Visible to Intel only — GUID: GUID-757BE972-B551-445B-B31E-109FC3382E49
Convolution
Inner Product
Matrix Multiplication
RNN
Batch Normalization
Binary
General
Execution Arguments
Implementation Details
Implementation Limitations
Performance Tips
Examples
Concat
Eltwise
Layer Normalization
Local Response Normalization (LRN)
Pooling
PReLU
Resampling
Shuffle
Softmax
Sum
Reorder
Reduction
Abs
AbsBackward
Add
AvgPool
AvgPoolBackward
BatchNormForwardTraining
BatchNormInference
BatchNormTrainingBackward
BiasAdd
BiasAddBackward
Clamp
ClampBackward
Concat
Convolution
ConvolutionBackwardData
ConvolutionBackwardWeights
ConvTranspose
ConvTransposeBackwardData
ConvTransposeBackwardWeights
Dequantize
Divide
DynamicDequantize
DynamicQuantize
Elu
EluBackward
End
Exp
GELU
GELUBackward
HardSwish
HardSwishBackward
Interpolate
InterpolateBackward
LayerNorm
LayerNormBackward
LeakyReLU
Log
LogSoftmax
LogSoftmaxBackward
MatMul
Maximum
MaxPool
MaxPoolBackward
Minimum
Mish
MishBackward
Multiply
PReLU
PReLUBackward
Quantize
Reciprocal
ReduceL1
ReduceL2
ReduceMax
ReduceMean
ReduceMin
ReduceProd
ReduceSum
ReLU
ReLUBackward
Reorder
Round
Sigmoid
SigmoidBackward
Softmax
SoftmaxBackward
SoftPlus
SoftPlusBackward
Sqrt
SqrtBackward
Square
SquaredDifference
StaticReshape
StaticTranspose
Subtract
Tanh
TanhBackward
TypeCast
Wildcard
enum dnnl_alg_kind_t
enum dnnl_normalization_flags_t
enum dnnl_primitive_kind_t
enum dnnl_prop_kind_t
enum dnnl_query_t
enum dnnl::normalization_flags
enum dnnl::query
struct dnnl_exec_arg_t
struct dnnl_primitive
struct dnnl_primitive_desc
struct dnnl::primitive
struct dnnl::primitive_desc
struct dnnl::primitive_desc_base
enum dnnl_rnn_direction_t
enum dnnl_rnn_flags_t
enum dnnl::rnn_direction
enum dnnl::rnn_flags
struct dnnl::augru_backward
struct dnnl::augru_forward
struct dnnl::gru_backward
struct dnnl::gru_forward
struct dnnl::lbr_augru_backward
struct dnnl::lbr_augru_forward
struct dnnl::lbr_gru_backward
struct dnnl::lbr_gru_forward
struct dnnl::lstm_backward
struct dnnl::lstm_forward
struct dnnl::rnn_primitive_desc_base
struct dnnl::vanilla_rnn_backward
struct dnnl::vanilla_rnn_forward
Visible to Intel only — GUID: GUID-757BE972-B551-445B-B31E-109FC3382E49
Binary
General
The binary primitive computes the result of a binary elementwise operation between tensors source 0 and source 1 (the variable names follow the standard Naming Conventions):
where 
is one of addition, subtraction, multiplication, division, greater than or equal to, greater than, less than or equal to, less than, equal to, not equal to, get maximum value, and get minimum value.
The binary 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 |
|---|---|
|
DNNL_ARG_SRC_0 |
|
DNNL_ARG_SRC_1 |
|
DNNL_ARG_DST |
|
DNNL_ARG_ATTR_MULTIPLE_POST_OP(binary_post_op_position) | DNNL_ARG_SRC_1 |
|
DNNL_ARG_ATTR_SCALES | DNNL_ARG_SRC_0 |
|
DNNL_ARG_ATTR_SCALES | DNNL_ARG_SRC_1 |
Implementation Details
General Notes
The binary primitive requires all source and destination tensors to have the same number of dimensions.
The binary primitive supports implicit broadcast semantics for source 0 and source 1. This means that if a dimension size is one, that single value will be broadcast (used to compute an operation with each point of the other source) for that dimension. It is recommended to use broadcast for source 1 to get better performance. Generally it should match the syntax below: {N,1}x{C,1}x{D,1}x{H,1}x{W,1}:{N,1}x{C,1}x{D,1}x{H,1}x{W,1} -> NxCxDxHxW. It is consistent with PyTorch broadcast semantic.
The dimensions of both sources must match unless either is equal to one.
- and memory formats 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 0 tensor. The tensor dimensions must match the ones of the source 0 and source 1 tensors (except for broadcast dimensions).
The binary primitive supports in-place operations, meaning that source 0 tensor may be used as the destination, in which case its data will be overwritten. In-place mode requires the
and source 0 data types to be the same. Different data types will unavoidably lead to correctness issues.
Post-Ops and Attributes
The following attributes are supported:
Type |
Operation |
Description |
Restrictions |
|---|---|---|---|
Attribute |
Scales the corresponding input tensor by the given scale factor(s). |
Only one scale per tensor is supported. Input tensors only. |
|
Post-op |
Adds the operation result to the destination tensor instead of overwriting it. |
||
Post-op |
Applies an Eltwise operation to the result. |
||
Post-op |
Applies a Binary operation to the result |
General binary post-op restrictions |
Data Types Support
The source and destination tensors may have f32, bf16, f16 or s8/u8 data types. The binary primitive supports the following combinations of data types:
Source 0 / 1 |
Destination |
|---|---|
f32, bf16, f16, u8, s8 |
f32, bf16, f16, u8, s8 |
WARNING:
There might be hardware and/or implementation specific restrictions. Check Implementation Limitations section below.
Data Representation
Sources, Destination
The binary primitive works with arbitrary data tensors. There is no special meaning associated with any of tensors dimensions.
Implementation Limitations
Refer to Data Types for limitations related to data types support.
GPU
Only tensors of 6 or fewer dimensions are supported.
Performance Tips
Whenever possible, avoid specifying different memory formats for source tensors.
Examples
Binary Primitive Example
This C++ API example demonstrates how to create and execute a Binary primitive.
Key optimizations included in this example:
In-place primitive execution;
Primitive attributes with fused post-ops.
Bnorm u8 by Binary Post-Ops Example
Bnorm u8 via binary postops example.