update by comment

paddle/fluid/framework/parallel_executor.cc

@@ -110,23 +110,30 @@ ParallelExecutor::ParallelExecutor(
 // Bcast Parameters to all GPUs
 #if defined(PADDLE_WITH_CUDA) && !defined(_WIN32)
     auto *nccl_id_var = scope->FindVar(NCCL_ID_VARNAME);
-    std::unique_ptr<ncclUniqueId> nccl_id = nullptr;
+    ncclUniqueId *nccl_id = nullptr;
     bool need_group_call = true;
-    if (nccl_id_var != nullptr) {
+    if (exec_strategy.type_ == ExecutionStrategy::kParallelGraph) {
-      nccl_id.reset(nccl_id_var->GetMutable<ncclUniqueId>());
+      // parallel graph mode should initialize nccl by ncclCommInitRank since
-    } else if (exec_strategy.type_ == ExecutionStrategy::kParallelGraph) {
+      // it call nccl operator per device per thread.
-      nccl_id.reset(new ncclUniqueId());
+      if (nccl_id_var == nullptr) {
-      PADDLE_ENFORCE(platform::dynload::ncclGetUniqueId(nccl_id.get()));
+        nccl_id = new ncclUniqueId();
-      *member_->global_scope_->Var(NCCL_ID_VARNAME)
+        PADDLE_ENFORCE(platform::dynload::ncclGetUniqueId(nccl_id));
-           ->GetMutable<ncclUniqueId>() = *nccl_id.get();
+        *member_->global_scope_->Var(NCCL_ID_VARNAME)
+             ->GetMutable<ncclUniqueId>() = *nccl_id;
+      } else {
+        nccl_id = nccl_id_var->GetMutable<ncclUniqueId>();
+      }
       need_group_call = false;
+    } else if (nccl_id_var != nullptr) {  // the other executor type.
+      // the distributed training with nccl mode would initialize the nccl id in
+      // startup_program.
+      nccl_id = nccl_id_var->GetMutable<ncclUniqueId>();
     } else {
-      // init nccl_id in NCCLContextMap
+      // initlize NCCL by ncclCommInitAll, do not need nccl_id.
     }
 
     member_->nccl_ctxs_.reset(new platform::NCCLContextMap(
-        member_->places_, nccl_id.get(), num_trainers, trainer_id,
+        member_->places_, nccl_id, num_trainers, trainer_id, need_group_call));
-        need_group_call));
 #else
     PADDLE_THROW("Not compiled with CUDA");
 #endif