diff --git a/paddle/fluid/framework/details/all_reduce_op_handle.cc b/paddle/fluid/framework/details/all_reduce_op_handle.cc
index 700c73c745bad72637d77385f5cd38c494501c86..bf493a3fa44e48deec734250d04b2a413c3ed9da 100644
--- a/paddle/fluid/framework/details/all_reduce_op_handle.cc
+++ b/paddle/fluid/framework/details/all_reduce_op_handle.cc
@@ -17,6 +17,7 @@
 #include "paddle/fluid/framework/details/container_cast.h"
 #include "paddle/fluid/framework/details/reduce_and_gather.h"
 #include "paddle/fluid/framework/details/variable_visitor.h"
+#include "paddle/fluid/platform/profiler.h"
 
 namespace paddle {
 namespace framework {
@@ -45,6 +46,7 @@ AllReduceOpHandle::AllReduceOpHandle(ir::Node *node,
 #endif
 
 void AllReduceOpHandle::RunImpl() {
+  platform::RecordEvent r("all_reduce", nullptr);
   if (NoDummyInputSize() == 1) {
     return;  // No need to all reduce when GPU count = 1;
   } else {
diff --git a/paddle/fluid/framework/details/reduce_op_handle.cc b/paddle/fluid/framework/details/reduce_op_handle.cc
index 7160e346dad0615e2fd32b70c096880af0359e1a..68bdfbaf52120d19d05d156529626f42adda630d 100644
--- a/paddle/fluid/framework/details/reduce_op_handle.cc
+++ b/paddle/fluid/framework/details/reduce_op_handle.cc
@@ -16,12 +16,14 @@
 #include "paddle/fluid/framework/details/container_cast.h"
 #include "paddle/fluid/framework/details/reduce_and_gather.h"
 #include "paddle/fluid/framework/details/variable_visitor.h"
+#include "paddle/fluid/platform/profiler.h"
 
 namespace paddle {
 namespace framework {
 namespace details {
 
 void ReduceOpHandle::RunImpl() {
+  platform::RecordEvent r("reduce", nullptr);
   if (places_.size() == 1) return;
   // the input and output may have dummy var.
   auto in_var_handles = DynamicCast<VarHandle>(inputs_);
diff --git a/paddle/fluid/framework/details/scope_buffered_ssa_graph_executor.cc b/paddle/fluid/framework/details/scope_buffered_ssa_graph_executor.cc
index 1d80bab90f513139f807b57258177c6b2ac53ac0..5bd974d6b789a2f085c0a69de5e133187342f587 100644
--- a/paddle/fluid/framework/details/scope_buffered_ssa_graph_executor.cc
+++ b/paddle/fluid/framework/details/scope_buffered_ssa_graph_executor.cc
@@ -17,6 +17,7 @@
 #include <string>
 #include <vector>
 #include "paddle/fluid/framework/executor.h"
+#include "paddle/fluid/platform/profiler.h"
 
 namespace paddle {
 namespace framework {
@@ -62,6 +63,7 @@ FeedFetchList ScopeBufferedSSAGraphExecutor::Run(
     eptr = std::current_exception();
   }
 
+  platform::RecordEvent e("ScopeBufferedSSAGraphExecutorAfterRun", nullptr);
   drop_scope_counter_ += 1;
   if (!fetch_tensors.empty() ||
       drop_scope_counter_ == strategy_.num_iteration_per_drop_scope_) {
diff --git a/paddle/fluid/framework/details/threaded_ssa_graph_executor.cc b/paddle/fluid/framework/details/threaded_ssa_graph_executor.cc
index e556c84b0219eba2b92c456c205e03947171626b..0eaf9a9c951991a5775604eb8d0e7535f81a4ae2 100644
--- a/paddle/fluid/framework/details/threaded_ssa_graph_executor.cc
+++ b/paddle/fluid/framework/details/threaded_ssa_graph_executor.cc
@@ -15,6 +15,7 @@
 #include "paddle/fluid/framework/details/threaded_ssa_graph_executor.h"
 
 #include "paddle/fluid/framework/details/ssa_graph_builder.h"
+#include "paddle/fluid/platform/profiler.h"
 
 namespace paddle {
 namespace framework {
@@ -34,6 +35,8 @@ ThreadedSSAGraphExecutor::ThreadedSSAGraphExecutor(
 
 FeedFetchList ThreadedSSAGraphExecutor::Run(
     const std::vector<std::string> &fetch_tensors) {
+  std::unique_ptr<platform::RecordEvent> event(
+      new platform::RecordEvent("ThreadedSSAGraphExecutorPrepare", nullptr));
   std::unordered_map<OpHandleBase *, size_t> pending_ops;
   std::unordered_set<VarHandleBase *> pending_vars;
   BlockingQueue<VarHandleBase *> ready_vars;
@@ -84,6 +87,7 @@ FeedFetchList ThreadedSSAGraphExecutor::Run(
   // Clean run context
   run_op_futures_.clear();
   exception_holder_.Clear();
+  event.reset(nullptr);
 
   // Step 3. Execution
   while (!pending_vars.empty()) {
diff --git a/paddle/fluid/operators/distributed/send_recv.proto b/paddle/fluid/operators/distributed/send_recv.proto
new file mode 100644
index 0000000000000000000000000000000000000000..2b3536b33af530b469c4e7f8ab06d604efdaf59f
--- /dev/null
+++ b/paddle/fluid/operators/distributed/send_recv.proto
@@ -0,0 +1,84 @@
+
+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. Licensed under
+the Apache License, Version 2.0 (the "License"); you may not use this file
+except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+syntax = "proto3";
+package sendrecv;
+
+option cc_generic_services = false;
+
+service SendRecvService {
+  // For parameter server round-robin like hashing, do not split tensors.
+  // Send and recv only one tensor
+  // TODO(typhoonzero): add streaming API
+  rpc SendVariable(VariableMessage) returns (VoidMessage) {}
+  // Argument VariableMessage for GetVariable should only contain varname.
+  rpc GetVariable(VariableMessage) returns (VariableMessage) {}
+  // pre-fetch variable by given variable name and Ids
+  rpc PrefetchVariable(VariableMessage) returns (VariableMessage) {}
+
+  rpc CheckpointNotify(VariableMessage) returns (VoidMessage) {}
+}
+
+// VariableMessage is serialized paddle variable message.
+// It can be:
+// LoDTensor
+// SelectedRows
+enum VarType {
+  LOD_TENSOR = 0;
+  SELECTED_ROWS = 1;
+  NCCL_ID = 2;
+}
+
+// NOTICE(gongwb):don't modify this proto if you are not
+//   not familar with how we serialize in sendrecvop_utils.h
+//   and deserilize it in  variable_response.h.
+message VariableMessage {
+  enum Type {
+    // Pod Types
+    BOOL = 0;
+    INT16 = 1;
+    INT32 = 2;
+    INT64 = 3;
+    FP16 = 4;
+    FP32 = 5;
+    FP64 = 6;
+  }
+
+  message LodData { repeated int64 lod_data = 1; }
+  string varname = 1;
+  // TODO(Yancey1989): reference framework::proto::VarDesc::VarType
+  VarType type = 2;
+  // bool persistable is not needed for sending.
+  // tensor info:
+  Type data_type = 3;
+  repeated int64 dims = 4;
+
+  // lod details:
+  int64 lod_level = 5;
+  repeated LodData lod = 6;
+  // selected_rows height, aka. original dim0
+  int64 slr_height = 7;
+  // tensor data
+  bytes serialized = 8;
+  // selected_rows data
+  bytes rows = 9;
+  // Look up table block execution output variable name.
+  string out_varname = 10;
+  // If 1, the ps server will start profiling, the ps
+  // server stops profiling and generates a profile to /tmp/profile_ps_*
+  // when profile switches from 1 to 2.
+  int64 profile = 11;
+}
+
+message VoidMessage {}
diff --git a/paddle/fluid/operators/parallel_do_op.cc b/paddle/fluid/operators/parallel_do_op.cc
index c9744db3d0654ef63357963d9a9a3cb946f56e2d..916cdad3fd288d1f3ffb19bc769ab827dd1e9103 100644
--- a/paddle/fluid/operators/parallel_do_op.cc
+++ b/paddle/fluid/operators/parallel_do_op.cc
@@ -18,7 +18,6 @@ limitations under the License. */
 #include "paddle/fluid/framework/op_registry.h"
 #include "paddle/fluid/framework/threadpool.h"
 #include "paddle/fluid/operators/detail/safe_ref.h"
-#include "paddle/fluid/platform/profiler.h"
 
 namespace paddle {
 namespace operators {
@@ -166,8 +165,6 @@ class ParallelDoOp : public framework::OperatorBase {
 
       workers.emplace_back(
           framework::Async([program, cur_scope, place, block, place_idx] {
-            // Give the thread an id to distinguish parallel block with same id.
-            platform::RecordThread rt(static_cast<int>(place_idx) + 1);
             framework::Executor executor(place);
             executor.Run(*program, cur_scope, block->ID(),
                          false /*create_local_scope*/);
@@ -244,8 +241,6 @@ class ParallelDoGradOp : public framework::OperatorBase {
       // execute
       workers.emplace_back(
           framework::Async([program, cur_scope, place, block, i] {
-            // Give the thread an id to distinguish parallel block with same id.
-            platform::RecordThread rt(static_cast<int>(i) + 1);
             framework::Executor executor(place);
             executor.Run(*program, cur_scope, block->ID(),
                          false /*create_local_scope*/);
diff --git a/paddle/fluid/platform/device_tracer.cc b/paddle/fluid/platform/device_tracer.cc
index d9e2afadaf8ec439d158e57c94d3e6e684bce116..8fa8dbd67c936439840cffa073b6fa6693dd31a1 100644
--- a/paddle/fluid/platform/device_tracer.cc
+++ b/paddle/fluid/platform/device_tracer.cc
@@ -30,9 +30,6 @@ limitations under the License. */
 namespace paddle {
 namespace platform {
 namespace {
-// Current thread's id. Note, we don't distinguish nested threads
-// for now.
-thread_local int cur_thread_id = 0;
 // Tracking the nested block stacks of each thread.
 thread_local std::deque<int> block_id_stack;
 // Tracking the nested event stacks.
@@ -413,12 +410,5 @@ void SetCurBlock(int block_id) { block_id_stack.push_back(block_id); }
 void ClearCurBlock() { block_id_stack.pop_back(); }
 
 int BlockDepth() { return block_id_stack.size(); }
-
-void SetCurThread(int thread_id) { cur_thread_id = thread_id; }
-
-void ClearCurThread() { cur_thread_id = 0; }
-
-int CurThread() { return cur_thread_id; }
-
 }  // namespace platform
 }  // namespace paddle
diff --git a/paddle/fluid/platform/device_tracer.h b/paddle/fluid/platform/device_tracer.h
index 0375c7439c29d4122e8ff6b58734dad4f504b7a2..d2a571f4345b544ad5e74f4629c3967593d6d628 100644
--- a/paddle/fluid/platform/device_tracer.h
+++ b/paddle/fluid/platform/device_tracer.h
@@ -99,9 +99,5 @@ std::string CurAnnotation();
 void SetCurBlock(int block_id);
 void ClearCurBlock();
 int BlockDepth();
-
-void SetCurThread(int thread_id);
-void ClearCurThread();
-int CurThread();
 }  // namespace platform
 }  // namespace paddle
diff --git a/paddle/fluid/platform/profiler.cc b/paddle/fluid/platform/profiler.cc
index 01de9d7041bf3eb40884e2a6295027cccfaebd2a..f4a0e2a86134096985ba169e3e28d9980f21c83c 100644
--- a/paddle/fluid/platform/profiler.cc
+++ b/paddle/fluid/platform/profiler.cc
@@ -190,7 +190,7 @@ RecordEvent::~RecordEvent() {
   DeviceTracer* tracer = GetDeviceTracer();
   if (tracer) {
     tracer->AddCPURecords(CurAnnotation(), start_ns_, PosixInNsec(),
-                          BlockDepth(), CurThread());
+                          BlockDepth(), g_thread_id);
   }
   ClearCurAnnotation();
   PopEvent(name_, dev_ctx_);
@@ -211,21 +211,11 @@ RecordBlock::~RecordBlock() {
     // We try to put all blocks at the same nested depth in the
     // same timeline lane. and distinguish the using thread_id.
     tracer->AddCPURecords(name_, start_ns_, PosixInNsec(), BlockDepth(),
-                          CurThread());
+                          g_thread_id);
   }
   ClearCurBlock();
 }
 
-RecordThread::RecordThread(int thread_id) {
-  if (g_state == ProfilerState::kDisabled) return;
-  SetCurThread(thread_id);
-}
-
-RecordThread::~RecordThread() {
-  if (g_state == ProfilerState::kDisabled) return;
-  ClearCurThread();
-}
-
 void EnableProfiler(ProfilerState state) {
   PADDLE_ENFORCE(state != ProfilerState::kDisabled,
                  "Can't enbale profling, since the input state is ",
diff --git a/paddle/fluid/platform/profiler.h b/paddle/fluid/platform/profiler.h
index bf43925373a12cd9ff2155d68c42d0266ba4df60..c99d9c807d1bfb45d1ce0725b84b9fff09049511 100644
--- a/paddle/fluid/platform/profiler.h
+++ b/paddle/fluid/platform/profiler.h
@@ -95,11 +95,6 @@ struct RecordBlock {
   uint64_t start_ns_;
 };
 
-struct RecordThread {
-  explicit RecordThread(int thread_id);
-  ~RecordThread();
-};
-
 // Return the event list of all threads. Assumed the returned value calls
 // event_lists, event_lists[i][j] represents the j-th Event of i-th thread.
 std::vector<std::vector<Event>> GetAllEvents();