Reduction
General
The reduction primitive performs reduction operation on arbitrary data. Each element in the destination is the result of reduction operation with specified algorithm along one or multiple source tensor dimensions:
where 
can be max, min, sum, mul, mean, Lp-norm and Lp-norm-power-p, 
is an index in an idle dimension and 
is an index in a reduction dimension.
can be max, min, sum, mul, mean, Lp-norm and Lp-norm-power-p, is an index in an idle dimension and is an index in a reduction dimension.Mean:
where 
is the size of a reduction dimension.
is the size of a reduction dimension.Lp-norm:
where 
can be max and sum.
can be max and sum.Lp-norm-power-p:
where can be max and sum.
can be max and sum.Notes
- The reduction primitive requires the source and destination tensors to have the same number of dimensions.
- Reduction dimensions are of size 1 in a destination tensor.
- The reduction primitive does not have a notion of forward or backward propagations.
Execution Arguments
When executed, the inputs and outputs should be mapped to an execution argument index as specified by the following table.
Primitive input/output | Execution argument index |
|---|
Implementation Details
General Notes
- The
memory format can be either specified explicitly or by dnnl::memory::format_tag::any (recommended), in which case the primitive will derive the most appropriate memory format based on the format of the source tensor.
Post-Ops and Attributes
The following attributes are supported:
Data Types Support
The source and destination tensors may have
f32
, bf16
, or int8
data types. See Data Types page for more details.Data Representation
Sources, Destination
The reduction primitive works with arbitrary data tensors. There is no special meaning associated with any of the dimensions of a tensor.
Performance Tips
- Whenever possible, avoid specifying different memory formats for source and destination tensors.