DML_SCATTER_OPERATOR_DESC structure (directml.h)

Copies the whole input tensor to the output, then overwrites selected indices with corresponding values from the updates tensor. This operator performs the following pseudocode.

output = input
output[index[i, j, k, ...], j, k, ...] = updates[i, j, k, ...] // if axis == 0
output[i, index[i, j, k, ...], k, ...] = updates[i, j, k, ...] // if axis == 1
output[i, j, index[i, j, k, ...], ...] = updates[i, j, k, ...] // if axis == 2
...

If two output element indices overlap (which is invalid), then there's no guarantee of which last write wins.

DML_SCATTER_OPERATOR_DESC is the inverse of DML_GATHER_OPERATOR_DESC.

Syntax

struct DML_SCATTER_OPERATOR_DESC {
  const DML_TENSOR_DESC *InputTensor;
  const DML_TENSOR_DESC *IndicesTensor;
  const DML_TENSOR_DESC *UpdatesTensor;
  const DML_TENSOR_DESC *OutputTensor;
  UINT                  Axis;
};

Members

InputTensor

Type: const DML_TENSOR_DESC*

The tensor to read from.

IndicesTensor

Type: const DML_TENSOR_DESC*

A tensor containing the indices into the output tensor. The Sizes must match InputTensor.Sizes for every dimension except Axis.

Starting with DML_FEATURE_LEVEL_3_0, this operator supports negative index values when using a signed integral type with this tensor. Negative indices are interpreted as being relative to the end of the axis dimension. For example, an index of -1 refers to the last element along that dimension.

UpdatesTensor

Type: const DML_TENSOR_DESC*

A tensor containing the new values to replace the existing input values at the corresponding indices. The Sizes of this tensor must match IndicesTensor.Sizes. The DataType must match InputTensor.DataType.

OutputTensor

Type: const DML_TENSOR_DESC*

The tensor to write the results to. The Sizes and DataType of this tensor must match InputTensor.

Axis

Type: UINT

The axis dimension to use for indexing in OutputTensor, ranging [0, OutputTensor.DimensionCount).

Examples

Example 1. 1D scatter

Axis: 0

InputTensor: (Sizes:{5}, DataType:FLOAT32)
    [0,1,2,3,4]

IndicesTensor: (Sizes:{4}, DataType:UINT32)
    [3,1,3,0]

UpdatesTensor: (Sizes:{4}, DataType:FLOAT32)
    [5,6,7,8]

// output = input
// output[indices[x]] = updates[x]
OutputTensor: (Sizes:{5}, DataType:FLOAT32)
    [8,6,2,7,4]

Example 2. 2D scatter

Axis: 0

InputTensor: (Sizes:{2,3}, DataType:FLOAT32)
    [[0.0, 0.0, 0.0],
     [0.0, 0.0, 0.0],
     [0.0, 0.0, 0.0]]

IndicesTensor: (Sizes:{2,3}, DataType:UINT32)
    [[1, 0, 2],
     [0, 2, 1]]

UpdatesTensor: (Sizes:{2,3}, DataType:FLOAT32)
    [[10, 11, 12],
     [20, 21, 22]]

// output = input
// output[indices[y, x], x] = updates[y, x]
OutputTensor: (Sizes:{3,3}, DataType:FLOAT32)
    [[20, 11,  0],
     [10,  0, 22],
     [ 0, 21, 12]]

Remarks

DML_SCATTER_OPERATOR_DESC has been more properly aliased to the name DML_SCATTER_ELEMENTS_OPERATOR_DESC as the proper counterpart to DML_GATHER_ELEMENTS_OPERATOR_DESC. This is because DML_SCATTER_OPERATOR_DESC was not really symmetric to DML_GATHER_OPERATOR_DESC.

Availability

This operator was introduced in DML_FEATURE_LEVEL_1_0.

Tensor constraints

• IndicesTensor, InputTensor, OutputTensor, and UpdatesTensor must have the same DimensionCount.
• InputTensor, OutputTensor, and UpdatesTensor must have the same DataType.

Tensor support

DML_FEATURE_LEVEL_4_1 and above

DML_FEATURE_LEVEL_3_0 and above

DML_FEATURE_LEVEL_2_1 and above

DML_FEATURE_LEVEL_1_0 and above

Requirements