aclnnMaxUnpool3d

Applicable Products

Function

• Description: The inverse operation of aclnnMaxPool in 3D. The D, H, and W axis sizes of outRef are determined by outputSize, and the element values of self are filled into outRef based on the indices index, with all other positions set to 0.

• Formula:

  • When the input is 4-dimensional, with dimensions N, D, H, and W (where N (Batch) represents the batch size, D (Depth) represents the feature map depth, H (Height) represents the feature map height, and W (Width) represents the feature map width):

    $$outRef[N][indices[N][i]] = self[N][i]$$

  • When the input is 5-dimensional, with dimensions N, C, D, H, and W (where C (Channels) represents the feature map channels):

    $$outRef[N][C][indices[N][C][i]] = self[N][C][i]$$

    Here, outRef, indices, and self are obtained by merging the last two axes into one axis through reshape, where i ∈ [0, DHW).

Function Prototype

• Each operator is divided into a two-phase API. The "aclnnMaxUnpool3dGetWorkspaceSize" API must be called first to obtain the workspace size required for computation and the executor that encapsulates the operator computation flow. Then, the "aclnnMaxUnpool3d" API is called to perform the computation.

aclnnStatus aclnnMaxUnpool3dGetWorkspaceSize(
  const aclTensor*     self,
  const aclTensor*     indices,
  const aclIntArray*   outputSize,
  const aclIntArray*   stride,
  const aclIntArray*   padding,
  aclTensor*           outRef,
  uint64_t*            workspaceSize,
  aclOpExecutor**      executor)

aclnnStatus aclnnMaxUnpool3d(
  void*          workspace,
  uint64_t       workspaceSize,
  aclOpExecutor* executor,
  aclrtStream    stream)

aclnnMaxUnpool3dGetWorkspaceSize

• Parameters:

  Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (Shape)	Non-contiguous Tensor
  self (aclTensor*)	Input	self in the formula, representing the target tensor to be converted.	The data type is the same as that of outRef. The shape is the same as that of indices. When the dimension is 4, the dimensions represent N, D, H, and W in sequence. When the dimension is 5, the dimensions represent N, C, D, H, and W in sequence.	FLOAT, FLOAT16, INT16, INT32, INT64, INT8, UINT8, DOUBLE	ND	4-5	√
  indices (aclTensor*)	Input	Indices in the formula, indicating the index positions of the input self elements in the output result.	The shape must be consistent with self. When the dimension is 4, each dimension represents N, D, H, and W in sequence. When the dimension is 5, each dimension represents N, C, D, H, and W in sequence.	INT64, INT32	ND	4-5	✓
  outputSize (aclIntArray*)	Input	Specifies the spatial size of the output in the D, H, and W dimensions.	The size is 3, and the product of the three elements must be greater than or equal to the product of the size of self in the D, H, and W dimensions.	-	-	-	-
  stride (aclIntArray*)	Input	Stride of the max pooling kernel along the D, H, and W dimensions.	Reserved parameter. This parameter is not used in the current version. Pass a host-side aclIntArray with a size of 3 and values greater than 0.	-	-	-	-
  padding (aclIntArray*)	Input	Specifies the padding values for the max pooling window in the D, H, and W dimensions.	Reserved parameter. This parameter is not used in computation in the current version. Pass a host-side aclIntArray with a size of 3.	-	-	-	-
  outRef (aclTensor*)	Output	outRef in the formula.	The data type is the same as that of self. When the dimension is 4, each dimension represents N, D, H, and W in sequence. When the dimension is 5, each dimension represents N, C, D, H, and W in sequence.	FLOAT, FLOAT16, INT16, INT32, INT64, INT8, UINT8, DOUBLE	ND	4-5	√
  workspaceSize (uint64_t*)	Output	Returns the workspace size to be applied for on the Device side.	-	-	-	-	-
  executor (aclOpExecutor**)	Output	Returns the op executor, which contains the operator computation flow.	-	-	-	-	-

• Return value

  aclnnStatus: return code. For details, see aclnn Return Code.

  The first-phase API performs input parameter validation and returns an error in the following scenarios:

  Return Value	Error Code	Description
  ACLNN_ERR_PARAM_NULLPTR	161001	Indicates that self, indices, outputSize, stride, padding, or outRef is a null pointer.
  ACLNN_ERR_PARAM_INVALID	161002	The data type of self and indices is not within the supported range.
  		The data type of self and outRef is inconsistent.
  		The dimension of self is not 4D or 5D.
  		The shape of self and indices is inconsistent.
  		The size of outputSize is not equal to 3.
  		The product of the three elements of outputSize is less than the product of the sizes of self in the D, H, and W dimensions.
  		The size of stride is not equal to 3.
  		The size of padding is not equal to 3.
  		The size of self in the C, D, H, and W dimensions is not greater than 0.
  		The element value of stride is not greater than 0.
  		The size of outRef in the N and C dimensions is not exactly the same as that of self.
  		The size of outRef in the D, H, and W dimensions is not equal to the three element values in outputSize.

aclnnMaxUnpool3d

• Parameters:

  Parameter	Input/Output	Description
  workspace	Input	Workspace memory address applied for on the Device side.
  workspaceSize	Input	Workspace size applied for on the Device side, obtained from the first-phase API aclnnLogAddExpGetWorkspaceSize.
  executor	Input	Operator executor, which contains the operator computation flow.
  stream	Input	Specifies the stream for task execution.

• Return value

  aclnnStatus: return code. For details, see aclnn Return Code.

Constraints

• Deterministic computation:

  • aclnnMaxUnpool3d defaults to a deterministic implementation.

Example

The following provides example code for reference only. For details about compilation and running, see Compile and Run Samples.

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_max_unpool3d.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, aclrtStream* stream) {
  // Boilerplate for resource initialization.
  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::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr,
                    aclDataType dataType, aclTensor** tensor) {
  auto size = GetShapeSize(shape) * sizeof(T);
  // Call aclrtMalloc to apply for device-side memory.
  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);
  // Copy host-side data to the device memory using aclrtMemcpy.
  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);

  // Calculate the strides of a contiguous tensor.
  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];
  }

  // Call the aclCreateTensor API to create an aclTensor.
  *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
                            shape.data(), shape.size(), *deviceAddr);
  return 0;
}

int main() {
  // 1. (Boilerplate) Initialize the device/stream. See the ACL API manual.
  // Set deviceId based on the actual device.
  int32_t 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);

  // 2. Construct the input and output. Customize the construction based on the API interface.
  std::vector<int64_t> selfShape = {1, 1, 2, 2};
  std::vector<int64_t> outShape = {1, 1, 4, 4};
  void* selfDeviceAddr = nullptr;
  void* indicesDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclTensor* out = nullptr;
  aclTensor* indices = nullptr;
  std::vector<float> selfHostData = {1, 2, 3, 4};
  std::vector<float> outHostData = {0, 0, 0, 0.0, 0, 0, 0, 0,
                                    0, 0, 0, 0.0, 0, 0, 0, 0};
  std::vector<int64_t> indicesHostData = {3, 8, 11, 13};
  // Create the self aclTensor.
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  // Create the indices aclTensor.
  ret = CreateAclTensor(indicesHostData, selfShape, &indicesDeviceAddr, aclDataType::ACL_INT64, &indices);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  // Create the out aclTensor.
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  std::vector<int64_t> arraySize1 = {1, 4, 4};
  const aclIntArray *outputSize = aclCreateIntArray(arraySize1.data(), arraySize1.size());
  CHECK_RET(outputSize != nullptr, return ACL_ERROR_INTERNAL_ERROR);

  std::vector<int64_t> arraySize2 = {1, 2, 3};
  const aclIntArray *stride = aclCreateIntArray(arraySize2.data(), arraySize2.size());
  CHECK_RET(stride != nullptr, return ACL_ERROR_INTERNAL_ERROR);
  const aclIntArray *padding = aclCreateIntArray(arraySize2.data(), arraySize2.size());
  CHECK_RET(padding != nullptr, return ACL_ERROR_INTERNAL_ERROR);

  // 3. Call the CANN operator library API. Replace with the specific API name.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnMaxUnpool3d.
  ret = aclnnMaxUnpool3dGetWorkspaceSize(self, indices, outputSize, stride, padding, out, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMaxUnpool3dGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // Apply for device memory based on the workspaceSize calculated by the first-phase API.
  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);
  }
  // Call the second-phase API of aclnnMaxUnpool3d.
  ret = aclnnMaxUnpool3d(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMaxUnpool3d failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate) Synchronize and wait for task execution to complete.
  ret = aclrtSynchronizeStream(stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);

  // 5. Obtain the output value and copy the result from the device memory to the host memory. Modify this based on the specific API definition.
  auto size = GetShapeSize(outShape);
  std::vector<float> outData(size, 0);
  ret = aclrtMemcpy(outData.data(), outData.size() * sizeof(outData[0]), outDeviceAddr,
                    size * sizeof(outData[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("out[%ld] is: %f\n", i, outData[i]);
  }

  // 6. Destroy aclTensor and aclScalar. Modify based on the specific API definition.
  aclDestroyTensor(self);
  aclDestroyTensor(out);
  aclDestroyTensor(indices);
  aclDestroyIntArray(outputSize);
  aclDestroyIntArray(stride);
  aclDestroyIntArray(padding);

  // 7. Release device resources.
  aclrtFree(selfDeviceAddr);
  aclrtFree(indicesDeviceAddr);
  aclrtFree(outDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
  return 0;
}