parallel_executor.cc 18.2 KB
Newer Older
Y
Yang Yang 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.

Licensed 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. */

#include "paddle/fluid/framework/parallel_executor.h"
Y
Yu Yang 已提交
16 17
#include "ThreadPool.h"
#include "executor.h"
Y
Yu Yang 已提交
18 19
#include "lod_tensor.h"
#include "op_registry.h"
Y
Yang Yang 已提交
20 21

namespace paddle {
Y
Yu Yang 已提交
22 23
namespace framework {

Y
Yu Yang 已提交
24 25 26 27 28 29
#ifdef PADDLE_WITH_CUDA

// FIXME: CHECK the return value of x;
#define NCCL_INVOKE(x) x
#endif

Y
Yu Yang 已提交
30 31
struct OpHandle;

Y
Yu Yang 已提交
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
struct VarHandleBase {
  virtual ~VarHandleBase() {}
  virtual std::string DebugString() const = 0;

  OpHandle *generated_op_;
  std::vector<OpHandle *> pending_ops_;
};

struct VarHandle : public VarHandleBase {
  std::string DebugString() const override {
    std::stringstream ss;
    ss << name_ << ":" << place_;
    return ss.str();
  }

Y
Yu Yang 已提交
47 48 49
  size_t version_;
  std::string name_;
  platform::Place place_;
Y
Yu Yang 已提交
50
};
Y
Yu Yang 已提交
51

Y
Yu Yang 已提交
52
struct DependencyVarHandle : public VarHandleBase {
Y
Yu Yang 已提交
53
  std::string DebugString() const override { return "Dependency Variable"; }
Y
Yu Yang 已提交
54 55 56
};

struct OpHandle {
Y
Yu Yang 已提交
57 58 59 60 61
  std::vector<VarHandleBase *> inputs_;
  std::vector<VarHandleBase *> outputs_;
  std::unordered_map<platform::Place, platform::DeviceContext *,
                     platform::PlaceHash>
      dev_ctx_;
Y
Yu Yang 已提交
62 63 64 65 66

  std::string DebugString() {
    std::stringstream ss;
    ss << "(";
    for (auto *var : inputs_) {
Y
Yu Yang 已提交
67
      ss << var->DebugString() << ", ";
Y
Yu Yang 已提交
68 69 70
    }
    ss << ") --> (";
    for (auto *var : outputs_) {
Y
Yu Yang 已提交
71
      ss << var->DebugString() << ", ";
Y
Yu Yang 已提交
72 73 74 75 76 77
    }
    ss << ")\n";
    return ss.str();
  }

  virtual ~OpHandle() {}
Y
Yu Yang 已提交
78

Y
Yu Yang 已提交
79
  virtual void Run() { PADDLE_THROW("Not implemented"); }
Y
Yu Yang 已提交
80
  virtual void Wait() {}
Y
Yu Yang 已提交
81 82 83 84
};

struct ComputationOpHandle : public OpHandle {
  std::unique_ptr<OperatorBase> op_;
Y
Yu Yang 已提交
85 86
  Scope *scope_;
  platform::Place place_;
Y
Yu Yang 已提交
87

Y
Yu Yang 已提交
88 89
  explicit ComputationOpHandle(const OpDesc &op_desc, Scope *scope,
                               platform::Place place)
Y
Yu Yang 已提交
90
      : op_(framework::OpRegistry::CreateOp(op_desc)),
Y
Yu Yang 已提交
91
        scope_(scope),
Y
Yu Yang 已提交
92 93 94 95
        place_(place) {}

  void Run() override {
    // Wait other op if necessary
Y
Yu Yang 已提交
96
    LOG(INFO) << "Run " << this << " " << DebugString();
Y
Yu Yang 已提交
97 98 99 100 101 102 103 104
    auto *cur_ctx = dev_ctx_[place_];
    for (auto *in : inputs_) {
      if (in->generated_op_ && in->generated_op_->dev_ctx_[place_] != cur_ctx) {
        in->generated_op_->Wait();
      }
    }

    op_->Run(*scope_, place_);
Y
Yu Yang 已提交
105
    LOG(INFO) << "Done " << this;
Y
Yu Yang 已提交
106
  }
Y
Yu Yang 已提交
107 108
};

Y
Yu Yang 已提交
109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
struct ScaleLossGradOpHandle : public OpHandle {
  float coeff_;
  Scope *scope_;
  platform::Place place_;

  explicit ScaleLossGradOpHandle(size_t num_dev, Scope *scope,
                                 platform::Place place)
      : coeff_(static_cast<float>(1.0 / num_dev)),
        scope_(scope),
        place_(place) {}

  void Run() override {
    LOG(INFO) << "Run Scale Loss Grad";

    std::string var_name = static_cast<VarHandle *>(this->outputs_[0])->name_;
Y
Yu Yang 已提交
124

Y
Yu Yang 已提交
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
    float *tmp = scope_->FindVar(var_name)
                     ->GetMutable<framework::LoDTensor>()
                     ->mutable_data<float>(make_ddim({1}), place_);

    if (platform::is_cpu_place(place_)) {
      *tmp = coeff_;
    } else {
      memory::Copy(
          boost::get<platform::CUDAPlace>(place_), tmp, platform::CPUPlace(),
          &coeff_, sizeof(float),
          static_cast<platform::CUDADeviceContext *>(this->dev_ctx_[place_])
              ->stream());
    }
  }
};

struct NCCLAllReduceOpHandle : public OpHandle {
  void Run() override {
    if (this->inputs_.size() == 1) {
      return;  // No need to all reduce when GPU count = 1;
    }
  }
};
Y
Yu Yang 已提交
148 149 150

class ParallelExecutorPrivate {
 public:
Y
Yu Yang 已提交
151 152 153
  explicit ParallelExecutorPrivate(size_t num_threads = 12)
      : pool_(num_threads) {}

Y
Yu Yang 已提交
154 155
  std::unordered_map<platform::Place, Scope *, platform::PlaceHash>
      local_scopes_;
Y
Yu Yang 已提交
156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200

#ifdef PADDLE_WITH_CUDA
  struct NCCLContext {
    std::unique_ptr<platform::CUDADeviceContext> ctx_;
    ncclComm_t comm;

    explicit NCCLContext(int dev_id) {
      ctx_.reset(new platform::CUDADeviceContext(platform::CUDAPlace(dev_id)));
    }

    cudaStream_t stream() const { return ctx_->stream(); }

    int device_id() const {
      return boost::get<platform::CUDAPlace>(ctx_->GetPlace()).device;
    }

    static void InitNCCLContext(std::map<int, NCCLContext> &contexts) {
      std::vector<ncclComm_t> comms;
      std::vector<int> devs;
      comms.resize(contexts.size());
      devs.reserve(contexts.size());

      for (auto &ctx : contexts) {
        devs.push_back(ctx.first);
      }

      NCCL_INVOKE(platform::dynload::ncclCommInitAll(
          &comms[0], static_cast<int>(contexts.size()), &devs[0]));

      int i = 0;
      for (auto &ctx : contexts) {
        ctx.second.comm = comms[i++];
      }
    }
  };

  std::map<int, NCCLContext> communication_streams_;

  NCCLContext &GetNCCLCtx(platform::Place p) {
    int dev_id = boost::get<platform::CUDAPlace>(p).device;
    return communication_streams_.at(dev_id);
  }

#endif

Y
Yu Yang 已提交
201 202 203 204 205 206 207 208 209 210 211 212 213
  platform::DeviceContext *CommunicationDevCtx(const platform::Place &place) {
    if (platform::is_cpu_place(place) || local_scopes_.size() == 1) {
      return const_cast<platform::DeviceContext *>(
          platform::DeviceContextPool::Instance().Get(place));
    } else {
#ifdef PADDLE_WITH_CUDA
      return GetNCCLCtx(place).ctx_.get();
#else
      PADDLE_THROW("Not compiled with CUDA")
#endif
    }
  }

Y
Yu Yang 已提交
214 215 216 217 218 219
  platform::Place main_place_;

  std::unordered_map<platform::Place,
                     std::unordered_map<std::string, std::map<int, VarHandle>>,
                     platform::PlaceHash>
      vars_;
Y
Yu Yang 已提交
220 221
  std::unordered_set<std::unique_ptr<VarHandleBase>> dep_vars_;

Y
Yu Yang 已提交
222
  std::vector<std::unique_ptr<OpHandle>> ops_;
Y
Yu Yang 已提交
223

Y
Yu Yang 已提交
224
  // Use a simpler thread pool, might be faster.
Y
Yu Yang 已提交
225
  ThreadPool pool_;
Y
Yu Yang 已提交
226 227

  std::unique_ptr<platform::EnforceNotMet> exception_;
Y
Yu Yang 已提交
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
};

// TODO(yy): Move this function somewhere
ncclDataType_t ToNCCLDataType(std::type_index type) {
  // FIXME!!
  return ncclFloat;
}

ParallelExecutor::ParallelExecutor(
    const std::vector<platform::Place> &places,
    const std::unordered_set<std::string> &params,
    const ProgramDesc &startup_program, const ProgramDesc &main_program,
    const std::string &loss_var_name, Scope *scope)
    : member_(new ParallelExecutorPrivate()) {
  // Step 1. RunStartupProgram and Bcast the params to devs.
  Executor exe(places[0]);
  exe.Run(startup_program, scope, 0);
  // Create local scopes
  for (auto &place : places) {
    member_->local_scopes_[place] = &scope->NewScope();
  }
  member_->main_place_ = places[0];

  // Bcast Parameters to all GPUs
Y
Yu Yang 已提交
252 253 254 255
  if (platform::is_gpu_place(member_->main_place_) &&
      member_->local_scopes_.size() != 1) {  // Is CUDA
    BuildNCCLCommunicator();
    BCastParamsToGPUs(startup_program);
Y
Yu Yang 已提交
256 257 258 259 260 261
  }
  // Startup Program has been run. All local scopes has correct parameters.

  // Step 2. Convert main_program to SSA form and dependency graph. Also, insert
  // ncclOp
  ConstructDependencyGraph(params, main_program, loss_var_name);
Y
Yu Yang 已提交
262 263 264 265 266 267 268 269 270 271 272 273 274

  // Step 3. Create vars in each scope;
  for (auto &pair : member_->local_scopes_) {
    auto *scope = pair.second;

    for (auto *var : main_program.Block(0).AllVars()) {
      if (scope->FindVar(var->Name()) != nullptr) {
        continue;
      }

      InitializeVariable(scope->Var(var->Name()), var->GetType());
    }
  }
Y
Yu Yang 已提交
275 276 277 278 279
}

void ParallelExecutor::ConstructDependencyGraph(
    const std::unordered_set<std::string> &params,
    const ProgramDesc &main_program, const std::string &loss_var_name) const {
Y
Yu Yang 已提交
280
  std::unordered_set<std::string> grads;
Y
Yu Yang 已提交
281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
  for (auto &each_param : params) {
    grads.insert(each_param + "@GRAD");
  }

  bool is_forwarding = true;
  for (auto *op : main_program.Block(0).AllOps()) {
    bool change_forward = false;

    if (!is_forwarding) {
      // FIXME(yy): Do not hard code like this
      if (op->OutputArgumentNames().size() == 1 &&
          op->OutputArgumentNames()[0] == loss_var_name + "@GRAD") {
        continue;  // Drop fill 1. for backward coeff;
      }
    }

    for (auto &pair : member_->local_scopes_) {
Y
Yu Yang 已提交
298 299
      member_->ops_.emplace_back(
          new ComputationOpHandle(*op, pair.second, pair.first));
Y
Yu Yang 已提交
300
      auto *op_handle = member_->ops_.back().get();
Y
Yu Yang 已提交
301 302
      op_handle->dev_ctx_[pair.first] = const_cast<platform::DeviceContext *>(
          platform::DeviceContextPool::Instance().Get(pair.first));
Y
Yu Yang 已提交
303 304 305 306 307 308 309

      auto var_names = op->InputArgumentNames();

      for (auto &each_var_name : var_names) {
        auto &place = pair.first;
        VarHandle *var = GetVarHandle(each_var_name, place);
        op_handle->inputs_.emplace_back(var);
Y
Yu Yang 已提交
310
        var->pending_ops_.emplace_back(op_handle);
Y
Yu Yang 已提交
311 312 313 314 315 316 317 318 319 320 321
      }
      var_names = op->OutputArgumentNames();

      for (auto &each_var_name : var_names) {
        auto &place = pair.first;
        GenerateVar(op_handle, each_var_name, place);
      }

      if (is_forwarding) {
        if (var_names.size() == 1 && var_names[0] == loss_var_name) {
          // Insert ScaleCost OpHandle
Y
Yu Yang 已提交
322 323
          member_->ops_.emplace_back(new ScaleLossGradOpHandle(
              this->member_->local_scopes_.size(), pair.second, pair.first));
Y
Yu Yang 已提交
324
          op_handle = member_->ops_.back().get();
Y
Yu Yang 已提交
325 326 327 328

          op_handle->dev_ctx_[pair.first] =
              member_->CommunicationDevCtx(pair.first);

Y
Yu Yang 已提交
329
          auto &place = pair.first;
Y
Yu Yang 已提交
330 331 332 333 334 335
          // FIXME: Currently ScaleLossGradOp only use device_count as scale
          // factor. So it does not depend on any other operators.
          // VarHandle *loss = GetVarHandle(loss_var_name, place);
          // loss->pending_ops_.emplace_back(op_handle);
          // op_handle->inputs_.emplace_back(loss);

Y
Yu Yang 已提交
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
          GenerateVar(op_handle, loss_var_name + "@GRAD", place);
          change_forward = true;
          LOG(INFO) << "Scale Loss " << op_handle->DebugString();
        }
      }
    }

    if (change_forward) {
      is_forwarding = false;
    }

    if (!is_forwarding) {
      auto var_names = op->OutputArgumentNames();
      for (auto &og : var_names) {
        if (grads.count(og) != 0) {  // is param grad
          // Insert NCCL AllReduce Op
          member_->ops_.emplace_back(new NCCLAllReduceOpHandle());
          auto *op_handle = member_->ops_.back().get();

          for (auto &pair : member_->local_scopes_) {
            auto &place = pair.first;
            auto &vars = member_->vars_[place][og];

            if (vars.empty()) {  // This device has no data. continue.
              continue;
            }
            auto *prev_grad = &vars[vars.size() - 1];
            op_handle->inputs_.emplace_back(prev_grad);
Y
Yu Yang 已提交
364
            prev_grad->pending_ops_.emplace_back(op_handle);
Y
Yu Yang 已提交
365 366 367 368 369 370
            auto &var = vars[vars.size()];
            var.place_ = place;
            var.generated_op_ = op_handle;
            var.name_ = og;
            var.version_ = vars.size() - 1;
            op_handle->outputs_.emplace_back(&var);
Y
Yu Yang 已提交
371 372 373 374 375

            for (auto &pair : member_->local_scopes_) {
              op_handle->dev_ctx_[pair.first] =
                  member_->CommunicationDevCtx(pair.first);
            }
Y
Yu Yang 已提交
376 377 378 379 380
          }
        }
      }
    }
  }
Y
Yu Yang 已提交
381

Y
Yu Yang 已提交
382 383 384
  /*
    Dependency graph has been constructed. However, there are still data
    harzaeds need to be handled.
Y
Yu Yang 已提交
385
   */
Y
Yu Yang 已提交
386 387
  PolishGraphToSupportDataHarzaeds();
}
Y
Yu Yang 已提交
388

Y
Yu Yang 已提交
389 390 391 392 393 394 395 396
/**
 * We only handle write after read(WAR), since it should not have a write
 * after write in program. If there are write after write operators, we need
 * prune them.
 *
 * https://en.wikipedia.org/wiki/Hazard_(computer_architecture)#Write_after_read_(WAR)
 */
void ParallelExecutor::PolishGraphToSupportDataHarzaeds() const {
Y
Yu Yang 已提交
397 398 399 400 401 402 403 404 405 406 407
  for (auto &place_pair : member_->vars_) {
    for (auto &name_pair : place_pair.second) {
      if (name_pair.second.size() <= 1) {
        return;
      }
      auto it_new = name_pair.second.rbegin();
      auto it_old = name_pair.second.rbegin();
      ++it_old;
      for (; it_old != name_pair.second.rend(); it_new = it_old, ++it_old) {
        auto *write_op = it_new->second.generated_op_;
        auto &read_ops = it_old->second.pending_ops_;
Y
Yu Yang 已提交
408 409 410 411 412 413 414 415 416
        auto *ex_write_op = it_old->second.generated_op_;

        if (ex_write_op == nullptr) {  // Nobody write this var.
          continue;
        }

        LOG(INFO) << "Link " << it_new->second.DebugString() << " From "
                  << it_old->second.version_ << " To "
                  << it_new->second.version_;
Y
Yu Yang 已提交
417 418 419

        for (auto *read_op : read_ops) {
          // Manually add a dependency var from read_op to write_op;
Y
Yu Yang 已提交
420 421 422 423
          if (read_op == write_op) {
            // Read Write is the same op.
            continue;
          }
Y
Yu Yang 已提交
424 425

          auto *dep_var = new DependencyVarHandle();
Y
Yu Yang 已提交
426

Y
Yu Yang 已提交
427 428 429 430 431 432 433 434 435 436
          dep_var->generated_op_ = read_op;
          read_op->outputs_.emplace_back(dep_var);

          dep_var->pending_ops_.emplace_back(write_op);
          write_op->inputs_.emplace_back(dep_var);
          member_->dep_vars_.emplace(dep_var);
        }
      }
    }
  }
Y
Yu Yang 已提交
437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471
}

void ParallelExecutor::GenerateVar(OpHandle *op_handle,
                                   const std::string &each_var_name,
                                   const platform::Place &place) const {
  auto &vars = member_->vars_[place][each_var_name];
  size_t version = vars.size();
  auto &var = vars[version];
  var.version_ = version;
  var.generated_op_ = op_handle;
  var.name_ = each_var_name;
  var.place_ = place;
  op_handle->outputs_.emplace_back(&var);
}

VarHandle *ParallelExecutor::GetVarHandle(const std::string &each_var_name,
                                          const platform::Place &place) const {
  auto &var_holders = member_->vars_[place];
  auto &var_holder = var_holders[each_var_name];
  VarHandle *var = nullptr;
  if (var_holder.empty()) {
    auto &init_var = var_holder[0];
    init_var.place_ = place;
    init_var.name_ = each_var_name;
    init_var.generated_op_ = nullptr;
    init_var.version_ = 0;
    var = &init_var;
  } else {
    var = &var_holder.rbegin()->second;
  }
  return var;
}

void ParallelExecutor::BCastParamsToGPUs(
    const ProgramDesc &startup_program) const {
Y
Yu Yang 已提交
472
#ifdef PADDLE_WITH_CUDA
Y
Yu Yang 已提交
473
  auto *main_scope = member_->local_scopes_[member_->main_place_];
Y
Yu Yang 已提交
474

Y
Yu Yang 已提交
475 476 477 478 479 480 481
  for (auto *var_desc : startup_program.Block(0).AllVars()) {
    if (var_desc->GetType() == proto::VarType::LOD_TENSOR) {
      auto &main_tensor =
          main_scope->FindVar(var_desc->Name())->Get<LoDTensor>();
      ncclDataType_t data_type = ToNCCLDataType(main_tensor.type());
      auto &dims = main_tensor.dims();
      size_t numel = main_tensor.numel();
Y
Yu Yang 已提交
482 483 484 485
      std::vector<std::pair<void *, ParallelExecutorPrivate::NCCLContext *>>
          mems;
      mems.emplace_back(const_cast<void *>(main_tensor.data<void>()),
                        &member_->GetNCCLCtx(member_->main_place_));
Y
Yu Yang 已提交
486 487 488 489 490 491 492 493 494 495

      for (auto &pair : member_->local_scopes_) {
        if (pair.first == member_->main_place_) {
          continue;
        }

        auto local_scope = pair.second;
        auto *t = local_scope->Var(var_desc->Name())->GetMutable<LoDTensor>();
        t->Resize(dims);
        mems.emplace_back(t->mutable_data(pair.first, main_tensor.type()),
Y
Yu Yang 已提交
496
                          &member_->GetNCCLCtx(member_->main_place_));
Y
Yu Yang 已提交
497 498 499 500 501 502 503
      }

      // TODO(yy): Invoke ncclBCast here. mems, numel, data_type. The mems[0]
      // is the src, rests are dests.

      (void)(data_type);
      (void)(numel);
Y
Yu Yang 已提交
504 505 506 507 508 509
    }
  }
#else
  PADDLE_THROW("Not compiled with CUDA");
#endif
}
Y
Yu Yang 已提交
510

Y
Yu Yang 已提交
511 512 513 514 515
void ParallelExecutor::BuildNCCLCommunicator() const {
#ifdef PADDLE_WITH_CUDA
  for (auto &place_pair : member_->local_scopes_) {
    auto place = place_pair.first;
    int dev_id = boost::get<platform::CUDAPlace>(place).device;
Y
Yu Yang 已提交
516

Y
Yu Yang 已提交
517 518
    member_->communication_streams_.emplace(
        dev_id, ParallelExecutorPrivate::NCCLContext(dev_id));
Y
Yu Yang 已提交
519
  }
Y
Yu Yang 已提交
520 521 522 523

  ParallelExecutorPrivate::NCCLContext::InitNCCLContext(
      member_->communication_streams_);
#endif
Y
Yu Yang 已提交
524 525 526 527 528
}

std::vector<LoDTensor> ParallelExecutor::Run(
    const std::vector<std::string> &fetch_tensors) {
  // Version --> VarHandle
Y
Yu Yang 已提交
529
  member_->exception_.reset();
Y
Yu Yang 已提交
530
  std::unordered_map<VarHandleBase *, bool> pending_vars;
Y
Yu Yang 已提交
531 532 533 534 535
  std::unordered_map<OpHandle *, size_t> pending_ops;

  for (auto &place_pair : member_->vars_) {
    for (auto &name_pair : place_pair.second) {
      for (auto &version_pair : name_pair.second) {
Y
Yu Yang 已提交
536 537
        pending_vars[&version_pair.second] =
            version_pair.second.generated_op_ == nullptr;
Y
Yu Yang 已提交
538 539 540 541
      }
    }
  }

Y
Yu Yang 已提交
542 543 544 545
  for (auto &var : member_->dep_vars_) {
    pending_vars[var.get()] = var->generated_op_ == nullptr;
  }

Y
Yu Yang 已提交
546 547
  std::vector<OpHandle *> to_run;

Y
Yu Yang 已提交
548
  for (auto &op : member_->ops_) {
Y
Yu Yang 已提交
549 550 551 552 553 554 555 556 557
    if (op->inputs_.empty()) {  // Special case, Op has no input.
      to_run.emplace_back(op.get());
    } else {
      pending_ops.insert({op.get(), op->inputs_.size()});
    }
  }

  for (auto *op : to_run) {
    RunOp(pending_vars, op);
Y
Yu Yang 已提交
558 559
  }

Y
Yu Yang 已提交
560
  while (!pending_ops.empty()) {
Y
Yu Yang 已提交
561
    VarHandleBase *ready_var = nullptr;
Y
Yu Yang 已提交
562 563 564
    for (auto &pair : pending_vars) {
      if (pair.second) {
        ready_var = pair.first;
Y
Yu Yang 已提交
565 566
      }
    }
Y
Yu Yang 已提交
567 568

    if (ready_var == nullptr) {
Y
Yu Yang 已提交
569 570 571 572 573 574 575
      // FIXME use conditional var instead of busy wait.

      if (member_->exception_) {
        throw * member_->exception_;
      }

      std::this_thread::yield();
Y
Yu Yang 已提交
576
      continue;
Y
Yu Yang 已提交
577 578
    }

Y
Yu Yang 已提交
579 580
    pending_vars.erase(ready_var);

Y
Yu Yang 已提交
581
    to_run.clear();
Y
Yu Yang 已提交
582 583 584 585 586 587

    for (auto *op : ready_var->pending_ops_) {
      auto &deps = pending_ops[op];
      --deps;
      if (deps == 0) {
        to_run.emplace_back(op);
Y
Yu Yang 已提交
588 589 590 591 592
      }
    }

    for (auto *op : to_run) {
      pending_ops.erase(op);
Y
Yu Yang 已提交
593
      RunOp(pending_vars, op);
Y
Yu Yang 已提交
594 595 596 597
    }
  }
  return std::vector<LoDTensor>();
}
Y
Yu Yang 已提交
598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622

void ParallelExecutor::RunOp(
    std::unordered_map<VarHandleBase *, bool> &pending_vars,
    OpHandle *op) const {
  std::vector<bool *> ready_buffer;
  for (auto *var : op->outputs_) {
    ready_buffer.emplace_back(&pending_vars[var]);
  }

  auto op_run = [ready_buffer, op, this] {
    try {
      // TODO(yy) Check Previous Op has same dev ctx.
      op->Run();
      for (auto *ready : ready_buffer) {
        *ready = true;
      }
    } catch (platform::EnforceNotMet ex) {
      member_->exception_.reset(new platform::EnforceNotMet(ex));
    } catch (...) {
      LOG(FATAL) << "Unknown exception catched";
    }
  };

  member_->pool_.enqueue(op_run);
}
Y
Yu Yang 已提交
623
}  // namespace framework
Y
Yang Yang 已提交
624
}  // namespace paddle