task_scheduler.h

Go to the documentation of this file.

/*
 * Licensed to the Apache Software Foundation (ASF) under one
 * or more contributor license agreements.  See the NOTICE file
 * distributed with this work for additional information
 * regarding copyright ownership.  The ASF licenses this file
 * to you under the Apache License, Version 2.0 (the
 * "License"); you may not use this file except in compliance
 * with the License.  You may obtain a copy of the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
 * KIND, either express or implied.  See the License for the
 * specific language governing permissions and limitations
 * under the License.
 */
#ifndef TVM_META_SCHEDULE_TASK_SCHEDULER_H_
#define TVM_META_SCHEDULE_TASK_SCHEDULER_H_
 
#include <tvm/ffi/container/array.h>
#include <tvm/ffi/function.h>
#include <tvm/ffi/optional.h>
#include <tvm/ffi/reflection/registry.h>
#include <tvm/meta_schedule/builder.h>
#include <tvm/meta_schedule/cost_model.h>
#include <tvm/meta_schedule/measure_callback.h>
#include <tvm/meta_schedule/runner.h>
#include <tvm/meta_schedule/tune_context.h>
#include <tvm/runtime/object.h>
#include <tvm/support/random_engine.h>
 
#include <string>
#include <vector>
 
namespace tvm {
namespace meta_schedule {
 
class TaskRecordNode : public runtime::Object {
 public:
  TuneContext ctx{ffi::UnsafeInit()};
  double task_weight{1.0};
  double flop{1.0};
  bool is_terminated = false;
  int build_error_count = 0;
  int run_error_count = 0;
  std::vector<double> latency_ms = {};
  ffi::Optional<ffi::Array<MeasureCandidate>> measure_candidates = std::nullopt;
  ffi::Optional<ffi::Array<BuilderResult>> builder_results = std::nullopt;
  ffi::Optional<ffi::Array<RunnerFuture>> runner_futures = std::nullopt;
 
  static void RegisterReflection() {
    namespace refl = tvm::ffi::reflection;
    refl::ObjectDef<TaskRecordNode>()
        .def_ro("ctx", &TaskRecordNode::ctx)
        .def_ro("task_weight", &TaskRecordNode::task_weight)
        .def_ro("flop", &TaskRecordNode::flop)
        .def_ro("is_terminated", &TaskRecordNode::is_terminated)
        .def_ro("build_error_count", &TaskRecordNode::build_error_count)
        .def_ro("run_error_count", &TaskRecordNode::run_error_count)
        .def_ro("measure_candidates", &TaskRecordNode::measure_candidates)
        .def_ro("builder_results", &TaskRecordNode::builder_results)
        .def_ro("runner_futures", &TaskRecordNode::runner_futures);
  }
 
  static constexpr const bool _type_mutable = true;
  TVM_FFI_DECLARE_OBJECT_INFO_FINAL("meta_schedule.TaskRecord", TaskRecordNode, Object);
};
 
class TaskRecord : public runtime::ObjectRef {
 public:
  explicit TaskRecord(TuneContext task, double task_weight);
 
  TVM_FFI_DEFINE_OBJECT_REF_METHODS_NOTNULLABLE(TaskRecord, ObjectRef, TaskRecordNode);
};
 
class TaskSchedulerNode : public runtime::Object {
 public:
  ffi::Function logger;
  ffi::Array<TaskRecord> tasks_;
  ffi::Array<MeasureCallback> measure_callbacks_;
  ffi::Optional<Database> database_;
  ffi::Optional<CostModel> cost_model_;
  int remaining_tasks_;
 
  virtual ~TaskSchedulerNode() = default;
 
  static void RegisterReflection() {
    namespace refl = tvm::ffi::reflection;
    refl::ObjectDef<TaskSchedulerNode>()
        .def_ro("tasks_", &TaskSchedulerNode::tasks_)
        .def_ro("measure_callbacks_", &TaskSchedulerNode::measure_callbacks_)
        .def_ro("database_", &TaskSchedulerNode::database_)
        .def_ro("cost_model_", &TaskSchedulerNode::cost_model_)
        .def_ro("remaining_tasks_", &TaskSchedulerNode::remaining_tasks_);
  }
 
  virtual int NextTaskId() = 0;
  virtual ffi::Array<RunnerResult> JoinRunningTask(int task_id);
  virtual void Tune(ffi::Array<TuneContext> tasks,                  //
                    ffi::Array<FloatImm> task_weights,              //
                    int max_trials_global,                          //
                    int max_trials_per_task,                        //
                    int num_trials_per_iter,                        //
                    Builder builder,                                //
                    Runner runner,                                  //
                    ffi::Array<MeasureCallback> measure_callbacks,  //
                    ffi::Optional<Database> database,               //
                    ffi::Optional<CostModel> cost_model);
  void TerminateTask(int task_id);
  void TouchTask(int task_id);
  void PrintTuningStatistics();
 
  static constexpr const bool _type_mutable = true;
  TVM_FFI_DECLARE_OBJECT_INFO("meta_schedule.TaskScheduler", TaskSchedulerNode, Object);
};
 
class TaskScheduler;
 
class PyTaskSchedulerNode : public TaskSchedulerNode {
 public:
  using FNextTaskId = ffi::TypedFunction<int()>;
  using FJoinRunningTask = ffi::TypedFunction<ffi::Array<RunnerResult>(int)>;
  using FTune = ffi::TypedFunction<void(ffi::Array<TuneContext> tasks,                  //
                                        ffi::Array<FloatImm> task_weights,              //
                                        int max_trials_global,                          //
                                        int max_trials_per_task,                        //
                                        int num_trials_per_iter,                        //
                                        Builder builder,                                //
                                        Runner runner,                                  //
                                        ffi::Array<MeasureCallback> measure_callbacks,  //
                                        ffi::Optional<Database> database,               //
                                        ffi::Optional<CostModel> cost_model)>;
 
  FNextTaskId f_next_task_id;
  FJoinRunningTask f_join_running_task;
  FTune f_tune;
 
  static void RegisterReflection() {
    namespace refl = tvm::ffi::reflection;
    refl::ObjectDef<PyTaskSchedulerNode>();
  }
 
  int NextTaskId() final;
  ffi::Array<RunnerResult> JoinRunningTask(int task_id) final;
  void Tune(ffi::Array<TuneContext> tasks, ffi::Array<FloatImm> task_weights, int max_trials_global,
            int max_trials_per_task, int num_trials_per_iter, Builder builder, Runner runner,
            ffi::Array<MeasureCallback> measure_callbacks, ffi::Optional<Database> database,
            ffi::Optional<CostModel> cost_model) final;
  TVM_FFI_DECLARE_OBJECT_INFO_FINAL("meta_schedule.PyTaskScheduler", PyTaskSchedulerNode,
                                    TaskSchedulerNode);
};
 
class TaskScheduler : public runtime::ObjectRef {
 public:
  explicit TaskScheduler(ObjectPtr<TaskSchedulerNode> data) : runtime::ObjectRef(data) {
    TVM_FFI_ICHECK(data != nullptr);
  }
  TVM_DLL static TaskScheduler RoundRobin(ffi::Function logger);
  TVM_DLL static TaskScheduler GradientBased(ffi::Function logger, double alpha, int window_size,
                                             support::LinearCongruentialEngine::TRandState seed);
  TVM_DLL static TaskScheduler PyTaskScheduler(
      ffi::Function logger, PyTaskSchedulerNode::FNextTaskId f_next_task_id,
      PyTaskSchedulerNode::FJoinRunningTask f_join_running_task, PyTaskSchedulerNode::FTune f_tune);
  TVM_FFI_DEFINE_OBJECT_REF_METHODS_NOTNULLABLE(TaskScheduler, ObjectRef, TaskSchedulerNode);
};
 
}  // namespace meta_schedule
}  // namespace tvm
 
#endif  // TVM_META_SCHEDULE_TASK_SCHEDULER_H_