llvm-mirror/lib/Analysis/DevelopmentModeInlineAdvisor.cpp

//===- DevelopmentModeInlineAdvisor.cpp - runtime-loadable model runner  --===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements a model runner using Tensorflow C APIs, allowing the
// loading of a model from a command line option.
//
//===----------------------------------------------------------------------===//
#include "llvm/Config/config.h"
#if defined(LLVM_HAVE_TF_API)

#include "llvm/Analysis/CallGraph.h"
#include "llvm/Analysis/InlineSizeEstimatorAnalysis.h"
#include "llvm/Analysis/MLInlineAdvisor.h"
#include "llvm/Analysis/Utils/TFUtils.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/ManagedStatic.h"

#include <vector>

using namespace llvm;

static cl::opt<std::string> TrainingLog(
    "training-log", cl::Hidden,
    cl::desc("Path where the development - mode inlining log is saved."));

static cl::opt<std::string> TFModelUnderTrainingPath(
    "ml-inliner-model-under-training", cl::Hidden,
    cl::desc("Path to SavedModel from the previous training iteration."));

static cl::opt<std::string> TFFeedPrefix("ml-inliner-trained-model-feed-prefix",
                                         cl::Hidden, cl::init("action_"),
                                         cl::desc("Prefix for feature names."));

static cl::opt<std::string> TFDecisionName(
    "ml-inliner-trained-model-decision-name", cl::Hidden,
    cl::init("StatefulPartitionedCall"),
    cl::desc("Name of the graph operation representing the decision."));

namespace {
/// An InlineEvent, used by TrainingLogger.
struct InlineEvent {
  /// What the default policy's decision would have been.
  bool DefaultDecision = false;

  /// What we advised. When training off the default policy, this is the same as
  /// DefaultDecision.
  bool AdvisedDecision = false;

  /// What actually happened. This would be 'false' in the case of an inline
  /// error, even if AdvisedDecision were true, otherwise it agrees with
  /// AdvisedDecision.
  bool Effect = false;

  /// What the change in size was: size_after - size_before
  int64_t Reward = 0;
};

/// Collect data we may use for training a model, and write it as a textual
/// Tensorflow SequenceExample
/// (https://www.tensorflow.org/api_docs/python/tf/train/SequenceExample)
/// protobuf (https://developers.google.com/protocol-buffers).
/// Because this is a protobuf, we cannot just stream the events as they come.
/// Internally, TrainingLogger stores data in column-major format, because that
/// lines up with how TF SequenceExample represents it.
class TrainingLogger final {
public:
  TrainingLogger();

  /// Log one inlining event.
  void logInlineEvent(const InlineEvent &Event,
                      const MLModelRunner &ModelRunner);

  /// Print the stored tensors.
  void print(raw_fd_ostream &OutFile);

private:
  /// Write the values of one tensor as a list.
  template <typename T>
  void writeTensorValues(raw_fd_ostream &OutFile, const char *TensorData,
                         size_t ElemCount) const {
    OutFile << "[";
    const T *TypedData = reinterpret_cast<const T *>(TensorData);
    for (size_t I = 0; I < ElemCount; ++I) {
      if (I > 0)
        OutFile << ", ";
      OutFile << TypedData[I];
    }
    OutFile << "]";
  }

  /// Write a list of tensors as a sequence of TensorFlow FeatureList protobufs.
  /// The tensors are assumed to be stored contiguously, in row-major format,
  /// in the TensorData buffer. Each tensor has the shape given by Spec. The
  /// feature name in the output is either the provided LoggingName, if
  /// specified, otherwise it's the name of the tensor (as given by Spec).
  template <typename T>
  void
  writeTensorsAsFeatureLists(raw_fd_ostream &OutFile, const TensorSpec &Spec,
                             const T *TensorData, size_t TensorCount,
                             Optional<StringRef> LoggingName = None) const {
    writeRawTensorsAsFeatureLists(OutFile, Spec,
                                  reinterpret_cast<const char *>(TensorData),
                                  TensorCount, LoggingName);
  }

  /// Untyped implementation of the API above.
  void
  writeRawTensorsAsFeatureLists(raw_fd_ostream &OutFile, const TensorSpec &Spec,
                                const char *TensorData, size_t TensorCount,
                                Optional<StringRef> LoggingName = None) const {
    const char *FieldName = "<invalid>";
    std::function<void(const char *)> ValueWriter;
    // The 'Feature' protobuf only has 3 possible fields: float_list,
    // int64_list, or bytes_list, so we capture int32 values as int64. We don't
    // support any other types.
    if (Spec.isElementType<int64_t>()) {
      FieldName = "int64_list";
      ValueWriter = [&](const char *Data) {
        writeTensorValues<int64_t>(OutFile, Data, Spec.getElementCount());
      };
    } else if (Spec.isElementType<int32_t>()) {
      FieldName = "int64_list";
      ValueWriter = [&](const char *Data) {
        writeTensorValues<int32_t>(OutFile, Data, Spec.getElementCount());
      };

    } else if (Spec.isElementType<float>()) {
      FieldName = "float_list";
      ValueWriter = [&](const char *Data) {
        writeTensorValues<float>(OutFile, Data, Spec.getElementCount());
      };

    } else
      llvm_unreachable("Unsupported tensor type.");

    OutFile << "  feature_list: {\n";
    OutFile << "    key: "
            << "\"" << (LoggingName ? *LoggingName : Spec.name()) << "\" ";
    OutFile << "value: {\n";
    size_t TensorByteSize = Spec.getElementCount() * Spec.getElementByteSize();
    for (const char *P = TensorData,
                    *E = TensorData + TensorByteSize * TensorCount;
         P < E; P += TensorByteSize) {
      OutFile << "      feature: { " << FieldName << ": { value: ";
      ValueWriter(P);
      OutFile << " } }\n";
    }
    OutFile << "    }\n";
    OutFile << "  }\n";
  }

  std::vector<InlineFeatures> Features;
  std::vector<int64_t> DefaultDecisions;
  std::vector<int64_t> Decisions;
  std::vector<bool> Effects;
  std::vector<int64_t> Rewards;
};

/// An extension of the MLInlineAdvisor for the 'development' mode, targeting
/// the offline training scenario. Note that training happens outside of the
/// compiler, this facility is concerned with producing training data ("logs").
/// This InlineAdvisor can operate in the following modes:
///
/// 1) collect logs for the default policy. This is useful for bootstrapping
/// training, which will be considerably faster by starting from a reasonable
/// policy.
///
/// 2) collect logs for the ML policy, using a model from a previous
/// training. Potentially, that model uses internally some small random
/// perturbation of its weights, to induce exploration (setting this up is the
/// responsibility of the training algorithm). The logs would then be used to
/// retrain and improve on this model.
///
/// 3) use the provided model, with no logging. This is useful for end to end
/// validation - the model, in this case, is a release candidate and shouldn't
/// have random perturbations. It is a convenience feature: rather than needing
/// to take the release candidate model and compile it in 'release' mode,
/// validate it, then potentially discard it, it's easier to just pass the model
/// to the compiler, albeit compilation would be slower, as a one-off. Once the
/// model behaves satisfactorily, it can be compiled AOT, for efficiency, in
/// release mode. The expectation is that a well-trained model provides a good
/// policy over a sufficiently diverse codebase, over many changes (i.e.
/// training happens seldom).
class DevelopmentModeMLInlineAdvisor : public MLInlineAdvisor {
public:
  DevelopmentModeMLInlineAdvisor(
      Module &M, ModuleAnalysisManager &MAM,
      std::unique_ptr<MLModelRunner> ModelRunner,
      std::function<bool(CallBase &)> GetDefaultAdvice, bool IsDoingInference);

  size_t getTotalSizeEstimate();

  virtual ~DevelopmentModeMLInlineAdvisor();
  void updateNativeSizeEstimate(int64_t Change) { CurrentNativeSize += Change; }
  void resetNativeSize(Function *F) {
    FAM.invalidate<InlineSizeEstimatorAnalysis>(*F);
  }

  std::unique_ptr<MLInlineAdvice>
  getMandatoryAdvice(CallBase &CB, OptimizationRemarkEmitter &ORE) override;
  std::unique_ptr<MLInlineAdvice>
  getAdviceFromModel(CallBase &CB, OptimizationRemarkEmitter &ORE) override;

  size_t getNativeSizeEstimate(const Function &F) const;

private:
  bool isLogging() const { return !TrainingLog.empty(); }

  std::function<bool(CallBase &)> GetDefaultAdvice;
  TrainingLogger Logger;
  const bool IsDoingInference;

  const int32_t InitialNativeSize;
  int32_t CurrentNativeSize = 0;
};

/// A variant of MLInlineAdvice that tracks all non-trivial inlining
/// decisions, for training/logging.
class LoggingMLInlineAdvice : public MLInlineAdvice {
public:
  LoggingMLInlineAdvice(DevelopmentModeMLInlineAdvisor *Advisor, CallBase &CB,
                        OptimizationRemarkEmitter &ORE, bool Recommendation,
                        TrainingLogger &Logger, size_t CallerSizeEstimateBefore,
                        size_t CalleeSizeEstimateBefore, bool DefaultDecision,
                        bool Mandatory = false)
      : MLInlineAdvice(Advisor, CB, ORE, Recommendation), Logger(Logger),
        CallerSizeEstimateBefore(CallerSizeEstimateBefore),
        CalleeSizeEstimateBefore(CalleeSizeEstimateBefore),
        DefaultDecision(DefaultDecision), Mandatory(Mandatory) {}

  virtual ~LoggingMLInlineAdvice() = default;

private:
  DevelopmentModeMLInlineAdvisor *getAdvisor() const {
    return static_cast<DevelopmentModeMLInlineAdvisor *>(Advisor);
  }
  void recordInliningImpl() override {
    MLInlineAdvice::recordInliningImpl();
    getAdvisor()->resetNativeSize(Caller);
    int Reward = std::numeric_limits<int>::max();
    if (!getAdvisor()->isForcedToStop()) {
      int NativeSizeAfter = getAdvisor()->getNativeSizeEstimate(*Caller) +
                            CalleeSizeEstimateBefore;
      Reward = NativeSizeAfter -
               (CallerSizeEstimateBefore + CalleeSizeEstimateBefore);
      getAdvisor()->updateNativeSizeEstimate(Reward);
    }
    log(Reward, /*Success=*/true);
  }

  void recordInliningWithCalleeDeletedImpl() override {
    MLInlineAdvice::recordInliningWithCalleeDeletedImpl();
    getAdvisor()->resetNativeSize(Caller);
    if (!getAdvisor()->isForcedToStop()) {
      int NativeSizeAfter = getAdvisor()->getNativeSizeEstimate(*Caller);
      int Reward = NativeSizeAfter -
                   (CallerSizeEstimateBefore + CalleeSizeEstimateBefore);
      getAdvisor()->updateNativeSizeEstimate(Reward);
      log(Reward, /*Success=*/true);
    }
  }

  void recordUnsuccessfulInliningImpl(const InlineResult &Result) override {
    MLInlineAdvice::recordUnsuccessfulInliningImpl(Result);
    log(NoReward, /*Success=*/false);
  }

  void recordUnattemptedInliningImpl() override {
    MLInlineAdvice::recordUnattemptedInliningImpl();
    log(NoReward, /*Success=*/false);
  }

  void log(int64_t Reward, bool Success) {
    if (Mandatory)
      return;
    InlineEvent Event;
    Event.AdvisedDecision = isInliningRecommended();
    Event.DefaultDecision = DefaultDecision;
    Event.Effect = Success;
    Event.Reward = Reward;
    Logger.logInlineEvent(Event, getAdvisor()->getModelRunner());
  }

  static const int64_t NoReward = 0;
  TrainingLogger &Logger;
  const size_t CallerSizeEstimateBefore;
  const size_t CalleeSizeEstimateBefore;
  const bool DefaultDecision;
  const bool Mandatory;
};

/// A pseudo model runner. We use it to store feature values when collecting
/// logs for the default policy, but never ask it to 'run'.
class NoInferenceModelRunner : public MLModelRunner {
public:
  NoInferenceModelRunner(LLVMContext &Ctx)
      : MLModelRunner(Ctx), Features(NumberOfFeatures) {}
  void setFeature(FeatureIndex Index, int64_t Value) override {
    Features[static_cast<int>(Index)] = Value;
  }

  int64_t getFeature(int Index) const override { return Features[Index]; }
  bool run() override {
    llvm_unreachable("We shouldn't call run on this model runner.");
  }

private:
  InlineFeatures Features;
};

/// ModelUnderTrainingRunner - training mode implementation. It uses TF C APIs
/// to dynamically load and evaluate a TF SavedModel
/// (https://www.tensorflow.org/guide/saved_model). Runtime performance is
/// sacrificed for ease of use while training.
class ModelUnderTrainingRunner final : public MLModelRunner {
public:
  ModelUnderTrainingRunner(LLVMContext &Ctx, const std::string &ModelPath);

  bool run() override;

  // Disallows copy and assign.
  ModelUnderTrainingRunner(const ModelUnderTrainingRunner &) = delete;
  ModelUnderTrainingRunner &
  operator=(const ModelUnderTrainingRunner &) = delete;

  void setFeature(FeatureIndex Index, int64_t Value) override;
  int64_t getFeature(int Index) const override;
  bool isValid() const { return !!Evaluator; }

private:
  std::unique_ptr<TFModelEvaluator> Evaluator;

  // The training framework needs some additional features.
  const std::vector<TensorSpec> TrainingOnlyFeatures{
      TensorSpec::createSpec<int64_t>(TFFeedPrefix + "inlining_default", {1}),
      TensorSpec::createSpec<float>(TFFeedPrefix + "discount", {1}),
      TensorSpec::createSpec<float>(TFFeedPrefix + "reward", {1}),
      TensorSpec::createSpec<int32_t>(TFFeedPrefix + "step_type", {1})};
};
} // namespace

TrainingLogger::TrainingLogger() {
  for (size_t I = 0; I < NumberOfFeatures; ++I) {
    Features.push_back(InlineFeatures());
  }
}

/// Log one inlining event.
void TrainingLogger::logInlineEvent(const InlineEvent &Event,
                                    const MLModelRunner &ModelRunner) {
  for (size_t I = 0; I < NumberOfFeatures; ++I) {
    Features[I].push_back(ModelRunner.getFeature(I));
  }
  Decisions.push_back(Event.AdvisedDecision);
  Effects.push_back(Event.Effect);
  Rewards.push_back(Event.Reward);
  DefaultDecisions.push_back(Event.DefaultDecision);
}

void TrainingLogger::print(raw_fd_ostream &OutFile) {
  size_t NumberOfRecords = Decisions.size();
  if (NumberOfRecords == 0)
    return;

  OutFile << "feature_lists: {\n";
  for (size_t I = 0; I < Features.size(); ++I)
    writeTensorsAsFeatureLists(
        OutFile, TensorSpec::createSpec<int64_t>(FeatureNameMap.at(I), {1}),
        Features[I].data(), NumberOfRecords);

  writeTensorsAsFeatureLists(
      OutFile, TensorSpec::createSpec<int64_t>(DefaultDecisionName, {1}),
      DefaultDecisions.data(), NumberOfRecords);

  writeTensorsAsFeatureLists(OutFile,
                             TensorSpec::createSpec<int64_t>(DecisionName, {1}),
                             Decisions.data(), NumberOfRecords);
  writeTensorsAsFeatureLists(OutFile,
                             TensorSpec::createSpec<int64_t>(RewardName, {1}),
                             Rewards.data(), NumberOfRecords);

  OutFile << "}\n";
}

DevelopmentModeMLInlineAdvisor::DevelopmentModeMLInlineAdvisor(
    Module &M, ModuleAnalysisManager &MAM,
    std::unique_ptr<MLModelRunner> ModelRunner,
    std::function<bool(CallBase &)> GetDefaultAdvice, bool IsDoingInference)
    : MLInlineAdvisor(M, MAM, std::move(ModelRunner)),
      GetDefaultAdvice(GetDefaultAdvice), IsDoingInference(IsDoingInference),
      InitialNativeSize(isLogging() ? getTotalSizeEstimate() : 0),
      CurrentNativeSize(InitialNativeSize) {
  // We cannot have the case of neither inference nor logging.
  assert(IsDoingInference || isLogging());
}

DevelopmentModeMLInlineAdvisor::~DevelopmentModeMLInlineAdvisor() {
  if (TrainingLog.empty())
    return;
  std::error_code ErrorCode;
  raw_fd_ostream OutFile(TrainingLog, ErrorCode);
  Logger.print(OutFile);
}

size_t
DevelopmentModeMLInlineAdvisor::getNativeSizeEstimate(const Function &F) const {
  auto &R =
      FAM.getResult<InlineSizeEstimatorAnalysis>(const_cast<Function &>(F));
  if (!R) {
    F.getParent()->getContext().emitError(
        "Native size estimator is not present.");
    return 0;
  }
  return *R;
}

std::unique_ptr<MLInlineAdvice>
DevelopmentModeMLInlineAdvisor::getMandatoryAdvice(
    CallBase &CB, OptimizationRemarkEmitter &ORE) {
  if (!isLogging())
    return MLInlineAdvisor::getMandatoryAdvice(CB, ORE);
  return std::make_unique<LoggingMLInlineAdvice>(
      /*Advisor=*/this,
      /*CB=*/CB, /*ORE=*/ORE, /*Recommendation=*/true, /*Logger=*/Logger,
      /*CallerSizeEstimateBefore=*/getNativeSizeEstimate(*CB.getCaller()),
      /*CalleeSizeEstimateBefore=*/
      getNativeSizeEstimate(*CB.getCalledFunction()),
      /*DefaultDecision=*/true, /*Mandatory*/ true);
}

std::unique_ptr<MLInlineAdvice>
DevelopmentModeMLInlineAdvisor::getAdviceFromModel(
    CallBase &CB, OptimizationRemarkEmitter &ORE) {
  if (IsDoingInference && !isLogging())
    return MLInlineAdvisor::getAdviceFromModel(CB, ORE);

  bool DefaultAdvice = GetDefaultAdvice(CB);
  auto Recommendation = IsDoingInference ? ModelRunner->run() : DefaultAdvice;
  return std::make_unique<LoggingMLInlineAdvice>(
      /*Advisor=*/this,
      /*CB=*/CB, /*ORE=*/ORE, /*Recommendation=*/Recommendation,
      /*Logger=*/Logger,
      /*CallerSizeEstimateBefore=*/getNativeSizeEstimate(*CB.getCaller()),
      /*CalleeSizeEstimateBefore=*/
      getNativeSizeEstimate(*CB.getCalledFunction()),
      /*DefaultDecision=*/DefaultAdvice);
}

size_t DevelopmentModeMLInlineAdvisor::getTotalSizeEstimate() {
  size_t Ret = 0;
  for (auto &F : M) {
    if (F.isDeclaration())
      continue;
    if (isFunctionDeleted(&F))
      continue;
    Ret += getNativeSizeEstimate(F);
  }
  return Ret;
}

ModelUnderTrainingRunner::ModelUnderTrainingRunner(LLVMContext &Ctx,
                                                   const std::string &ModelPath)
    : MLModelRunner(Ctx) {
  std::vector<TensorSpec> InputSpecs;
  std::vector<TensorSpec> OutputSpecs;
  for (size_t I = 0; I < NumberOfFeatures; ++I)
    InputSpecs.push_back(
        TensorSpec::createSpec<int64_t>(TFFeedPrefix + FeatureNameMap[I], {1}));
  InputSpecs.insert(InputSpecs.end(), TrainingOnlyFeatures.begin(),
                    TrainingOnlyFeatures.end());
  OutputSpecs.push_back(TensorSpec::createSpec<int64_t>(TFDecisionName, {1}));

  Evaluator =
      std::make_unique<TFModelEvaluator>(ModelPath, InputSpecs, OutputSpecs);
  if (!Evaluator || !Evaluator->isValid()) {
    Ctx.emitError("Failed to create inliner saved model evaluator");
    Evaluator.reset();
    return;
  }
}

bool ModelUnderTrainingRunner::run() {
  auto ER = Evaluator->evaluate();
  if (!ER.hasValue()) {
    Ctx.emitError("Error evaluating model.");
    return false;
  }
  int64_t Decision = *ER->getTensorValue<int64_t>(0);
  return static_cast<bool>(Decision);
}

int64_t ModelUnderTrainingRunner::getFeature(int Index) const {
  return *Evaluator->getInput<int64_t>(Index);
}

void ModelUnderTrainingRunner::setFeature(FeatureIndex Index, int64_t Value) {
  size_t NumericIndex = static_cast<size_t>(Index);
  *(Evaluator->getInput<int64_t>(NumericIndex)) = Value;
}

std::unique_ptr<InlineAdvisor> llvm::getDevelopmentModeAdvisor(
    Module &M, ModuleAnalysisManager &MAM,
    std::function<bool(CallBase &)> GetDefaultAdvice) {
  auto &Ctx = M.getContext();
  if (TrainingLog.empty() !=
      !InlineSizeEstimatorAnalysis::isEvaluatorRequested()) {
    Ctx.emitError("For development mode, if training logs are requested, then "
                  "a size estimator must be available; either that, or neither "
                  "are specified.");
    return nullptr;
  }

  std::unique_ptr<MLModelRunner> Runner;

  bool IsDoingInference = false;
  if (TFModelUnderTrainingPath.empty())
    Runner.reset(new NoInferenceModelRunner(Ctx));
  else {
    Runner = std::make_unique<ModelUnderTrainingRunner>(
        Ctx, TFModelUnderTrainingPath);
    if (!Runner) {
      Ctx.emitError("Could not load the policy model from the provided path");
      return nullptr;
    }
    IsDoingInference = true;
  }
  return std::make_unique<DevelopmentModeMLInlineAdvisor>(
      M, MAM, std::move(Runner), GetDefaultAdvice, IsDoingInference);
}
#endif // defined(LLVM_HAVE_TF_API)