Edit

Share via

DML_GEMM_OPERATOR_DESC structure (directml.h)

  • Article

Performs a general matrix multiplication function of the form Output = FusedActivation(Alpha * TransA(A) x TransB(B) + Beta * C), where x denotes matrix multiplication, and * denotes multiplication with a scalar.

This operator requires 4D tensors with layout { BatchCount, ChannelCount, Height, Width }, and it will perform BatchCount * ChannelCount number of independent matrix multiplications.

For example, if ATensor has Sizes of { BatchCount, ChannelCount, M, K }, and BTensor has Sizes of { BatchCount, ChannelCount, K, N }, and OutputTensor has Sizes of { BatchCount, ChannelCount, M, N }, then this operator performs BatchCount * ChannelCount independent matrix multiplications of dimensions {M,K} x {K,N} = {M,N}.

Syntax

struct DML_GEMM_OPERATOR_DESC {
  const DML_TENSOR_DESC   *ATensor;
  const DML_TENSOR_DESC   *BTensor;
  const DML_TENSOR_DESC   *CTensor;
  const DML_TENSOR_DESC   *OutputTensor;
  DML_MATRIX_TRANSFORM    TransA;
  DML_MATRIX_TRANSFORM    TransB;
  FLOAT                   Alpha;
  FLOAT                   Beta;
  const DML_OPERATOR_DESC *FusedActivation;
};

Members

ATensor

Type: const DML_TENSOR_DESC*

A tensor containing the A matrix. This tensor's Sizes should be { BatchCount, ChannelCount, M, K } if TransA is DML_MATRIX_TRANSFORM_NONE, or { BatchCount, ChannelCount, K, M } if TransA is DML_MATRIX_TRANSFORM_TRANSPOSE.

BTensor

Type: const DML_TENSOR_DESC*

A tensor containing the B matrix. This tensor's Sizes should be { BatchCount, ChannelCount, K, N } if TransB is DML_MATRIX_TRANSFORM_NONE, or { BatchCount, ChannelCount, N, K } if TransB is DML_MATRIX_TRANSFORM_TRANSPOSE.

CTensor

Type: _Maybenull_ const DML_TENSOR_DESC*

A tensor containing the C matrix, or nullptr. Values default to 0 when not provided. If provided, this tensor's Sizes should be { BatchCount, ChannelCount, M, N }.

OutputTensor

Type: const DML_TENSOR_DESC*

The tensor to write the results to. This tensor's Sizes are { BatchCount, ChannelCount, M, N }.

TransA

Type: DML_MATRIX_TRANSFORM

The transform to be applied to ATensor; either a transpose, or no transform.

TransB

Type: DML_MATRIX_TRANSFORM

The transform to be applied to BTensor; either a transpose, or no transform.

Alpha

Type: FLOAT

The value of the scalar multiplier for the product of inputs ATensor and BTensor.

Beta

Type: FLOAT

The value of the scalar multiplier for the optional input CTensor. If CTensor is not provided, then this value is ignored.

FusedActivation

Type: _Maybenull_ const DML_OPERATOR_DESC*

An optional fused activation layer to apply after the GEMM. For more info, see Using fused operators for improved performance.

Availability

This operator was introduced in DML_FEATURE_LEVEL_1_0.

Tensor constraints

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

Tensor support

DML_FEATURE_LEVEL_4_0 and above

Tensor Kind Dimensions Supported dimension counts Supported data types
ATensor Input { [BatchCount], [ChannelCount], M, K } 2 to 4 FLOAT32, FLOAT16
BTensor Input { [BatchCount], [ChannelCount], K, N } 2 to 4 FLOAT32, FLOAT16
CTensor Optional input { [BatchCount], [ChannelCount], M, N } 2 to 4 FLOAT32, FLOAT16
OutputTensor Output { [BatchCount], [ChannelCount], M, N } 2 to 4 FLOAT32, FLOAT16

DML_FEATURE_LEVEL_1_0 and above

Tensor Kind Dimensions Supported dimension counts Supported data types
ATensor Input { BatchCount, ChannelCount, M, K } 4 FLOAT32, FLOAT16
BTensor Input { BatchCount, ChannelCount, K, N } 4 FLOAT32, FLOAT16
CTensor Optional input { BatchCount, ChannelCount, M, N } 4 FLOAT32, FLOAT16
OutputTensor Output { BatchCount, ChannelCount, M, N } 4 FLOAT32, FLOAT16

Requirements

Requirement Value
Header directml.h

See also