* Copyright (c) 2025 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 <algorithm>
#include <cstdint>
#include <iostream>
#include <vector>
#include <random>
#include "acl/acl.h"
#include "aclnn_add_custom.h"
namespace {
#define CHECK_RET(cond, return_expr) \
do { \
if (!(cond)) { \
return_expr; \
} \
} while (0)
#define LOG_PRINT(message, ...) \
do { \
printf(message, ##__VA_ARGS__); \
} while (0)
int64_t GetShapeSize(const std::vector<int64_t> &shape)
{
int64_t shapeSize = 1;
for (auto i : shape) {
shapeSize *= i;
}
return shapeSize;
}
int Init(int32_t deviceId, aclrtStream *stream)
{
auto ret = aclInit(nullptr);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return 1);
ret = aclrtSetDevice(deviceId);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return 1);
ret = aclrtCreateStream(stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return 1);
return 0;
}
template <typename T>
int CreateAclTensor(const std::vector<T> &hostData, const std::vector<int64_t> &shape, void **deviceAddr,
aclDataType dataType, aclTensor **tensor)
{
auto size = GetShapeSize(shape) * sizeof(T);
auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return 1);
ret = aclrtMemcpy(*deviceAddr, size, hostData.data(), size, ACL_MEMCPY_HOST_TO_DEVICE);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMemcpy failed. ERROR: %d\n", ret); return 1);
*tensor = aclCreateTensor(shape.data(), shape.size(), dataType, nullptr, 0, aclFormat::ACL_FORMAT_ND, shape.data(),
shape.size(), *deviceAddr);
return 0;
}
void DestroyResources(std::vector<void *> tensors, std::vector<void *> deviceAddrs, aclrtStream stream,
int32_t deviceId, void *workspaceAddr = nullptr)
{
for (uint32_t i = 0; i < tensors.size(); i++) {
if (tensors[i] != nullptr) {
aclDestroyTensor(reinterpret_cast<aclTensor *>(tensors[i]));
}
if (deviceAddrs[i] != nullptr) {
aclrtFree(deviceAddrs[i]);
}
}
if (workspaceAddr != nullptr) {
aclrtFree(workspaceAddr);
}
aclrtDestroyStream(stream);
aclrtResetDevice(deviceId);
aclFinalize();
}
void InitializeData(std::vector<float> &inputX, std::vector<float> &inputY, std::vector<float> &inputZ,
std::vector<float> &golden, std::vector<int64_t> &shape)
{
std::random_device rd;
std::mt19937 gen(rd());
std::uniform_real_distribution<float> dis(1.0f, 100.0f);
for (int i = 0; i < shape[0] * shape[1]; ++i) {
inputX[i] = dis(gen);
inputY[i] = dis(gen);
golden[i] = inputX[i] + inputY[i];
inputZ[i] = 0.0;
}
}
bool VerifyResults(const std::vector<float> &goldenData, const std::vector<float> &resultData)
{
int64_t len = 10;
LOG_PRINT("result is:\n");
for (int64_t i = 0; i < len; i++) {
LOG_PRINT("%.1f ", resultData[i]);
}
LOG_PRINT("\n");
if (std::equal(resultData.begin(), resultData.end(), goldenData.begin())) {
LOG_PRINT("test pass\n");
} else {
LOG_PRINT("test failed\n");
return false;
}
return true;
}
}
int main(int argc, char **argv)
{
int32_t deviceId = 0;
aclrtStream stream;
auto ret = Init(deviceId, &stream);
CHECK_RET(ret == 0, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return 1);
std::vector<int64_t> shape = {8, 2048};
void *inputXDeviceAddr = nullptr;
void *inputYDeviceAddr = nullptr;
void *outputZDeviceAddr = nullptr;
aclTensor *inputX = nullptr;
aclTensor *inputY = nullptr;
aclTensor *outputZ = nullptr;
std::vector<float> inputXHostData(shape[0] * shape[1]);
std::vector<float> inputYHostData(shape[0] * shape[1]);
std::vector<float> outputZHostData(shape[0] * shape[1]);
std::vector<float> goldenData(shape[0] * shape[1]);
InitializeData(inputXHostData, inputYHostData, outputZHostData, goldenData, shape);
std::vector<void *> tensors = {inputX, inputY, outputZ};
std::vector<void *> deviceAddrs = {inputXDeviceAddr, inputYDeviceAddr, outputZDeviceAddr};
ret = CreateAclTensor(inputXHostData, shape, &inputXDeviceAddr, aclDataType::ACL_FLOAT, &inputX);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId); return 1);
ret = CreateAclTensor(inputYHostData, shape, &inputYDeviceAddr, aclDataType::ACL_FLOAT, &inputY);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId); return 1);
ret = CreateAclTensor(outputZHostData, shape, &outputZDeviceAddr, aclDataType::ACL_FLOAT, &outputZ);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId); return 1);
uint64_t workspaceSize = 0;
aclOpExecutor *executor;
ret = aclnnAddCustomGetWorkspaceSize(inputX, inputY, outputZ, &workspaceSize, &executor);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnAddCustomGetWorkspaceSize failed. ERROR: %d\n", ret);
DestroyResources(tensors, deviceAddrs, stream, deviceId); return 1);
void *workspaceAddr = nullptr;
if (workspaceSize > 0U) {
ret = aclrtMalloc(&workspaceAddr, workspaceSize, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId, workspaceAddr); return 1);
}
ret = aclnnAddCustom(workspaceAddr, workspaceSize, executor, stream);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId, workspaceAddr); return 1);
ret = aclrtSynchronizeStream(stream);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId, workspaceAddr); return 1);
auto size = GetShapeSize(shape);
std::vector<float> resultData(size, 0);
ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outputZDeviceAddr,
size * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST);
CHECK_RET(ret == ACL_SUCCESS, DestroyResources(tensors, deviceAddrs, stream, deviceId, workspaceAddr); return 1);
DestroyResources(tensors, deviceAddrs, stream, deviceId, workspaceAddr);
if (!VerifyResults(goldenData, resultData)) {
return -1;
}
return 0;
}