The ExecutionPipeline class is used to duplicate task graphs, such that each duplicate executes concurrently. More...

#include <htgs/api/ExecutionPipeline.hpp>

Inheritance diagram for htgs::ExecutionPipeline< T, U >:

Collaboration diagram for htgs::ExecutionPipeline< T, U >:

Public Member Functions
	ExecutionPipeline (size_t numPipelines, TaskGraphConf< T, U > *graph, std::string name="Execution Pipeline", bool waitForInit=true)
	Creates an execution pipeline, which encapsulates a graph and duplicates it numPipelines times. More...

	ExecutionPipeline (size_t numPipelines, TaskGraphConf< T, U > *graph, std::shared_ptr< IRuleList< T, T >> rules, std::string name="Execution Pipeline", bool waitForInit=true)
	Creates an execution pipeline, which encapsulates a graph and duplicates it numPipelines times. More...

	~ExecutionPipeline ()
	Destructor that frees all memory allocated by the execution pipeline.

size_t	getNumGraphsSpawned () override
	Gets the number of graphs spawned by this ITask. More...

void	addInputRule (IRule< T, T > *rule)
	Adds an input rule, which can be used for domain decomposition. More...

void	addInputRule (std::shared_ptr< IRule< T, T >> rule)
	Adds an input rule, which can be used for domain decomposition. More...

void	initialize ()
	Initializes the execution pipeline and duplicates the task graph based on the number of pipelines. More...

void	shutdown ()
	Shuts down the execution pipeline. More...

void	shutdownParallel (int id)
	Waits for all runtimes in parallel to update overall runtime of each graph. More...

void	executeTask (std::shared_ptr< T > data)
	Executes the execution pipeline task on data and forwards that data to the input rules. More...

std::string	getName ()
	Gets the name for the execution pipeline. More...

ITask< T, U > *	copy ()
	Makes a copy of the execution pipeline. More...

void	debug ()
	Provides debugging output for the execution pipeline. More...

virtual void	gatherProfileData (std::map< AnyTaskManager , TaskManagerProfile > *taskManagerProfiles) override
	Gathers profile data. More...

void	printProfile () override
	Prints the profile data to std::out. More...

std::string	genCustomDot (ProfileUtils *profileUtils, int colorFlag) override
	Virtual function to generate customized dot file. More...

virtual std::string	genDotProducerEdgeToTask (std::map< std::shared_ptr< AnyConnector >, AnyITask *> &inputConnectorDotMap, int dotFlags) override

virtual std::string	genDotProducerEdgeFromConnector (std::shared_ptr< AnyConnector > connector, int flags)

virtual std::string	genDotConsumerEdgeFromConnector (std::shared_ptr< AnyConnector > connector, int flags) override

std::string	getConsumerDotIds () override

std::string	genDot (int flags, std::string dotId, std::shared_ptr< AnyConnector > input, std::shared_ptr< AnyConnector > output) override
	Virtual function that generates the input/output and per-task dot notation. More...

Public Member Functions inherited from htgs::ITask< T, U >
	ITask ()
	Creates an ITask with number of threads equal to 1.

	ITask (size_t numThreads)
	Constructs an ITask with a specified number of threads. More...

	ITask (size_t numThreads, bool isStartTask, bool poll, size_t microTimeoutTime)
	Constructs an ITask with a specified number of threads as well as additional scheduling options. More...

virtual bool	canTerminate (std::shared_ptr< AnyConnector > inputConnector) override
	Virtual function that is called when an ITask is checking if it can be terminated. More...

virtual void	executeTaskFinal () override
	Virtual function that is called just before the task has shutdown. More...

virtual std::string	getDotLabelName () override
	Virtual function to get the label name used for dot graph viz. More...

virtual std::string	getDotShapeColor () override
	Gets the color of the shape for graphviz dot. More...

virtual std::string	getDotFillColor () override
	Gets the color for filling the shape for graphviz dot. More...

virtual std::string	getDotShape () override
	Gets the shape for graphviz dot. More...

virtual std::string	getDotCustomProfile () override
	Adds the string text to the profiling of this task in the graphviz dot visualization. More...

ITask< T, U > *	copyITask (bool deep) override
	Copies the ITask (including a copy of all memory edges) More...

void	addResult (std::shared_ptr< U > result)
	Adds results to the output list to be sent to the next connected ITask in a TaskGraph. More...

void	addResult (U *result)
	Adds results to the output list to be sent to the next connected ITask in a TaskGraph. More...

void	initialize (size_t pipelineId, size_t numPipeline, TaskManager< T, U > *ownerTask)
	Function that is called when an ITask is being initialized by it's owner thread. More...

template<class V >
m_data_t< V >	getMemory (std::string name, IMemoryReleaseRule *releaseRule)
	Retrieves memory from a memory edge. More...

template<class V >
m_data_t< V >	getDynamicMemory (std::string name, IMemoryReleaseRule *releaseRule, size_t numElems)
	Retrieves memory from a memory edge. More...

template<class V >
void	releaseMemory (m_data_t< V > memory)
	Releases memory onto a memory edge, which is transferred by the graph communicator. More...

void	resetProfile ()
	Resets profile data.

size_t	getThreadID ()
	Gets the thread ID associated with this task. More...

unsigned long long int	getTaskComputeTime () const
	Gets the task's compute time. More...

std::string	inTypeName () override final
	Gets the demangled input type name of the connector. More...

std::string	outTypeName () override final
	Gets the demangled output type name of the connector. More...

std::string	getAddress () override final
	Gets the address from the owner task, which is the address of the task graph. More...

void	setTaskManager (TaskManager< T, U > *ownerTask)
	Sets the owner task manager for this ITask. More...

TaskManager< T, U > *	getOwnerTaskManager ()
	Gets the owner task manager for this ITask. More...

Public Member Functions inherited from htgs::AnyITask
	AnyITask ()
	Creates an ITask with number of threads equal to 1.

	AnyITask (size_t numThreads)
	Constructs an ITask with a specified number of threads. More...

	AnyITask (size_t numThreads, bool isStartTask, bool poll, size_t microTimeoutTime)
	Constructs an ITask with a specified number of threads as well as additional scheduling options. More...

virtual	~AnyITask ()
	Destructor.

virtual std::string	genDot (int flags, std::string dotId, std::shared_ptr< htgs::AnyConnector > input, std::shared_ptr< htgs::AnyConnector > output)
	Virtual function that generates the input/output and per-task dot notation. More...

virtual std::string	getProducerDotIds ()

virtual std::string	genDot (int flags, std::string dotId)
	Virtual function that adds additional dot attributes to this node. More...

virtual std::string	debugDotNode ()
	Provides debug output for a node in the dot graph. More...

virtual void	profile ()
	Virtual function that is called to provide profile output for the ITask. More...

virtual std::string	profileStr ()
	Virtual function that is called after executionTask is called. More...

void	initialize (size_t pipelineId, size_t numPipeline)
	Virtual function that is called when an ITask is being initialized by it's owner thread. More...

void	setPipelineId (size_t pipelineId)
	Sets the pipeline Id for this ITask. More...

size_t	getPipelineId ()
	Gets the pipeline ID. More...

void	setNumPipelines (size_t numPipelines)
	Sets the number of pipelines that this ITask belongs too. More...

size_t	getNumPipelines () const
	Sets the task graph communicator. More...

size_t	getNumThreads () const
	Gets the number of threads associated with this ITask. More...

bool	isStartTask () const
	Gets whether this ITask is a starting task. More...

bool	isPoll () const
	Gets whether this ITask is polling for data or not. More...

size_t	getMicroTimeoutTime () const
	Gets the timeout time for polling. More...

void	copyMemoryEdges (AnyITask *iTaskCopy)
	Copies the memory edges from this AnyITask to another AnyITask. More...

std::string	genDot (int flags, std::shared_ptr< AnyConnector > input, std::shared_ptr< AnyConnector > output)
	Creates a dot notation representation for this task. More...

void	profileITask ()
	Provides profile output for the ITask,. More...

std::string	getDotId ()
	Gets the id used for dot nodes. More...

std::string	getNameWithPipelineId ()
	Gets the name of the ITask with it's pipeline ID. More...

const std::shared_ptr< ConnectorMap > &	getMemoryEdges () const
	Gets the memory edges for the task. More...

const std::shared_ptr< ConnectorMap > &	getReleaseMemoryEdges () const
	Gets the memory edges for releasing memory for the memory manager, used to shutdown the memory manager. More...

bool	hasMemoryEdge (std::string name)
	Checks whether this ITask contains a memory edge for a specified name. More...

void	attachMemoryEdge (std::string name, std::shared_ptr< AnyConnector > getMemoryConnector, std::shared_ptr< AnyConnector > releaseMemoryConnector, MMType type)
	Attaches a memory edge to this ITask to get memory. More...

unsigned long long int	getMemoryWaitTime () const
	Gets the amount of time the task was waiting for memory. More...

void	incMemoryWaitTime (unsigned long long int val)
	Increments memory wait time. More...

Private Member Functions
std::string	cleanupVisualization (TaskGraphConf< T, U > *graph, std::string str)
	Moves the output connector outside of the execution pipeline graphs to cleanup how the graph looks during graph visualization. More...

Private Attributes
size_t	numPipelinesExec
	The number of pipelines that will spawn from the ExecutionPipeline.

Bookkeeper< T > *	inputBk
	The input Bookkeeper for the ExecutionPipeline.

TaskGraphConf< T, U > *	graph
	The graph that the ExecutionPipeline manages, duplicates, and executes.

std::shared_ptr< IRuleList< T, T > >	inputRules
	The rules associated with the input Bookkeeper for decomposition.

std::vector< TaskGraphRuntime * > *	runtimes
	The list of Runtimes that will execute the TaskGraphs (one for each duplicate TaskGraph)

std::vector< TaskGraphConf< T, U > * > *	graphs
	The list of duplicate TaskGraphs.

bool	waitForInit
	Flag whether to wait for initialization of sub-graphs to complete or not.

std::string	name
	The name given to the execution pipeline task.

Detailed Description

template<class T, class U>
class htgs::ExecutionPipeline< T, U >

The ExecutionPipeline class is used to duplicate task graphs, such that each duplicate executes concurrently.

Each task within the task graph is an exact duplicate of the original's. The only difference is the pipelineId passed to each task. The pipelineId indicates which task graph the task belongs too. This value can be used as a rank to distribute computation and determine if additional functionality is needed to process data (such as copying data from one GPU to another).

An ExecutionPipeline can be used to execute across multiple GPUs on a single system. Any ICudaTask will automatically be bound to a separate GPU based on the CUContext array passed to the ICudaTask. The size of the CUContext array should match the number of execution pipelines specified. If data exists between two GPUs, then additional copying may be required, see ICudaTask for details and example usage for copying data between GPUs.

The execution pipeline can be used to distribute data among each task graph by adding a rule that uses the pipelineId parameter in an IRule. See addInputRule(IRule <T, T> *rule)

If you wish to share a rule with multiple execution pipelines or bookkeepers, you must wrap the rule into a std::shared_ptr prior to calling the addInputRule function.

Example usage:

htgs::TaskGraphConf<MatrixData, MatrixData> *subGraph = new htgs::TaskGraphConf<MatrixData, MatrixData>();
... build subGraph
int numPipelines = 3;
htgs::ExecutionPipeline<MatrixData, MatrixData> *execPipeline = new htgs::ExecutionPipeline<MatrixData, MatrixData>(numPipelines, subGraph);
execPipeline->addInputRule(new DecompRule());
PostProcessTask *taskOutsideExecPipeline = new PostProcessTask();
htgs::TaskGraph<MatrixData, VoidData> *mainGraph = new htgs::TaskGraph<MatrixData, VoidData>();
// Declares that the execPipeline connects to the input of the mainGraph
mainGraph->setGraphConsumerTask(execPipeline);
mainGraph->addEdge(execPipeline, taskOutsideExecPipeline);
htgs::TaskGraphRuntime *runTime = new htgs::TaskGraphRuntime(mainGraph);
while(hasDataToAdd)
{
  // Add data to graph
  mainGraph->produceData(data);
}
mainGraph->finishProducingData();
// Executes the main graph (spawns sub graphs for execPipeline)
runTime->executeAndWaitForRunTime();

Template Parameters

T	the input type for the ExecutionPipeline ITask, must derive from IData.
U	the output type for the ExecutionPipeline ITask, must derive from IData.

Constructor & Destructor Documentation

◆ ExecutionPipeline() [1/2]

template<class T, class U>

htgs::ExecutionPipeline< T, U >::ExecutionPipeline	(	size_t	numPipelines,
		TaskGraphConf< T, U > *	graph,
		std::string	name = `"Execution Pipeline"`,
		bool	waitForInit = `true`
	)

inline

Creates an execution pipeline, which encapsulates a graph and duplicates it numPipelines times.

Parameters

numPipelines	the number of times to duplicate the graph
graph	the graph that the execution pipeline manages
name	the name of the execution pipeline
waitForInit	Indicates that the execution pipeline will wait for its underlying graphs to initialize prior to marking that it is initialized.

◆ ExecutionPipeline() [2/2]

template<class T, class U>

htgs::ExecutionPipeline< T, U >::ExecutionPipeline	(	size_t	numPipelines,
		TaskGraphConf< T, U > *	graph,
		std::shared_ptr< IRuleList< T, T >>	rules,
		std::string	name = `"Execution Pipeline"`,
		bool	waitForInit = `true`
	)

inline

Creates an execution pipeline, which encapsulates a graph and duplicates it numPipelines times.

Parameters

numPipelines	the number of times to duplicate the graph
graph	the graph that the execution pipeline manages
rules	the list of decomposition rules that will be used for this pipeline
name	the name of the execution pipeline
waitForInit	Indicates that the execution pipeline will wait for its underlying graphs to initialize prior to marking that it is initialized.

Member Function Documentation

◆ addInputRule() [1/2]

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::addInputRule ( IRule< T, T > * rule )

inline

Adds an input rule, which can be used for domain decomposition.

This rule should use the pipelineId parameter in the IRule::applyRule function to aid in distributing data to the appropriate execution pipeline TaskGraph

Parameters

rule	the rule to handle distributing data between pipelines

◆ addInputRule() [2/2]

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::addInputRule ( std::shared_ptr< IRule< T, T >> rule )

inline

Adds an input rule, which can be used for domain decomposition.

This rule should use the pipelineId parameter in the IRule::applyRule function to aid in distributing data to the appropriate execution pipeline TaskGraph This variant should be used if the rule is intended to be shared with multiple bookkeepers or other execution pipelines.

Parameters

rule	the rule as a shared_ptr to handle distributing data between pipelines

◆ cleanupVisualization()

template<class T, class U>

std::string htgs::ExecutionPipeline< T, U >::cleanupVisualization	(	TaskGraphConf< T, U > *	graph,
		std::string	str
	)

inlineprivate

Moves the output connector outside of the execution pipeline graphs to cleanup how the graph looks during graph visualization.

Parameters

graph	the graph
str	the dot file string to be cleaned up

Returns: the improved dot file text

◆ copy()

template<class T, class U>

ITask<T, U>* htgs::ExecutionPipeline< T, U >::copy ( )

inlinevirtual

Makes a copy of the execution pipeline.

Copies the graph and reuses the same input decomposition rules

Returns: the copied execution pipeline task

Note: This function should only be called by the HTGS API

Implements htgs::ITask< T, U >.

◆ debug()

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::debug ( )

inlinevirtual

Provides debugging output for the execution pipeline.

Note: #define DEBUG_FLAG to enable debugging

Reimplemented from htgs::AnyITask.

◆ executeTask()

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::executeTask ( std::shared_ptr< T > data )

inlinevirtual

Executes the execution pipeline task on data and forwards that data to the input rules.

The input rules should parse the data and determine the correct pipelineId to forward the data to.

Parameters

data	the data to be forwarded to the proper execution pipeline

Note: This function should only be called by the HTGS API

Implements htgs::ITask< T, U >.

◆ gatherProfileData()

template<class T, class U>

virtual void htgs::ExecutionPipeline< T, U >::gatherProfileData ( std::map< AnyTaskManager *, TaskManagerProfile *> * taskManagerProfiles )

inlineoverridevirtual

Gathers profile data.

Parameters

taskManagerProfiles the mapping between the task manager and the profile data for the task manager.

Reimplemented from htgs::ITask< T, U >.

◆ genCustomDot()

template<class T, class U>

std::string htgs::ExecutionPipeline< T, U >::genCustomDot	(	ProfileUtils *	profileUtils,
		int	colorFlag
	)

inlineoverridevirtual

Virtual function to generate customized dot file.

Parameters

profiler the profiler that contains color maps

Returns: the dot file text that will be added near the end of the dot file

Reimplemented from htgs::AnyITask.

◆ genDot()

template<class T, class U>

std::string htgs::ExecutionPipeline< T, U >::genDot	(	int	flags,
		std::string	dotId,
		std::shared_ptr< AnyConnector >	input,
		std::shared_ptr< AnyConnector >	output
	)

inlineoverride

Virtual function that generates the input/output and per-task dot notation.

Parameters

flags	the DOTGEN flags
dotId	the id for this task
input	the input connector for this task
output	the output connector for this task

Returns: the dot that represents the interaction between the input/output and the internal custom dot notation

Note: This function will generate the dot notation for all sub-graphs within the execution pipeline.; If the dot notation is generated prior to execution, then a virtual pipeline is created. Generating the dot notation after execution will show the actual sub-graphs.

◆ getName()

template<class T, class U>

std::string htgs::ExecutionPipeline< T, U >::getName ( )

inlinevirtual

Gets the name for the execution pipeline.

Returns: the name of the execution pipeline

Reimplemented from htgs::ITask< T, U >.

◆ getNumGraphsSpawned()

template<class T, class U>

size_t htgs::ExecutionPipeline< T, U >::getNumGraphsSpawned ( )

inlineoverridevirtual

Gets the number of graphs spawned by this ITask.

Returns: the number of graphs spawned

Reimplemented from htgs::ITask< T, U >.

◆ initialize()

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::initialize ( )

inlinevirtual

Initializes the execution pipeline and duplicates the task graph based on the number of pipelines.

If wait for initialization is set to true, then this function will only return once all threads from all sub-graphs have been spawned and initialized.

Note: This function should only be called by the HTGS API

Reimplemented from htgs::ITask< T, U >.

◆ printProfile()

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::printProfile ( )

inlineoverridevirtual

Prints the profile data to std::out.

Reimplemented from htgs::ITask< T, U >.

◆ shutdown()

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::shutdown ( )

inlinevirtual

Shuts down the execution pipeline.

Note: This function should only be called by the HTGS API

Reimplemented from htgs::ITask< T, U >.

◆ shutdownParallel()

template<class T, class U>

void htgs::ExecutionPipeline< T, U >::shutdownParallel ( int id )

inline

Waits for all runtimes in parallel to update overall runtime of each graph.

Parameters

id	the thread ID

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

/home/tjb3/work/HTGS/src/htgs/api/ExecutionPipeline.hpp

Public Member Functions

Private Member Functions

Private Attributes

Detailed Description

template<class T, class U> class htgs::ExecutionPipeline< T, U >

Constructor & Destructor Documentation

◆ ExecutionPipeline() [1/2]

◆ ExecutionPipeline() [2/2]

Member Function Documentation

◆ addInputRule() [1/2]

◆ addInputRule() [2/2]

◆ cleanupVisualization()

◆ copy()

◆ debug()

◆ executeTask()

◆ gatherProfileData()

◆ genCustomDot()

◆ genDot()

◆ getName()

◆ getNumGraphsSpawned()

◆ initialize()

◆ printProfile()

◆ shutdown()

◆ shutdownParallel()

template<class T, class U>
class htgs::ExecutionPipeline< T, U >