aclnnMedianDim

📄 View Source Code

Applicable Products

Function

Computes the median and its index along a specified dimension of a tensor (if the number of elements along the specified dimension is size, the median corresponds to the sorted index (size - 1) // 2).

Function Prototype

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

aclnnStatus aclnnMedianDimGetWorkspaceSize(
  const aclTensor *self,
  int64_t          dim,
  bool             keepDim,
  aclTensor       *valuesOut,
  aclTensor       *indicesOut,
  uint64_t        *workspaceSize,
  aclOpExecutor  **executor)

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

aclnnMedianDimGetWorkspaceSize

• Parameters:

  Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (shape)	Non-contiguous Tensor
  self (const aclTensor*)	Input	The input tensor to be computed.	Must have the same data type as valuesOut.	FLOAT, FLOAT16, UINT8, INT8, INT16, INT32, INT64, BFLOAT16	ND	-	√
  dim (int64_t)	Input	The specified dimension.	Supports the INT64 data type.	INT64	-	-	-
  keepDim (bool)	Input	Whether to retain the dimension of the reduction axis.	If True, the output valuesOut and indicesOut have the same number of dimensions as the input self, but the shape of the dimension corresponding to dim in the output aclTensor is 1. Otherwise, the dimension corresponding to dim is squeezed, resulting in the output having one fewer dimension than the input.	BOOL	-	-	-
  valuesOut (aclTensor*)	Output	Output tensor, the median along the specified dimension.	Must have the same data type as self, and the shape must correspond to self based on keepDim.	FLOAT, FLOAT16, UINT8, INT8, INT16, INT32, INT64, BFLOAT16	ND	-	√
  indicesOut (aclTensor*)	Output	Output tensor, representing the positions of the median values along the specified dimension.	The shape needs to correspond to self based on keepDim.	INT64	ND	-	√
  workspaceSize (uint64_t*)	Output	Returns the workspace size to be requested on the Device side.	-	-	-	-	-
  executor（aclOpExecutor**）	Output	Returns the op executor, which contains the operator computation flow.	-	-	-	-	-

  • Atlas inference products, Atlas training products: The BFLOAT16 data type is not supported.

• 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	The passed self, valuesOut, or indicesOut is a null pointer.
  ACLNN_ERR_PARAM_INVALID	161002	The data type of self, valuesOut, or indicesOut is not within the supported range.
  		The data type of valuesOut differs from that of self.
  		The dim value exceeds the dimension range of the input self.
  		The shape of valuesOut or indicesOut does not match the actual output shape.

aclnnMedianDim

• Parameters:

  Parameter	Input/Output	Description
  workspace	Input	Address of the workspace memory applied on the Device side.
  workspaceSize	Input	Size of the workspace applied on the Device side, obtained through the first-phase API aclnnMedianDimGetWorkspaceSize.
  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:

  • aclnnMedianDim defaults to a deterministic implementation.

• When the data type of self is not FLOAT, FLOAT16, or BFLOAT16, an excessively large tensor size may cause the operator execution to time out (aicpu error, with the error reason being [aicpu timeout]). The maximum size limits for specific types (strongly dependent on the actual remaining memory of the machine) are as follows:

  • INT64: 150000000

  • UINT8, INT8, INT16, INT32: 725000000

• When self is of the BFLOAT16 data type, the dimension where dim is located being 1 is not supported, and scalars are not supported.

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_median.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 shape_size = 1;
  for (auto i : shape) {
    shape_size *= i;
  }
  return shape_size;
}

int Init(int32_t deviceId, aclrtStream* stream) {
  // Boilerplate: initialize resources.
  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 memory on the device side.
  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.
  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 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. Refer to the acl API manual.
  // Fill in the deviceId based on your actual device.
  int32_t deviceId = 0;
  aclrtStream stream;
  auto ret = Init(deviceId, &stream);
  // Handle the check based on your needs.
  CHECK_RET(ret == 0, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);
  // 2. Construct inputs and outputs. Customize the construction based on the API interface.
  int64_t dim = 1;
  bool keepDim = false;
  std::vector<int64_t> selfShape = {4, 2};
  std::vector<int64_t> valuesOutShape = {4};
  std::vector<int64_t> indicesOutShape = {4};
  void* selfDeviceAddr = nullptr;
  void* valuesOutDeviceAddr = nullptr;
  void* indicesOutDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclTensor* valuesOut = nullptr;
  aclTensor* indicesOut = nullptr;
  std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
  std::vector<float> valuesOutHostData = {0, 0, 0, 0};
  std::vector<float> indicesOutHostData = {0, 0, 0, 0};
  // Create self aclTensor.
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the valuesOut aclTensor.
  ret = CreateAclTensor(valuesOutHostData, valuesOutShape, &valuesOutDeviceAddr, aclDataType::ACL_FLOAT, &valuesOut);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the indicesOut aclTensor.
  ret = CreateAclTensor(indicesOutHostData, indicesOutShape, &indicesOutDeviceAddr, aclDataType::ACL_INT64,
                        &indicesOut);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  // 3. Call the CANN operator library API. Replace with the specific API.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnMedianDim.
  ret = aclnnMedianDimGetWorkspaceSize(self, dim, keepDim, valuesOut, indicesOut, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMedianDimGetWorkspaceSize 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 aclnnMedianDim.
  ret = aclnnMedianDim(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMedianDim failed. ERROR: %d\n", ret); return ret);
  // 4. (Boilerplate) Synchronize and wait for task execution to finish.
  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(valuesOutShape);
  std::vector<float> valuesOutData(size, 0);
  ret = aclrtMemcpy(valuesOutData.data(), valuesOutData.size() * sizeof(valuesOutData[0]), valuesOutDeviceAddr,
                    size * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy valuesOut 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, valuesOutData[i]);
  }

  std::vector<float> indicesOutData(size, 0);
  ret = aclrtMemcpy(indicesOutData.data(), indicesOutData.size() * sizeof(indicesOutData[0]), indicesOutDeviceAddr,
                    size * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy indicesOut 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, indicesOutData[i]);
  }

  // 6. Release aclTensor and aclScalar. Modify based on the specific API definition.
  aclDestroyTensor(self);
  aclDestroyTensor(valuesOut);
  aclDestroyTensor(indicesOut);

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