cann/sip列方向逐点乘算子

Colwise_mul【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asdBlasMakeColwiseMulPlan初始化该句柄对应的ColwiseMul算子配置。asdBlasColwiseMul复数矩阵与复数向量按列逐点乘返回一个和输入矩阵同样形状大小的复数矩阵。计算公式 示例输入“A”为[ [ 11i, 11i ],[ 22i, 22i ] ]输入“X”为[ 11i, 22i ]调用“asdBlasColwiseMul”算子后输出“result”为[ [ 02i, 02i ],[ 08i, 08i ] ]函数原型AspbStatus asdBlasMakeColwiseMulPlan( asdBlasHandle handle)AspbStatus asdBlasColwiseMul( asdBlasHandle handle, const int64_t m, const int64_t n, aclTensor * mat, aclTensor * vec, aclTensor * result)asdBlasMakeColwiseMulPlan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。asdBlasColwiseMul参数说明参数名输入/输出描述handleasdBlasHandle输入算子的句柄mataclTensor*输入输入向量对应公式中的A。数据类型支持COMPLEX64。数据格式支持ND。shape为[m,n]。mint64_t输入矩阵mat的行数向量vec的元素个数。nint64_t输入矩阵mat的列数。vecaclTensor*输入输入向量对应公式中的X。数据类型支持COMPLEX64。数据格式支持ND。shape为[m]。resultaclTensor*输出输出向量对应公式中的result。数据类型支持COMPLEX64。数据格式支持ND。shape为[m, n]。返回值返回状态码具体参见SiP返回码。约束说明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。#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 rows, int64_t cols) { for (int64_t i 0; i rows; i) { for (int64_t j 0; j cols; j) { std::cout tensorData[i * cols j] ; } 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 m 3; int64_t n 2; int64_t matSize m * n; std::vectorstd::complexfloat tensorInMatData; tensorInMatData.reserve(matSize); for (int64_t i 0; i m; i) { for (int64_t j 0; j n; j) { tensorInMatData[n * i j] (std::complexfloat){2.0, -2.0}; } } int64_t vecSize m; std::vectorstd::complexfloat tensorInVecData; tensorInVecData.reserve(vecSize); for (int64_t i 0; i vecSize; i) { tensorInVecData[i] (std::complexfloat){3.0, -4.0}; } int64_t resultSize m * n; std::vectorstd::complexfloat resultData; resultData.reserve(resultSize); std::cout ------- input mat ------- std::endl; printTensor(tensorInMatData.data(), m, n); std::cout ------- input vec ------- std::endl; printTensor(tensorInVecData.data(), m, 1); std::vectorint64_t matShape {m, n}; std::vectorint64_t vecShape {m}; std::vectorint64_t resultShape {m, n}; aclTensor *inputMat nullptr; aclTensor *inputVec nullptr; aclTensor *outputResult nullptr; void *inputMatDeviceAddr nullptr; void *inputVecDeviceAddr nullptr; void *outputResultDeviceAddr nullptr; ret CreateAclTensor(tensorInMatData, matShape, inputMatDeviceAddr, aclDataType::ACL_COMPLEX64, inputMat); CHECK_RET(ret ::ACL_SUCCESS, return ret); ret CreateAclTensor(tensorInVecData, vecShape, inputVecDeviceAddr, aclDataType::ACL_COMPLEX64, inputVec); CHECK_RET(ret ::ACL_SUCCESS, return ret); ret CreateAclTensor(resultData, resultShape, outputResultDeviceAddr, aclDataType::ACL_COMPLEX64, outputResult); CHECK_RET(ret ::ACL_SUCCESS, return ret); asdBlasHandle handle; asdBlasCreate(handle); size_t lwork 0; void *buffer nullptr; asdBlasMakeColwiseMulPlan(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(asdBlasColwiseMul(handle, m, n, inputMat, inputVec, outputResult)); asdBlasSynchronize(handle); asdBlasDestroy(handle); buffer nullptr; ret aclrtMemcpy(resultData.data(), resultSize * sizeof(std::complexfloat), outputResultDeviceAddr, resultSize * sizeof(std::complexfloat), 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(), m, n); aclDestroyTensor(inputMat); aclDestroyTensor(inputVec); aclDestroyTensor(outputResult); aclrtFree(inputMatDeviceAddr); aclrtFree(inputVecDeviceAddr); aclrtFree(outputResultDeviceAddr); 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