dnnl::stream Struct Reference

An execution stream. More...

#include <dnnl.hpp>

Inheritance diagram for dnnl::stream:

Collaboration diagram for dnnl::stream:

Public Types
enum class	flags : unsigned
	Stream flags. Can be combined using the bitwise OR operator. More...

Public Member Functions
	stream ()=default
	Constructs an empty stream.
	stream (const engine &engine, flags flags=flags::default_flags, const stream_attr &attr=stream_attr())
	Constructs a stream for the specified engine and with behavior controlled by the specified flags.
	stream (const engine &engine, cl_command_queue queue)
	Constructs a stream for the specified engine and the OpenCL queue.
cl_command_queue	get_ocl_command_queue () const
	Returns the underlying OpenCL queue object.
stream &	wait ()
	Waits for all primitives executing in the stream to finish.
	handle ()=default
	Constructs an empty handle object.
	handle (const handle< T, traits > &)=default
	Copy constructor.
	handle (handle< T, traits > &&)=default
	Move constructor.
	handle (T t, bool weak=false)
	Constructs a handle wrapper object from a C API handle.
Public Member Functions inherited from dnnl::handle< dnnl_stream_t >
	handle ()=default
	Constructs an empty handle object.
handle< dnnl_stream_t, handle_traits< dnnl_stream_t > > &	operator= (const handle< dnnl_stream_t, handle_traits< dnnl_stream_t > > &)=default
	Assignment operator.
void	reset (dnnl_stream_t t, bool weak=false)
	Resets the handle wrapper objects to wrap a new C API handle.
dnnl_stream_t	get (bool allow_empty=false) const
	Returns the underlying C API handle.
	operator dnnl_stream_t () const
	Converts a handle to the underlying C API handle type.
	operator bool () const
	Checks whether the object is empty.

Detailed Description

An execution stream.

Examples

batch_normalization.cpp, binary.cpp, cnn_inference_f32.cpp, cnn_inference_int8.cpp, cnn_training_f32.cpp, concat.cpp, convolution.cpp, cpu_cnn_training_bf16.cpp, cpu_matmul_quantization.cpp, cpu_rnn_inference_f32.cpp, cpu_rnn_inference_int8.cpp, cpu_sgemm_and_matmul.cpp, cross_engine_reorder.cpp, eltwise.cpp, getting_started.cpp, gpu_opencl_interop.cpp, inference_int8_matmul.cpp, inner_product.cpp, layer_normalization.cpp, logsoftmax.cpp, lrn.cpp, lstm.cpp, matmul.cpp, memory_format_propagation.cpp, performance_profiling.cpp, pooling.cpp, reorder.cpp, resampling.cpp, rnn_training_f32.cpp, shuffle.cpp, softmax.cpp, and sum.cpp.

Member Enumeration Documentation

flags

enum class dnnl::stream::flags : unsigned

strong

Stream flags. Can be combined using the bitwise OR operator.

Enumerator
default_order	Default order execution. Either in-order or out-of-order depending on the engine runtime.
in_order	In-order execution.
out_of_order	Out-of-order execution.
default_flags	Default stream configuration.

Constructor & Destructor Documentation

stream() [1/3]

dnnl::stream::stream ( )

default

Constructs an empty stream.

An empty stream cannot be used in any operations.

stream() [2/3]

dnnl::stream::stream ( const engine & engine, flags flags = flags::default_flags, const stream_attr & attr = stream_attr() )

inline

Constructs a stream for the specified engine and with behavior controlled by the specified flags.

Parameters

engine	Engine to create the stream on.
flags	Flags controlling stream behavior.
attr	Stream attributes.

stream() [3/3]

dnnl::stream::stream ( const engine & engine, cl_command_queue queue )

inline

Constructs a stream for the specified engine and the OpenCL queue.

Parameters

engine	Engine to create the stream on.
queue	OpenCL queue to use for the stream.

Member Function Documentation

get_ocl_command_queue()

cl_command_queue dnnl::stream::get_ocl_command_queue ( ) const

inline

Returns the underlying OpenCL queue object.

Returns

OpenCL queue.

wait()

stream & dnnl::stream::wait ( )

inline

Waits for all primitives executing in the stream to finish.

Returns

The stream itself.

Examples

batch_normalization.cpp, binary.cpp, cnn_inference_f32.cpp, cnn_inference_int8.cpp, cnn_training_f32.cpp, concat.cpp, convolution.cpp, cpu_cnn_training_bf16.cpp, eltwise.cpp, getting_started.cpp, inner_product.cpp, layer_normalization.cpp, logsoftmax.cpp, lrn.cpp, lstm.cpp, matmul.cpp, memory_format_propagation.cpp, pooling.cpp, reorder.cpp, resampling.cpp, rnn_training_f32.cpp, shuffle.cpp, softmax.cpp, and sum.cpp.

handle() [1/2]

dnnl::handle< T, traits >::handle ( )

default

Constructs an empty handle object.

Warning

Uninitialized object cannot be used in most library calls and is equivalent to a null pointer. Any attempt to use its methods, or passing it to the other library function, will cause an exception to be thrown.

handle() [2/2]

dnnl::handle< T, traits >::handle ( T t, bool weak = false )

inlineexplicit

Constructs a handle wrapper object from a C API handle.

Parameters

t	The C API handle to wrap.
weak	A flag specifying whether to construct a weak wrapper; defaults to false.

---

The documentation for this struct was generated from the following file:

• include/dnnl.hpp