CANN/sip BLAS Swap算子文档

Swap【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asdBlasMakeSwapPlan初始化该句柄对应的Swap算子配置。asdBlasSswap交换两个实数向量。asdBlasCswap交换两个复数向量。计算公式asdBlasSswap的公式$$ t x, x y, yt $$示例 输入“x”为 [1.0, 2.0] 输入“y”为 [3.0, 4.0] 调用asdBlasSswap算子后输出“x”为 [3.0, 4.0] 输出“y”为 [1.0, 2.0]asdBlasCswap的公式$$ t x, x y, yt $$示例 输入“x”为[1.0 1.0i, 2.0 2.0i] 输入“y”为[3.0 3.0i, 4.0 4.0i] 调用asdBlasCswap算子后输出“x”为 [3.0 3.0i, 4.0 4.0i] 输出“y”为 [1.0 1.0i, 2.0 2.0i]函数原型AspbStatus asdBlasMakeSwapPlan( asdBlasHandle handle)AspbStatus asdBlasSswap( asdBlasHandle handle, const int64_t n, aclTensor * x, const int64_t incx, aclTensor * y, const int64_t incy)AspbStatus asdBlasCswap( asdBlasHandle handle, const int64_t n, aclTensor *x, const int64_t incx, aclTensor *y, const int64_t incy)asdBlasMakeSwapPlan参数说明参数名输入/输出描述handleasdBlasHandle输入算子的句柄返回值返回状态码具体参见SiP返回码。asdBlasSswap asdBlasCswap参数说明参数名输入/输出描述handleasdBlasHandle输入算子的句柄。nint64_t输入向量x或者向量y中元素的个数。xaclTensor *输入/输出对应公式中的x。asdBlasSswap支持的数据类型支持FLOAT32。asdBlasCswap支持的数据类型支持COMPLEX64。数据格式支持ND。shape为[n]。incxint64_t输入x相邻元素间的内存地址偏移量当前约束为1。yaclTensor *输入/输出对应公式中的y。asdBlasSswap支持的数据类型支持FLOAT32。asdBlasCswap支持的数据类型支持COMPLEX64。数据格式支持ND。shape为[n]。incyint64_t输入y相邻元素间的内存地址偏移量当前约束为1。返回值返回状态码具体参见SiP返回码。约束说明输入和输出数据格式、数据类型、shape保持一致。输入的元素个数n当前覆盖支持[16.71e07]。算子输入shape为[n]、[n]输出shape为[n]、[n]。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。asdBlasSswap#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::vectorint64_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::vectorT hostData, const std::vectorint64_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::vectorint64_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) { 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 incx 1; int64_t incy 1; const int64_t xSize 8; std::vectorfloat tensorInXData; tensorInXData.reserve(xSize); for (int64_t i 0; i xSize; i) { tensorInXData[i] 1.0; } const int64_t ySize 8; std::vectorfloat tensorInYData; tensorInYData.reserve(ySize); for (int64_t i 0; i ySize; i) { tensorInYData[i] 2.0; } 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; std::vectorint64_t xShape {xSize}; std::vectorint64_t yShape {ySize}; aclTensor *inputX nullptr; aclTensor *inputY nullptr; void *inputXDeviceAddr nullptr; void *inputYDeviceAddr 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); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork 0; void *buffer nullptr; asdBlasMakeSwapPlan(handle); asdBlasGetWorkspaceSize(handle, lwork); std::cout lwork lwork std::endl; if (lwork 0) { ret aclrtMalloc(buffer, static_castint64_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(asdBlasSswap(handle, n, inputX, incx, inputY, incy)); asdBlasSynchronize(handle); asdBlasDestroy(handle); ret aclrtMemcpy(tensorInXData.data(), xSize * sizeof(float), inputXDeviceAddr, xSize * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret ::ACL_SUCCESS, LOG_PRINT(copy tensor x from device to host failed. ERROR: %d\n, ret); return ret); ret aclrtMemcpy(tensorInYData.data(), ySize * sizeof(float), inputYDeviceAddr, ySize * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret ::ACL_SUCCESS, LOG_PRINT(copy tensor y from device to host failed. ERROR: %d\n, ret); return ret); std::cout ------- output X ------- std::endl; for (int64_t i 0; i xSize; i) { std::cout tensorInXData[i] ; } std::cout std::endl; std::cout ------- output Y ------- std::endl; for (int64_t i 0; i ySize; i) { std::cout tensorInYData[i] ; } std::cout std::endl; std::cout Execute successfully. std::endl; aclDestroyTensor(inputX); aclDestroyTensor(inputY); aclrtFree(inputXDeviceAddr); aclrtFree(inputYDeviceAddr); aclrtDestroyStream(stream); aclrtResetDevice(deviceId); aclFinalize(); return 0; }asdBlasCswap#include iostream #include vector #include asdsip.h #include cmath #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) #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::vectorint64_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::vectorT hostData, const std::vectorint64_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::vectorint64_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) { 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 6; int64_t incx 1; int64_t incy 1; int64_t xSize n; std::vectorstd::complexfloat tensorInXData; tensorInXData.reserve(xSize); for (int64_t i 0; i xSize; i) { tensorInXData[i] (std::complexfloat){2.0, -2.0}; } int64_t ySize n; std::vectorstd::complexfloat tensorInYData; tensorInYData.reserve(ySize); for (int64_t i 0; i ySize; i) { tensorInYData[i] (std::complexfloat){4.0, -4.0}; } std::cout ------- input TensorInX ------- std::endl; for (int64_t i 0; i xSize; i) { std::cout tensorInXData[i] ; } std::cout std::endl; std::cout ------- input TensorInY ------- std::endl; for (int64_t i 0; i ySize; i) { std::cout tensorInYData[i] ; } std::cout std::endl; std::vectorint64_t xShape {xSize}; std::vectorint64_t yShape {ySize}; aclTensor *inputX nullptr; aclTensor *inputY nullptr; void *inputXDeviceAddr nullptr; void *inputYDeviceAddr 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); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork 0; void *buffer nullptr; asdBlasMakeSwapPlan(handle); asdBlasGetWorkspaceSize(handle, lwork); std::cout lwork lwork std::endl; if (lwork 0) { ret aclrtMalloc(buffer, static_castint64_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(asdBlasCswap(handle, n, inputX, incx, inputY, incy)); asdBlasSynchronize(handle); asdBlasDestroy(handle); ret aclrtMemcpy(tensorInXData.data(), xSize * sizeof(std::complexfloat), inputXDeviceAddr, xSize * sizeof(std::complexfloat), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret ::ACL_SUCCESS, LOG_PRINT(copy tensor x from device to host failed. ERROR: %d\n, ret); return ret); ret aclrtMemcpy(tensorInYData.data(), ySize * sizeof(std::complexfloat), inputYDeviceAddr, ySize * sizeof(std::complexfloat), ACL_MEMCPY_DEVICE_TO_HOST); CHECK_RET(ret ::ACL_SUCCESS, LOG_PRINT(copy tensor y from device to host failed. ERROR: %d\n, ret); return ret); std::cout ------- output TensorInX ------- std::endl; for (int64_t i 0; i xSize; i) { std::cout tensorInXData[i] ; } std::cout std::endl; std::cout ------- output TensorInY ------- std::endl; for (int64_t i 0; i ySize; i) { std::cout tensorInYData[i] ; } std::cout std::endl; aclDestroyTensor(inputX); aclDestroyTensor(inputY); aclrtFree(inputXDeviceAddr); aclrtFree(inputYDeviceAddr); 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