Base node for computation. More...

#include "abstract_task.h"

Inheritance diagram for hh::AbstractTask< TaskOutput, TaskInputs >:

Collaboration diagram for hh::AbstractTask< TaskOutput, TaskInputs >:

Public Member Functions
	AbstractTask ()
	AbstractTask default constructor. More...

	AbstractTask (std::string_view const &name)
	Create a task with a custom name. More...

	AbstractTask (std::string_view const &name, size_t numberThreads)
	Create a task with a custom name and the task's cluster size. More...

	AbstractTask (std::string_view const &name, size_t numberThreads, bool automaticStart)
	Create a task with a custom name, the task's cluster size, and the automatic start. More...

	AbstractTask (std::string_view const name, size_t numberThreads, core::NodeType nodeType, bool automaticStart)
	Internal constructor needed to create nodes that derive from AbstractTask. More...

	AbstractTask (AbstractTask< TaskOutput, TaskInputs ... > *rhs)
	Copy constructor. More...

virtual	~AbstractTask ()=default
	Default destructor.

void	addResult (std::shared_ptr< TaskOutput > output)
	Add an output data. More...

std::string_view	name ()
	Task's name accessor. More...

size_t	numberThreads ()
	Task's number of threads accessor. More...

bool	automaticStart ()
	Task's automatic start accessor. More...

core::NodeType	nodeType ()
	Task's node type accessor. More...

int	deviceId ()
	Task's device ID accessor. More...

int	graphId ()
	Task's graph ID accessor. More...

std::shared_ptr< core::CoreNode >	core () final
	Task's core accessor. More...

std::shared_ptr< AbstractMemoryManager< TaskOutput > > const &	memoryManager () const
	Task's memory manager accessor. More...

virtual std::shared_ptr< AbstractTask< TaskOutput, TaskInputs... > >	copy ()
	Default copy overload, fail if cluster need to be copied. More...

virtual void	initialize ()
	Initialize method called before AbstractTask::Execute loop.

virtual void	shutdown ()
	Shutdown method called after AbstractTask::Execute loop, when AbstractTask::canTerminate evaluates to true.

void	connectMemoryManager (std::shared_ptr< AbstractMemoryManager< TaskOutput >> mm)
	Connect a memory manager to the task. More...

std::shared_ptr< TaskOutput >	getManagedMemory ()
	Memory manager accessor. More...

bool	canTerminate () override
	Method called to test the task;s termination, by default, the test is: no input nodes connected and, no data waiting to be treated. More...

Public Member Functions inherited from hh::behavior::Node
virtual std::string	extraPrintingInformation () const
	Adds node information to print in the dot file. More...

Public Member Functions inherited from hh::behavior::Execute< TaskInputs >
virtual void	execute (std::shared_ptr< TaskInputs >)=0
	Virtual declaration of execute function for a data of type Input.

Private Attributes
std::shared_ptr< core::CoreTask< TaskOutput, TaskInputs... > >	taskCore_ = nullptr
	Task Core.

std::shared_ptr< AbstractMemoryManager< TaskOutput > >	mm_ = nullptr
	Task's memory manager.

Additional Inherited Members
Public Types inherited from hh::behavior::MultiReceivers< TaskInputs... >
using	inputs_t = std::tuple< Inputs... >
	Tuple with the list of input types.

Public Types inherited from hh::behavior::Sender< TaskOutput >
using	output_t = TaskOutput
	Output Type.

Detailed Description

template<class TaskOutput, class ... TaskInputs>
class hh::AbstractTask< TaskOutput, TaskInputs >

Base node for computation.

Hedgehog Graph's node made for holding a computation kernel into their execute overload method from Execute::execute.

An AbstractTask can be bound to multiple threads, forming a cluster of tasks. The base AbstractTask will be copied n times and each of them will be bound to a thread. The AbstractTask::copy method must be overloaded to use this functionality. Also, if the AbstractTask is part of Graph that will be duplicated with an AbstractExecutionPipeline, AbstractTask::copy method needs to be overloaded.

An AbstractMemoryManager could be linked to a task through a graph (from tutorial 4, example shows Cuda memory usage):

auto productTask = std::make_shared<CudaProductTask<MatrixType>>(p, numberThreadProduct);
auto cudaMemoryManagerA = std::make_shared<CudaMemoryManager<MatrixType, 'a'>>(nBlocks + 4, blockSize);
productTask->connectMemoryManager(cudaMemoryManagerProduct);

Linking an AbstractMemoryManager requires that the AbstractTask's TaskOutput match the template type of the AbstractMemoryManager

The default order of execution is:

The cluster is created, for each task:
The thread is spawned,
The AbstractTask::initialize is called,
The memory manager, if it exists, is bound to the task (the device Id is shared, and is initialized),
Execute is called, while AbstractTask::canTerminate is True:
- If the task is set to start automatically Execute::execute is called with nullptr,
- If not, the task will wait for a data to come and Execute::execute is called with the received data,
AbstractTask::shutdown is called, and is signal to the linked nodes.

Only Execute::execute method need to be overloaded for every AbstractTask input types.

Virtual functions: Execute::execute (one for each of TaskInputs)
AbstractTask::initialize
AbstractTask::shutdown
AbstractTask::copy (mandatory if usage of clusters or AbstractExecutionPipeline)
Node::canTerminate (mandatory if cycle in the graph)
Node::extraPrintingInformation

Attention: In case of cycle in the graph AbstractTask::canTerminate needs to be overloaded or the graph will deadlock. By default, AbstractTask::canTerminate will be true if there is no "input node" connected AND no data available in the task input queue (CF tutorial 3).

Template Parameters

TaskOutput	Output task data type.
TaskInputs	Inputs task data type.

Definition at line 76 of file abstract_task.h.

Constructor & Destructor Documentation

◆ AbstractTask() [1/6]

template<class TaskOutput, class ... TaskInputs>

hh::AbstractTask< TaskOutput, TaskInputs >::AbstractTask ( )

inline

AbstractTask default constructor.

Create a task with the name "Task", one task in the cluster (no copy overload needed), and no automatic start.

Definition at line 92 of file abstract_task.h.

◆ AbstractTask() [2/6]

template<class TaskOutput, class ... TaskInputs>

hh::AbstractTask< TaskOutput, TaskInputs >::AbstractTask ( std::string_view const & name )

inlineexplicit

Create a task with a custom name.

Create a task with one task in the cluster (no copy overload needed), and no automatic start.

Parameters

name Task name

Definition at line 103 of file abstract_task.h.

◆ AbstractTask() [3/6]

template<class TaskOutput, class ... TaskInputs>

hh::AbstractTask< TaskOutput, TaskInputs >::AbstractTask	(	std::string_view const &	name,
		size_t	numberThreads
	)

inlineexplicit

Create a task with a custom name and the task's cluster size.

Parameters

name	Task name
numberThreads	Task's cluster size

Definition at line 111 of file abstract_task.h.

◆ AbstractTask() [4/6]

template<class TaskOutput, class ... TaskInputs>

hh::AbstractTask< TaskOutput, TaskInputs >::AbstractTask	(	std::string_view const &	name,
		size_t	numberThreads,
		bool	automaticStart
	)

inlineexplicit

Create a task with a custom name, the task's cluster size, and the automatic start.

Parameters

name	Task name
numberThreads	Task's cluster size
automaticStart	Task's automatic start

Definition at line 124 of file abstract_task.h.

◆ AbstractTask() [5/6]

template<class TaskOutput, class ... TaskInputs>

hh::AbstractTask< TaskOutput, TaskInputs >::AbstractTask	(	std::string_view const	name,
		size_t	numberThreads,
		core::NodeType	nodeType,
		bool	automaticStart
	)

inline

Internal constructor needed to create nodes that derive from AbstractTask.

Attention: Do not use

Parameters

name	Task name
numberThreads	Task's cluster size
nodeType	Task node type
automaticStart	Task's automatic start

Definition at line 136 of file abstract_task.h.

◆ AbstractTask() [6/6]

template<class TaskOutput, class ... TaskInputs>

hh::AbstractTask< TaskOutput, TaskInputs >::AbstractTask ( AbstractTask< TaskOutput, TaskInputs ... > * rhs )

inlineexplicit

Copy constructor.

Parameters

rhs	Task to copy

Definition at line 143 of file abstract_task.h.

Member Function Documentation

◆ addResult()

template<class TaskOutput, class ... TaskInputs>

void hh::AbstractTask< TaskOutput, TaskInputs >::addResult ( std::shared_ptr< TaskOutput > output )

inline

Add an output data.

Parameters

output Data to send to the output nodes

Definition at line 157 of file abstract_task.h.

◆ automaticStart()

template<class TaskOutput, class ... TaskInputs>

bool hh::AbstractTask< TaskOutput, TaskInputs >::automaticStart ( )

inline

Task's automatic start accessor.

Returns: Task automatic start property

Definition at line 169 of file abstract_task.h.

Here is the caller graph for this function:

◆ canTerminate()

template<class TaskOutput, class ... TaskInputs>

bool hh::AbstractTask< TaskOutput, TaskInputs >::canTerminate ( )

inlineoverridevirtual

Method called to test the task;s termination, by default, the test is: no input nodes connected and, no data waiting to be treated.

Returns: True if the task can be terminated, else False

Reimplemented from hh::behavior::Node.

Definition at line 236 of file abstract_task.h.

◆ connectMemoryManager()

template<class TaskOutput, class ... TaskInputs>

void hh::AbstractTask< TaskOutput, TaskInputs >::connectMemoryManager ( std::shared_ptr< AbstractMemoryManager< TaskOutput >> mm )

inline

Connect a memory manager to the task.

Parameters

mm	Memory manager to connect

Definition at line 203 of file abstract_task.h.

◆ copy()

template<class TaskOutput, class ... TaskInputs>

virtual std::shared_ptr<AbstractTask<TaskOutput, TaskInputs...> > hh::AbstractTask< TaskOutput, TaskInputs >::copy ( )

inlinevirtual

Default copy overload, fail if cluster need to be copied.

Returns: A copy of the current AbstractTask

Reimplemented in hh::StateManager< StateOutput, StateInputs >.

Definition at line 193 of file abstract_task.h.

Here is the caller graph for this function:

◆ core()

template<class TaskOutput, class ... TaskInputs>

std::shared_ptr<core::CoreNode> hh::AbstractTask< TaskOutput, TaskInputs >::core ( )

inlinefinalvirtual

Task's core accessor.

Returns: Task's core

Implements hh::behavior::Node.

Definition at line 185 of file abstract_task.h.

Here is the caller graph for this function:

◆ deviceId()

template<class TaskOutput, class ... TaskInputs>

int hh::AbstractTask< TaskOutput, TaskInputs >::deviceId ( )

inline

Task's device ID accessor.

Returns: Task's device ID

Definition at line 177 of file abstract_task.h.

◆ getManagedMemory()

template<class TaskOutput, class ... TaskInputs>

std::shared_ptr<TaskOutput> hh::AbstractTask< TaskOutput, TaskInputs >::getManagedMemory ( )

inline

Memory manager accessor.

Returns: Connected memory manager

Definition at line 212 of file abstract_task.h.

◆ graphId()

template<class TaskOutput, class ... TaskInputs>

int hh::AbstractTask< TaskOutput, TaskInputs >::graphId ( )

inline

Task's graph ID accessor.

Returns: Task's graph ID

Definition at line 181 of file abstract_task.h.

◆ memoryManager()

template<class TaskOutput, class ... TaskInputs>

std::shared_ptr<AbstractMemoryManager<TaskOutput> > const& hh::AbstractTask< TaskOutput, TaskInputs >::memoryManager ( ) const

inline

Task's memory manager accessor.

Returns: Task's memory manager

Definition at line 189 of file abstract_task.h.

Here is the caller graph for this function:

◆ name()

template<class TaskOutput, class ... TaskInputs>

std::string_view hh::AbstractTask< TaskOutput, TaskInputs >::name ( )

inline

Task's name accessor.

Returns: Task name

Definition at line 161 of file abstract_task.h.

Here is the caller graph for this function:

◆ nodeType()

template<class TaskOutput, class ... TaskInputs>

core::NodeType hh::AbstractTask< TaskOutput, TaskInputs >::nodeType ( )

inline

Task's node type accessor.

Returns: Task's node type property

Definition at line 173 of file abstract_task.h.

Here is the caller graph for this function:

◆ numberThreads()

template<class TaskOutput, class ... TaskInputs>

size_t hh::AbstractTask< TaskOutput, TaskInputs >::numberThreads ( )

inline

Task's number of threads accessor.

Returns: Task number of threads

Definition at line 165 of file abstract_task.h.

Here is the caller graph for this function:

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

/home/tjb3/work/hedgehog-nist/hedgehog/api/abstract_task.h

Public Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

template<class TaskOutput, class ... TaskInputs> class hh::AbstractTask< TaskOutput, TaskInputs >

Constructor & Destructor Documentation

◆ AbstractTask() [1/6]

◆ AbstractTask() [2/6]

◆ AbstractTask() [3/6]

◆ AbstractTask() [4/6]

◆ AbstractTask() [5/6]

◆ AbstractTask() [6/6]

Member Function Documentation

◆ addResult()

◆ automaticStart()

◆ canTerminate()

◆ connectMemoryManager()

◆ copy()

◆ core()

◆ deviceId()

◆ getManagedMemory()

◆ graphId()

◆ memoryManager()

◆ name()

◆ nodeType()

◆ numberThreads()

template<class TaskOutput, class ... TaskInputs>
class hh::AbstractTask< TaskOutput, TaskInputs >