Cal【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asdBlasMakeCalPlan初始化该句柄对应的Cal算子配置。asdBlasSscal对一个实数向量进行缩放即将向量中的每个元素乘以一个实数alpha。asdBlasCsscal对一个复数向量进行缩放即将向量中的每个元素乘以一个实数alpha。asdBlasCscal对一个复数向量进行缩放即将向量中的每个元素乘以一个复数alpha。计算公式asdBlasSscal的公式$$ x alpha * x $$示例 输入“x”为 [3.0, 4.0] 输入“alpha”为 2.0 调用asdBlasSscal算子后输出“x”为 [6.0, 8.0]asdBlasCsscal的公式$$ x alpha * x $$示例 输入“x”为 [3.0 4.0j, 4.0 4.0j] 输入“alpha”为 2.0 调用asdBlasCsscal算子后输出“x”为 [6.0 8.0j, 8.0 8.0j]asdBlasCscal的公式 $$ x alpha * x $$ 示例 输入“x”为 [3.0 4.0j, 4.0 4.0j] 输入“alpha”为 [3.0 4.0j] 调用asdBlasCscal算子后输出“x”为 [-7.0 24.0j, -4.0 28.0j]函数原型AspbStatus asdBlasMakeCalPlan(asdBlasHandle handle)AspbStatus asdBlasSscal( asdBlasHandle handle, const int64_t n, const float * alpha, aclTensor * x, const int64_t incx)AspbStatus asdBlasCsscal( asdBlasHandle handle, const int64_t n, const float * alpha, aclTensor * x, const int64_t incx)AspbStatus asdBlasCscal( asdBlasHandle handle, const int64_t n, const std::complexfloat * alpha, aclTensor * x, const int64_t incx)asdBlasMakeCalPlan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。asdBlasSscal asdBlasCsscal参数说明参数名输入/输出描述handleasdBlasHandle输入算子的句柄。nint64_t输入表示总的元素个数。xaclTensor *输入/输出表示输入/输出的向量对应公式中的x。数据类型支持FLOAT32数据格式支持ND。shape为[n]incxint64_t输入相邻元素间的内存地址偏移量当前约束为1。alphafloat *输入向量的缩放因子。返回值返回状态码具体参见SiP返回码。asdBlasCscal参数说明参数名输入/输出描述handleasdBlasHandle输入算子的句柄。nint64_t输入表示总的元素个数。xaclTensor *输入/输出表示输入/输出的向量对应公式中的x。数据类型支持FLOAT32数据格式支持ND。shape为[n]incxint64_t输入相邻元素间的内存地址偏移量当前约束为1。alphaconst std::complexltfloat *输入向量的缩放因子。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当前覆盖支持[16.71e06]。算子输入shape为[n]输出shape为[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。asdBlasSscal#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 7; int64_t incx 1; float alpha 10.0; int64_t xSize 7; std::vectorfloat tensorInXData; tensorInXData.reserve(xSize); for (int64_t i 0; i xSize; i) { tensorInXData[i] 1.0 i; } std::cout alpha alpha std::endl; std::cout ------- input X ------- std::endl; for (int64_t i 0; i xSize; i) { std::cout tensorInXData[i] ; } std::cout std::endl; std::vectorint64_t xShape {xSize}; aclTensor *inputX nullptr; void *inputXDeviceAddr nullptr; ret CreateAclTensor(tensorInXData, xShape, inputXDeviceAddr, aclDataType::ACL_FLOAT, inputX); CHECK_RET(ret ::ACL_SUCCESS, return ret); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork 0; void *buffer nullptr; asdBlasMakeCalPlan(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(asdBlasSscal(handle, n, alpha, inputX, incx)); 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); std::cout ------- output X ------- std::endl; for (int64_t i 0; i xSize; i) { std::cout tensorInXData[i] ; } std::cout std::endl; std::cout Execute successfully. std::endl; aclDestroyTensor(inputX); aclrtFree(inputXDeviceAddr); aclrtDestroyStream(stream); aclrtResetDevice(deviceId); aclFinalize(); return 0; }asdBlasCscal#include iostream #include vector #include asdsip.h #include complex #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; } void printTensor(const std::complexfloat *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 5; int64_t incx 1; std::complexfloat alpha (std::complexfloat){2, 3}; int64_t xSize 5; std::vectorstd::complexfloat tensorInXData; tensorInXData.reserve(xSize); for (int64_t i 0; i xSize; i) { tensorInXData[i] {(float)(1.0 i), (float)(2.0 i)}; } std::cout alpha alpha std::endl; std::cout ------- input TensorInX ------- std::endl; printTensor(tensorInXData.data(), xSize); std::vectorint64_t xShape {xSize}; aclTensor *inputX nullptr; void *inputXDeviceAddr nullptr; ret CreateAclTensor(tensorInXData, xShape, inputXDeviceAddr, aclDataType::ACL_COMPLEX64, inputX); CHECK_RET(ret ::ACL_SUCCESS, return ret); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork 0; void *buffer nullptr; asdBlasMakeCalPlan(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(asdBlasCscal(handle, n, alpha, inputX, incx)); 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); std::cout ------- output TensorInX ------- std::endl; printTensor(tensorInXData.data(), xSize); aclDestroyTensor(inputX); aclrtFree(inputXDeviceAddr); aclrtDestroyStream(stream); aclrtResetDevice(deviceId); aclFinalize(); return 0; }asdBlasCsscal#include iostream #include vector #include cmath #include random #include asdsip.h #include complex #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 5; int incx 1; float alpha 2.0; int64_t xSize 5; std::vectorstd::complexfloat tensorInXData; tensorInXData.reserve(xSize); for (int64_t i 0; i xSize; i) { tensorInXData[i] {(float)(1.0 i), (float)(2.0 i)}; } std::cout alpha alpha std::endl; std::cout ------- input TensorInX ------- std::endl; for (int64_t i 0; i n; i) { std::cout tensorInXData[i] ; } std::cout std::endl; std::vectorint64_t xShape {xSize}; aclTensor *inputX nullptr; void *inputXDeviceAddr nullptr; ret CreateAclTensor(tensorInXData, xShape, inputXDeviceAddr, aclDataType::ACL_COMPLEX64, inputX); CHECK_RET(ret ::ACL_SUCCESS, return ret); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork 0; void *buffer nullptr; asdBlasMakeCalPlan(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(asdBlasCsscal(handle, n, alpha, inputX, incx)); 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); std::cout ------- output TensorInX ------- std::endl; for (int64_t i 0; i xSize; i) { std::cout tensorInXData[i] ; } std::cout std::endl; aclDestroyTensor(inputX); aclrtFree(inputXDeviceAddr); 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