Edit

Share via

DML_ELEMENT_WISE_LOGICAL_AND_OPERATOR_DESC structure (directml.h)

  • Article

Performs a logical AND on each pair of corresponding elements of the input tensors, placing the result (1 for true, 0 for false) into the corresponding element of OutputTensor.

f(a, b) = (a && b)

This operator supports in-place execution, meaning that OutputTensor is permitted to alias one of the input tensors during binding.

Syntax

struct DML_ELEMENT_WISE_LOGICAL_AND_OPERATOR_DESC {
  const DML_TENSOR_DESC *ATensor;
  const DML_TENSOR_DESC *BTensor;
  const DML_TENSOR_DESC *OutputTensor;
};

Members

ATensor

Type: const DML_TENSOR_DESC*

A tensor containing the left-hand side inputs.

BTensor

Type: const DML_TENSOR_DESC*

A tensor containing the right-hand side inputs.

OutputTensor

Type: const DML_TENSOR_DESC*

The output tensor to write the results to.

Availability

This operator was introduced in DML_FEATURE_LEVEL_1_0.

Tensor constraints

ATensor, BTensor, and OutputTensor must have the same DataType, DimensionCount, and Sizes.

Tensor support

DML_FEATURE_LEVEL_3_0 and above

Tensor Kind Supported dimension counts Supported data types
ATensor Input 1 to 8 UINT32, UINT8
BTensor Input 1 to 8 UINT32, UINT8
OutputTensor Output 1 to 8 UINT32, UINT8

DML_FEATURE_LEVEL_2_0 and above

Tensor Kind Supported dimension counts Supported data types
ATensor Input 4 UINT32, UINT8
BTensor Input 4 UINT32, UINT8
OutputTensor Output 4 UINT32, UINT8

DML_FEATURE_LEVEL_1_0 and above

Tensor Kind Supported dimension counts Supported data types
ATensor Input 4 UINT32
BTensor Input 4 UINT32
OutputTensor Output 4 UINT32

Requirements

Requirement Value
Header directml.h