aclnnLogSumExp

📄 View Source Code

Applicable Products

Function

• Description: Returns the logarithm of the sum of exponentials of the input tensor along a specified dimension.

• Formula:

  In the formula, i is the dimension specified by dim, and j is the element index of the input along the specified dimension.

  $$logsumexp(x)_i = log\sum_{j} exp(x_{ij} )$$

• Example

  Example1：
    self: [2, 3, 4]       # self_shape=[2, 3, 4];
    dim: [2]              # dim_shape=[2], dim_data = {1, 2}, specifies the dimension.
    keepDim: false
    out: [2]              # out_shape=[2];
  
  Example2：
    self: [2, 3, 4]       # self_shape=[2, 3, 4];
    dim: [2]              # dim_shape=[2], dim_data = {1, 2}, specifies the dimension.
    keepDim: true
    out: [2, 1, 1]        # out_shape=[2, 1, 1];

Function Prototype

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

aclnnStatus aclnnLogSumExpGetWorkspaceSize(
  const aclTensor*     self,
  const aclIntArray*   dim,
  bool                 keepDim,
  aclTensor*           out,
  uint64_t*            workspaceSize,
  aclOpExecutor**      executor)

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

aclnnLogSumExpGetWorkspaceSize

• Parameters:

  Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (Shape)	Non-contiguous Tensor
  self (aclTensor*)	Input	x in the formula.	The data type must be convertible to the data type of out (see Conversion Relationship).	FLOAT, FLOAT16, BFLOAT16, INT32, INT64, INT16, INT8, UINT8, BOOL	ND	0-8	√
  dim (aclIntArray*)	Input	Dimension along which the computation is performed, i in the formula.	Value range: [-self.dim(), self.dim()-1].	INT64	-	-	-
  keepDim	Input	Determines whether the dimension of the reduced axis is retained.	-	BOOL	-	-	-
  out (aclTensor*)	Output	logsumexp(x) in the formula.	If keepDim is true, all dimensions except the dimension specified by dim, which has a size of 1, must have the same shape as self. If keepDim is false, the reduce axis dimension is not retained, and the remaining dimensions must have the same shape as self. The data type must be convertible to the data type of self (see Conversion Relationship).	FLOAT, FLOAT16, BFLOAT16	ND	0–8	√
  workspaceSize (uint64_t*)	Output	Returns the workspace size to be applied on the Device side.	-	-	-	-	-
  executor (aclOpExecutor**)	Output	Returns the op executor, which contains the operator computation flow.	-	-	-	-	-

  • Atlas inference products and Atlas training products: The BFLOAT16 data type is not 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	Error Code	Description
  ACLNN_ERR_PARAM_NULLPTR	161001	The input self, out, and dim are null pointers.
  ACLNN_ERR_PARAM_INVALID	161002	The data type of self, out, or dim is not within the supported range.
  		A dimension in the dim array exceeds the dimension range of the input tensor.
  		Elements in the dim array are duplicated.
  		The number of dimensions of self or out exceeds 8.
  		The data type of self cannot be converted to the data type of out.
  		The shape of out does not match the shape derived from self, dim, and keepDim.

aclnnLogSumExp

• Parameters:

  Parameter	Input/Output	Description
  workspace	Input	Specifies the workspace memory address applied on the Device side.
  workspaceSize	Input	Specifies the workspace size applied on the Device side, obtained by the first-phase API aclnnLogSumExpGetWorkspaceSize.
  executor	Input	Operator executor that includes the operator computation flow.
  stream	Input	Specifies the stream for executing the task.

• Return value

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

Constraints

• Deterministic computation:

  • aclnnLogSumExp 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_logsumexp.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 side 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);

  // Compute strides for 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;
}

template <typename T>
int CreateAclIntArray(const std::vector<T>& hostData, aclIntArray** intArray) {
  auto size = GetShapeSize(hostData) * sizeof(T);

  // Call the aclCreateIntArray API to create an aclIntArray.
  *intArray = aclCreateIntArray(hostData.data(), hostData.size());
  return 0;
}

int main() {
  // 1. (Boilerplate) Initialize the device/stream. See the acl API manual.
  // Set deviceId based on your 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 = {2, 2, 2};
  std::vector<int64_t> outShape = {2};

  void* selfDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclIntArray* dim = nullptr;
  aclTensor* out = nullptr;

  std::vector<float> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7};
  std::vector<float> outHostData = {2, 3};
  std::vector<int64_t> dimData = {1, 2};
  bool keepdim = false;

  // Create the self aclTensor.
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the dim aclIntArray.
  ret = CreateAclIntArray(dimData, &dim);
  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);

  // 3. Call the CANN operator library API.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnLogSumExp.
  ret = aclnnLogSumExpGetWorkspaceSize(self, dim, keepdim, out, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnLogSumExpGetWorkspaceSize 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 aclnnLogSumExp.
  ret = aclnnLogSumExp(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnLogSumExp failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate) Synchronously 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 side.
  auto size = GetShapeSize(outShape);
  std::vector<float> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]),
                    outDeviceAddr, size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy resultData 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.
  aclDestroyTensor(self);
  aclDestroyIntArray(dim);
  aclDestroyTensor(out);

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