* Copyright (c) 2026 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
#include <iostream>
#include <fstream>
#include <string.h>
#include <stdint.h>
#include <vector>
#include <string>
#include <map>
#include "assert.h"
#include "graph.h"
#include "types.h"
#include "tensor.h"
#include "ge_error_codes.h"
#include "ge_api_types.h"
#include "ge_api.h"
#include "array_ops.h"
#include "ge_ir_build.h"
#include "experiment_ops.h"
#include "nn_other.h"
#include "../../op_graph/avg_pool_proto.h"
#define FAILED -1
#define SUCCESS 0
using namespace ge;
using std::map;
using std::string;
using std::vector;
#define ADD_INPUT(inputIndex, inputName, inputDtype, inputShape) \
vector<int64_t> placeholder##inputIndex##_shape = inputShape; \
auto placeholder##inputIndex = op::Data("placeholder" + inputIndex).set_attr_index(0); \
TensorDesc placeholder##inputIndex##_desc = \
TensorDesc(ge::Shape(placeholder##inputIndex##_shape), FORMAT_ND, inputDtype); \
placeholder##inputIndex##_desc.SetPlacement(ge::kPlacementHost); \
placeholder##inputIndex##_desc.SetFormat(FORMAT_ND); \
Tensor tensor_placeholder##inputIndex; \
ret = GenOnesData( \
placeholder##inputIndex##_shape, tensor_placeholder##inputIndex, placeholder##inputIndex##_desc, inputDtype, \
2); \
if (ret != SUCCESS) { \
printf("%s - ERROR - [XIR]: Generate input data failed\n", GetTime().c_str()); \
return FAILED; \
} \
placeholder##inputIndex.update_input_desc_x(placeholder##inputIndex##_desc); \
input.push_back(tensor_placeholder##inputIndex); \
graph.AddOp(placeholder##inputIndex); \
add1.set_input_##inputName(placeholder##inputIndex); \
inputs.push_back(placeholder##inputIndex)
#define ADD_OUTPUT(outputIndex, outputName, outputDtype, outputShape) \
TensorDesc outputName##outputIndex##_desc = TensorDesc(ge::Shape(outputShape), FORMAT_ND, outputDtype); \
add1.update_output_desc_##outputName(outputName##outputIndex##_desc)
#define LOG_PRINT(message, ...) \
do { \
printf(message, ##__VA_ARGS__); \
} while (0)
#define ADD_INPUT_ATTR(attrName, attrValue) add1.set_attr_##attrName(attrValue)
string GetTime()
{
time_t timep;
time(&timep);
char tmp[64];
strftime(tmp, sizeof(tmp), "%Y-%m-%d %H:%M:%S,000", localtime(&timep));
return tmp;
}
uint32_t GetDataTypeSize(DataType dt)
{
uint32_t dilation = 1;
uint32_t oneByte = 1;
uint32_t twoByte = 2;
uint32_t fourByte = 4;
uint32_t eightByte = 8;
if (dt == ge::DT_FLOAT) {
dilation = fourByte;
} else if (dt == ge::DT_FLOAT16) {
dilation = twoByte;
} else if (dt == ge::DT_BF16) {
dilation = twoByte;
} else if (dt == ge::DT_INT16) {
dilation = twoByte;
} else if (dt == ge::DT_UINT16) {
dilation = twoByte;
} else if (dt == ge::DT_INT32) {
dilation = fourByte;
} else if (dt == ge::DT_UINT32) {
dilation = fourByte;
} else if (dt == ge::DT_INT64) {
dilation = eightByte;
} else if (dt == ge::DT_UINT64) {
dilation = eightByte;
} else if (dt == ge::DT_INT8) {
dilation = oneByte;
}
return dilation;
}
int32_t GenOnesData(
vector<int64_t> shapes, Tensor& input_tensor, TensorDesc& input_tensor_desc, DataType data_type, int value)
{
input_tensor_desc.SetRealDimCnt(shapes.size());
size_t size = 1;
for (uint32_t i = 0; i < shapes.size(); i++) {
size *= shapes[i];
}
uint32_t data_len = size * GetDataTypeSize(data_type);
int32_t* pData = new (std::nothrow) int32_t[data_len];
for (uint32_t i = 0; i < size; ++i) {
*(pData + i) = value;
}
input_tensor = Tensor(input_tensor_desc, reinterpret_cast<uint8_t*>(pData), data_len);
return SUCCESS;
}
int32_t WriteDataToFile(string bin_file, uint64_t data_size, uint8_t* inputData)
{
FILE* fp = fopen(bin_file.c_str(), "w");
fwrite(inputData, sizeof(uint8_t), data_size, fp);
fclose(fp);
return SUCCESS;
}
int CreateOppInGraph(
DataType inDtype, std::vector<ge::Tensor>& input, std::vector<Operator>& inputs, std::vector<Operator>& outputs,
Graph& graph)
{
Status ret = SUCCESS;
auto add1 = op::AvgPool("avg_pool");
vector<vector<int64_t>> shapes = {
{4, 1, 4, 4}
};
vector<vector<int64_t>> attrs = {
{1, 1, 2, 2},
{1, 1, 2, 2}
};
ADD_INPUT(1, x, inDtype, shapes[0]);
ADD_OUTPUT(2, y, inDtype, shapes[0]);
ADD_INPUT_ATTR(ksize, attrs[0]);
ADD_INPUT_ATTR(strides, attrs[1]);
ADD_INPUT_ATTR(padding, "VALID");
ADD_INPUT_ATTR(data_format, "NCHW");
outputs.push_back(add1);
return SUCCESS;
}
bool InitEnv()
{
printf("%s - INFO - [XIR]: Start to initialize ge using ge global options\n", GetTime().c_str());
std::map<AscendString, AscendString> global_options = {{"ge.exec.deviceId", "0"}, {"ge.graphRunMode", "1"}};
Status ret = ge::GEInitialize(global_options);
if (ret != SUCCESS) {
printf("%s - INFO - [XIR]: Initialize ge using ge global options failed\n", GetTime().c_str());
return false;
}
printf("%s - INFO - [XIR]: Initialize ge using ge global options success\n", GetTime().c_str());
return true;
}
bool CreateAndConfigGraph(Graph& graph, std::vector<ge::Tensor>& input)
{
printf("%s - INFO - [XIR]: Start to CreateAndConfigGraph\n", GetTime().c_str());
std::vector<Operator> inputs{};
std::vector<Operator> outputs{};
DataType inDtype = DT_FLOAT16;
std::cout << inDtype << std::endl;
Status ret = CreateOppInGraph(inDtype, input, inputs, outputs, graph);
if (ret != SUCCESS) {
printf("%s - ERROR - [XIR]: Create ir session using build options failed\n", GetTime().c_str());
return false;
}
if (!inputs.empty() && !outputs.empty()) {
graph.SetInputs(inputs).SetOutputs(outputs);
}
return true;
}
bool AddGraphToSession(ge::Session* session, Graph& graph, uint32_t graph_id)
{
printf("%s - INFO - [XIR]: Create ir session using build options success\n", GetTime().c_str());
printf("%s - INFO - [XIR]: Start to add compute graph to ir session\n", GetTime().c_str());
std::map<AscendString, AscendString> graph_options = {
};
Status ret = session->AddGraph(graph_id, graph, graph_options);
if (ret != SUCCESS) {
printf("%s - ERROR - [XIR]: Add graph to session failed\n", GetTime().c_str());
return false;
}
return true;
}
bool RunGraph(ge::Session* session, uint32_t graph_id, std::vector<ge::Tensor>& input, std::vector<ge::Tensor>& output)
{
printf("%s - INFO - [XIR]: Start to run graph\n", GetTime().c_str());
Status ret = session->RunGraph(graph_id, input, output);
if (ret != SUCCESS) {
printf("%s - ERROR - [XIR]: Run graph failed\n", GetTime().c_str());
return false;
}
return true;
}
bool FinalizeEnv()
{
printf("%s - INFO - [XIR]: Start to finalize ge\n", GetTime().c_str());
Status ret = ge::GEFinalize();
if (ret != SUCCESS) {
printf("%s - INFO - [XIR]: Finalize ge failed\n", GetTime().c_str());
return false;
}
printf("%s - INFO - [XIR]: Finalize ge success\n", GetTime().c_str());
return true;
}
int main(int argc, char* argv[])
{
std::vector<ge::Tensor> input{};
std::vector<ge::Tensor> output{};
Graph graph("avg_pool_graph");
if (!InitEnv()) {
return FAILED;
}
if (!CreateAndConfigGraph(graph, input)) {
return FAILED;
}
std::map<ge::AscendString, ge::AscendString> session_options = {};
ge::Session* session = new ge::Session(session_options);
if (session == nullptr) {
return FAILED;
}
if (!AddGraphToSession(session, graph, 0)) {
return FAILED;
}
if (!RunGraph(session, 0, input, output)) {
return FAILED;
}
printf("%s - INFO - [XIR]: Session run ir compute graph success\n", GetTime().c_str());
int input_num = input.size();
for (int i = 0; i < input_num; i++) {
std::cout << "input " << i << " dtype : " << input[i].GetTensorDesc().GetDataType() << std::endl;
string input_file = "./avg_pool_npu_input_" + std::to_string(i) + ".bin";
uint8_t* input_data_i = input[i].GetData();
int64_t input_shape = input[i].GetTensorDesc().GetShape().GetShapeSize();
std::cout << "this is " << i << "th input, input shape size =" << input_shape << std::endl;
uint32_t data_size = input_shape * GetDataTypeSize(input[i].GetTensorDesc().GetDataType());
WriteDataToFile((const char*)input_file.c_str(), data_size, input_data_i);
}
int output_num = output.size();
for (int i = 0; i < output_num; i++) {
std::cout << "output " << i << " dtype : " << output[i].GetTensorDesc().GetDataType() << std::endl;
string output_file = "./avg_pool_npu_output_" + std::to_string(i) + ".bin";
uint8_t* output_data_i = output[i].GetData();
int64_t output_shape = output[i].GetTensorDesc().GetShape().GetShapeSize();
std::cout << "this is " << i << "th output, output shape size =" << output_shape << std::endl;
uint32_t data_size = output_shape * GetDataTypeSize(output[i].GetTensorDesc().GetDataType());
WriteDataToFile((const char*)output_file.c_str(), data_size, output_data_i);
std::cout << "output y (avg_pool result):" << std::endl;
if (output[i].GetTensorDesc().GetDataType() == DT_FLOAT16) {
uint16_t* data = reinterpret_cast<uint16_t*>(output_data_i);
for (int j = 0; j < std::min(output_shape, (int64_t)16); j++) {
printf(" y[%d] = %u (float16)\n", j, data[j]);
}
}
}
ge::AscendString error_msg = ge::GEGetErrorMsgV2();
std::string error_str(error_msg.GetString());
std::cout << "Error message: " << error_str << std::endl;
ge::AscendString warning_msg = ge::GEGetWarningMsgV2();
std::string warning_str(warning_msg.GetString());
std::cout << "Warning message: " << warning_str << std::endl;
printf("%s - INFO - [XIR]: Precision is ok\n", GetTime().c_str());
if (!FinalizeEnv()) {
return FAILED;
}
if (session != nullptr) {
delete session;
}
printf("%s - INFO - [XIR]: Test case passed successfully\n", GetTime().c_str());
return SUCCESS;
}