* 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.
*/
* \file test_geir_xdivy.cpp
* \brief
*/
#include <iostream>
#include <fstream>
#include <string.h>
#include <stdint.h>
#include <vector>
#include <string>
#include <map>
#include <cmath>
#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 "../op_graph/xdivy_proto.h"
#define FAILED -1
#define SUCCESS 0
using namespace ge;
using std::map;
using std::string;
using std::vector;
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)
{
if (dt == ge::DT_FLOAT) return 4;
if (dt == ge::DT_FLOAT16 || dt == ge::DT_BF16) return 2;
if (dt == ge::DT_INT32 || dt == ge::DT_UINT32) return 4;
if (dt == ge::DT_INT64 || dt == ge::DT_UINT64) return 8;
if (dt == ge::DT_INT8 || dt == ge::DT_UINT8) return 1;
return 1;
}
int32_t GenFloatData(
vector<int64_t> shapes, Tensor& input_tensor, TensorDesc& input_tensor_desc,
DataType data_type, const vector<float>& values)
{
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 * sizeof(float);
float* pData = new (std::nothrow) float[size];
for (size_t i = 0; i < size; ++i) pData[i] = values[i % values.size()];
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 xdivyOp = op::Xdivy("xdivy1");
std::vector<int64_t> shape = {2, 3};
{
auto data = op::Data("x1_data").set_attr_index(0);
TensorDesc desc = TensorDesc(ge::Shape(shape), FORMAT_ND, inDtype);
desc.SetPlacement(ge::kPlacementHost);
desc.SetFormat(FORMAT_ND);
Tensor tensor;
vector<float> xVals = {1.0f, 2.0f, 3.0f, 0.0f, 5.0f, 6.0f};
ret = GenFloatData(shape, tensor, desc, inDtype, xVals);
if (ret != SUCCESS) {
printf("%s - ERROR: Generate x1 data failed\n", GetTime().c_str());
return FAILED;
}
data.update_input_desc_x(desc);
input.push_back(tensor);
graph.AddOp(data);
xdivyOp.set_input_x1(data);
inputs.push_back(data);
}
{
auto data = op::Data("x2_data").set_attr_index(1);
TensorDesc desc = TensorDesc(ge::Shape(shape), FORMAT_ND, inDtype);
desc.SetPlacement(ge::kPlacementHost);
desc.SetFormat(FORMAT_ND);
Tensor tensor;
vector<float> yVals = {2.0f, 3.0f, 4.0f, 5.0f, 1.0f, 2.0f};
ret = GenFloatData(shape, tensor, desc, inDtype, yVals);
if (ret != SUCCESS) {
printf("%s - ERROR: Generate x2 data failed\n", GetTime().c_str());
return FAILED;
}
data.update_input_desc_x(desc);
input.push_back(tensor);
graph.AddOp(data);
xdivyOp.set_input_x2(data);
inputs.push_back(data);
}
TensorDesc outDesc = TensorDesc(ge::Shape(shape), FORMAT_ND, inDtype);
xdivyOp.update_output_desc_y(outDesc);
outputs.push_back(xdivyOp);
return SUCCESS;
}
int main(int argc, char* argv[])
{
const char* graph_name = "tc_ge_irrun_xdivy_test";
Graph graph(graph_name);
std::vector<ge::Tensor> input;
printf("%s - [XIR]: Initializing GE\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 - [XIR]: GEInitialize failed\n", GetTime().c_str());
return FAILED;
}
printf("%s - [XIR]: GEInitialize success\n", GetTime().c_str());
std::vector<Operator> inputs{};
std::vector<Operator> outputs{};
DataType inDtype = DT_FLOAT;
ret = CreateOppInGraph(inDtype, input, inputs, outputs, graph);
if (ret != SUCCESS) {
printf("%s - [XIR]: CreateOppInGraph failed\n", GetTime().c_str());
return FAILED;
}
if (!inputs.empty() && !outputs.empty()) {
graph.SetInputs(inputs).SetOutputs(outputs);
}
std::map<AscendString, AscendString> build_options = {};
printf("%s - [XIR]: Creating IR session\n", GetTime().c_str());
ge::Session* session = new Session(build_options);
if (session == nullptr) {
printf("%s - [XIR]: Create session failed\n", GetTime().c_str());
return FAILED;
}
std::map<AscendString, AscendString> graph_options = {};
uint32_t graph_id = 0;
ret = session->AddGraph(graph_id, graph, graph_options);
printf("%s - [XIR]: AddGraph success\n", GetTime().c_str());
std::string file_path = "./dump_xdivy";
aclgrphDumpGraph(graph, file_path.c_str(), file_path.length());
printf("%s - [XIR]: Running xdivy graph\n", GetTime().c_str());
std::vector<ge::Tensor> output;
ret = session->RunGraph(graph_id, input, output);
if (ret != SUCCESS) {
printf("%s - [XIR]: RunGraph failed\n", GetTime().c_str());
ge::AscendString error_msg = ge::GEGetErrorMsgV2();
std::string error_str(error_msg.GetString());
std::cout << "GE Error: " << error_str << std::endl;
delete session;
GEFinalize();
return FAILED;
}
printf("%s - [XIR]: RunGraph success\n", GetTime().c_str());
int input_num = input.size();
for (int i = 0; i < input_num; i++) {
string input_file = "./xdivy_input_" + std::to_string(i) + ".bin";
uint8_t* data = input[i].GetData();
int64_t shape_size = input[i].GetTensorDesc().GetShape().GetShapeSize();
uint32_t data_size = shape_size * GetDataTypeSize(input[i].GetTensorDesc().GetDataType());
WriteDataToFile(input_file, data_size, data);
}
int output_num = output.size();
printf("\n========== GEIR NPU Output (xdivy: x1/x2, x1==0 -> 0) ==========\n");
for (int i = 0; i < output_num; i++) {
string output_file = "./xdivy_output_" + std::to_string(i) + ".bin";
uint8_t* data = output[i].GetData();
int64_t shape_size = output[i].GetTensorDesc().GetShape().GetShapeSize();
uint32_t data_size = shape_size * GetDataTypeSize(output[i].GetTensorDesc().GetDataType());
WriteDataToFile(output_file, data_size, data);
float* floatData = reinterpret_cast<float*>(data);
printf("Output tensor[%d] (%ld elements):\n", i, shape_size);
for (int64_t j = 0; j < shape_size; j++) {
printf(" [%ld] NPU=%.6f\n", j, floatData[j]);
}
}
printf("========================================\n");
printf("\n========== Golden Reference (x1/x2) ==========\n");
float x1Vals[] = {1.0f, 2.0f, 3.0f, 0.0f, 5.0f, 6.0f};
float x2Vals[] = {2.0f, 3.0f, 4.0f, 5.0f, 1.0f, 2.0f};
int allMatch = 1;
for (int i = 0; i < 6; i++) {
float golden = (x1Vals[i] == 0.0f) ? 0.0f : x1Vals[i] / x2Vals[i];
float npuVal = reinterpret_cast<float*>(output[0].GetData())[i];
int match = (fabs(npuVal - golden) < 1e-5f);
if (!match) allMatch = 0;
printf(" [%d] x1=%.1f x2=%.1f -> NPU=%.6f golden=%.6f %s\n",
i, x1Vals[i], x2Vals[i], npuVal, golden, match ? "OK" : "MISMATCH");
}
printf("========================================\n");
printf("Result: %s\n", allMatch ? "ALL PASS" : "FAIL");
ge::AscendString error_msg = ge::GEGetErrorMsgV2();
std::string error_str(error_msg.GetString());
std::cout << "GE Error: " << error_str << std::endl;
printf("%s - [XIR]: Finalizing GE\n", GetTime().c_str());
ret = ge::GEFinalize();
if (ret != SUCCESS) {
printf("%s - [XIR]: GEFinalize failed\n", GetTime().c_str());
return FAILED;
}
printf("%s - [XIR]: Success\n", GetTime().c_str());
return allMatch ? SUCCESS : FAILED;
}
