diff --git a/paddle/fluid/framework/parallel_executor.cc b/paddle/fluid/framework/parallel_executor.cc
index 8d8004fc6d4d8ca8b296b64a19d01b07a85b7af9..fce1bf47244316bd099d227789afcb4cea20a469 100644
--- a/paddle/fluid/framework/parallel_executor.cc
+++ b/paddle/fluid/framework/parallel_executor.cc
@@ -782,10 +782,6 @@ void ParallelExecutor::Run(const std::vector<std::string> &fetch_tensors,
     }
   }
 
-  for (auto &p : member_->places_) {
-    platform::DeviceContextPool::Instance().Get(p)->Wait();
-  }
-
   for (auto &fetch_op : fetch_ops) {
     fetch_op.WaitAndMergeCPUTensors();
   }