aclnnConstantPadNd

Applicable Products

Product	Supported
Ascend 950PR/Ascend 950DT	√
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	×
Atlas inference products	√
Atlas training products	√

Function

Description: Pads the input tensor self with the constant value specified by value, based on the pad parameter.
Formula:
- Derivation formula for the shape of the out tensor:
  $\begin{aligned} &\text{Assume the shape of input self is:} \\ &[dim0_{in},dim1_{in}, dim2_{in}, dim3_{in}] &\text{Assume} \\ &pad = &[dim3_{begin},dim3_{end}, \\&&dim2_{begin},dim2_{end}, \\&&dim1_{begin},dim1_{end}, \\&&dim0_{begin},dim0_{end}] \end{aligned}$ $\begin{aligned} &\text{Then the shape of output out is:} \\ &[dim0_{out}, dim1_{out}, dim2_{out}, dim3_{out}] = &[dim0_{begin}+dim0_{in}+dim0_{end}, \\&&dim1_{begin}+dim1_{in}+dim1_{end}, \\&&dim2_{begin}+dim2_{in}+dim2_{end}, \\&&dim3_{begin}+dim3_{in}+dim3_{end}] \end{aligned}$
- Example 1:
  
  (When the pad array length equals twice the number of self dimensions)
  $\begin{aligned} selfShape &= [1, 1, 1, 1, 1]\\ pad &= \lbrace 0, 1, 2, 3, 4, 5, 6, 7, 8, 9\rbrace \\ outputShape &= [8+1+9, 6+1+7, 4+1+5, 2+1+3, 0+1+1]\\ &= [18,14,10,6,2] \end{aligned}$
- Example 2:
  
  (When the length of the pad array is less than twice the number of self dimensions)
  $\begin{aligned} selfShape &= [1, 1, 1, 1, 1]\\ pad &= \lbrace 0, 1, 2, 3, 4, 5\rbrace \\ outputShape &= [0+1+0, 0+1+0, 4+1+5, 2+1+3, 0+1+1]\\ &= [1,1,10,6,2] \end{aligned}$

Function Prototype

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

  aclnnStatus aclnnConstantPadNdGetWorkspaceSize(
    const aclTensor*   self,
    const aclIntArray* pad,
    const aclScalar*   value,
    aclTensor*         out,
    uint64_t*          workspaceSize,
    aclOpExecutor**    executor)

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

aclnnConstantPadNdGetWorkspaceSize

Parameters:

Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (shape)	Non-contiguous Tensor
self (aclTensor*)	Input	Original input data to be padded.	-	FLOAT, FLOAT16, INT32, INT64, INT16, INT8, UINT8, UINT16, UINT32, UINT64, BOOL, DOUBLE, COMPLEX64, COMPLEX128, BFLOAT16, HIFLOAT8, FLOAT8_E5M2, FLOAT8_E4M3FN, FLOAT8_E8M0.	ND	0-8	√
pad (aclIntArray*)	Input	Dimensions to pad on each axis of the input.	The array length must be an even number and cannot exceed twice the number of self dimensions.	INT64	-	-	-
value (aclScalar*)	Input	Padding value for the padded region.	-	Data types that can be converted to the data type of self.	-	-	-
out (aclTensor*)	Output	Output tensor	Output result after padding.	Same as self.	ND	The shape is consistent with self.	√
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.	-	-	-	-	-

Atlas training products and Atlas inference products: The data types BFLOAT16, HIFLOAT8, FLOAT8_E5M2, FLOAT8_E4M3FN, and FLOAT8_E8M0 are not supported.
Atlas A3 training products/Atlas A3 inference products and Atlas A2 training products/Atlas A2 inference products: The data types HIFLOAT8, FLOAT8_E5M2, FLOAT8_E4M3FN, and FLOAT8_E8M0 are not supported.
The data types of value and self must satisfy the data type deduction rules (see Deduction Relationship).
When the data type of self is HIFLOAT8, FLOAT8_E5M2, FLOAT8_E4M3FN, or FLOAT8_E8M0, only value with all-zero bits is supported.

Return Value

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

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

Return Value	Return Code	Description
ACLNN_ERR_PARAM_NULLPTR	161001	The input self, pad, value, or out is a null pointer.
ACLNN_ERR_PARAM_INVALID	161002	The data type of self, value, or out is not within the supported range.
		The data type of self is inconsistent with that of out.
		The data types of self and value do not satisfy the data type deduction rules.
		The shape derived from the shape of self and the pad input is inconsistent with the shape of out.
		The number of elements in pad is not even or exceeds twice the number of dimensions of self.
		The number of dimensions of self or out exceeds 8.
		Each value in pad must not cause the shape of out to be less than 0. If pad contains positive values, the shape of out must not contain 0.
		When the data format of self is not ND, the data format of out is inconsistent with that of self.
		When the data type of self is fp8, the elements in pad must not be negative.
		When the data type of self is HIFLOAT8, FLOAT8_E5M2, FLOAT8_E4M3FN, or FLOAT8_E8M0, value must not be 0.

aclnnConstantPadNd

Parameters

Parameter	Input/Output	Description
workspace	Input	The workspace memory address applied for on the Device side.
workspaceSize	Input	Size of the workspace applied for on the Device side, obtained through the first-phase API aclnnConstantPadNdGetWorkspaceSize.
executor	Input	Operator executor, which contains the operator computation process.
stream	Input	Specifies the stream on which the task is executed.

Return value

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

Constraints

Deterministic computation:
- aclnnConstantPadNd 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_constant_pad_nd.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: 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);
  // Call aclrtMemcpy to copy data from the host to the device memory.
  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 the 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 aclCreateTensor 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 = {3, 3};
  std::vector<int64_t> outShape = {5, 5};
  void* selfDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclIntArray* pad = nullptr;
  aclScalar* value = nullptr;
  aclTensor* out = nullptr;
  std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7, 8};
  std::vector<float> outHostData(25, 0);
  float valueValue = 0.0f;
  std::vector<int64_t> padData = {1,1,1,1};
  // Create the self aclTensor.
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the pad array.
  pad = aclCreateIntArray(padData.data(), 4);
  CHECK_RET(pad != nullptr, return ret);
  // Create the value aclScalar.
  value = aclCreateScalar(&valueValue, aclDataType::ACL_FLOAT);
  CHECK_RET(value != nullptr, return ret);
  // Create the out aclTensor.
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  // 3. Call the CANN operator library API. Replace with the actual API name.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnConstantPadNd.
  ret = aclnnConstantPadNdGetWorkspaceSize(self, pad, value, out, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnConstantPadNdGetWorkspaceSize 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 aclnnConstantPadNd.
  ret = aclnnConstantPadNd(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnConstantPadNd failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate) Wait for the task execution to complete synchronously.
  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. Modify based on the specific API definition.
  auto size = GetShapeSize(outShape);
  std::vector<float> 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. Release aclTensor and aclScalar. Modify based on the specific API definition.
  aclDestroyTensor(self);
  aclDestroyIntArray(pad);
  aclDestroyScalar(value);
  aclDestroyTensor(out);

  // 7. Release device resources. Modify based on the specific API definition.
  aclrtFree(selfDeviceAddr);
  aclrtFree(outDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();

  return 0;
}