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

Method RecvOutputs

tensorflow/core/distributed_runtime/graph_mgr.cc:374–390  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

372}
373
374Status GraphMgr::RecvOutputs(const int64 step_id, NamedTensors* out) {
375 Rendezvous* rendezvous = worker_env_->rendezvous_mgr->Find(step_id);
376 Status s = RecvOutputsFromRendezvous(rendezvous, out, Rendezvous::Args());
377 rendezvous->Unref();
378 if (!s.ok()) {
379 // Failing to fetch the outputs should not be possible, so rewrite the error
380 // status to an INTERNAL error.
381 s = errors::Internal("Failed to fetch outputs for step ", step_id,
382 ". (Original error message: ", s.ToString(), ")");
383 }
384 size_t output_size = 0;
385 for (auto& p : *out) {
386 output_size += p.second.AllocatedBytes();
387 }
388 metrics::RecordGraphOutputTensors(output_size);
389 return s;
390}
391
392Status GraphMgr::RecvOutputs(const int64 step_id, NamedTensors* out,
393 std::map<std::string, bool>* is_out_dead) {

Callers 2

StarRunGraphAsyncMethod · 0.80
DoRunGraphMethod · 0.80

Calls 9

InternalFunction · 0.85
RecordGraphOutputTensorsFunction · 0.85
ArgsClass · 0.50
FindMethod · 0.45
UnrefMethod · 0.45
okMethod · 0.45
ToStringMethod · 0.45
AllocatedBytesMethod · 0.45

Tested by

no test coverage detected