news 2026/5/9 11:38:30

CANN/sip BLAS点积算子文档

作者头像

张小明

前端开发工程师

1.2k 24
文章封面图
CANN/sip BLAS点积算子文档

Dot

【免费下载链接】sip本项目是CANN提供的一款高效、可靠的高性能信号处理算子加速库，基于华为Ascend AI处理器，专门为信号处理领域而设计。项目地址: https://gitcode.com/cann/sip

产品支持情况

产品是否支持
Atlas 200I/500 A2 推理产品×
Atlas 推理系列产品×
Atlas 训练系列产品×
Atlas A3 训练系列产品/Atlas A3 推理系列产品√
Atlas A2 训练系列产品/Atlas A2 推理系列产品√
Ascend 950PR/Ascend 950DT×

功能说明

  • 接口功能：
    asdBlasMakeDotPlan：初始化该句柄对应的Dot算子配置。
    asdBlasSdot：计算两个实数向量的点积。
    asdBlasCdotu：计算两个复数向量的点积。
    asdBlasCdotc：计算一个复数向量取共轭后和另一个复数向量的点积。

  • 计算公式：

    • asdBlasSdot的公式

    $$ result=\sum _{i=1}^n(x[i] * y[i]) $$

    示例： 输入“x”为： [1.0, 2.0] 输入“y”为： [1.0, 2.0] 调用asdBlasSdot算子后，输出“result”为： 5.0
    • asdBlasCdotu的公式

    $$ result=\sum _{i=1}^n(conj(x[i]) * y[i]) $$ 其中，x[i]和y[i]是复数。 示例： 输入“x”为： [ 0.1554+0.8840j, -0.3564-0.2552j] 输入“y”为： [-0.1404+1.3380j, -0.4876+0.1842j] 调用asdBlasCdotu算子后，输出“result”为： 1.2877-0.1420j

  • asdBlasCdotc的公式

    $$ result=\sum _{i=1}^n(conj(x[i]) * y[i]) $$ 其中，x[i]和y[i]是复数,conj共轭操作。 示例： 输入“x”为： [ 0.1554+0.8840j, -0.3564-0.2552j] 输入“y”为： [-0.1404+1.3380j, -0.4876+0.1842j] 调用asdBlasCdotc算子后，输出“result”为： 1.2877-0.1420j

函数原型

AspbStatus asdBlasMakeDotPlan( asdBlasHandle handle)
AspbStatus asdBlasSdot( asdBlasHandle handle, const int64_t n, aclTensor * x, const int64_t incx, aclTensor * y, const int64_t incy, aclTensor * result)
AspbStatus asdBlasCdotu( asdBlasHandle handle, const int64_t n, aclTensor * x, const int64_t incx, aclTensor * y, const int64_t incy, aclTensor * result)
AspbStatus asdBlasCdotc( asdBlasHandle handle, const int64_t n, aclTensor * x, const int64_t incx, aclTensor * y, const int64_t incy, aclTensor * result)

asdBlasMakeDotPlan

  • 参数说明：

    参数名输入/输出描述
    handle（asdBlasHandle）输入算子的句柄

  • 返回值：

    返回状态码，具体参见SiP返回码。

asdBlasSdot & asdBlasCdotu & asdBlasCdotc

  • 参数说明：

    参数名输入/输出描述
    handle（asdBlasHandle）输入算子的句柄
    n（int64_t）输入向量x或向量y中的元素个数。
    x（aclTensor *）输入
    • 对应公式中的'x'。
    • asdBlasSdot支持的数据类型支持FLOAT32。
    • asdBlasCdotu & asdBlasCdotc支持的数据类型支持COMPLEX64。
    • 数据格式支持ND。
    • shape为[n]。
    incx（int64_t）输入向量x相邻元素间的内存地址偏移量（当前约束为1）。
    y（aclTensor *）输入
    • 对应公式中的'y'。
    • asdBlasSdot支持的数据类型支持FLOAT32。
    • asdBlasCdotu & asdBlasCdotc支持的数据类型支持COMPLEX64。
    • 数据格式支持ND。
    • shape为[n]。
    incy（int64_t）输入向量y相邻元素间的内存地址偏移量（当前约束为1）。
    result（aclTensor *）输出
    • 表示输出的结果，对应公式中的'result'。
    • 数据类型支持FLOAT32，只包含一个元素。
    • 数据格式支持ND。
    • shape为[1]。

  • 返回值：

    返回状态码，具体参见SiP返回码。

约束说明

  • 输入的元素个数n当前覆盖支持[1，6.71e+06]。
  • 算子输入shape为[n]，输出shape为[1]。
  • 算子实际计算时，不支持ND高维度运算（不支持维度≥3的运算）。

调用示例

示例代码如下，该样例旨在提供快速上手、开发和调试算子的最小化实现，其核心目标是使用最精简的代码展示算子的核心功能，而非提供生产级的安全保障。不推荐用户直接将示例代码作为业务代码，若用户将示例代码应用在自身的真实业务场景中且发生了安全问题，则需用户自行承担。

  • asdBlasSdot
#include <iostream> #include <vector> #include "asdsip.h" #include "acl/acl.h" #include "acl_meta.h" using namespace AsdSip; #define ASD_STATUS_CHECK(err) \ do { \ AsdSip::AspbStatus err_ = (err); \ if (err_ != AsdSip::ErrorType::ACL_SUCCESS) { \ std::cout << "Execute failed." << std::endl; \ exit(-1); \ } \ } while (0) #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) { // 固定写法，acl初始化 auto ret = aclInit(nullptr); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret); ret = aclrtSetDevice(deviceId); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret); ret = aclrtCreateStream(stream); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret); 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); // 调用aclrtMalloc申请device侧内存 auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret); // 调用aclrtMemcpy将host侧数据复制到device侧内存上 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 ret); // 计算连续tensor的strides std::vector<int64_t> strides(shape.size(), 1); for (int64_t i = shape.size() - 2; i >= 0; i--) { strides[i] = shape[i + 1] * strides[i + 1]; } // 调用aclCreateTensor接口创建aclTensor *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND, shape.data(), shape.size(), *deviceAddr); return 0; } int main(int argc, char **argv) { // 设置算子使用的device id int deviceId = 0; //（固定写法）创造执行流 aclrtStream stream; auto ret = Init(deviceId, &stream); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret); // 创造tensor的Host侧数据 int64_t n = 5; int64_t incx = 1; int64_t incy = 1; int64_t xSize = 5; std::vector<float> tensorInXData; tensorInXData.reserve(xSize); for (int64_t i = 0; i < xSize; i++) { tensorInXData[i] = 1.0 + i; } int64_t ySize = 5; std::vector<float> tensorInYData; tensorInYData.reserve(xSize); for (int64_t i = 0; i < ySize; i++) { tensorInYData[i] = 10.0 + i; } int64_t resultSize = 1; std::vector<float> resultData; resultData.reserve(resultSize); std::cout << "------- input x -------" << std::endl; for (int64_t i = 0; i < xSize; i++) { std::cout << tensorInXData[i] << " "; } std::cout << std::endl; std::cout << "------- input y -------" << std::endl; for (int64_t i = 0; i < ySize; i++) { std::cout << tensorInYData[i] << " "; } std::cout << std::endl; // 创造输入/输出tensor std::vector<int64_t> xShape = {xSize}; std::vector<int64_t> yShape = {ySize}; std::vector<int64_t> resultShape = {resultSize}; aclTensor *inputX = nullptr; aclTensor *inputY = nullptr; aclTensor *result = nullptr; void *inputXDeviceAddr = nullptr; void *inputYDeviceAddr = nullptr; void *resultDeviceAddr = nullptr; ret = CreateAclTensor(tensorInXData, xShape, &inputXDeviceAddr, aclDataType::ACL_FLOAT, &inputX); CHECK_RET(ret == ::ACL_SUCCESS, return ret); ret = CreateAclTensor(tensorInYData, yShape, &inputYDeviceAddr, aclDataType::ACL_FLOAT, &inputY); CHECK_RET(ret == ::ACL_SUCCESS, return ret); ret = CreateAclTensor(resultData, resultShape, &resultDeviceAddr, aclDataType::ACL_FLOAT, &result); CHECK_RET(ret == ::ACL_SUCCESS, return ret); // 创建算子执行句柄 asdBlasHandle handle; asdBlasCreate(handle); // 创造算子执行所需workspace size_t lwork = 0; void *buffer = nullptr; asdBlasMakeDotPlan(handle); asdBlasGetWorkspaceSize(handle, lwork); std::cout << "lwork = " << lwork << std::endl; if (lwork > 0) { ret = aclrtMalloc(&buffer, static_cast<int64_t>(lwork), ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret); } asdBlasSetWorkspace(handle, buffer); // 配置算子执行信息 asdBlasSetStream(handle, stream); // 调用接口执行算子（固定调用逻辑） ASD_STATUS_CHECK(asdBlasSdot(handle, n, inputX, incx, inputY, incy, result)); asdBlasSynchronize(handle); // 调度算子后销毁算子句柄 asdBlasDestroy(handle); // 将输出tensor的Device侧数据复制到Host侧内存上 ret = aclrtMemcpy(resultData.data(), resultSize * sizeof(float), resultDeviceAddr, resultSize * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret); std::cout << "------- result -------" << std::endl; for (int64_t i = 0; i < 1; i++) { std::cout << resultData[i] << " "; } std::cout << std::endl; std::cout << "Execute successfully." << std::endl; // 资源释放 aclDestroyTensor(inputX); aclDestroyTensor(inputY); aclDestroyTensor(result); aclrtFree(inputXDeviceAddr); aclrtFree(inputYDeviceAddr); aclrtFree(resultDeviceAddr); aclrtDestroyStream(stream); aclrtResetDevice(deviceId); aclFinalize(); return 0; }
  • asdBlasCdotu
#include <iostream> #include <vector> #include <cmath> #include <random> #include <complex> #include "asdsip.h" #include "acl/acl.h" #include "acl_meta.h" using namespace AsdSip; #define ASD_STATUS_CHECK(err) \ do { \ AsdSip::AspbStatus err_ = (err); \ if (err_ != AsdSip::ErrorType::ACL_SUCCESS) { \ std::cout << "Execute failed." << std::endl; \ exit(-1); \ } else { \ std::cout << "Execute successfully." << std::endl; \ } \ } while (0) void printTensor(const std::complex<float> *tensorData, int64_t tensorSize) { for (int64_t i = 0; i < tensorSize; i++) { std::cout << tensorData[i] << " "; } std::cout << std::endl; } #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) { // 固定写法，acl初始化 auto ret = aclInit(nullptr); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret); ret = aclrtSetDevice(deviceId); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret); ret = aclrtCreateStream(stream); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret); 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); // 调用aclrtMalloc申请device侧内存 auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret); // 调用aclrtMemcpy将host侧数据复制到device侧内存上 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 ret); // 计算连续tensor的strides std::vector<int64_t> strides(shape.size(), 1); for (int64_t i = shape.size() - 2; i >= 0; i--) { strides[i] = shape[i + 1] * strides[i + 1]; } // 调用aclCreateTensor接口创建aclTensor *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND, shape.data(), shape.size(), *deviceAddr); return 0; } void printTensor(std::vector<std::complex<float>> tensorData, int64_t tensorSize) { for (int64_t i = 0; i < tensorSize; i++) { std::cout << tensorData[i] << " "; } std::cout << std::endl; } int main(int argc, char **argv) { int deviceId = 0; aclrtStream stream; auto ret = Init(deviceId, &stream); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret); int64_t n = 8; int64_t xSize = 8; int64_t ySize = 8; std::vector<std::complex<float>> tensorInXData; tensorInXData.reserve(xSize); for (int64_t i = 0; i < xSize; i++) { tensorInXData[i] = {2.0, (float)(1.0 + i)}; } std::vector<std::complex<float>> tensorInYData; tensorInYData.reserve(ySize); for (int64_t i = 0; i < ySize; i++) { tensorInYData[i] = {3.0, 4.0}; } int64_t resultSize = 1; std::vector<std::complex<float>> resultData; resultData.reserve(resultSize); std::cout << "------- input TensorInX -------" << std::endl; printTensor(tensorInXData.data(), xSize); std::cout << "------- input TensorInY -------" << std::endl; printTensor(tensorInYData.data(), ySize); std::vector<int64_t> xShape = {xSize}; std::vector<int64_t> yShape = {ySize}; std::vector<int64_t> resultShape = {resultSize}; aclTensor *inputX = nullptr; aclTensor *inputY = nullptr; aclTensor *result = nullptr; void *inputXDeviceAddr = nullptr; void *inputYDeviceAddr = nullptr; void *resultDeviceAddr = nullptr; ret = CreateAclTensor(tensorInXData, xShape, &inputXDeviceAddr, aclDataType::ACL_COMPLEX64, &inputX); CHECK_RET(ret == ::ACL_SUCCESS, return ret); ret = CreateAclTensor(tensorInYData, yShape, &inputYDeviceAddr, aclDataType::ACL_COMPLEX64, &inputY); CHECK_RET(ret == ::ACL_SUCCESS, return ret); ret = CreateAclTensor(resultData, resultShape, &resultDeviceAddr, aclDataType::ACL_COMPLEX64, &result); CHECK_RET(ret == ::ACL_SUCCESS, return ret); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork = 0; void *buffer = nullptr; asdBlasMakeDotPlan(handle); asdBlasGetWorkspaceSize(handle, lwork); std::cout << "lwork = " << lwork << std::endl; if (lwork > 0) { ret = aclrtMalloc(&buffer, static_cast<int64_t>(lwork), ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret); } asdBlasSetWorkspace(handle, buffer); asdBlasSetStream(handle, stream); ASD_STATUS_CHECK(asdBlasCdotu(handle, n, inputX, 1, inputY, 1, result)); asdBlasSynchronize(handle); asdBlasDestroy(handle); ret = aclrtMemcpy(resultData.data(), resultSize * sizeof(std::complex<float>), resultDeviceAddr, resultSize * sizeof(std::complex<float>), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret); std::cout << "------- result -------" << std::endl; printTensor(resultData.data(), resultSize); aclDestroyTensor(inputX); aclDestroyTensor(inputY); aclDestroyTensor(result); aclrtFree(inputXDeviceAddr); aclrtFree(inputYDeviceAddr); aclrtFree(resultDeviceAddr); aclrtDestroyStream(stream); aclrtResetDevice(deviceId); aclFinalize(); return 0; }
  • asdBlasCdotc
#include <iostream> #include <vector> #include <complex> #include <cmath> #include <random> #include "asdsip.h" #include "acl/acl.h" #include "acl_meta.h" using namespace AsdSip; #define ASD_STATUS_CHECK(err) \ do { \ AsdSip::AspbStatus err_ = (err); \ if (err_ != AsdSip::ErrorType::ACL_SUCCESS) { \ std::cout << "Execute failed." << std::endl; \ exit(-1); \ } else { \ std::cout << "Execute successfully." << std::endl; \ } \ } while (0) void printTensor(const std::complex<float> *tensorData, int64_t tensorSize) { for (int64_t i = 0; i < tensorSize; i++) { std::cout << tensorData[i] << " "; } std::cout << std::endl; } #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) { // 固定写法，acl初始化 auto ret = aclInit(nullptr); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret); ret = aclrtSetDevice(deviceId); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret); ret = aclrtCreateStream(stream); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret); 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); // 调用aclrtMalloc申请device侧内存 auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret); // 调用aclrtMemcpy将host侧数据复制到device侧内存上 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 ret); // 计算连续tensor的strides std::vector<int64_t> strides(shape.size(), 1); for (int64_t i = shape.size() - 2; i >= 0; i--) { strides[i] = shape[i + 1] * strides[i + 1]; } // 调用aclCreateTensor接口创建aclTensor *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND, shape.data(), shape.size(), *deviceAddr); return 0; } void printTensor(std::vector<std::complex<float>> tensorData, int64_t tensorSize) { for (int64_t i = 0; i < tensorSize; i++) { std::cout << tensorData[i] << " "; } std::cout << std::endl; } int main(int argc, char **argv) { int deviceId = 0; aclrtStream stream; auto ret = Init(deviceId, &stream); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret); int64_t n = 8; int64_t xSize = 8; int64_t ySize = 8; std::vector<std::complex<float>> tensorInXData; tensorInXData.reserve(xSize); for (int64_t i = 0; i < xSize; i++) { tensorInXData[i] = {2.0, (float)(1.0 + i)}; } std::vector<std::complex<float>> tensorInYData; tensorInYData.reserve(ySize); for (int64_t i = 0; i < ySize; i++) { tensorInYData[i] = {3.0, 4.0}; } int64_t resultSize = 1; std::vector<std::complex<float>> resultData; resultData.reserve(resultSize); std::cout << "------- input TensorInX -------" << std::endl; printTensor(tensorInXData.data(), xSize); std::cout << "------- input TensorInY -------" << std::endl; printTensor(tensorInYData.data(), ySize); std::vector<int64_t> xShape = {xSize}; std::vector<int64_t> yShape = {ySize}; std::vector<int64_t> resultShape = {resultSize}; aclTensor *inputX = nullptr; aclTensor *inputY = nullptr; aclTensor *result = nullptr; void *inputXDeviceAddr = nullptr; void *inputYDeviceAddr = nullptr; void *resultDeviceAddr = nullptr; ret = CreateAclTensor(tensorInXData, xShape, &inputXDeviceAddr, aclDataType::ACL_COMPLEX64, &inputX); CHECK_RET(ret == ::ACL_SUCCESS, return ret); ret = CreateAclTensor(tensorInYData, yShape, &inputYDeviceAddr, aclDataType::ACL_COMPLEX64, &inputY); CHECK_RET(ret == ::ACL_SUCCESS, return ret); ret = CreateAclTensor(resultData, resultShape, &resultDeviceAddr, aclDataType::ACL_COMPLEX64, &result); CHECK_RET(ret == ::ACL_SUCCESS, return ret); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork = 0; void *buffer = nullptr; asdBlasMakeDotPlan(handle); asdBlasGetWorkspaceSize(handle, lwork); std::cout << "lwork = " << lwork << std::endl; if (lwork > 0) { ret = aclrtMalloc(&buffer, static_cast<int64_t>(lwork), ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret); } asdBlasSetWorkspace(handle, buffer); asdBlasSetStream(handle, stream); ASD_STATUS_CHECK(asdBlasCdotc(handle, n, inputX, 1, inputY, 1, result)); asdBlasSynchronize(handle); asdBlasDestroy(handle); ret = aclrtMemcpy(resultData.data(), resultSize * sizeof(std::complex<float>), resultDeviceAddr, resultSize * sizeof(std::complex<float>), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret == ::ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret); std::cout << "------- result -------" << std::endl; printTensor(resultData.data(), resultSize); aclDestroyTensor(inputX); aclDestroyTensor(inputY); aclDestroyTensor(result); aclrtFree(inputXDeviceAddr); aclrtFree(inputYDeviceAddr); aclrtFree(resultDeviceAddr); aclrtDestroyStream(stream); aclrtResetDevice(deviceId); aclFinalize(); return 0; }

【免费下载链接】sip本项目是CANN提供的一款高效、可靠的高性能信号处理算子加速库，基于华为Ascend AI处理器，专门为信号处理领域而设计。项目地址: https://gitcode.com/cann/sip

创作声明：本文部分内容由AI辅助生成（AIGC），仅供参考

版权声明: 本文来自互联网用户投稿，该文观点仅代表作者本人，不代表本站立场。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如若内容造成侵权/违法违规/事实不符，请联系邮箱：809451989@qq.com进行投诉反馈，一经查实，立即删除！
网站建设 2026/5/9 11:35:49

CANN/asc-tools NPU检查工具

npu_check 【免费下载链接】asc-tools Ascend C Tools仓是CANN基于Ascend C编程语言推出的配套调试工具仓。 项目地址: https://gitcode.com/cann/asc-tools 概述 Ascend C Tools提供的孪生调试分为debug功能和npu check功能&#xff0c;debug功能包含诸如是否合法使用…

作者头像 李华
网站建设 2026/5/9 11:34:29

小红书内容采集终极指南：XHS-Downloader 5分钟快速上手

小红书内容采集终极指南&#xff1a;XHS-Downloader 5分钟快速上手 【免费下载链接】XHS-Downloader 小红书&#xff08;XiaoHongShu、RedNote&#xff09;链接提取/作品采集工具&#xff1a;提取账号发布、收藏、点赞、专辑作品链接&#xff1b;提取搜索结果作品、用户链接&am…

作者头像 李华
网站建设 2026/5/9 11:33:31

2025最权威的十大降AI率平台推荐榜单

Ai论文网站排名&#xff08;开题报告、文献综述、降aigc率、降重综合对比&#xff09; TOP1. 千笔AI TOP2. aipasspaper TOP3. 清北论文 TOP4. 豆包 TOP5. kimi TOP6. deepseek 维普AIGC检测系统的主要目的乃精确辨认学术文本里那帮通过人工智能给弄出来的内容。在用户使…

作者头像 李华
网站建设 2026/5/9 11:33:31

通过curl命令快速测试Taotoken各大模型接口响应与功能

&#x1f680; 告别海外账号与网络限制&#xff01;稳定直连全球优质大模型&#xff0c;限时半价接入中。 &#x1f449; 点击领取海量免费额度 通过curl命令快速测试Taotoken各大模型接口响应与功能 对于需要在无SDK环境或进行底层接口调试的开发者而言&#xff0c;直接使用c…

作者头像 李华
网站建设 2026/5/9 11:31:29

如何用Python自动化工具轻松完成智慧树课程学习：Autovisor终极指南

如何用Python自动化工具轻松完成智慧树课程学习&#xff1a;Autovisor终极指南 【免费下载链接】Autovisor 2025智慧树刷课脚本 基于Python Playwright的自动化程序 [有免安装版] 项目地址: https://gitcode.com/gh_mirrors/au/Autovisor 还在为智慧树平台繁琐的手动学习…

作者头像 李华
网站建设 2026/5/9 11:31:27

如何让你的Atom编辑器说中文：三步实现完整中文汉化体验

如何让你的Atom编辑器说中文&#xff1a;三步实现完整中文汉化体验 【免费下载链接】atom-simplified-chinese-menu Atom 的简体中文汉化扩展,目前最全的汉化包。包含菜单汉化、右键菜单汉化以及设置汉化 项目地址: https://gitcode.com/gh_mirrors/at/atom-simplified-chine…

作者头像 李华