aclnnAminmax

📄 View source code

Supported Products

Function Description

Returns the minimum and maximum values of each row of the input tensor along the specified dimension.

Prototype

Each operator has two-phase API calls. First, aclnnAminmaxGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnAminmax is called to perform computation.

aclnnStatus aclnnAminmaxGetWorkspaceSize(
const aclTensor   *self, 
const aclIntArray *dim, 
bool               keepDim, 
aclTensor         *minOut, 
aclTensor         *maxOut, 
uint64_t          *workspaceSize, 
aclOpExecutor    **executor)

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

aclnnAminmaxGetWorkspaceSize

• Parameters:

  Name	Input/Output	Description	Usage Notes	Data Type	Data Format	Dimension (Shape)	Non-contiguous Tensor
  self	Input	Input tensor.	-	FLOAT, BFLOAT16, FLOAT16, DOUBLE, INT8, INT16, INT32, INT64, UINT8, or BOOL	ND	-	√
  dim	Input	Specifies the dimension to be reduced.	The value range is [–self.dim(), self.dim() – 1]	INT64	-	-	-
  keepdim	Input	Indicates whether to retain the reduction axes.	-	BOOL	-	-	-
  minOut	Output	An output tensor storing the minimum values.	Its data type must be the same as that of self.	FLOAT, BFLOAT16, FLOAT16, DOUBLE, INT8, INT16, INT32, INT64, UINT8, or BOOL	ND	-	√
  maxOut	Output	An output tensor storing the maximum values.	Its data type must be the same as that of self.	FLOAT, BFLOAT16, FLOAT16, DOUBLE, INT8, INT16, INT32, INT64, UINT8, or BOOL	ND	-	√
  workspaceSize	Output	Size of the workspace required to be allocated on the device.	-	-	-	-	-
  executor	Output	Operator executor, covering the operator computation process.	-	-	-	-	-

  • Atlas inference products and Atlas training products: The data type cannot be BFLOAT16.

• Returns:

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

  The first-phase API implements input parameter verification. The following errors may be thrown.

  Return	Error Code	Description
  ACLNN_ERR_PARAM_NULLPTR	161001	The input self, dim, minOut, or maxOut is a null pointer.
  ACLNN_ERR_PARAM_INVALID	161002	The data type of self is not supported.
  		The data type of minOut or maxOut is different from that of self.
  		self, minOut, or maxOut has more than 8 dimensions.
  		The value of dim is out of the range.
  		The number of elements in dim is 1 and the shape of this specified dimension in self is 0.
  		The number of elements in dim is not 1 and self is an empty tensor.

aclnnAminmax

• Parameters:

  Name	Input/Output	Description
  workspace	Input	Address of the workspace to be allocated on the device.
  workspaceSize	Input	Size of the workspace to be allocated on the device, obtained by calling the first-phase API aclnnAminmaxGetWorkspaceSize.
  executor	Input	Operator executor, covering the operator computation process.
  stream	Input	Stream for executing the task.

• Returns:

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

Constraints

• Deterministic computation:
  • aclnnAminmax defaults to a deterministic implementation.

Example

The following example is for reference only. For details, see Compilation and Running Sample.

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_aminmax.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 code 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 allocate memory on the device.
  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 the 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);

  // Compute 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 code for device/stream initialization. For details, see the ACL API manual.
  // Set the device ID in use.
  int32_t deviceId = 0;
  aclrtStream stream;
  auto ret = Init(deviceId, &stream);
  // Customize the error handling as needed.
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);

  // 2. Construct the inputs and outputs based on the API definition.
  std::vector<int64_t> selfShape = {2, 3, 2};
  std::vector<int64_t> outShape = {1, 3, 2};
  void* selfDeviceAddr = nullptr;
  void* minOutDeviceAddr = nullptr;
  void* maxOutDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclTensor* minOut = nullptr;
  aclTensor* maxOut = nullptr;
  std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11};
  std::vector<float> minOutHostData = {0, 0, 0, 0, 0, 0};
  std::vector<float> maxOutHostData = {0, 0, 0, 0, 0, 0};
  std::vector<int64_t> dimData = {0};
  bool keepDim = true;
  // Create a self aclTensor.
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create a dim aclIntArray.
  aclIntArray *dim = aclCreateIntArray(dimData.data(), dimData.size());
  // Create an out aclTensor.
  ret = CreateAclTensor(minOutHostData, outShape, &minOutDeviceAddr, aclDataType::ACL_FLOAT, &minOut);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(maxOutHostData, outShape, &maxOutDeviceAddr, aclDataType::ACL_FLOAT, &maxOut);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  // 3. Call the CANN operator library API, which needs to be replaced with the actual one.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnAminmax.
  ret = aclnnAminmaxGetWorkspaceSize(self, dim, keepDim, minOut, maxOut, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnAminmaxGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // Allocate 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 aclnnAminmax.
  ret = aclnnAminmax(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnAminmax failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate code) Wait until the task execution is 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 to the host. Modify the code based on the API definition.
  auto size = GetShapeSize(outShape);
  std::vector<float> minResultData(size, 0);
  ret = aclrtMemcpy(minResultData.data(), minResultData.size() * sizeof(minResultData[0]), minOutDeviceAddr,
                    size * sizeof(minResultData[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("result[%ld] is: %f\n", i, minResultData[i]);
  }
  std::vector<float> maxResultData(size, 0);
  ret = aclrtMemcpy(maxResultData.data(), maxResultData.size() * sizeof(maxResultData[0]), maxOutDeviceAddr,
                    size * sizeof(maxResultData[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("result[%ld] is: %f\n", i, maxResultData[i]);
  }

  // 6. Destroy aclTensor. Modify the code based on the API definition.
  aclDestroyTensor(self);
  aclDestroyIntArray(dim);
  aclDestroyTensor(minOut);
  aclDestroyTensor(maxOut);

  // 7. Free device resources. Modify the code based on the API definition.
  aclrtFree(selfDeviceAddr);
  aclrtFree(minOutDeviceAddr);
  aclrtFree(maxOutDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
  return 0;
}