--- title: aclnnAdaptiveAvgPool2d description: "每个算子有两段接口，必须先调用“aclnnXxxGetWorkspaceSize”接口获取入参并根据计算流程计算所需workspace大小，再调用“aclnnXxx”接口执行计算。两段式接口如下：" url: https://www.hiascend.com/document/detail/zh/Atlas200IDKA2DeveloperKit/latest/Application%20Development%20Guide/aadgc/aclcppdevg_03_0192.html sourcePath: /source/zh/Atlas200IDKA2DeveloperKit/23.0.RC2/Application Development Guide/aadgc/aclcppdevg_03_0192.html indexId: 61e2cb7f19c405877f9508f47afd4d5dc78ba0cfd04f969ca8b6f2b6f404c840105 --- # aclnnAdaptiveAvgPool2d #### 接口原型 每个算子有两段接口，必须先调用“aclnnXxxGetWorkspaceSize”接口获取入参并根据计算流程计算所需workspace大小，再调用“aclnnXxx”接口执行计算。两段式接口如下： - 第一段接口：aclnnStatus aclnnAdaptiveAvgPool2dGetWorkspaceSize(const aclTensor *self, const aclIntArray *outputSize, const aclTensor *out, uint64_t *workspaceSize, aclOpExecutor **executor) - 第二段接口：aclnnStatus aclnnAdaptiveAvgPool2d(void *workspace, uint64_t* workspaceSize, aclOpExecutor *executor, const aclrtStream stream) #### 功能描述 算子功能：aclnnAvgPool2d的自适应操作算子，为2D平均池化。不同的是，本算子只需指定输出的大小，即可自动推导出kernel的大小与对应的步长。 #### aclnnAdaptiveAvgPool2dGetWorkspaceSize - 接口定义： aclnnStatus aclnnAdaptiveAvgPool2dGetWorkspaceSize(const aclTensor *self, const aclIntArray *outputSize, const aclTensor *out, uint64_t *workspaceSize, aclOpExecutor **executor) - 参数说明： - self：Device侧的aclTensor，数据类型支持FLOAT16、FLOAT，支持非连续Tensor，数据格式支持NCHW、CHW。 - outputSize：输出Tensor的Size，数据类型为INT数组，数组长度不能超过2。 - out：Device侧的aclTensor，数据类型支持FLOAT16、FLOAT；shape由self和outputSize共同指定，数据格式支持NCHW、CHW。 - workspaceSize：返回用户需要在Device侧申请的workspace大小。 - executor：返回op执行器，包含了算子计算流程。 - 返回值： 返回aclnnStatus状态码，具体参见aclnn返回码。 第一段接口完成入参校验，出现以下场景时报错： 返回161001（ACLNN_ERR_PARAM_NULLPTR）：传入的self或out是空指针。 返回161002（ACLNN_ERR_PARAM_INVALID）： 参数self的数据类型和数据格式不在支持的范围内。 参数self和out数据类型不一致。 参数outputsize非法。 参数self的维度不等于3或4。 #### aclnnAdaptiveAvgPool2d - 接口定义： aclnnStatus aclnnAdaptiveAvgPool2d(void *workspace, uint64_t* workspaceSize, aclOpExecutor *executor, const aclrtStream stream) - 参数说明： - workspace：在Device侧申请的workspace内存起址。 - workspaceSize：在Device侧申请的workspace大小，由第一段接口aclnnAdaptiveAvgPool2dGetWorkspaceSize获取。 - executor：op执行器，包含了算子计算流程。 - stream：指定执行任务的AscendCL stream流。 - 返回值： 返回aclnnStatus状态码，具体参见aclnn返回码。 #### 调用示例 ``` #include #include #include "acl/acl.h" #include "aclnnop/aclnn_adaptive_avg_pool2d.h" #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& shape) { int64_t shape_size = 1; for (auto i : shape) { shape_size *= i; } return shape_size; } int Init(int32_t deviceId, aclrtContext* context, aclrtStream* stream) { // 固定写法，AscendCL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 = aclrtCreateContext(context, deviceId); CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateContext failed. ERROR: %d\n", ret); return ret); ret = aclrtSetCurrentContext(*context); CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetCurrentContext 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 int CreateAclTensor(const std::vector& hostData, const std::vector& 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侧数据拷贝到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 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_NCHW, shape.data(), shape.size(), *deviceAddr); return 0; } int main() { // 1. （固定写法）device/context/stream初始化, 参考AscendCL对外接口列表 // 根据自己的实际device填写deviceId int32_t deviceId = 0; aclrtContext context; aclrtStream stream; auto ret = Init(deviceId, &context, &stream); // check根据自己的需要处理 CHECK_RET(ret == 0, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret); // 2. 构造输入与输出，需要根据API的接口自定义构造 std::vector inputShape = {2, 2, 3, 3}; std::vector outShape = {2, 2, 2, 2}; std::vector outSize = {2, 2}; void* inputDeviceAddr = nullptr; void* outDeviceAddr = nullptr; aclTensor* input = nullptr; aclTensor* out = nullptr; aclIntArray* outputSize = aclCreateIntArray(outSize.data(), 2); std::vector inputHostData = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36}; std::vector outHostData = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; // 创建input aclTensor ret = CreateAclTensor(inputHostData, inputShape, &inputDeviceAddr, aclDataType::ACL_FLOAT, &input); CHECK_RET(ret == ACL_SUCCESS, return ret); // 创建out aclTensor ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out); CHECK_RET(ret == ACL_SUCCESS, return ret); // 3.调用CANN算子库API，需要修改为具体的算子接口 uint64_t workspaceSize = 0; aclOpExecutor* executor; // 调用aclnnAdaptiveAvgPool2d第一段接口 ret = aclnnAdaptiveAvgPool2dGetWorkspaceSize(input, outputSize, out, &workspaceSize, &executor); CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnAdaptiveAvgPool2dGetWorkspaceSize failed. ERROR: %d\n", ret); return ret); // 根据第一段接口计算出的workspaceSize申请device内存 void* workspaceAddr = nullptr; if (workspaceSize > 0) { ret = aclrtMalloc(&workspaceAddr, workspaceSize, ACL_MEM_MALLOC_HUGE_FIRST); CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret;); } // 调用aclnnAdaptiveAvgPool2d第二段接口 ret = aclnnAdaptiveAvgPool2d(workspaceAddr, workspaceSize, executor, stream); CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnAdaptiveAvgPool2d failed. ERROR: %d\n", ret); return ret); // 4. （固定写法）同步等待任务执行结束 ret = aclrtSynchronizeStream(stream); CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret); // 5. 获取输出的值，将device侧内存上的结果拷贝至Host侧，需要根据具体API的接口定义修改 auto size = GetShapeSize(outShape); std::vector resultData(size, 0); ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * 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); for (int64_t i = 0; i < size; i++) { LOG_PRINT("result[%ld] is: %f\n", i, resultData[i]); } // 6. 释放aclTensor和aclScalar，需要根据具体API的接口定义修改 aclDestroyTensor(input); aclDestroyTensor(out); return 0; } ```