MCPcopy Create free account
hub / github.com/DeepRec-AI/DeepRec / Start

Method Start

tensorflow/core/distributed_runtime/master.cc:235–252  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

233 }
234
235 void Start() {
236 {
237 mutex_lock l(mu_);
238 num_pending_ = targets_.size();
239 if (num_pending_ == 0) {
240 pending_zero_.notify_all();
241 }
242 }
243 // Talk to all workers to get the list of available devices.
244 using std::placeholders::_1;
245 using std::placeholders::_2;
246 for (size_t i = 0; i < targets_.size(); ++i) {
247 // TODO(mrry): Propagate a timeout here, since `this->WhenFound()` may
248 // never be called.
249 NewRemoteDevices(env_->env, worker_cache_, targets_[i],
250 std::bind(&ME::WhenFound, this, i, _1, _2));
251 }
252 }
253
254 // Every `kLoggingPeriodMs`, while the DeviceFinder is still waiting
255 // to hear from workers, log a list of the workers who have not

Callers 5

RecvFromPeerMethod · 0.45
GetRemoteDevicesMethod · 0.45
MakeGRPCClusterFunction · 0.45

Calls 3

NewRemoteDevicesFunction · 0.85
notify_allMethod · 0.80
sizeMethod · 0.45

Tested by 1

MakeGRPCClusterFunction · 0.36