Thread token generator for Tango sources. More...

#include <cutthreadtokengen.h>

Inheritance diagram for CuTThreadTokenGen:

Public Member Functions
	CuTThreadTokenGen (int pool_siz, const std::string &tok_prefix="tthread_")
	CuTThreadTokenGen::CuTThreadTokenGen constructor. More...

	~CuTThreadTokenGen ()

std::string	generate (const std::string &dev)
	This function implements the CuThreadTokenI::generate method and is invoked by Cumbia::threadToken function to provide a token (in the form of std::string) to be used to group threads together. More...

void	map (const std::string &device, int which)
	User defined associations between device names and thread number. More...

int	which_thread (const std::string &device) const
	CuTThreadTokenGen::which_thread returns which thread has been assigned to the given device. More...

int	which_from_token (const std::string &tok) const
	Removes the prefix from tok and returns the. More...

std::list< std::string >	devicesForToken (const std::string &token) const
	Returns the list of devices whose sources belong to the thread with the given token. More...

std::string	tok_prefix () const
	Returns the prefix used for the token name. More...

int	threadPoolSize () const
	CuTThreadTokenGen::threadPoolSize returns the upper limit on the number of threads created and managed by CuThreadService. More...

void	setThreadPoolSize (int siz)
	CuTThreadTokenGen::setThreadPoolSize change the thread pool size. More...

Detailed Description

Thread token generator for Tango sources.

A thread token determines how threads are grouped in Cumbia. When Cumbia::registerActivity is called, the thread token will assign an activity to the thread with the given token. Activities with the same token are executed in the same thread.

Tokens for grouping Tango sources can be either generated automatically, with an imposed limit to the set, or through a user defined map. When the limit is reached, the first method starts reusing threads, ensuring at the same time that sources referring to the same devices belong to the same thread. The automatic mode is employed if no user map has been defined. The CuTThreadTokenGen::generate method provides more details.

Use setThreadPoolSize to fix a limit on the number of desired threads.

After instantiating and configuring CuTThreadTokenGen, call Cumbia::setThreadTokenGenerator from your application to use the token generator.

Recommended readings

CuThreadTokenGenI
Cumbia::setThreadTokenGenerator
Cumbia::threadToken

Constructor & Destructor Documentation

◆ CuTThreadTokenGen()

CuTThreadTokenGen::CuTThreadTokenGen	(	int	pool_siz,
		const std::string &	tok_prefix = `"tthread_"`
	)

CuTThreadTokenGen::CuTThreadTokenGen constructor.

Build a Tango thread token generator. A thread token determines how threads are grouped in Cumbia. When Cumbia::registerActivity is called, the thread token will assign an activity to the thread with the given token.

Parameters

pool_siz	the desired upper limit on the thread count.
tok_prefix	a string prefix to prepend to the thread number in the token returned by generate

See also: CuTThreadTokenGen::generate; Cumbia::setThreadTokenGenerator; Cumbia::threadToken; Cumbia::registerActivity

References tok_prefix().

◆ ~CuTThreadTokenGen()

CuTThreadTokenGen::~CuTThreadTokenGen ( )

Member Function Documentation

◆ devicesForToken()

std::list< std::string > CuTThreadTokenGen::devicesForToken ( const std::string & token ) const

Returns the list of devices whose sources belong to the thread with the given token.

Parameters

token the thread token

Returns: list of devices whose sources live in the same thread with the given token

References which_from_token().

◆ generate()

std::string CuTThreadTokenGen::generate ( const std::string & dev )

This function implements the CuThreadTokenI::generate method and is invoked by Cumbia::threadToken function to provide a token (in the form of std::string) to be used to group threads together.

User defined mapping: If a user defined map has been defined by calling CuTThreadTokenGen::map, it is used to group threads accordingly. In that case, the defined thread pool size is ignored.

Automatic token generation

If the user defined map is empty and the thread pool size is greater than zero, automatic device grouping into threads takes place, so that:

sources with the same device belong to the same thread
the total number of threads is at most threadPoolSize.

If the in parameter is null, then an empty string is returned.

If the device value in the input argument is not in the user map (i.e. the user only partially defines the device/thread association), then the generation will be hybrid.

Parameters

in

Returns

◆ map()

void CuTThreadTokenGen::map	(	const std::string &	device,
		int	which
	)

User defined associations between device names and thread number.

Parameters

device	the device name
which	the thread, as integer, to associate to the given device

Note: If the user defined map is not empty, its elements only are used by the generate function. In that case, no automatic generation will take place and if a device is not in the user map the token will be left untouched.

See also: generate

◆ setThreadPoolSize()

void CuTThreadTokenGen::setThreadPoolSize ( int siz )

CuTThreadTokenGen::setThreadPoolSize change the thread pool size.

Parameters

siz	the upper limit on the number of threads created and managed by CuThreadService

Note: Changes to the thread pool size after activities have been registered do not affect the threads already running. For example, their number will not be reduced and their grouping will not be changed.

◆ threadPoolSize()

int CuTThreadTokenGen::threadPoolSize ( ) const

CuTThreadTokenGen::threadPoolSize returns the upper limit on the number of threads created and managed by CuThreadService.

Returns: the thread pool size, initialised either in the constructor or with setThreadPoolSize

◆ tok_prefix()

std::string CuTThreadTokenGen::tok_prefix ( ) const

Returns the prefix used for the token name.

Returns: the string used as prefix to name the token

Referenced by CuTThreadTokenGen().

◆ which_from_token()

int CuTThreadTokenGen::which_from_token ( const std::string & tok ) const

Removes the prefix from tok and returns the.

Parameters

tok

Returns: the integer obtained from the given token after removing the prefix, or -1 if tok does not exactly match prefix + number

Referenced by devicesForToken().

◆ which_thread()

int CuTThreadTokenGen::which_thread ( const std::string & device ) const

CuTThreadTokenGen::which_thread returns which thread has been assigned to the given device.

Parameters

device the name of the Tango device

Returns: an integer representing which thread has been assigned to the given device

The user defined map is searched first, the automatically generated map afterwards. If the device has not been mapped, -1 is returned.

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

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ CuTThreadTokenGen()

◆ ~CuTThreadTokenGen()

Member Function Documentation

◆ devicesForToken()

◆ generate()

◆ map()

◆ setThreadPoolSize()

◆ threadPoolSize()

◆ tok_prefix()

◆ which_from_token()

◆ which_thread()