"
<< " {0, 0, 0}, // Invalid\n";
for (unsigned RAIdx = 1, RAEnd = SchedTables.ReadAdvanceEntries.size();
RAIdx != RAEnd; ++RAIdx) {
MCReadAdvanceEntry &RAEntry = SchedTables.ReadAdvanceEntries[RAIdx];
OS << " {" << RAEntry.UseIdx << ", "
<< format("%2d", RAEntry.WriteResourceID) << ", "
<< format("%2d", RAEntry.Cycles) << "}";
if (RAIdx + 1 < RAEnd)
OS << ',';
OS << " // #" << RAIdx << '\n';
}
OS << "}; // " << Target << "ReadAdvanceTable\n";
// Emit a SchedClass table for each processor.
for (CodeGenSchedModels::ProcIter PI = SchedModels.procModelBegin(),
PE = SchedModels.procModelEnd(); PI != PE; ++PI) {
if (!PI->hasInstrSchedModel())
continue;
std::vector<MCSchedClassDesc> &SCTab =
SchedTables.ProcSchedClasses[1 + (PI - SchedModels.procModelBegin())];
OS << "\n// {Name, NumMicroOps, BeginGroup, EndGroup,"
<< " WriteProcResIdx,#, WriteLatencyIdx,#, ReadAdvanceIdx,#}\n";
OS << "static const llvm::MCSchedClassDesc "
<< PI->ModelName << "SchedClasses[] = {\n";
// The first class is always invalid. We no way to distinguish it except by
// name and position.
assert(SchedModels.getSchedClass(0).Name == "NoInstrModel"
&& "invalid class not first");
OS << " {DBGFIELD(\"InvalidSchedClass\") "
<< MCSchedClassDesc::InvalidNumMicroOps
<< ", false, false, 0, 0, 0, 0, 0, 0},\n";
for (unsigned SCIdx = 1, SCEnd = SCTab.size(); SCIdx != SCEnd; ++SCIdx) {
MCSchedClassDesc &MCDesc = SCTab[SCIdx];
const CodeGenSchedClass &SchedClass = SchedModels.getSchedClass(SCIdx);
OS << " {DBGFIELD(\"" << SchedClass.Name << "\") ";
if (SchedClass.Name.size() < 18)
OS.indent(18 - SchedClass.Name.size());
OS << MCDesc.NumMicroOps
<< ", " << ( MCDesc.BeginGroup ? "true" : "false" )
<< ", " << ( MCDesc.EndGroup ? "true" : "false" )
<< ", " << format("%2d", MCDesc.WriteProcResIdx)
<< ", " << MCDesc.NumWriteProcResEntries
<< ", " << format("%2d", MCDesc.WriteLatencyIdx)
<< ", " << MCDesc.NumWriteLatencyEntries
<< ", " << format("%2d", MCDesc.ReadAdvanceIdx)
<< ", " << MCDesc.NumReadAdvanceEntries
<< "}, // #" << SCIdx << '\n';
}
OS << "}; // " << PI->ModelName << "SchedClasses\n";
}
}
void SubtargetEmitter::EmitProcessorModels(raw_ostream &OS) {
// For each processor model.
for (const CodeGenProcModel &PM : SchedModels.procModels()) {
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca This patch allows the description of register files in processor scheduling models. This addresses PR36662. A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td. Targets can optionally describe register files for their processors using that class. In particular, class RegisterFile allows to specify: - The total number of physical registers. - Which target registers are accessible through the register file. - The cost of allocating a register at register renaming stage. Example (from this patch - see file X86/X86ScheduleBtVer2.td) def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]> Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar (btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM register definitions only cost 1 physical register. The syntax allows to specify an empty set of register classes. An empty set of register classes means: this register file models all the registers specified by the Target. For each register class, users can specify an optional register cost. By default, register costs default to 1. A value of 0 for the number of physical registers means: "this register file has an unbounded number of physical registers". This patch is structured in two parts. * Part 1 - MC/Tablegen * A first part adds the tablegen definition of RegisterFile, and teaches the SubtargetEmitter how to emit information related to register files. Information about register files is accessible through an instance of MCExtraProcessorInfo. The idea behind this design is to logically partition the processor description which is only used by external tools (like llvm-mca) from the processor information used by the llvm machine schedulers. I think that this design would make easier for targets to get rid of the extra processor information if they don't want it. * Part 2 - llvm-mca related * The second part of this patch is related to changes to llvm-mca. The main differences are: 1) class RegisterFile now needs to take into account the "cost of a register" when allocating physical registers at register renaming stage. 2) Point 1. triggered a minor refactoring which lef to the removal of the "maximum 32 register files" restriction. 3) The BackendStatistics view has been updated so that we can print out extra details related to each register file implemented by the processor. The effect of point 3. is also visible in tests register-files-[1..5].s. Differential Revision: https://reviews.llvm.org/D44980 llvm-svn: 329067
2018-04-03 15:36:24 +02:00
// Emit extra processor info if available.
if (PM.hasExtraProcessorInfo())
EmitExtraProcessorInfo(PM, OS);
// Emit processor resource table.
if (PM.hasInstrSchedModel())
EmitProcessorResources(PM, OS);
else if(!PM.ProcResourceDefs.empty())
PrintFatalError(PM.ModelDef->getLoc(), "SchedMachineModel defines "
"ProcResources without defining WriteRes SchedWriteRes");
// Begin processor itinerary properties
OS << "\n";
OS << "static const llvm::MCSchedModel " << PM.ModelName << " = {\n";
EmitProcessorProp(OS, PM.ModelDef, "IssueWidth", ',');
EmitProcessorProp(OS, PM.ModelDef, "MicroOpBufferSize", ',');
EmitProcessorProp(OS, PM.ModelDef, "LoopMicroOpBufferSize", ',');
EmitProcessorProp(OS, PM.ModelDef, "LoadLatency", ',');
EmitProcessorProp(OS, PM.ModelDef, "HighLatency", ',');
EmitProcessorProp(OS, PM.ModelDef, "MispredictPenalty", ',');
bool PostRAScheduler =
(PM.ModelDef ? PM.ModelDef->getValueAsBit("PostRAScheduler") : false);
OS << " " << (PostRAScheduler ? "true" : "false") << ", // "
<< "PostRAScheduler\n";
bool CompleteModel =
(PM.ModelDef ? PM.ModelDef->getValueAsBit("CompleteModel") : false);
OS << " " << (CompleteModel ? "true" : "false") << ", // "
<< "CompleteModel\n";
OS << " " << PM.Index << ", // Processor ID\n";
if (PM.hasInstrSchedModel())
OS << " " << PM.ModelName << "ProcResources" << ",\n"
<< " " << PM.ModelName << "SchedClasses" << ",\n"
<< " " << PM.ProcResourceDefs.size()+1 << ",\n"
<< " " << (SchedModels.schedClassEnd()
- SchedModels.schedClassBegin()) << ",\n";
else
OS << " nullptr, nullptr, 0, 0,"
<< " // No instruction-level machine model.\n";
if (PM.hasItineraries())
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca This patch allows the description of register files in processor scheduling models. This addresses PR36662. A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td. Targets can optionally describe register files for their processors using that class. In particular, class RegisterFile allows to specify: - The total number of physical registers. - Which target registers are accessible through the register file. - The cost of allocating a register at register renaming stage. Example (from this patch - see file X86/X86ScheduleBtVer2.td) def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]> Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar (btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM register definitions only cost 1 physical register. The syntax allows to specify an empty set of register classes. An empty set of register classes means: this register file models all the registers specified by the Target. For each register class, users can specify an optional register cost. By default, register costs default to 1. A value of 0 for the number of physical registers means: "this register file has an unbounded number of physical registers". This patch is structured in two parts. * Part 1 - MC/Tablegen * A first part adds the tablegen definition of RegisterFile, and teaches the SubtargetEmitter how to emit information related to register files. Information about register files is accessible through an instance of MCExtraProcessorInfo. The idea behind this design is to logically partition the processor description which is only used by external tools (like llvm-mca) from the processor information used by the llvm machine schedulers. I think that this design would make easier for targets to get rid of the extra processor information if they don't want it. * Part 2 - llvm-mca related * The second part of this patch is related to changes to llvm-mca. The main differences are: 1) class RegisterFile now needs to take into account the "cost of a register" when allocating physical registers at register renaming stage. 2) Point 1. triggered a minor refactoring which lef to the removal of the "maximum 32 register files" restriction. 3) The BackendStatistics view has been updated so that we can print out extra details related to each register file implemented by the processor. The effect of point 3. is also visible in tests register-files-[1..5].s. Differential Revision: https://reviews.llvm.org/D44980 llvm-svn: 329067
2018-04-03 15:36:24 +02:00
OS << " " << PM.ItinsDef->getName() << ",\n";
else
OS << " nullptr, // No Itinerary\n";
if (PM.hasExtraProcessorInfo())
OS << " &" << PM.ModelName << "ExtraInfo,\n";
else
OS << " nullptr // No extra processor descriptor\n";
OS << "};\n";
}
}
//
// EmitProcessorLookup - generate cpu name to itinerary lookup table.
//
void SubtargetEmitter::EmitProcessorLookup(raw_ostream &OS) {
// Gather and sort processor information
std::vector<Record*> ProcessorList =
Records.getAllDerivedDefinitions("Processor");
llvm::sort(ProcessorList.begin(), ProcessorList.end(), LessRecordFieldName());
// Begin processor table
OS << "\n";
OS << "// Sorted (by key) array of itineraries for CPU subtype.\n"
<< "extern const llvm::SubtargetInfoKV "
<< Target << "ProcSchedKV[] = {\n";
2011-04-01 03:56:55 +02:00
// For each processor
for (Record *Processor : ProcessorList) {
StringRef Name = Processor->getValueAsString("Name");
const std::string &ProcModelName =
SchedModels.getModelForProc(Processor).ModelName;
2011-04-01 03:56:55 +02:00
// Emit as { "cpu", procinit },
OS << " { \"" << Name << "\", (const void *)&" << ProcModelName << " },\n";
}
2011-04-01 03:56:55 +02:00
// End processor table
OS << "};\n";
}
//
// EmitSchedModel - Emits all scheduling model tables, folding common patterns.
//
void SubtargetEmitter::EmitSchedModel(raw_ostream &OS) {
OS << "#ifdef DBGFIELD\n"
<< "#error \"<target>GenSubtargetInfo.inc requires a DBGFIELD macro\"\n"
<< "#endif\n"
<< "#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)\n"
<< "#define DBGFIELD(x) x,\n"
<< "#else\n"
<< "#define DBGFIELD(x)\n"
<< "#endif\n";
if (SchedModels.hasItineraries()) {
std::vector<std::vector<InstrItinerary>> ProcItinLists;
// Emit the stage data
EmitStageAndOperandCycleData(OS, ProcItinLists);
EmitItineraries(OS, ProcItinLists);
}
OS << "\n// ===============================================================\n"
<< "// Data tables for the new per-operand machine model.\n";
SchedClassTables SchedTables;
for (const CodeGenProcModel &ProcModel : SchedModels.procModels()) {
GenSchedClassTables(ProcModel, SchedTables);
}
EmitSchedClassTables(SchedTables, OS);
// Emit the processor machine model
EmitProcessorModels(OS);
// Emit the processor lookup data
EmitProcessorLookup(OS);
OS << "\n#undef DBGFIELD";
}
static void emitPredicateProlog(const RecordKeeper &Records, raw_ostream &OS) {
std::string Buffer;
raw_string_ostream Stream(Buffer);
// Collect all the PredicateProlog records and print them to the output
// stream.
std::vector<Record *> Prologs =
Records.getAllDerivedDefinitions("PredicateProlog");
llvm::sort(Prologs.begin(), Prologs.end(), LessRecord());
for (Record *P : Prologs)
Stream << P->getValueAsString("Code") << '\n';
Stream.flush();
OS << Buffer;
}
static void emitPredicates(const CodeGenSchedTransition &T,
const CodeGenSchedClass &SC,
PredicateExpander &PE,
raw_ostream &OS) {
std::string Buffer;
raw_string_ostream StringStream(Buffer);
formatted_raw_ostream FOS(StringStream);
FOS.PadToColumn(6);
FOS << "if (";
for (RecIter RI = T.PredTerm.begin(), RE = T.PredTerm.end(); RI != RE; ++RI) {
if (RI != T.PredTerm.begin()) {
FOS << "\n";
FOS.PadToColumn(8);
FOS << "&& ";
}
const Record *Rec = *RI;
if (Rec->isSubClassOf("MCSchedPredicate"))
PE.expandPredicate(FOS, Rec->getValueAsDef("Pred"));
else
FOS << "(" << Rec->getValueAsString("Predicate") << ")";
}
FOS << ")\n";
FOS.PadToColumn(8);
FOS << "return " << T.ToClassIdx << "; // " << SC.Name << '\n';
FOS.flush();
OS << Buffer;
}
void SubtargetEmitter::emitSchedModelHelpersImpl(
raw_ostream &OS, bool OnlyExpandMCInstPredicates) {
// Collect Variant Classes.
IdxVec VariantClasses;
for (const CodeGenSchedClass &SC : SchedModels.schedClasses()) {
if (SC.Transitions.empty())
continue;
VariantClasses.push_back(SC.Index);
}
if (!VariantClasses.empty()) {
OS << " switch (SchedClass) {\n";
for (unsigned VC : VariantClasses) {
// Emit code for each variant scheduling class.
const CodeGenSchedClass &SC = SchedModels.getSchedClass(VC);
IdxVec ProcIndices;
for (const CodeGenSchedTransition &T : SC.Transitions) {
if (OnlyExpandMCInstPredicates &&
!all_of(T.PredTerm, [](const Record *Rec) {
return Rec->isSubClassOf("MCSchedPredicate");
}))
continue;
IdxVec PI;
std::set_union(T.ProcIndices.begin(), T.ProcIndices.end(),
ProcIndices.begin(), ProcIndices.end(),
std::back_inserter(PI));
ProcIndices.swap(PI);
}
if (ProcIndices.empty())
continue;
OS << " case " << VC << ": // " << SC.Name << '\n';
PredicateExpander PE;
PE.setByRef(false);
PE.setExpandForMC(OnlyExpandMCInstPredicates);
for (unsigned PI : ProcIndices) {
OS << " ";
if (PI != 0) {
OS << (OnlyExpandMCInstPredicates
? "if (CPUID == "
: "if (SchedModel->getProcessorID() == ");
OS << PI << ") ";
}
OS << "{ // " << (SchedModels.procModelBegin() + PI)->ModelName << '\n';
for (const CodeGenSchedTransition &T : SC.Transitions) {
if (PI != 0 && !count(T.ProcIndices, PI))
continue;
PE.setIndentLevel(4);
emitPredicates(T, SchedModels.getSchedClass(T.ToClassIdx), PE, OS);
}
OS << " }\n";
if (PI == 0)
break;
}
if (SC.isInferred())
OS << " return " << SC.Index << ";\n";
OS << " break;\n";
}
// Add a default case to avoid generating a potentially empty switch.
OS << " default : break;\n"
<< " };\n";
}
if (OnlyExpandMCInstPredicates) {
OS << " // Don't know how to resolve this scheduling class.\n"
<< " return 0;\n";
return;
}
OS << " report_fatal_error(\"Expected a variant SchedClass\");\n";
}
void SubtargetEmitter::EmitSchedModelHelpers(const std::string &ClassName,
raw_ostream &OS) {
OS << "unsigned " << ClassName
<< "\n::resolveSchedClass(unsigned SchedClass, const MachineInstr *MI,"
<< " const TargetSchedModel *SchedModel) const {\n";
// Emit the predicate prolog code.
emitPredicateProlog(Records, OS);
// Emit target predicates.
emitSchedModelHelpersImpl(OS);
OS << "} // " << ClassName << "::resolveSchedClass\n";
}
void SubtargetEmitter::EmitHwModeCheck(const std::string &ClassName,
raw_ostream &OS) {
const CodeGenHwModes &CGH = TGT.getHwModes();
assert(CGH.getNumModeIds() > 0);
if (CGH.getNumModeIds() == 1)
return;
OS << "unsigned " << ClassName << "::getHwMode() const {\n";
for (unsigned M = 1, NumModes = CGH.getNumModeIds(); M != NumModes; ++M) {
const HwMode &HM = CGH.getMode(M);
OS << " if (checkFeatures(\"" << HM.Features
<< "\")) return " << M << ";\n";
}
OS << " return 0;\n}\n";
}
//
// ParseFeaturesFunction - Produces a subtarget specific function for parsing
// the subtarget features string.
//
void SubtargetEmitter::ParseFeaturesFunction(raw_ostream &OS,
unsigned NumFeatures,
unsigned NumProcs) {
std::vector<Record*> Features =
Records.getAllDerivedDefinitions("SubtargetFeature");
llvm::sort(Features.begin(), Features.end(), LessRecord());
2011-04-01 03:56:55 +02:00
OS << "// ParseSubtargetFeatures - Parses features string setting specified\n"
<< "// subtarget options.\n"
<< "void llvm::";
OS << Target;
OS << "Subtarget::ParseSubtargetFeatures(StringRef CPU, StringRef FS) {\n"
<< " LLVM_DEBUG(dbgs() << \"\\nFeatures:\" << FS);\n"
<< " LLVM_DEBUG(dbgs() << \"\\nCPU:\" << CPU << \"\\n\\n\");\n";
if (Features.empty()) {
OS << "}\n";
return;
}
OS << " InitMCProcessorInfo(CPU, FS);\n"
<< " const FeatureBitset& Bits = getFeatureBits();\n";
for (Record *R : Features) {
// Next record
StringRef Instance = R->getName();
StringRef Value = R->getValueAsString("Value");
StringRef Attribute = R->getValueAsString("Attribute");
if (Value=="true" || Value=="false")
OS << " if (Bits[" << Target << "::"
<< Instance << "]) "
<< Attribute << " = " << Value << ";\n";
else
OS << " if (Bits[" << Target << "::"
<< Instance << "] && "
<< Attribute << " < " << Value << ") "
<< Attribute << " = " << Value << ";\n";
}
OS << "}\n";
}
void SubtargetEmitter::emitGenMCSubtargetInfo(raw_ostream &OS) {
OS << "struct " << Target
<< "GenMCSubtargetInfo : public MCSubtargetInfo {\n";
OS << " " << Target << "GenMCSubtargetInfo(const Triple &TT, \n"
<< " StringRef CPU, StringRef FS, ArrayRef<SubtargetFeatureKV> PF,\n"
<< " ArrayRef<SubtargetFeatureKV> PD,\n"
<< " const SubtargetInfoKV *ProcSched,\n"
<< " const MCWriteProcResEntry *WPR,\n"
<< " const MCWriteLatencyEntry *WL,\n"
<< " const MCReadAdvanceEntry *RA, const InstrStage *IS,\n"
<< " const unsigned *OC, const unsigned *FP) :\n"
<< " MCSubtargetInfo(TT, CPU, FS, PF, PD, ProcSched,\n"
<< " WPR, WL, RA, IS, OC, FP) { }\n\n"
<< " unsigned resolveVariantSchedClass(unsigned SchedClass,\n"
<< " const MCInst *MI, unsigned CPUID) const override {\n";
emitSchedModelHelpersImpl(OS, /* OnlyExpandMCPredicates */ true);
OS << " }\n";
OS << "};\n";
}
//
// SubtargetEmitter::run - Main subtarget enumeration emitter.
//
void SubtargetEmitter::run(raw_ostream &OS) {
emitSourceFileHeader("Subtarget Enumeration Source Fragment", OS);
OS << "\n#ifdef GET_SUBTARGETINFO_ENUM\n";
OS << "#undef GET_SUBTARGETINFO_ENUM\n\n";
OS << "namespace llvm {\n";
Enumeration(OS);
OS << "} // end namespace llvm\n\n";
OS << "#endif // GET_SUBTARGETINFO_ENUM\n\n";
OS << "\n#ifdef GET_SUBTARGETINFO_MC_DESC\n";
OS << "#undef GET_SUBTARGETINFO_MC_DESC\n\n";
OS << "namespace llvm {\n";
#if 0
OS << "namespace {\n";
#endif
unsigned NumFeatures = FeatureKeyValues(OS);
OS << "\n";
unsigned NumProcs = CPUKeyValues(OS);
OS << "\n";
EmitSchedModel(OS);
OS << "\n";
#if 0
OS << "} // end anonymous namespace\n\n";
#endif
// MCInstrInfo initialization routine.
emitGenMCSubtargetInfo(OS);
OS << "\nstatic inline MCSubtargetInfo *create" << Target
<< "MCSubtargetInfoImpl("
<< "const Triple &TT, StringRef CPU, StringRef FS) {\n";
OS << " return new " << Target << "GenMCSubtargetInfo(TT, CPU, FS, ";
if (NumFeatures)
OS << Target << "FeatureKV, ";
else
OS << "None, ";
if (NumProcs)
OS << Target << "SubTypeKV, ";
else
OS << "None, ";
OS << '\n'; OS.indent(22);
OS << Target << "ProcSchedKV, "
<< Target << "WriteProcResTable, "
<< Target << "WriteLatencyTable, "
<< Target << "ReadAdvanceTable, ";
OS << '\n'; OS.indent(22);
if (SchedModels.hasItineraries()) {
OS << Target << "Stages, "
<< Target << "OperandCycles, "
<< Target << "ForwardingPaths";
} else
OS << "nullptr, nullptr, nullptr";
OS << ");\n}\n\n";
OS << "} // end namespace llvm\n\n";
OS << "#endif // GET_SUBTARGETINFO_MC_DESC\n\n";
OS << "\n#ifdef GET_SUBTARGETINFO_TARGET_DESC\n";
OS << "#undef GET_SUBTARGETINFO_TARGET_DESC\n\n";
OS << "#include \"llvm/Support/Debug.h\"\n";
OS << "#include \"llvm/Support/raw_ostream.h\"\n\n";
ParseFeaturesFunction(OS, NumFeatures, NumProcs);
OS << "#endif // GET_SUBTARGETINFO_TARGET_DESC\n\n";
// Create a TargetSubtargetInfo subclass to hide the MC layer initialization.
OS << "\n#ifdef GET_SUBTARGETINFO_HEADER\n";
OS << "#undef GET_SUBTARGETINFO_HEADER\n\n";
std::string ClassName = Target + "GenSubtargetInfo";
OS << "namespace llvm {\n";
OS << "class DFAPacketizer;\n";
OS << "struct " << ClassName << " : public TargetSubtargetInfo {\n"
<< " explicit " << ClassName << "(const Triple &TT, StringRef CPU, "
<< "StringRef FS);\n"
<< "public:\n"
<< " unsigned resolveSchedClass(unsigned SchedClass, "
<< " const MachineInstr *DefMI,"
<< " const TargetSchedModel *SchedModel) const override;\n"
<< " DFAPacketizer *createDFAPacketizer(const InstrItineraryData *IID)"
<< " const;\n";
if (TGT.getHwModes().getNumModeIds() > 1)
OS << " unsigned getHwMode() const override;\n";
OS << "};\n"
<< "} // end namespace llvm\n\n";
OS << "#endif // GET_SUBTARGETINFO_HEADER\n\n";
OS << "\n#ifdef GET_SUBTARGETINFO_CTOR\n";
OS << "#undef GET_SUBTARGETINFO_CTOR\n\n";
OS << "#include \"llvm/CodeGen/TargetSchedule.h\"\n\n";
OS << "namespace llvm {\n";
OS << "extern const llvm::SubtargetFeatureKV " << Target << "FeatureKV[];\n";
OS << "extern const llvm::SubtargetFeatureKV " << Target << "SubTypeKV[];\n";
OS << "extern const llvm::SubtargetInfoKV " << Target << "ProcSchedKV[];\n";
OS << "extern const llvm::MCWriteProcResEntry "
<< Target << "WriteProcResTable[];\n";
OS << "extern const llvm::MCWriteLatencyEntry "
<< Target << "WriteLatencyTable[];\n";
OS << "extern const llvm::MCReadAdvanceEntry "
<< Target << "ReadAdvanceTable[];\n";
if (SchedModels.hasItineraries()) {
OS << "extern const llvm::InstrStage " << Target << "Stages[];\n";
OS << "extern const unsigned " << Target << "OperandCycles[];\n";
2012-07-07 05:59:48 +02:00
OS << "extern const unsigned " << Target << "ForwardingPaths[];\n";
}
OS << ClassName << "::" << ClassName << "(const Triple &TT, StringRef CPU, "
<< "StringRef FS)\n"
<< " : TargetSubtargetInfo(TT, CPU, FS, ";
if (NumFeatures)
OS << "makeArrayRef(" << Target << "FeatureKV, " << NumFeatures << "), ";
else
OS << "None, ";
if (NumProcs)
OS << "makeArrayRef(" << Target << "SubTypeKV, " << NumProcs << "), ";
else
OS << "None, ";
OS << '\n'; OS.indent(24);
OS << Target << "ProcSchedKV, "
<< Target << "WriteProcResTable, "
<< Target << "WriteLatencyTable, "
<< Target << "ReadAdvanceTable, ";
OS << '\n'; OS.indent(24);
if (SchedModels.hasItineraries()) {
OS << Target << "Stages, "
<< Target << "OperandCycles, "
<< Target << "ForwardingPaths";
} else
OS << "nullptr, nullptr, nullptr";
OS << ") {}\n\n";
EmitSchedModelHelpers(ClassName, OS);
EmitHwModeCheck(ClassName, OS);
OS << "} // end namespace llvm\n\n";
OS << "#endif // GET_SUBTARGETINFO_CTOR\n\n";
}
namespace llvm {
void EmitSubtarget(RecordKeeper &RK, raw_ostream &OS) {
CodeGenTarget CGTarget(RK);
SubtargetEmitter(RK, CGTarget).run(OS);
}
} // end namespace llvm