Locks implementation on base of system synchronization mechanisms and OpenMP locks. details Implementation of locks (including scoped lock) on base of system synchronization mechanisms and OpenMP locks.
. More...

#include <math_define.h>

Classes
class	CommonMutex
	Lock class. More...

class	ParallelRegionGuard
	Scoped guard of parallel region. More...

class	ScopedLock
	Scoped lock. Can accept a null pointer to a mutex. Locking occurs if the pointer to the mutex is nonzero and the code runs in parallel. More...

class	ScopedRecursiveLock
	Recursive scoped lock. Can accept a null pointer to a mutex. Locking occurs if the pointer to the mutex is nonzero and the code runs in parallel. More...

class	MbSyncItem
	Base synchronization object with lazy initialization. More...

class	MbNestSyncItem
	Base synchronization object with lazy initialization which supports nested locks. More...

class	MbPersistentSyncItem
	Base object providing means of synchronization. More...

class	MbPersistentNestSyncItem
	Base synchronization object with support of nested locks. More...

Macros
#define	CommonRecursiveMutex CommonMutex
	Same implementation of CommonMutex and CommonRecursiveMutex.

#define	ENTER_PARALLEL(useParallel) if ( useParallel ) EnterParallelRegion();
	If useParallel == true, notifies the kernel about entering a parallel region. More...

#define	EXIT_PARALLEL(useParallel) if ( useParallel ) ExitParallelRegion();
	If useParallel == true, notifies the kernel about exiting a parallel region. More...

#define	ENTER_PARALLEL_FORCED EnterParallelRegion();
	Notifies the kernel about entering a parallel region. More...

#define	EXIT_PARALLEL_FORCED ExitParallelRegion();
	Notifiesotifies the kernel about exiting a parallel region. More...

#define	SET_MEMORY_SCOPED_LOCK ScopedRecursiveLock memScopedLock( GetGlobalRecursiveLock() );
	Set scoped lock for memory allocation and deallocation operations.

Enumerations
enum	MbeMultithreadedMode { mtm_Off = 0 , mtm_Standard = 1 , mtm_SafeItems = 2 , mtm_Items = 3 , mtm_Max = 31 }
	Multithreading modes. More...

Functions
void	EnterParallelRegion ()
	The function notifies the kernel about entering a parallel region. The function call should be placed before the start of a parallel block.

void	ExitParallelRegion ()
	The function notifies the kernel about exiting a parallel region. The function call should be placed after the end of the parallel block.

bool	IsInParallel ()
	The function determines whether the code is executed in parallel.

bool	CacheCleanupAllowed ()
	The function determines whether the caches cleanup is allowed.

bool	IsSafeMultithreading ()
	Whether is enabled a safe multithreading mode (used in CacheManager).

bool	LocksEnabled ()
	Whether locks are enabled (multithreading is on and code is executed in parallel).

unsigned int	GetThreadKey ()
	Get a current thread pseudo-identifier.

CommonMutex *	GetGlobalLock ()
	Get a pointer to the global mutex (used in CacheManager).

CommonMutex *	GetGlobalRecursiveLock ()
	Get a pointer to the global recursive mutex (used for memory allocation and deallocation operations).

Detailed Description

Locks implementation on base of system synchronization mechanisms and OpenMP locks. details Implementation of locks (including scoped lock) on base of system synchronization mechanisms and OpenMP locks.
.

Macro Definition Documentation

◆ ENTER_PARALLEL

#define ENTER_PARALLEL ( useParallel ) if ( useParallel ) EnterParallelRegion();

If useParallel == true, notifies the kernel about entering a parallel region.

If useParallel == true, notifies the kernel about entering a parallel region. The call should be placed before the start of a parallel block (before OpenMP pragma). Using a macro speeds up parallel OpenMP cycles significantly.

◆ EXIT_PARALLEL

#define EXIT_PARALLEL ( useParallel ) if ( useParallel ) ExitParallelRegion();

If useParallel == true, notifies the kernel about exiting a parallel region.

If useParallel == true, notifies the kernel about exiting a parallel region. The call should be placed after the end of the parallel block. Using a macro speeds up parallel OpenMP cycles significantly.

◆ ENTER_PARALLEL_FORCED

#define ENTER_PARALLEL_FORCED EnterParallelRegion();

Notifies the kernel about entering a parallel region.

Notifies the kernel about entering a parallel region. The call should be placed before the start of a parallel block (before OpenMP pragma). Using a macro speeds up parallel OpenMP cycles significantly.

◆ EXIT_PARALLEL_FORCED

#define EXIT_PARALLEL_FORCED ExitParallelRegion();

Notifiesotifies the kernel about exiting a parallel region.

Notifies the kernel about exiting a parallel region. The call should be placed after the end of the parallel block. Using a macro speeds up parallel OpenMP cycles significantly.

Classes

Macros