diff --git a/mindspore/lite/micro/cmake/file_list.cmake b/mindspore/lite/micro/cmake/file_list.cmake
index 80b6f3d413..7417c87d5a 100644
--- a/mindspore/lite/micro/cmake/file_list.cmake
+++ b/mindspore/lite/micro/cmake/file_list.cmake
@@ -78,6 +78,7 @@ set(CODER_OPCODERS_SRC
         ${MICRO_DIR}/coder/opcoders/nnacl/fp32/tile_fp32_coder.cc
         ${MICRO_DIR}/coder/opcoders/nnacl/fp32/transpose_fp32_coder.cc
         #### nnacl int8 coder
+        ${MICRO_DIR}/coder/opcoders/nnacl/int8/add_int8_coder.cc
         ${MICRO_DIR}/coder/opcoders/nnacl/int8/concat_int8_coder.cc
         ${MICRO_DIR}/coder/opcoders/nnacl/int8/fullconnection_int8_coder.cc
         ${MICRO_DIR}/coder/opcoders/nnacl/int8/matmul_int8_coder.cc
diff --git a/mindspore/lite/micro/cmake/wrapper.cmake b/mindspore/lite/micro/cmake/wrapper.cmake
index fa24daa627..5c2fd0c21f 100644
--- a/mindspore/lite/micro/cmake/wrapper.cmake
+++ b/mindspore/lite/micro/cmake/wrapper.cmake
@@ -4,9 +4,9 @@ set(MICRO_WRAPPER_SRC
         ${LITE_DIR}/src/runtime/thread_pool.c
         ${MICRO_DIR}/wrapper/fp32/matmul_fp32_wrapper.c
         ${MICRO_DIR}/wrapper/int8/matmul_int8_wrapper.c
-        ${MICRO_DIR}/wrapper/int8/conv_init_int8.c
-        ${MICRO_DIR}/wrapper/int8/conv1x1_init_int8.c
-        ${MICRO_DIR}/wrapper/int8/conv1x1_run_int8.c
+        ${MICRO_DIR}/wrapper/int8/conv_init_int8_wrapper.c
+        ${MICRO_DIR}/wrapper/int8/conv1x1_init_int8_wrapper.c
+        ${MICRO_DIR}/wrapper/int8/conv1x1_run_int8_wrapper.c
         )
 
 list(APPEND FILE_SET ${MICRO_WRAPPER_SRC})
\ No newline at end of file
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.cc b/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.cc
index a9dd2ea0ae..be9041267b 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.cc
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.cc
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.h b/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.h
index e16a0e3db3..6946459d42 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.h
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/fp32/convolution_depthwise_fp32_coder.h
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/add_int8_coder.cc b/mindspore/lite/micro/coder/opcoders/nnacl/int8/add_int8_coder.cc
new file mode 100644
index 0000000000..afe7085610
--- /dev/null
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/add_int8_coder.cc
@@ -0,0 +1,156 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ *
+ * 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 "micro/coder/opcoders/nnacl/int8/add_int8_coder.h"
+#include <algorithm>
+#include <type_traits>
+#include "nnacl/int8/quantize.h"
+#include "micro/coder/log.h"
+#include "micro/coder/opcoders/serializers/nnacl_serializer/nnacl_int8_serializer.h"
+#include "micro/coder/opcoders/file_collector.h"
+
+using mindspore::schema::PrimitiveType_Add;
+
+namespace mindspore::lite::micro {
+
+int AddInt8Coder::Prepare(CoderContext *const context) {
+  input0 = input_tensors().at(0);
+  input1 = input_tensors().at(1);
+  MS_CHECK_PTR(input0);
+  MS_CHECK_PTR(input1);
+
+  MS_CHECK_RET_CODE(Init(), "Init failed");
+  MS_CHECK_RET_CODE(ReSize(), "ReSize failed");
+
+  return RET_OK;
+}
+
+int AddInt8Coder::DoCode(CoderContext *const context) {
+  Collect(context, {"wrapper/int8/conv1x1_init_int8.h"}, {"add_int8_wrapper.c", "add_int8.c", "thread_pool.c"});
+
+  nnacl::NNaclInt8Serializer code;
+
+  code.CodeStruct("para", para_);
+  code.CodeStruct("arith_para", *arith_para_);
+  code.CodeBaseStruct("AddArgs", "args", "para", "arith_para", in_size_, out_size_, thread_num_s_, elements_num_,
+                      support_opt_add_, input0, input1, output_tensor_);
+
+  if (arith_para_->broadcasting_) {
+    code.CodeFunction("ParallelLaunch", "THREAD_POOL_DEFAULT", "AddBroadcastRun", "&args", thread_num_s_);
+  } else {
+    code.CodeFunction("ParallelLaunch", "THREAD_POOL_DEFAULT", "AddRun", "&args", thread_num_s_);
+  }
+
+  return RET_OK;
+}
+
+int AddInt8Coder::Init() {
+  para_.in0_args_.zp_ = input0->quant_params().front().zeroPoint * -1;
+  para_.in1_args_.zp_ = input1->quant_params().front().zeroPoint * -1;
+  para_.out_zp_ = output_tensor_->quant_params().front().zeroPoint;
+
+  const double in0_scale = input0->quant_params().front().scale;
+  const double in1_scale = input1->quant_params().front().scale;
+  const double out_scale = output_tensor_->quant_params().front().scale;
+
+  para_.left_shift_ = 20;
+  const double twice_max_input_scale = 2 * std::max(in0_scale, in1_scale);
+  const double in0_multiplier = in0_scale / twice_max_input_scale;
+  const double in1_multiplier = in1_scale / twice_max_input_scale;
+  const double out_multiplier = twice_max_input_scale / ((1 << para_.left_shift_) * out_scale);
+
+  QuantizeMultiplierSmallerThanOne(in0_multiplier, &para_.in0_args_.multiplier_, &para_.in0_args_.left_shift_);
+  QuantizeMultiplierSmallerThanOne(in1_multiplier, &para_.in1_args_.multiplier_, &para_.in1_args_.left_shift_);
+  QuantizeMultiplierSmallerThanOne(out_multiplier, &para_.out_multiplier_, &para_.out_left_shift_);
+
+  para_.in0_args_.right_shift_ = -para_.in0_args_.left_shift_ > 0 ? 0 : para_.in0_args_.left_shift_;
+  para_.in1_args_.right_shift_ = -para_.in1_args_.left_shift_ > 0 ? 0 : para_.in1_args_.left_shift_;
+  para_.out_right_shift_ = -para_.out_left_shift_ > 0 ? 0 : para_.out_left_shift_;
+
+  para_.in0_args_.left_shift_ = -para_.in0_args_.left_shift_ > 0 ? -para_.in0_args_.left_shift_ : 0;
+  para_.in1_args_.left_shift_ = -para_.in1_args_.left_shift_ > 0 ? -para_.in1_args_.left_shift_ : 0;
+  para_.out_left_shift_ = -para_.out_left_shift_ > 0 ? -para_.out_left_shift_ : 0;
+
+  auto act = arith_para_->activation_type_;
+  CalculateActivationRangeQuantized(act == ActType_Relu, act == ActType_Relu6, para_.out_zp_,
+                                    static_cast<float>(out_scale), &para_.min_, &para_.max_);
+  return RET_OK;
+}
+
+int AddInt8Coder::ReSize() {
+  support_opt_add_ = (input0->ElementsNum() == 1) || (input1->ElementsNum() == 1);
+  if (support_opt_add_) {
+    arith_para_->broadcasting_ = false;
+  }
+
+  elements_num_ = output_tensor_->ElementsNum();
+
+  arith_para_->in_elements_num0_ = input_tensors_[0]->ElementsNum();
+  arith_para_->in_elements_num1_ = input_tensors_[1]->ElementsNum();
+  arith_para_->out_elements_num_ = output_tensors_[0]->ElementsNum();
+
+  for (size_t i = 0; i < input_tensors_.at(0)->shape().size(); i++) {
+    if (arith_para_->in_shape0_[i] == -1) {
+      MS_CHECK_RET_CODE(memcpy_s(arith_para_->in_shape0_, std::extent<decltype(arith_para_->in_shape0_)>::value,
+                                 input0->shape().data(), input0->shape().size() * sizeof(int)),
+                        "memcpy failed");
+      break;
+    }
+  }
+  for (size_t i = 0; i < input_tensors_.at(1)->shape().size(); i++) {
+    if (arith_para_->in_shape1_[i] == -1) {
+      MS_CHECK_RET_CODE(memcpy_s(arith_para_->in_shape1_, std::extent<decltype(arith_para_->in_shape1_)>::value,
+                                 input1->shape().data(), input1->shape().size() * sizeof(int)),
+                        "memcpy failed");
+      break;
+    }
+  }
+  for (size_t i = 0; i < output_tensor_->shape().size(); i++) {
+    if (arith_para_->out_shape_[i] == -1) {
+      MS_CHECK_RET_CODE(memcpy_s(arith_para_->out_shape_, std::extent<decltype(arith_para_->out_shape_)>::value,
+                                 output_tensor_->shape().data(), output_tensor_->shape().size() * sizeof(int)),
+                        "memcpy failed");
+      break;
+    }
+  }
+
+  if (arith_para_->broadcasting_) {
+    size_t break_pos_ = 0;
+    for (auto i = arith_para_->ndim_ - 1; i >= 0; --i) {
+      if (arith_para_->in_shape0_[i] != arith_para_->in_shape1_[i]) {
+        break_pos_ = i;
+        break;
+      }
+    }
+    in_size_ = 1;
+    out_size_ = 1;
+    for (size_t i = 0; i < arith_para_->ndim_; i++) {
+      if (i > break_pos_) {
+        in_size_ *= arith_para_->out_shape_[i];
+      } else {
+        out_size_ *= arith_para_->out_shape_[i];
+      }
+    }
+
+    ComputeStrides(arith_para_->in_shape0_, arith_para_->in_strides0_, arith_para_->ndim_);
+    ComputeStrides(arith_para_->in_shape1_, arith_para_->in_strides1_, arith_para_->ndim_);
+    ComputeStrides(arith_para_->out_shape_, arith_para_->out_strides_, arith_para_->ndim_);
+  }
+  return RET_OK;
+}
+
+REG_OPERATOR_CODER(kAllTargets, kNumberTypeInt8, PrimitiveType_Add, CPUOpCoderCreator<AddInt8Coder>)
+}  // namespace mindspore::lite::micro
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/add_int8_coder.h b/mindspore/lite/micro/coder/opcoders/nnacl/int8/add_int8_coder.h
new file mode 100644
index 0000000000..c1064a89a9
--- /dev/null
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/add_int8_coder.h
@@ -0,0 +1,53 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ *
+ * 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.
+ */
+
+#ifndef MINDSPORE_LITE_MICRO_CODER_OPCODERS_NNACL_ADD_INT8_CODER_H_
+#define MINDSPORE_LITE_MICRO_CODER_OPCODERS_NNACL_ADD_INT8_CODER_H_
+
+#include <vector>
+#include "micro/coder/opcoders/op_coder.h"
+#include "nnacl/int8/add_int8.h"
+
+namespace mindspore::lite::micro {
+class AddInt8Coder : public OperatorCoder {
+ public:
+  AddInt8Coder(const std::vector<Tensor *> &in_tensors, const std::vector<Tensor *> &out_tensors,
+               const Model::Node *node, size_t node_index, Target target)
+      : OperatorCoder(in_tensors, out_tensors, node, node_index, target) {
+    arith_para_ = reinterpret_cast<ArithmeticParameter *>(parameter_);
+  }
+
+  ~AddInt8Coder() override = default;
+
+  int Prepare(CoderContext *const context) override;
+
+  int DoCode(CoderContext *const context) override;
+
+ private:
+  int Init();
+  int ReSize();
+
+  AddQuantParameter para_;
+  ArithmeticParameter *arith_para_{nullptr};
+  Tensor *input0{nullptr};
+  Tensor *input1{nullptr};
+  int in_size_{0};
+  int out_size_{0};
+  int elements_num_{0};
+  bool support_opt_add_{false};
+};
+}  // namespace mindspore::lite::micro
+#endif  // MINDSPORE_LITE_MICRO_CODER_OPCODERS_NNACL_ADD_INT8_CODER_H_
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.cc b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.cc
index 8966f4715e..6b9c1c6608 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.cc
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.cc
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -45,8 +45,8 @@ int Conv2D1x1Int8Coder::DoCode(CoderContext *const context) {
   Collect(context,
           {"nnacl/int8/conv1x1_int8.h", "nnacl/common_func.h", "wrapper/int8/conv1x1_init_int8.h",
            "wrapper/int8/conv1x1_run_int8.h"},
-          {"common_func.c", "pack.c", "conv1x1_int8.c", "matmul_int8.c", "fixed_point.c", "conv1x1_init_int8.c",
-           "conv1x1_run_int8.c"});
+          {"common_func.c", "pack.c", "conv1x1_int8.c", "matmul_int8.c", "fixed_point.c", "conv1x1_init_int8_wrapper.c",
+           "conv1x1_run_int8_wrapper.c", "thread_pool.c"});
 
   nnacl::NNaclInt8Serializer code;
 
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.h b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.h
index 04a232484b..0f1d9e144a 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.h
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_1x1_int8_coder.h
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.cc b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.cc
index 176e5ddbdd..30ccf2f976 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.cc
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.cc
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.h b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.h
index eb05af43bf..d32a7aa4c6 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.h
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_3x3_int8_coder.h
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.cc b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.cc
index 8d9a8f2220..0c60ecc62d 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.cc
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.cc
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -182,7 +182,7 @@ int Conv2DINT8Coder::Resize() {
 int Conv2DINT8Coder::DoCode(CoderContext *const context) {
   Collect(context, {"nnacl/int8/conv_int8.h", "nnacl/common_func.h", "nnacl/kernel/int8/conv_init_int8.h"},
           {"common_func.c", "pack.c", "conv_int8.c", "winograd_transform.c", "matmul_int8.c", "fixed_point.c",
-           "conv_init_int8.c"});
+           "conv_init_int8_wrapper.c", "thread_pool.c"});
   // call the op function
   nnacl::NNaclInt8Serializer code;
   code.precision(kPrecision);
diff --git a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.h b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.h
index 52859be068..c89496ce39 100644
--- a/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.h
+++ b/mindspore/lite/micro/coder/opcoders/nnacl/int8/conv2d_int8_coder.h
@@ -1,5 +1,5 @@
 /**
- * Copyright 2020 Huawei Technologies Co., Ltd
+ * Copyright 2021 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
diff --git a/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_int8_serializer.cc b/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_int8_serializer.cc
index ea814afb42..fc0e3093fc 100644
--- a/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_int8_serializer.cc
+++ b/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_int8_serializer.cc
@@ -70,6 +70,13 @@ void NNaclInt8Serializer::CodeStruct(const std::string &name, const MatMulParame
                  matmul_parameter.b_const_, matmul_parameter.act_type_);
 }
 
+void NNaclInt8Serializer::CodeStruct(const std::string &name, const AddQuantParameter &add_quant_parameter) {
+  CodeBaseStruct("AddQuantParameter", name, add_quant_parameter.left_shift_, add_quant_parameter.min_,
+                 add_quant_parameter.max_, add_quant_parameter.in0_args_, add_quant_parameter.in1_args_,
+                 add_quant_parameter.out_zp_, add_quant_parameter.out_left_shift_, add_quant_parameter.out_right_shift_,
+                 add_quant_parameter.out_multiplier_);
+}
+
 void NNaclInt8Serializer::CodeStruct(const std::string &name, const ArithmeticParameter &arithmetic_parameter) {
   CodeBaseStruct("ArithmeticParameter", name, arithmetic_parameter.op_parameter_, arithmetic_parameter.broadcasting_,
                  arithmetic_parameter.ndim_, arithmetic_parameter.activation_type_,
diff --git a/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_stream_utils.h b/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_stream_utils.h
index 9fe57a7eb7..3d59511609 100644
--- a/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_stream_utils.h
+++ b/mindspore/lite/micro/coder/opcoders/serializers/nnacl_serializer/nnacl_stream_utils.h
@@ -20,6 +20,7 @@
 #include <string>
 #include "nnacl/pooling_parameter.h"
 #include "nnacl/softmax_parameter.h"
+#include "nnacl/int8/add_int8.h"
 #include "nnacl/int8/quantize.h"
 
 namespace mindspore::lite::micro {
@@ -35,6 +36,11 @@ inline std::ostream &operator<<(std::ostream &code, const OpParameter &tile) {
   return code;
 }
 
+inline std::ostream &operator<<(std::ostream &code, const AddQuantQrgs &args) {
+  code << "{" << args.zp_ << ", " << args.left_shift_ << ", " << args.right_shift_ << ", " << args.multiplier_ << "}";
+  return code;
+}
+
 inline std::ostream &operator<<(std::ostream &code, PoolMode pool_mode) {
   code << "(PoolMode)"
        << "(" << static_cast<int>(pool_mode) << ")";
diff --git a/mindspore/lite/micro/wrapper/int8/add_int8_wrapper.c b/mindspore/lite/micro/wrapper/int8/add_int8_wrapper.c
new file mode 100644
index 0000000000..a179c668d5
--- /dev/null
+++ b/mindspore/lite/micro/wrapper/int8/add_int8_wrapper.c
@@ -0,0 +1,69 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ *
+ * 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 "wrapper/int8/add_int8_wrapper.h"
+#include "nnacl/errorcode.h"
+
+int AddBroadcastRun(void *cdata, int task_id) {
+  AddArgs *args = (AddArgs *)(cdata);
+  int stride = UP_DIV(args->out_size_, args->thread_count_);
+  int real_out_count = MSMIN(stride, args->out_size_ - stride * task_id);
+  if (real_out_count <= 0) {
+    return NNACL_OK;
+  }
+  int8_t *cur_in0 = NULL;
+  int8_t *cur_in1 = NULL;
+  int8_t *cur_out = NULL;
+  for (int i = 0; i < real_out_count; i++) {
+    if (args->arith_para_->in_elements_num0_ == args->arith_para_->out_elements_num_) {
+      cur_in0 = args->input0_data_ + task_id * stride * args->in_size_ + i * args->in_size_;
+      cur_in1 = args->input1_data_;
+      cur_out = args->output_data_ + task_id * stride * args->in_size_ + i * args->in_size_;
+    } else {
+      cur_in0 = args->input0_data_;
+      cur_in1 = args->input1_data_ + task_id * stride * args->in_size_ + i * args->in_size_;
+      cur_out = args->output_data_ + task_id * stride * args->in_size_ + i * args->in_size_;
+    }
+    AddInt8(cur_in0, cur_in1, cur_out, args->in_size_, &args->para_);
+  }
+  return NNACL_OK;
+}
+
+int AddRun(void *cdata, int task_id) {
+  AddArgs *args = (AddArgs *)(cdata);
+  /* no need broadcast */
+  int stride = UP_DIV(args->elements_num_, args->thread_count_);
+  int rest_count = args->elements_num_ - task_id * stride;
+  int real_count = MSMIN(stride, rest_count);
+  if (real_count <= 0) {
+    return NNACL_OK;
+  }
+  int8_t *cur_in0 = args->input0_data_ + stride * task_id;
+  int8_t *cur_in1 = args->input1_data_ + stride * task_id;
+  int8_t *cur_out = args->output_data_ + stride * task_id;
+  if (args->support_opt_add_) {
+    int8_t *ptr_in = args->arith_para_->in_elements_num0_ == 1 ? cur_in1 : cur_in0;
+    int8_t element_in = args->arith_para_->in_elements_num0_ == 1 ? args->input0_data_[0] : args->input1_data_[0];
+    AddQuantQrgs *ptr_args =
+      args->arith_para_->in_elements_num0_ == 1 ? &args->para_.in1_args_ : &args->para_.in0_args_;
+    AddQuantQrgs *ele_args =
+      args->arith_para_->in_elements_num0_ == 1 ? &args->para_.in0_args_ : &args->para_.in1_args_;
+    AddOptInt8(ptr_in, element_in, cur_out, rest_count, &args->para_, ptr_args, ele_args);
+  } else {
+    AddInt8(cur_in0, cur_in1, cur_out, rest_count, &args->para_);
+  }
+  return NNACL_OK;
+}
diff --git a/mindspore/lite/micro/wrapper/int8/add_int8_wrapper.h b/mindspore/lite/micro/wrapper/int8/add_int8_wrapper.h
new file mode 100644
index 0000000000..c23e32e3ae
--- /dev/null
+++ b/mindspore/lite/micro/wrapper/int8/add_int8_wrapper.h
@@ -0,0 +1,50 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ *
+ * 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.
+ */
+
+#ifndef MINDSPORE_LITE_MICRO_WRAPPER_INT8_ADD_INT8_WRAPPER_H_
+#define MINDSPORE_LITE_MICRO_WRAPPER_INT8_ADD_INT8_WRAPPER_H_
+#include <string.h>
+#include "nnacl/int8/matmul_int8.h"
+#include "src/runtime/thread_pool.h"
+#include "nnacl/int8/add_int8.h"
+#include "nnacl/arithmetic.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct {
+  AddQuantParameter para_;
+  ArithmeticParameter *arith_para_;
+  int in_size_;
+  int out_size_;
+  int thread_count_;
+  int elements_num_;
+  bool support_opt_add_;
+  int8_t *input0_data_;
+  int8_t *input1_data_;
+  int8_t *output_data_;
+} AddArgs;
+
+int AddBroadcastRun(void *cdata, int task_id);
+
+int AddRun(void *cdata, int task_id);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif  // MINDSPORE_LITE_MICRO_WRAPPER_INT8_ADD_INT8_WRAPPER_H_
diff --git a/mindspore/lite/micro/wrapper/int8/conv1x1_init_int8.c b/mindspore/lite/micro/wrapper/int8/conv1x1_init_int8_wrapper.c
similarity index 98%
rename from mindspore/lite/micro/wrapper/int8/conv1x1_init_int8.c
rename to mindspore/lite/micro/wrapper/int8/conv1x1_init_int8_wrapper.c
index 315efceb2b..109532eb48 100644
--- a/mindspore/lite/micro/wrapper/int8/conv1x1_init_int8.c
+++ b/mindspore/lite/micro/wrapper/int8/conv1x1_init_int8_wrapper.c
@@ -14,7 +14,7 @@
  * limitations under the License.
  */
 
-#include "wrapper/int8/conv1x1_init_int8.h"
+#include "wrapper/int8/conv1x1_init_int8_wrapper.h"
 #include <memory.h>
 #include "nnacl/int8/matmul_int8.h"
 #include "nnacl/errorcode.h"
diff --git a/mindspore/lite/micro/wrapper/int8/conv1x1_init_int8.h b/mindspore/lite/micro/wrapper/int8/conv1x1_init_int8_wrapper.h
similarity index 100%
rename from mindspore/lite/micro/wrapper/int8/conv1x1_init_int8.h
rename to mindspore/lite/micro/wrapper/int8/conv1x1_init_int8_wrapper.h
diff --git a/mindspore/lite/micro/wrapper/int8/conv1x1_run_int8.c b/mindspore/lite/micro/wrapper/int8/conv1x1_run_int8_wrapper.c
similarity index 99%
rename from mindspore/lite/micro/wrapper/int8/conv1x1_run_int8.c
rename to mindspore/lite/micro/wrapper/int8/conv1x1_run_int8_wrapper.c
index 82e42bf125..ddaa1ecee8 100644
--- a/mindspore/lite/micro/wrapper/int8/conv1x1_run_int8.c
+++ b/mindspore/lite/micro/wrapper/int8/conv1x1_run_int8_wrapper.c
@@ -14,7 +14,7 @@
  * limitations under the License.
  */
 
-#include "wrapper/int8/conv1x1_run_int8.h"
+#include "wrapper/int8/conv1x1_run_int8_wrapper.h"
 #include "nnacl/base/conv1x1_base.h"
 #include "nnacl/int8/matmul_int8.h"
 #include "nnacl/int8/pack_int8.h"
diff --git a/mindspore/lite/micro/wrapper/int8/conv1x1_run_int8.h b/mindspore/lite/micro/wrapper/int8/conv1x1_run_int8_wrapper.h
similarity index 100%
rename from mindspore/lite/micro/wrapper/int8/conv1x1_run_int8.h
rename to mindspore/lite/micro/wrapper/int8/conv1x1_run_int8_wrapper.h
diff --git a/mindspore/lite/micro/wrapper/int8/conv_init_int8.c b/mindspore/lite/micro/wrapper/int8/conv_init_int8_wrapper.c
similarity index 98%
rename from mindspore/lite/micro/wrapper/int8/conv_init_int8.c
rename to mindspore/lite/micro/wrapper/int8/conv_init_int8_wrapper.c
index 276e9d9902..71e8db0fee 100644
--- a/mindspore/lite/micro/wrapper/int8/conv_init_int8.c
+++ b/mindspore/lite/micro/wrapper/int8/conv_init_int8_wrapper.c
@@ -14,7 +14,7 @@
  * limitations under the License.
  */
 
-#include "wrapper/int8/conv_init_int8.h"
+#include "wrapper/int8/conv_init_int8_wrapper.h"
 #include <memory.h>
 #include "nnacl/op_base.h"
 #include "nnacl/int8/matmul_int8.h"
diff --git a/mindspore/lite/micro/wrapper/int8/conv_init_int8.h b/mindspore/lite/micro/wrapper/int8/conv_init_int8_wrapper.h
similarity index 100%
rename from mindspore/lite/micro/wrapper/int8/conv_init_int8.h
rename to mindspore/lite/micro/wrapper/int8/conv_init_int8_wrapper.h