|
| 1 | +/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. |
| 2 | +
|
| 3 | +Licensed under the Apache License, Version 2.0 (the "License"); |
| 4 | +you may not use this file except in compliance with the License. |
| 5 | +You may obtain a copy of the License at |
| 6 | +
|
| 7 | + http://www.apache.org/licenses/LICENSE-2.0 |
| 8 | +
|
| 9 | +Unless required by applicable law or agreed to in writing, software |
| 10 | +distributed under the License is distributed on an "AS IS" BASIS, |
| 11 | +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 12 | +See the License for the specific language governing permissions and |
| 13 | +limitations under the License. */ |
| 14 | + |
| 15 | +#include "ParameterServerController.h" |
| 16 | + |
| 17 | +namespace paddle { |
| 18 | + |
| 19 | +ParameterServerController::ParameterServerController( |
| 20 | + const ParameterServerConfig& config) { |
| 21 | + // round robin to load balance RDMA server ENGINE |
| 22 | + std::vector<std::string> devices; |
| 23 | + int rdmaCpu = 0; |
| 24 | + int onlineCpus = rdma::numCpus(); |
| 25 | + int numPorts = config.ports_num() + config.ports_num_for_sparse(); |
| 26 | + |
| 27 | + if (config.nics().empty()) { |
| 28 | + parameterServers_.resize(numPorts); |
| 29 | + for (int i = 0; i < numPorts; ++i) { |
| 30 | + if (config.rdma_tcp() == "rdma") { |
| 31 | + parameterServers_[i].reset( |
| 32 | + new ParameterServer2(std::string(), config.port() + i, rdmaCpu++)); |
| 33 | + rdmaCpu = rdmaCpu % onlineCpus; |
| 34 | + } else { |
| 35 | + parameterServers_[i].reset( |
| 36 | + new ParameterServer2(std::string(), config.port() + i)); |
| 37 | + } |
| 38 | + CHECK(parameterServers_[i]->init()) << "Fail to initialize parameter " |
| 39 | + "server on port " |
| 40 | + << config.port() + i; |
| 41 | + } |
| 42 | + } else { |
| 43 | + str::split(config.nics(), ',', &devices); |
| 44 | + parameterServers_.resize(devices.size() * numPorts); |
| 45 | + for (int i = 0; i < numPorts; ++i) { |
| 46 | + for (size_t j = 0; j < devices.size(); ++j) { |
| 47 | + if (config.rdma_tcp() == "rdma") { |
| 48 | + parameterServers_[i * devices.size() + j].reset(new ParameterServer2( |
| 49 | + getIpAddr(devices[j]), config.port() + i, rdmaCpu++)); |
| 50 | + rdmaCpu = rdmaCpu % onlineCpus; |
| 51 | + } else { |
| 52 | + parameterServers_[i * devices.size() + j].reset( |
| 53 | + new ParameterServer2(getIpAddr(devices[j]), config.port() + i)); |
| 54 | + } |
| 55 | + CHECK(parameterServers_[i * devices.size() + j]->init()) |
| 56 | + << "Fail to initialize parameter server with device " << devices[j] |
| 57 | + << config.port() + i; |
| 58 | + } |
| 59 | + } |
| 60 | + } |
| 61 | +} |
| 62 | + |
| 63 | +ParameterServerController::~ParameterServerController() { this->wait(); } |
| 64 | + |
| 65 | +ParameterServerController* ParameterServerController::createFromGflags() { |
| 66 | + ParameterServerConfig config; |
| 67 | + |
| 68 | + config.set_nics(FLAGS_nics); |
| 69 | + config.set_rdma_tcp(FLAGS_rdma_tcp); |
| 70 | + config.set_port(FLAGS_port); |
| 71 | + config.set_ports_num(FLAGS_ports_num); |
| 72 | + config.set_ports_num_for_sparse(FLAGS_ports_num_for_sparse); |
| 73 | + |
| 74 | + return create(config); |
| 75 | +} |
| 76 | + |
| 77 | +ParameterServerController* ParameterServerController::create( |
| 78 | + const ParameterServerConfig& config) { |
| 79 | + return new ParameterServerController(config); |
| 80 | +} |
| 81 | + |
| 82 | +void ParameterServerController::start() { |
| 83 | + LOG(INFO) << "number of parameterServer instances: " |
| 84 | + << parameterServers_.size(); |
| 85 | + int i = 0; |
| 86 | + for (const auto& parameterServer : parameterServers_) { |
| 87 | + LOG(INFO) << "Starting parameterServer[" << i << "]"; |
| 88 | + parameterServer->start(); |
| 89 | + i++; |
| 90 | + } |
| 91 | +} |
| 92 | + |
| 93 | +void ParameterServerController::wait() { |
| 94 | + int i = 0; |
| 95 | + for (const auto& parameterServer : parameterServers_) { |
| 96 | + LOG(INFO) << "Waiting parameterServer[" << i << "]"; |
| 97 | + parameterServer->join(); |
| 98 | + i++; |
| 99 | + } |
| 100 | +} |
| 101 | + |
| 102 | +} // namespace paddle |
0 commit comments