aclnnConvolutionBackward

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：

功能描述

aclnnConvolutionBackward

aclnnConvolutionBackwardGetWorkspaceSize

约束与限制

无

调用示例

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/level2/aclnn_convolution_backward.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<int64_t>& shape) {
  int64_t shapeSize = 1;
  for (auto i : shape) {
    shapeSize *= i;
  }
  return shapeSize;
}
int Init(int32_t deviceId, aclrtContext* context, aclrtStream* stream) {
  // 固定写法，acl初始化
  auto 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);
  ret = aclInit(nullptr);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclInit 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<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
  if (shape.size() == 4) {
    *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_NCHW,
                              shape.data(), shape.size(), *deviceAddr);
  } else {
    *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
                              shape.data(), shape.size(), *deviceAddr);   
  }
  return 0;
}
int main() {
  // 1. （固定写法）device/context/stream初始化，参考acl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_RET(ret == ACL_SUCCESS, 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<int64_t> gradOutputShape = {2, 2, 7, 7};
  std::vector<int64_t> inputShape = {2, 2, 7, 7};
  std::vector<int64_t> weightShape = {2, 2, 1, 1};
  std::vector<int64_t> bias = {2};
  std::vector<int64_t> stride = {1, 1};
  std::vector<int64_t> padding = {0, 0};
  std::vector<int64_t> dilation = {1, 1};
  bool transposed = false;
  std::vector<int64_t> outputPadding = {0, 0};
  int groups = 1;
  bool outputMask[3] = {true, true, true};
  int8_t cubeMathType = 0;
  std::vector<int64_t> gradInputShape = {2, 2, 7, 7};
  std::vector<int64_t> gradWeightShape = {2, 2, 1, 1};
  std::vector<int64_t> gradBiasShape = {2};
  // 创建gradOut aclTensor
  std::vector<float> gradOutputData(GetShapeSize(gradOutputShape) * 2, 1);
  aclTensor* gradOutput = nullptr;
  void *gradOutputdeviceAddr = nullptr;
  ret = CreateAclTensor(gradOutputData, gradOutputShape, &gradOutputdeviceAddr, aclDataType::ACL_FLOAT16, &gradOutput);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建input aclTensor
  std::vector<float> inputData(GetShapeSize(inputShape) * 2, 1);
  aclTensor* input = nullptr;
  void *inputdeviceAddr = nullptr;
  ret = CreateAclTensor(inputData, inputShape, &inputdeviceAddr, aclDataType::ACL_FLOAT16, &input);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建weight aclTensor
  std::vector<float> weightData(GetShapeSize(weightShape) * 2, 1);
  aclTensor* weight = nullptr;
  void *weightDeviceAddr = nullptr;
  ret = CreateAclTensor(weightData, weightShape, &weightDeviceAddr, aclDataType::ACL_FLOAT16, &weight);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建gradInput aclTensor
  std::vector<float> gradInputData(GetShapeSize(gradInputShape) * 2, 1);
  aclTensor* gradInput = nullptr;
  void *gradInputDeviceAddr = nullptr;
  ret = CreateAclTensor(gradInputData, gradInputShape, &gradInputDeviceAddr, aclDataType::ACL_FLOAT16, &gradInput);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建gradWeight aclTensor
  std::vector<float> gradWeightData(GetShapeSize(gradWeightShape) * 2, 1);
  aclTensor* gradWeight = nullptr;
  void *gradWeightDeviceAddr = nullptr;
  ret = CreateAclTensor(gradWeightData, gradWeightShape, &gradWeightDeviceAddr, aclDataType::ACL_FLOAT16, &gradWeight);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建gradBias aclTensor
  std::vector<float> gradBiasData(GetShapeSize(gradBiasShape) * 2, 1);
  aclTensor* gradBias = nullptr;
  void *gradBiasDeviceAddr = nullptr;
  ret = CreateAclTensor(gradBiasData, gradBiasShape, &gradBiasDeviceAddr, aclDataType::ACL_FLOAT16, &gradBias);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  aclIntArray *biasSizes = aclCreateIntArray(bias.data(), 1);
  aclIntArray *strides = aclCreateIntArray(stride.data(), 2);
  aclIntArray *pads = aclCreateIntArray(padding.data(), 2);
  aclIntArray *dilations = aclCreateIntArray(dilation.data(), 2);
  aclIntArray *outputPads = aclCreateIntArray(outputPadding.data(), 2);
  aclBoolArray *outMask = aclCreateBoolArray(outputMask, 3);

  // 3. 调用CANN算子库API，需要修改为具体的Api名称
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // 调用aclnnConvolutionBackward第一段接口
  ret = aclnnConvolutionBackwardGetWorkspaceSize(gradOutput, input, weight, biasSizes, strides, pads, dilations, transposed, outputPads, groups, outMask, cubeMathType, gradInput, gradWeight, gradBias, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnConvolutionBackwardGetWorkspaceSize 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aclnnConvolutionBackward第二段接口
  ret = aclnnConvolutionBackward(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnConvolutionBackward 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(gradInputShape);
  std::vector<float> gradInputResult(size, 0);
  ret = aclrtMemcpy(gradInputResult.data(), gradInputResult.size() * sizeof(gradInputResult[0]), gradInputDeviceAddr,
                    size * sizeof(gradInputResult[0]), 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("gradInputResult[%ld] is: %f\n", i, gradInputResult[i]);
  }
  size = GetShapeSize(gradWeightShape);
  std::vector<float> gradWeightResult(size, 0);
  ret = aclrtMemcpy(gradWeightResult.data(), gradWeightResult.size() * sizeof(gradWeightResult[0]), gradWeightDeviceAddr,
                    size * sizeof(gradWeightResult[0]), 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("gradWeightResult[%ld] is: %f\n", i, gradWeightResult[i]);
  }
  size = GetShapeSize(gradBiasShape);
  std::vector<float> gradBiasResult(size, 0);
  ret = aclrtMemcpy(gradBiasResult.data(), gradBiasResult.size() * sizeof(gradBiasResult[0]), gradInputDeviceAddr,
                    size * sizeof(gradBiasResult[0]), 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("gradBiasResult[%ld] is: %f\n", i, gradBiasResult[i]);
  }

  // 6. 释放aclTensor和aclScalar，需要根据具体API的接口定义修改
  aclDestroyTensor(gradOutput);
  aclDestroyTensor(input);
  aclDestroyTensor(weight);
  aclDestroyTensor(gradInput);
  aclDestroyTensor(gradWeight);
  aclDestroyTensor(gradBias);
  return 0;
}
父主题： NN类算子接口（aclnn）