Asum

Applicable Products

Hardware Model

Supported or Not

Atlas 200I/500 A2 inference products

Not supported

Atlas inference products

Not supported

Atlas training products

Not supported

Atlas A2 training products/Atlas A2 inference products

Supported

Atlas A3 inference products/Atlas A3 training products

Supported

Function Description

API function:

  • asdBlasMakeAsumPlan: initializes the Asum operator configuration corresponding to the handle.
  • asdBlasSasum: sums up the absolute values of all input elements.
  • asdBlasScasum: sums up the absolute values of all input elements. The input elements are complex numbers.

Formula:

  • asdBlasSasum formula:
    • Example:

      The input x is as follows:

      [1, 2, -3, 4]

      After the Sasum operator is called, the output result is as follows:

      10
  • asdBlasScasum formula:
    • Example:

      Input x:

      [1+2i, 2-2i, -3+3i, 4-3i]

      After the Scasum operator is called, the output result is as follows:

      20

Function Prototype

  • AspbStatus asdBlasMakeAsumPlan(asdBlasHandle handle)
  • AspbStatus asdBlasSasum(asdBlasHandle handle, const int64_t n, aclTensor *x, const int64_t incx, aclTensor *result)
  • AspbStatus asdBlasScasum(asdBlasHandle handle, const int64_t n, aclTensor *x, const int64_t incx, aclTensor *result)

Parameter Description

  • asdBlasMakeAsumPlan

    Parameter

    Input/Output

    Type

    Description

    handle

    Input

    asdBlasHandle

    Asum operator handle.

  • asdBlasSasum & asdBlasScasum

    Parameter

    Input/Output

    Type

    Description

    handle

    Input

    asdBlasHandle

    asum operator handle.

    n

    Input

    const int64_t

    Total number of elements.

    x

    Input

    aclTensor *

    x in the formula, which is a tensor on the device. The data format can be ND, and the shape is [n].

    • For Sasum, the supported data type is FLOAT32.
    • For Scasum, the supported data type is COMPLEX64.

    incx

    Input

    const int64_t

    Memory address increment between the consecutive elements (currently, the offset is 1).

    result

    Output

    aclTensor *

    result in the formula, which is a tensor on the device. The supported data type is FLOAT32. It contains only one element. The data format supports ND, and the shape is [1].

Return Value Description

For details about the return values, see Return Value.

Constraints

  • asdBlasMakeAsumPlan: none.
  • asdBlasSasum
    • The input element count n currently supports the range [1, 6.71e+06].
    • The input shape of the operator is [n], and the output shape is [1].
    • The operator does not support the calculation for three or more dimensions.
  • asdBlasScasum
    • The input element count n currently supports the range [1, 6.71e+06].
    • The input shape of the operator is [n], and the output shape is [1].
    • The operator does not support the calculation for three or more dimensions.

Calling Example

For details about the operator calling example, see Asum.