diff --git a/paddle/fluid/lite/api/cxx_api.cc b/paddle/fluid/lite/api/cxx_api.cc
index 14a526fe1a4ec7d4014269bd4ea75514f1d64502..07a216b63739d48ce68dc359c87245f812f97715 100644
--- a/paddle/fluid/lite/api/cxx_api.cc
+++ b/paddle/fluid/lite/api/cxx_api.cc
@@ -13,6 +13,10 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/api/cxx_api.h"
+#include <memory>
+#include <string>
+#include <utility>
+#include <vector>
 #ifndef LITE_WITH_LIGHT_WEIGHT_FRAMEWORK
 #include "paddle/fluid/platform/port.h"
 #endif
diff --git a/paddle/fluid/lite/api/cxx_api_test.cc b/paddle/fluid/lite/api/cxx_api_test.cc
index 373cd070407452493619a2ea2eaec235ffc3506c..3d929ffb38a9d014b37ac128ce3ce48f14986152 100644
--- a/paddle/fluid/lite/api/cxx_api_test.cc
+++ b/paddle/fluid/lite/api/cxx_api_test.cc
@@ -15,6 +15,7 @@
 #include "paddle/fluid/lite/api/cxx_api.h"
 #include <gflags/gflags.h>
 #include <gtest/gtest.h>
+#include <vector>
 #include "paddle/fluid/lite/core/mir/passes.h"
 #include "paddle/fluid/lite/core/op_registry.h"
 
diff --git a/paddle/fluid/lite/api/light_api.h b/paddle/fluid/lite/api/light_api.h
index 484af4d339bc4dac399798a7c7af4040ef43d1f4..9cd9f62a0b095090f1d7c40df902615347bf90b5 100644
--- a/paddle/fluid/lite/api/light_api.h
+++ b/paddle/fluid/lite/api/light_api.h
@@ -18,7 +18,9 @@
  */
 #pragma once
 
+#include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 #include "paddle/fluid/lite/core/program.h"
 #include "paddle/fluid/lite/core/types.h"
diff --git a/paddle/fluid/lite/core/CMakeLists.txt b/paddle/fluid/lite/core/CMakeLists.txt
index f6c3006844616511eee6a039fc552425941d3fc7..bf40d198dee47f958313894f0130edcd51341715 100644
--- a/paddle/fluid/lite/core/CMakeLists.txt
+++ b/paddle/fluid/lite/core/CMakeLists.txt
@@ -1,7 +1,6 @@
 cc_library(lite_gtest_main SRCS lite_gtest_main.cc DEPS gtest)
 cc_library(memory_lite SRCS memory.cc DEPS target_wrapper_lite target_wrapper_host)
 cc_library(target_wrapper_lite SRCS target_wrapper.cc)
-
 cc_library(lite_tensor SRCS lite_tensor.cc DEPS memory_lite target_wrapper_lite)
 if (NOT LITE_WITH_LIGHT_WEIGHT_FRAMEWORK)
     cc_library(hvy_tensor SRCS hvy_tensor.cc DEPS lod_tensor)
@@ -24,7 +23,6 @@ cc_library(op_lite SRCS op_lite.cc DEPS scope_lite op_registry_lite compatible_p
 cc_library(types_lite SRCS types.cc)
 cc_library(type_system SRCS type_system.cc DEPS ${tensor_lite} target_wrapper_lite)
 
-#cc_library(kernel_executor_lite SRCS kernel_executor.cc DEPS mir_ssa_graph kernel_lite)
 cc_library(program_lite SRCS program.cc DEPS op_lite kernel_lite)
 cc_library(optimizer_lite SRCS optimizer.cc DEPS mir_pass_manager model_parser_lite program_lite)
 
diff --git a/paddle/fluid/lite/core/context.h b/paddle/fluid/lite/core/context.h
index 3e40588b73ea8c152bf37b5e4127db4cba14fb83..d255645c72c2be24dae9a9f98959c9bb70e429fb 100644
--- a/paddle/fluid/lite/core/context.h
+++ b/paddle/fluid/lite/core/context.h
@@ -20,6 +20,7 @@
 #include "paddle/fluid/lite/cuda/cuda_utils.h"
 #endif
 #include <memory>
+#include <set>
 #include <vector>
 #include "paddle/fluid/lite/core/target_wrapper.h"
 
diff --git a/paddle/fluid/lite/core/kernel.cc b/paddle/fluid/lite/core/kernel.cc
index c86e38b3801340638c9627f29d9b97aa78e1ec3a..44b00f53d018ffe9431c7b481fb1bc1a6e1f7cdc 100644
--- a/paddle/fluid/lite/core/kernel.cc
+++ b/paddle/fluid/lite/core/kernel.cc
@@ -62,5 +62,6 @@ std::ostream &operator<<(std::ostream &os,
      << other.place.DebugString();
   return os;
 }
+
 }  // namespace lite
-}  // namespace paddle
\ No newline at end of file
+}  // namespace paddle
diff --git a/paddle/fluid/lite/core/kernel.h b/paddle/fluid/lite/core/kernel.h
index c86194c47839eab1e08c074b367e620aa64839df..6846dbb920d2b8ebef0ad1062ff3074ac9409e37 100644
--- a/paddle/fluid/lite/core/kernel.h
+++ b/paddle/fluid/lite/core/kernel.h
@@ -19,6 +19,7 @@
 #include <set>
 #include <sstream>
 #include <string>
+#include <utility>
 #include <vector>
 #include "paddle/fluid/lite/core/context.h"
 #include "paddle/fluid/lite/core/target_wrapper.h"
diff --git a/paddle/fluid/lite/core/kernel_executor.cc b/paddle/fluid/lite/core/kernel_executor.cc
deleted file mode 100644
index 4ca0d20d465a252ba5bc7e35caf7144af085a24a..0000000000000000000000000000000000000000
--- a/paddle/fluid/lite/core/kernel_executor.cc
+++ /dev/null
@@ -1,19 +0,0 @@
-// Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
-//
-// 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.
-
-#include "paddle/fluid/lite/core/kernel_executor.h"
-
-namespace paddle {
-namespace lite {}  // namespace lite
-}  // namespace paddle
\ No newline at end of file
diff --git a/paddle/fluid/lite/core/kernel_executor.h b/paddle/fluid/lite/core/kernel_executor.h
deleted file mode 100644
index b933c13bfc17bfbdcb28b776a85c69dae5f509b9..0000000000000000000000000000000000000000
--- a/paddle/fluid/lite/core/kernel_executor.h
+++ /dev/null
@@ -1,50 +0,0 @@
-// Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
-//
-// 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.
-
-#pragma once
-#include "paddle/fluid/lite/core/mir/ssa_graph.h"
-
-namespace paddle {
-namespace lite {
-
-/*
- * KernelExecutor executes a list of kernels.
- */
-class KernelExecutorBase {
- public:
-  KernelExecutorBase(std::unique_ptr<mir::Program>&& program);
-
-  // Prepare runtime context.
-  void PrepareWorkspace();
-
-  void Run();
-
- private:
-  lite::Scope* scope_{};
-  lite::Scope* exec_scope_{};
-};
-
-/*
- * KernelExecutor executes the kernels without concurrency, works in X86 place.
- */
-class SerialKernelExecutor : public KernelExecutorBase {};
-
-/*
- * KernelExecutor executes the kernels with CUDA like stream parallel support,
- * works in CUDA like devices.
- */
-class StreamKernelExecutor : public KernelExecutorBase {};
-
-}  // namespace lite
-}  // namespace paddle
diff --git a/paddle/fluid/lite/core/lite_tensor.h b/paddle/fluid/lite/core/lite_tensor.h
index ea31b3f9aa5e976ba28babf82dadb5d53c8e2725..3fe29cc33313e08fe3feca6771a6e7bf40416fd1 100644
--- a/paddle/fluid/lite/core/lite_tensor.h
+++ b/paddle/fluid/lite/core/lite_tensor.h
@@ -29,7 +29,7 @@ class DDimLite : public DDimBase<DDimLite> {
  public:
   DDimLite() = default;
 
-  DDimLite(const std::vector<value_type> &x) : DDimBase<DDimLite>() {
+  explicit DDimLite(const std::vector<value_type> &x) : DDimBase<DDimLite>() {
     ConstructFrom(x);
   }
 
diff --git a/paddle/fluid/lite/core/mir/argument_type_display_pass.cc b/paddle/fluid/lite/core/mir/argument_type_display_pass.cc
index e4db847d127152a6ee74988b2224d1237c50789e..69b9c5fa5a06491f0ba143e688277d8620f6e893 100644
--- a/paddle/fluid/lite/core/mir/argument_type_display_pass.cc
+++ b/paddle/fluid/lite/core/mir/argument_type_display_pass.cc
@@ -21,7 +21,7 @@ namespace mir {
 
 class ArgumentTypeDisplayPass : public DebugPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override {
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override {
     LOG(INFO) << "== Argument types ==";
     for (auto& node : graph->mutable_nodes()) {
       if (!node.IsArg()) continue;
diff --git a/paddle/fluid/lite/core/mir/demo_pass.cc b/paddle/fluid/lite/core/mir/demo_pass.cc
index 36542dcafa542a5aa62de3a4b2418c6395176dd3..d9e05adc6d3c58e333886d46fbdf56b97336e7c3 100644
--- a/paddle/fluid/lite/core/mir/demo_pass.cc
+++ b/paddle/fluid/lite/core/mir/demo_pass.cc
@@ -21,7 +21,7 @@ namespace mir {
 
 class DemoPass : public mir::DebugPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override {}
+  void Apply(const std::unique_ptr<SSAGraph> &graph) override {}
 };
 
 /*
diff --git a/paddle/fluid/lite/core/mir/generate_program_pass.cc b/paddle/fluid/lite/core/mir/generate_program_pass.cc
index ec671d7bbfa94031133ffca486a92563b99d53c9..3751b6f1f84a19e5ce7753e086e76f289e707e5d 100644
--- a/paddle/fluid/lite/core/mir/generate_program_pass.cc
+++ b/paddle/fluid/lite/core/mir/generate_program_pass.cc
@@ -13,6 +13,9 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/mir/generate_program_pass.h"
+#include <memory>
+#include <utility>
+#include <vector>
 #include "paddle/fluid/lite/core/mir/graph_visualize_pass.h"
 #include "paddle/fluid/lite/core/mir/pass_registry.h"
 
@@ -20,7 +23,7 @@ namespace paddle {
 namespace lite {
 namespace mir {
 
-void GenerateProgramPass::Apply(std::unique_ptr<mir::SSAGraph>& graph) {
+void GenerateProgramPass::Apply(const std::unique_ptr<SSAGraph>& graph) {
   LOG(INFO) << "final program \n" << Visualize(graph.get());
   for (auto& item : graph->StmtTopologicalOrder()) {
     if (item->IsStmt()) {
diff --git a/paddle/fluid/lite/core/mir/generate_program_pass.h b/paddle/fluid/lite/core/mir/generate_program_pass.h
index cd0e70b2c82c9506c4756e2c3e757b6212dfe409..0ef358ec23338364e9f837b2ffb56b8b83296cdf 100644
--- a/paddle/fluid/lite/core/mir/generate_program_pass.h
+++ b/paddle/fluid/lite/core/mir/generate_program_pass.h
@@ -28,7 +28,7 @@ namespace mir {
  */
 class GenerateProgramPass : public ProgramPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override;
+  void Apply(const std::unique_ptr<SSAGraph> &graph) override;
 
   std::unique_ptr<RuntimeProgram> GenProgram() {
     LOG(INFO) << "insts.size " << insts_.size();
diff --git a/paddle/fluid/lite/core/mir/graph_visualize_pass.cc b/paddle/fluid/lite/core/mir/graph_visualize_pass.cc
index 440dfd785bef505dc586c65d4c0a258e13ef2443..6a13bafd67ca710691f1a20a62ea411c90064e85 100644
--- a/paddle/fluid/lite/core/mir/graph_visualize_pass.cc
+++ b/paddle/fluid/lite/core/mir/graph_visualize_pass.cc
@@ -13,7 +13,9 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/mir/graph_visualize_pass.h"
+#include <memory>
 #include <set>
+#include <string>
 #include "paddle/fluid/lite/core/mir/pass_registry.h"
 
 namespace paddle {
@@ -22,7 +24,7 @@ namespace mir {
 
 using inference::analysis::Dot;
 
-void GraphVisualizePass::Apply(std::unique_ptr<mir::SSAGraph>& graph) {
+void GraphVisualizePass::Apply(const std::unique_ptr<SSAGraph>& graph) {
   Visualize(graph.get());
 }
 
diff --git a/paddle/fluid/lite/core/mir/graph_visualize_pass.h b/paddle/fluid/lite/core/mir/graph_visualize_pass.h
index 79dbd4080b604a7e198c74f90ed7cf7485d4fbce..f11cfaabf4a71c3991d23a515704fc5983a9f690 100644
--- a/paddle/fluid/lite/core/mir/graph_visualize_pass.h
+++ b/paddle/fluid/lite/core/mir/graph_visualize_pass.h
@@ -27,7 +27,7 @@ namespace mir {
  */
 class GraphVisualizePass : public DebugPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override;
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override;
 };
 
 std::string Visualize(mir::SSAGraph* graph);
diff --git a/paddle/fluid/lite/core/mir/io_copy_kernel_pick_pass.cc b/paddle/fluid/lite/core/mir/io_copy_kernel_pick_pass.cc
index 1978f50db1a814e7b790e8388421d369b2d8f9a9..ebf9e5a57bfb9395cbd661c4e69ec2980eebbd17 100644
--- a/paddle/fluid/lite/core/mir/io_copy_kernel_pick_pass.cc
+++ b/paddle/fluid/lite/core/mir/io_copy_kernel_pick_pass.cc
@@ -21,7 +21,7 @@ namespace mir {
 
 class IoCopyKernelPickPass : public StmtPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override {
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override {
     for (auto& node : graph->mutable_nodes()) {
       if (!node.IsStmt()) continue;
       auto& inst = node.AsStmt();
diff --git a/paddle/fluid/lite/core/mir/node.h b/paddle/fluid/lite/core/mir/node.h
index d827a021108793abd1737d2c164c8d4092dba658..e10472c5885982d500ee05756d5d1e14eb11c89a 100644
--- a/paddle/fluid/lite/core/mir/node.h
+++ b/paddle/fluid/lite/core/mir/node.h
@@ -17,6 +17,7 @@
 #include <list>
 #include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 #include "paddle/fluid/lite/core/kernel.h"
 #include "paddle/fluid/lite/core/op_lite.h"
diff --git a/paddle/fluid/lite/core/mir/pass.h b/paddle/fluid/lite/core/mir/pass.h
index 0f14f6defb68bdc18cc111794598daa741f4a27a..6935ec462eb1f3ca0efd505c8c8f08a6aad70a71 100644
--- a/paddle/fluid/lite/core/mir/pass.h
+++ b/paddle/fluid/lite/core/mir/pass.h
@@ -13,6 +13,8 @@
 // limitations under the License.
 
 #pragma once
+#include <memory>
+#include <string>
 #include "paddle/fluid/lite/core/mir/node.h"
 #include "paddle/fluid/lite/core/mir/ssa_graph.h"
 
@@ -32,9 +34,9 @@ class Pass {
     kDebug,
   };
 
-  Pass(Kind kind) : kind_(kind) {}
+  explicit Pass(Kind kind) : kind_(kind) {}
 
-  virtual void Apply(std::unique_ptr<mir::SSAGraph>& graph) = 0;
+  virtual void Apply(const std::unique_ptr<SSAGraph>& graph) = 0;
 
   void set_name(const std::string& name) { name_ = name; }
   const std::string& name() const { return name_; }
diff --git a/paddle/fluid/lite/core/mir/pass_manager.h b/paddle/fluid/lite/core/mir/pass_manager.h
index ba3fad4d9f242e0d921616ba8eced13d3df180dc..901702285ac75c33a9304634c5e8e89b2acb2d91 100644
--- a/paddle/fluid/lite/core/mir/pass_manager.h
+++ b/paddle/fluid/lite/core/mir/pass_manager.h
@@ -32,7 +32,7 @@ class PassManager {
 
   PassManager();
 
-  void Run(std::unique_ptr<SSAGraph>& graph) {
+  void Run(const std::unique_ptr<SSAGraph>& graph) {
     for (auto& pass : passes_) {
       LOG(INFO) << "Running MIR pass " << pass->name();
       pass->Apply(graph);
diff --git a/paddle/fluid/lite/core/mir/runtime_context_assign_pass.cc b/paddle/fluid/lite/core/mir/runtime_context_assign_pass.cc
index 8e89d922e0b0985881c80ac80c9bf8c80bec3338..3d2012306b9fd4c89a72f2e29223be75e847c204 100644
--- a/paddle/fluid/lite/core/mir/runtime_context_assign_pass.cc
+++ b/paddle/fluid/lite/core/mir/runtime_context_assign_pass.cc
@@ -27,7 +27,7 @@ class RuntimeContextAssignPass : public StmtPass {
 #endif
   }
 
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override {
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override {
     for (auto& node : graph->mutable_nodes()) {
       if (!node.IsStmt()) continue;
 
diff --git a/paddle/fluid/lite/core/mir/ssa_graph.cc b/paddle/fluid/lite/core/mir/ssa_graph.cc
index 10bd01ef601a790e1aab54db4ad1a579afb4bcd3..b565b6fd3ecbff36a95fb6e2f8a8d78d92a2a069 100644
--- a/paddle/fluid/lite/core/mir/ssa_graph.cc
+++ b/paddle/fluid/lite/core/mir/ssa_graph.cc
@@ -13,6 +13,10 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/mir/ssa_graph.h"
+#include <algorithm>
+#include <memory>
+#include <set>
+#include <utility>
 
 namespace paddle {
 namespace lite {
diff --git a/paddle/fluid/lite/core/mir/static_kernel_pick_pass.cc b/paddle/fluid/lite/core/mir/static_kernel_pick_pass.cc
index 4dfc15e3e195beb90a877aa78e4016b6cc4ac5f1..3709094ed297a891dedf3d3b98abf787c143b9e5 100644
--- a/paddle/fluid/lite/core/mir/static_kernel_pick_pass.cc
+++ b/paddle/fluid/lite/core/mir/static_kernel_pick_pass.cc
@@ -27,7 +27,7 @@ bool KernelScoreCmp(const std::pair<size_t, std::unique_ptr<KernelBase>>& a,
   return a.first > b.first;
 }
 
-void StaticKernelPickPass::Apply(std::unique_ptr<mir::SSAGraph>& graph) {
+void StaticKernelPickPass::Apply(const std::unique_ptr<SSAGraph>& graph) {
   CHECK(kernel_pick_factors_.AnyFactorConsidered())
       << "kernel_pick_factors should be specified first";
   CHECK(graph) << "graph not valid";
diff --git a/paddle/fluid/lite/core/mir/static_kernel_pick_pass.h b/paddle/fluid/lite/core/mir/static_kernel_pick_pass.h
index 5c48ba9ca02823859c61c168764de216544f3473..52d5dd08d26c0420b3bea4c9c4f79bfb709b7418 100644
--- a/paddle/fluid/lite/core/mir/static_kernel_pick_pass.h
+++ b/paddle/fluid/lite/core/mir/static_kernel_pick_pass.h
@@ -35,7 +35,7 @@ namespace mir {
  */
 class StaticKernelPickPass : public mir::StmtPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override;
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override;
 
   void SetPreferPlace(const Place& place) { place_ = place; }
   const Place& place() const { return place_; }
diff --git a/paddle/fluid/lite/core/mir/type_target_transform_pass.cc b/paddle/fluid/lite/core/mir/type_target_transform_pass.cc
index 9ac80b552e7b2d4986dcaa1bd64541718454c766..ddd079701662825e6b3da7c5d89d2714e6d3e195 100644
--- a/paddle/fluid/lite/core/mir/type_target_transform_pass.cc
+++ b/paddle/fluid/lite/core/mir/type_target_transform_pass.cc
@@ -14,7 +14,9 @@
 
 #include "paddle/fluid/lite/core/mir/type_target_transform_pass.h"
 #include <list>
+#include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 #include "paddle/fluid/lite/core/mir/graph_visualize_pass.h"
 #include "paddle/fluid/lite/core/mir/pass_registry.h"
@@ -23,7 +25,7 @@ namespace paddle {
 namespace lite {
 namespace mir {
 
-void TypeTargetTransformPass::Apply(std::unique_ptr<mir::SSAGraph>& graph) {
+void TypeTargetTransformPass::Apply(const std::unique_ptr<SSAGraph>& graph) {
   // Start from inputs of the graph, those should have place set.
   std::list<Node*> nodes;
   for (auto& node : graph->mutable_nodes()) {
diff --git a/paddle/fluid/lite/core/mir/type_target_transform_pass.h b/paddle/fluid/lite/core/mir/type_target_transform_pass.h
index 4f3f0c1c148f1dbe486600856968553cbc1af439..411090219fb3f47b5050689e2fde4b86effd5fe5 100644
--- a/paddle/fluid/lite/core/mir/type_target_transform_pass.h
+++ b/paddle/fluid/lite/core/mir/type_target_transform_pass.h
@@ -40,7 +40,7 @@ static void UpdateInputTo(framework::proto::OpDesc* desc,
  */
 class TypeTargetTransformPass : public ProgramPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override;
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override;
 
   void ComplementInputs(SSAGraph* graph, Node* inst_node, Node* in);
 
diff --git a/paddle/fluid/lite/core/mir/variable_place_inference_pass.cc b/paddle/fluid/lite/core/mir/variable_place_inference_pass.cc
index 6d4b2ce246204e435c08d80c682a8cdfffbdeef5..8ee0fb24ba285f3bd456721cd6aa67af460336ce 100644
--- a/paddle/fluid/lite/core/mir/variable_place_inference_pass.cc
+++ b/paddle/fluid/lite/core/mir/variable_place_inference_pass.cc
@@ -13,13 +13,14 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/mir/variable_place_inference_pass.h"
+#include <memory>
 #include "paddle/fluid/lite/core/mir/pass_registry.h"
 
 namespace paddle {
 namespace lite {
 namespace mir {
 
-void VariablePlaceInferencePass::Apply(std::unique_ptr<mir::SSAGraph>& graph) {
+void VariablePlaceInferencePass::Apply(const std::unique_ptr<SSAGraph> &graph) {
   MarkInputPlace(graph.get());
   InferenceArgumentPlace(graph.get());
   CheckAllArgumentTypeDetermined(graph.get());
diff --git a/paddle/fluid/lite/core/mir/variable_place_inference_pass.h b/paddle/fluid/lite/core/mir/variable_place_inference_pass.h
index ff129600ea6f67b1226a90303d54c3462a6d1646..79c977b331f85f4d580e3e79f01262554a106dbc 100644
--- a/paddle/fluid/lite/core/mir/variable_place_inference_pass.h
+++ b/paddle/fluid/lite/core/mir/variable_place_inference_pass.h
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 #pragma once
+#include <memory>
 #include "paddle/fluid/lite/core/mir/pass.h"
 #include "paddle/fluid/lite/core/target_wrapper.h"
 
@@ -26,7 +27,7 @@ namespace mir {
  */
 class VariablePlaceInferencePass : public DebugPass {
  public:
-  void Apply(std::unique_ptr<mir::SSAGraph>& graph) override;
+  void Apply(const std::unique_ptr<SSAGraph>& graph) override;
 
  private:
   // Mark the place of input arguments.
diff --git a/paddle/fluid/lite/core/op_lite.cc b/paddle/fluid/lite/core/op_lite.cc
index dd159becc785d0963e4f23e50bb9f92520d1d2c4..47e0c441e75443f066a5f6f69e6876030d649f81 100644
--- a/paddle/fluid/lite/core/op_lite.cc
+++ b/paddle/fluid/lite/core/op_lite.cc
@@ -13,7 +13,10 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/op_lite.h"
-#include "op_lite.h"
+#include <list>
+#include <set>
+#include <utility>
+#include <vector>
 #include "paddle/fluid/lite/core/op_registry.h"
 
 namespace paddle {
diff --git a/paddle/fluid/lite/core/op_registry.cc b/paddle/fluid/lite/core/op_registry.cc
index 4ac233ad2e63d5aeef0c830d487b102e84b53702..681cbdafcdeee33c50bbb60dbca88feaac832681 100644
--- a/paddle/fluid/lite/core/op_registry.cc
+++ b/paddle/fluid/lite/core/op_registry.cc
@@ -13,6 +13,8 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/op_registry.h"
+#include <list>
+#include <set>
 
 namespace paddle {
 namespace lite {
@@ -90,4 +92,4 @@ KernelRegistry &KernelRegistry::Global() {
 }
 
 }  // namespace lite
-}  // namespace paddle
\ No newline at end of file
+}  // namespace paddle
diff --git a/paddle/fluid/lite/core/op_registry.h b/paddle/fluid/lite/core/op_registry.h
index 1a66ed894432ac59bd649aee6b38d8500bcfa3a8..2321e3a250e5e644f3eeb5d229d13bbcc2f76a53 100644
--- a/paddle/fluid/lite/core/op_registry.h
+++ b/paddle/fluid/lite/core/op_registry.h
@@ -14,8 +14,10 @@
 
 #pragma once
 #include <memory>
+#include <set>
 #include <string>
 #include <unordered_map>
+#include <utility>
 #include "paddle/fluid/lite/core/kernel.h"
 #include "paddle/fluid/lite/core/op_lite.h"
 #include "paddle/fluid/lite/core/target_wrapper.h"
@@ -43,7 +45,7 @@ class LiteOpRegistry final : public Factory<OpLite, std::shared_ptr<OpLite>> {
 template <typename OpClass>
 class OpLiteRegistor : public Registor<OpClass> {
  public:
-  OpLiteRegistor(const std::string &op_type)
+  explicit OpLiteRegistor(const std::string &op_type)
       : Registor<OpClass>([&] {
           LiteOpRegistry::Global().Register(
               op_type, [op_type]() -> std::unique_ptr<OpLite> {
diff --git a/paddle/fluid/lite/core/optimizer.h b/paddle/fluid/lite/core/optimizer.h
index 34272f570d04d981644706e2a03c35fa3621d705..f585224a07f6ca63f99345a2460445e736e9feab 100644
--- a/paddle/fluid/lite/core/optimizer.h
+++ b/paddle/fluid/lite/core/optimizer.h
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 #pragma once
+#include <memory>
 #include <string>
 #include <vector>
 #include "paddle/fluid/lite/core/mir/generate_program_pass.h"
diff --git a/paddle/fluid/lite/core/optimizer_test.cc b/paddle/fluid/lite/core/optimizer_test.cc
index ab9a9cd024027e041866c11d9c95142a1b50657e..ae543dc1b19768a9147af1c3114b46c546318eb2 100644
--- a/paddle/fluid/lite/core/optimizer_test.cc
+++ b/paddle/fluid/lite/core/optimizer_test.cc
@@ -14,6 +14,8 @@
 
 #include "paddle/fluid/lite/core/optimizer.h"
 #include <gtest/gtest.h>
+#include <memory>
+#include <utility>
 #include "paddle/fluid/lite/core/mir/generate_program_pass.h"
 #include "paddle/fluid/lite/core/mir/pass_manager.h"
 #include "paddle/fluid/lite/core/mir/passes.h"
diff --git a/paddle/fluid/lite/core/program.h b/paddle/fluid/lite/core/program.h
index c38bf0630f5692c14f397fe4c6f959740ad33788..33f2f2a7789cbe5ace9c1bcd27a9f274a7684616 100644
--- a/paddle/fluid/lite/core/program.h
+++ b/paddle/fluid/lite/core/program.h
@@ -14,10 +14,11 @@
 
 #pragma once
 #include <list>
+#include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 #include "paddle/fluid/lite/core/kernel.h"
-#include "paddle/fluid/lite/core/kernel.h"
 #include "paddle/fluid/lite/core/mir/node.h"
 #include "paddle/fluid/lite/core/op_lite.h"
 #include "paddle/fluid/lite/core/op_registry.h"
@@ -25,7 +26,7 @@
 namespace paddle {
 namespace lite {
 
-static const std::string kKernelTypeAttr = "__@kernel_type_attr@__";
+static const char kKernelTypeAttr[] = "__@kernel_type_attr@__";
 
 // A program is used to represent a code program, in Paddle, a code program
 // contains:
diff --git a/paddle/fluid/lite/core/program_fake_utils.h b/paddle/fluid/lite/core/program_fake_utils.h
index 7d01c5a467edd0c8e826128a3b0eee6c0e2cdd10..6d24897ca6f5cd7e9347f76eb0af94ff39cf6257 100644
--- a/paddle/fluid/lite/core/program_fake_utils.h
+++ b/paddle/fluid/lite/core/program_fake_utils.h
@@ -13,7 +13,11 @@
 // limitations under the License.
 
 #pragma once
+#include <memory>
+#include <set>
 #include <string>
+#include <utility>
+#include <vector>
 #include "paddle/fluid/framework/program_desc.h"
 #include "paddle/fluid/lite/core/mir/ssa_graph.h"
 #include "paddle/fluid/lite/core/op_registry.h"
diff --git a/paddle/fluid/lite/core/target_wrapper.cc b/paddle/fluid/lite/core/target_wrapper.cc
index 4978f622f975b1f79ef4e53c255824e73202ddbc..5ce5f3942d84b88a9936bd23d86cd5e857687464 100644
--- a/paddle/fluid/lite/core/target_wrapper.cc
+++ b/paddle/fluid/lite/core/target_wrapper.cc
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/core/target_wrapper.h"
+#include <string>
 #include "paddle/fluid/lite/utils/all.h"
 
 namespace paddle {
@@ -43,4 +44,4 @@ std::string Place::DebugString() const {
 }
 
 }  // namespace lite
-}  // namespace paddle
\ No newline at end of file
+}  // namespace paddle
diff --git a/paddle/fluid/lite/core/target_wrapper.h b/paddle/fluid/lite/core/target_wrapper.h
index 7afd6e1926dba4dc83d9cb0ee097c7f06331db74..cf0632d387999356f762fa48dbec8025842f3173 100644
--- a/paddle/fluid/lite/core/target_wrapper.h
+++ b/paddle/fluid/lite/core/target_wrapper.h
@@ -82,11 +82,11 @@ struct Place {
   TargetType target{TARGET(kUnk)};
   PrecisionType precision{PRECISION(kUnk)};
   DataLayoutType layout{DATALAYOUT(kUnk)};
-  short device{0};  // device ID
+  int16_t device{0};  // device ID
 
   Place() = default;
   Place(TargetType target, PrecisionType precision,
-        DataLayoutType layout = DATALAYOUT(kNCHW), short device = 0)
+        DataLayoutType layout = DATALAYOUT(kNCHW), int16_t device = 0)
       : target(target), precision(precision), layout(layout), device(device) {}
 
   bool is_valid() const {
diff --git a/paddle/fluid/lite/core/variable.h b/paddle/fluid/lite/core/variable.h
index dc0a211d60b2d7ba7a0224c37c2e9b8358b86c1c..c83871446d254ab6464241bc5c2eb4db2447e27f 100644
--- a/paddle/fluid/lite/core/variable.h
+++ b/paddle/fluid/lite/core/variable.h
@@ -13,6 +13,8 @@
 // limitations under the License.
 
 #pragma once
+#include <set>
+#include <string>
 #include "paddle/fluid/lite/core/compatible_tensor.h"
 #include "paddle/fluid/lite/utils/all.h"
 
diff --git a/paddle/fluid/lite/cuda/target_wrapper.h b/paddle/fluid/lite/cuda/target_wrapper.h
index 6040f444f02c3c699f9ebe1df2a0d6f84a8d742b..4567db3e2165e187a86186b539161ea292a39eef 100644
--- a/paddle/fluid/lite/cuda/target_wrapper.h
+++ b/paddle/fluid/lite/cuda/target_wrapper.h
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#pragma once
 #include <cuda.h>
 #include <cuda_runtime.h>
 #include "paddle/fluid/lite/core/target_wrapper.h"
diff --git a/paddle/fluid/lite/kernels/host/relu_compute.h b/paddle/fluid/lite/kernels/host/relu_compute.h
index 5a1fd41c1724ade15526eb3114aecabeca59b973..aae9e161aabe30c22c7a12953b98d094a9c6637f 100644
--- a/paddle/fluid/lite/kernels/host/relu_compute.h
+++ b/paddle/fluid/lite/kernels/host/relu_compute.h
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 #pragma once
+#include <algorithm>
 #include "paddle/fluid/lite/core/kernel.h"
 #include "paddle/fluid/lite/core/op_registry.h"
 
diff --git a/paddle/fluid/lite/model_parser/model_parser.cc b/paddle/fluid/lite/model_parser/model_parser.cc
index 54e99cfb1a8367f9d5ee443bfb4de8f59bf67af2..2addf5ff911d9c4404a759e8b14b357b1fb95143 100644
--- a/paddle/fluid/lite/model_parser/model_parser.cc
+++ b/paddle/fluid/lite/model_parser/model_parser.cc
@@ -13,7 +13,9 @@
 // limitations under the License.
 
 #include "paddle/fluid/lite/model_parser/model_parser.h"
+#include <algorithm>
 #include <fstream>
+#include <limits>
 #include "paddle/fluid/lite/core/compatible_tensor.h"
 #include "paddle/fluid/lite/core/scope.h"
 #include "paddle/fluid/lite/core/variable.h"
@@ -218,8 +220,8 @@ void TensorToStream(std::ostream &os, const lite::Tensor &tensor) {
                                     tensor.data_size(), IoDirection::DtoH);
       os.write(static_cast<const char *>(tmp_buffer.get()),
                static_cast<std::streamsize>(size));
-    } else
-#endif  // LITE_WITH_CUDA
+    } else  // NOLINT
+#endif      // LITE_WITH_CUDA
     {
       os.write(static_cast<const char *>(tensor.data<void>()),
                static_cast<std::streamsize>(size));
diff --git a/paddle/fluid/lite/model_parser/model_parser.h b/paddle/fluid/lite/model_parser/model_parser.h
index 8cea4df2c86578bc4b2b4545fa7dccdb20151cce..de40c82ab7d6318acebaa7a2c2f4a394d1081de5 100644
--- a/paddle/fluid/lite/model_parser/model_parser.h
+++ b/paddle/fluid/lite/model_parser/model_parser.h
@@ -15,6 +15,7 @@
 // This file contains model format related operations, such as load a model,
 // parse an operator definitions and so on.
 
+#pragma once
 #include <memory>
 #include <string>
 #include <vector>
diff --git a/paddle/fluid/lite/model_parser/pb/block_desc.cc b/paddle/fluid/lite/model_parser/pb/block_desc.cc
index ce71e4de2b8f6ee4332a47e7d26876f75e0d75fd..2b9c2314cf0d3c5261692c40a9f31d9357bd12fa 100644
--- a/paddle/fluid/lite/model_parser/pb/block_desc.cc
+++ b/paddle/fluid/lite/model_parser/pb/block_desc.cc
@@ -11,3 +11,5 @@
 // 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.
+
+#include "paddle/fluid/lite/model_parser/pb/block_desc.h"
diff --git a/paddle/fluid/lite/model_parser/pb/op_desc.h b/paddle/fluid/lite/model_parser/pb/op_desc.h
index fd1613f78f02c7575359cb40c2a340f11529de0c..b1fbce54d865c9924fbdc686edd87662be4cec4f 100644
--- a/paddle/fluid/lite/model_parser/pb/op_desc.h
+++ b/paddle/fluid/lite/model_parser/pb/op_desc.h
@@ -21,6 +21,8 @@
  */
 
 #include <algorithm>
+#include <map>
+#include <set>
 #include <string>
 #include <unordered_map>
 #include <vector>
@@ -44,7 +46,7 @@ class OpDesc {
  public:
   OpDesc() {}
 
-  OpDesc(const framework::proto::OpDesc &desc) : desc_(desc) {}
+  explicit OpDesc(const framework::proto::OpDesc &desc) : desc_(desc) {}
 
   void CopyFrom(const OpDesc &op_desc) { desc_ = op_desc.ReadonlyProto(); }
 
@@ -127,13 +129,13 @@ class OpDesc {
     }
 
     size_t hash = typeid(T).hash_code();
-    if (hash == typeid(int).hash_code()) {
+    if (hash == typeid(int).hash_code()) {  // NOLINT
       it->set_type(framework::proto::INT);
       it->set_i(v);
-    } else if (hash == typeid(float).hash_code()) {
+    } else if (hash == typeid(float).hash_code()) {  // NOLINT
       it->set_type(framework::proto::FLOAT);
       it->set_f(v);
-    } else if (hash == typeid(bool).hash_code()) {
+    } else if (hash == typeid(bool).hash_code()) {  // NOLINT
       it->set_type(framework::proto::BOOLEAN);
       it->set_b(v);
     } else {
diff --git a/paddle/fluid/lite/model_parser/pb/program_desc.cc b/paddle/fluid/lite/model_parser/pb/program_desc.cc
index ce71e4de2b8f6ee4332a47e7d26876f75e0d75fd..744db6b53e81141aa5aaa56de4b6b6e18085ed8b 100644
--- a/paddle/fluid/lite/model_parser/pb/program_desc.cc
+++ b/paddle/fluid/lite/model_parser/pb/program_desc.cc
@@ -11,3 +11,5 @@
 // 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.
+
+#include "paddle/fluid/lite/model_parser/pb/program_desc.h"
diff --git a/paddle/fluid/lite/model_parser/pb/program_desc.h b/paddle/fluid/lite/model_parser/pb/program_desc.h
index ce71e4de2b8f6ee4332a47e7d26876f75e0d75fd..c718a48c8912ce816054c21102392256bd086d28 100644
--- a/paddle/fluid/lite/model_parser/pb/program_desc.h
+++ b/paddle/fluid/lite/model_parser/pb/program_desc.h
@@ -11,3 +11,5 @@
 // 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.
+
+#pragma once
diff --git a/paddle/fluid/lite/model_parser/pb/var_desc.cc b/paddle/fluid/lite/model_parser/pb/var_desc.cc
index 2aaf5ee14d05358dd944f98ff97c8d6dafefc53c..ac19f323b73f4aa3d977d4f275658675f76a2a5c 100644
--- a/paddle/fluid/lite/model_parser/pb/var_desc.cc
+++ b/paddle/fluid/lite/model_parser/pb/var_desc.cc
@@ -18,7 +18,7 @@ namespace paddle {
 namespace lite {
 namespace pb {
 
-using namespace framework;
+using namespace framework;  // NOLINT
 
 proto::VarType::Type VarDesc::GetType() const { return desc_.type().type(); }
 
diff --git a/paddle/fluid/lite/model_parser/pb/var_desc.h b/paddle/fluid/lite/model_parser/pb/var_desc.h
index a29bd223928b027b9f5439dab903854fe26f6d25..55b756c4a05f3db86fd9243c70409d6363c579b2 100644
--- a/paddle/fluid/lite/model_parser/pb/var_desc.h
+++ b/paddle/fluid/lite/model_parser/pb/var_desc.h
@@ -119,5 +119,5 @@ class VarDesc {
 };
 
 }  // namespace pb
-}  // namespace framework
+}  // namespace lite
 }  // namespace paddle
diff --git a/paddle/fluid/lite/model_parser/runtime.h b/paddle/fluid/lite/model_parser/runtime.h
index 38f710c941405c02a44f13aff9d1991116133bc3..22beeb41abdb4ae6b3c96db3256569b23cb944d4 100644
--- a/paddle/fluid/lite/model_parser/runtime.h
+++ b/paddle/fluid/lite/model_parser/runtime.h
@@ -13,6 +13,9 @@
 // limitations under the License.
 
 #pragma once
+#include <map>
+#include <string>
+#include <vector>
 #include "paddle/fluid/framework/framework.pb.h"
 #include "paddle/fluid/lite/utils/all.h"
 
@@ -44,7 +47,7 @@ class LoDTensorDesc {
   TensorDesc tensor;
   int lod_level{-1};
 
-  LoDTensorDesc(const framework::proto::VarType_LoDTensorDesc& proto) {
+  explicit LoDTensorDesc(const framework::proto::VarType_LoDTensorDesc& proto) {
     Parse(proto);
   }
 
diff --git a/paddle/fluid/lite/operators/relu_op.h b/paddle/fluid/lite/operators/relu_op.h
index 66929cef199ef0e4fc2f0db36eb9d598d8896af6..a6204a107d8c4b8fe0bf7f3a58dad20d909fda86 100644
--- a/paddle/fluid/lite/operators/relu_op.h
+++ b/paddle/fluid/lite/operators/relu_op.h
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#pragma once
 #include <string>
 #include <vector>
 #include "paddle/fluid/lite/core/op_lite.h"
@@ -25,7 +26,7 @@ namespace operators {
 class ReluOp : public OpLite {
  public:
   ReluOp() {}
-  ReluOp(const std::string &op_type) : OpLite(op_type) {}
+  explicit ReluOp(const std::string &op_type) : OpLite(op_type) {}
 
   bool CheckShape() const override;
 
diff --git a/paddle/fluid/lite/utils/cp_logging.h b/paddle/fluid/lite/utils/cp_logging.h
index 3fac352c1b2583b668a21ac3ac9a7a628cbc7aba..d356b337abd5f46a0ca30eb9743ee5aa47d4f31e 100644
--- a/paddle/fluid/lite/utils/cp_logging.h
+++ b/paddle/fluid/lite/utils/cp_logging.h
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#pragma once
 #ifdef LITE_WITH_LIGHT_WEIGHT_FRAMEWORK
 #include "paddle/fluid/lite/utils/logging.h"
 #else  // LITE_WITH_LIGHT_WEIGHT_FRAMEWORK
diff --git a/paddle/fluid/lite/utils/factory.h b/paddle/fluid/lite/utils/factory.h
index 97abce93747c98d24c7e5013fd8c6cea4c1e20aa..3f86df842f1190178b269a278b941b3ddfd597ea 100644
--- a/paddle/fluid/lite/utils/factory.h
+++ b/paddle/fluid/lite/utils/factory.h
@@ -17,7 +17,9 @@
 #include <list>
 #include <memory>
 #include <sstream>
+#include <string>
 #include <unordered_map>
+#include <utility>
 #include "paddle/fluid/lite/utils/cp_logging.h"
 
 namespace paddle {
@@ -83,7 +85,7 @@ class Factory {
 template <typename Type>
 class Registor {
  public:
-  Registor(std::function<void()>&& functor) { functor(); }
+  explicit Registor(std::function<void()>&& functor) { functor(); }
 
   // Touch will do nothing.
   int Touch() { return 0; }
diff --git a/paddle/fluid/lite/utils/logging.h b/paddle/fluid/lite/utils/logging.h
index 37294969920b26681e74f8e932fe41c07059b669..9c2b6903e4905a94ffeba92178f5e896ba92a768 100644
--- a/paddle/fluid/lite/utils/logging.h
+++ b/paddle/fluid/lite/utils/logging.h
@@ -34,6 +34,7 @@
 #define VLOG(level) LOG_INFO.stream()
 
 // CHECK()
+// NOLINT
 #define CHECK(x)                                             \
   if (!(x))                                                  \
   paddle::lite::LogMessageFatal(__FILE__, __LINE__).stream() \
diff --git a/paddle/fluid/lite/utils/varient.h b/paddle/fluid/lite/utils/varient.h
index 4a9eacabfdfe7c8b8f566062c646fa7bbadb560c..49714eea7f43915e178f127002e05eca9d1a4690 100644
--- a/paddle/fluid/lite/utils/varient.h
+++ b/paddle/fluid/lite/utils/varient.h
@@ -13,10 +13,12 @@
 // limitations under the License.
 
 #pragma once
+#include <algorithm>
 #include <exception>
 #include <memory>
 #include <type_traits>
 #include <typeinfo>
+#include <utility>
 #include "paddle/fluid/lite/utils/cp_logging.h"
 
 // This is an equivalent implementation of boost::any. We implement this to
diff --git a/paddle/fluid/lite/x86/target_wrapper.cc b/paddle/fluid/lite/x86/target_wrapper.cc
index 8fd941591ac3eb24d4667eb4a7fe172f5c84c0eb..76a108c6d1d4399ec55dc840e5fb71779c6e8ba8 100644
--- a/paddle/fluid/lite/x86/target_wrapper.cc
+++ b/paddle/fluid/lite/x86/target_wrapper.cc
@@ -15,6 +15,7 @@
 #include "paddle/fluid/lite/core/target_wrapper.h"
 #include <algorithm>
 #include "paddle/fluid/lite/utils/all.h"
+#include "paddle/fluid/lite/x86/target_wrapper.h"
 
 namespace paddle {
 namespace lite {