BiCGStabSolver< T > Class Template Reference

#include <BiCGStabSolver.H>

Inheritance diagram for BiCGStabSolver< T >:

[legend]

Collaboration diagram for BiCGStabSolver< T >:

[legend]

List of all members.

Detailed Description

template<class T>
class BiCGStabSolver< T >

Elliptic solver using the BiCGStab algorithm.


Public Member Functions
	BiCGStabSolver ()
virtual	~BiCGStabSolver ()
virtual void	setHomogeneous (bool a_homogeneous)
virtual void	define (LinearOp< T > *a_op, bool a_homogeneous)
virtual void	solve (T &a_phi, const T &a_rhs)
	solve the equation.
virtual void	setConvergenceMetrics (Real a_metric, Real a_tolerance)
	Set a convergence metric, along with solver tolerance, if desired.
Public Attributes
bool	m_homogeneous
LinearOp< T > *	m_op
int	m_imax
int	m_verbosity
Real	m_eps
Real	m_reps
Real	m_convergenceMetric
Real	m_hang
int	m_exitStatus
Real	m_small
int	m_numRestarts
int	m_normType

Constructor & Destructor Documentation

template<class T>

BiCGStabSolver< T >::BiCGStabSolver ( ) [inline]

template<class T>

BiCGStabSolver< T >::~BiCGStabSolver ( ) [inline, virtual]

References BiCGStabSolver< T >::m_op.

Member Function Documentation

template<class T>

virtual void BiCGStabSolver< T >::setHomogeneous ( bool a_homogeneous ) [inline, virtual]

reset whether the solver is homogeneous.

Implements LinearSolver< T >.

Reimplemented in MergeSolver< T >.

template<class T>

void BiCGStabSolver< T >::define	(	LinearOp< T > *	a_op,
		bool	a_homogeneous
	)			`[inline, virtual]`

define the solver. a_op is the linear operator. a_homogeneous is whether the solver uses homogeneous boundary conditions.

Implements LinearSolver< T >.

Reimplemented in MergeSolver< T >.

References BiCGStabSolver< T >::m_homogeneous, and BiCGStabSolver< T >::m_op.

template<class T>

void BiCGStabSolver< T >::solve	(	T &	a_phi,
		const T &	a_rhs
	)			`[inline, virtual]`

template<class T>

void BiCGStabSolver< T >::setConvergenceMetrics	(	Real	a_metric,
		Real	a_tolerance
	)			`[inline, virtual]`

Set a convergence metric, along with solver tolerance, if desired.

Default implementation does nothing, since there are probably cases (liked direct solves), where this has no real meaning.

Reimplemented from LinearSolver< T >.

Reimplemented in MergeSolver< T >.

References BiCGStabSolver< T >::m_convergenceMetric, and BiCGStabSolver< T >::m_eps.

Member Data Documentation

template<class T>

bool BiCGStabSolver< T >::m_homogeneous

public member data: whether the solver is restricted to homogeneous boundary conditions

Referenced by BiCGStabSolver< T >::define(), BiCGStabSolver< LevelData< T > >::setHomogeneous(), and BiCGStabSolver< T >::solve().

template<class T>

LinearOp<T>* BiCGStabSolver< T >::m_op

public member data: operator to solve.

Referenced by BiCGStabSolver< T >::define(), BiCGStabSolver< T >::solve(), and BiCGStabSolver< T >::~BiCGStabSolver().

template<class T>

int BiCGStabSolver< T >::m_imax

public member data: maximum number of iterations

Referenced by BiCGStabSolver< T >::solve().

template<class T>

int BiCGStabSolver< T >::m_verbosity

public member data: how much screen out put the user wants. set = 0 for no output.

Referenced by MultilevelLinearOp< T >::define(), and BiCGStabSolver< T >::solve().

template<class T>

Real BiCGStabSolver< T >::m_eps

public member data: solver tolerance

Referenced by BiCGStabSolver< T >::setConvergenceMetrics(), and BiCGStabSolver< T >::solve().

template<class T>

Real BiCGStabSolver< T >::m_reps

public member data: relative solver tolerance

Referenced by BiCGStabSolver< T >::solve().

template<class T>

Real BiCGStabSolver< T >::m_convergenceMetric

public member data: solver convergence metric -- if negative, use initial residual; if positive, then use m_convergenceMetric

Referenced by BiCGStabSolver< T >::setConvergenceMetrics(), and BiCGStabSolver< T >::solve().

template<class T>

Real BiCGStabSolver< T >::m_hang

public member data: minium norm of solution should change per iterations

Referenced by BiCGStabSolver< T >::solve().

template<class T>

int BiCGStabSolver< T >::m_exitStatus

public member data: set = -1 if solver exited for an unknown reason set = 1 if solver converged to tolerance set = 2 if rho = 0 set = 3 if max number of restarts was reached

Referenced by BiCGStabSolver< T >::solve().

template<class T>

Real BiCGStabSolver< T >::m_small

public member data: what the algorithm should consider "close to zero"

Referenced by BiCGStabSolver< T >::solve().

template<class T>

int BiCGStabSolver< T >::m_numRestarts

public member data: number of times the algorithm can restart

Referenced by BiCGStabSolver< T >::solve().

template<class T>

int BiCGStabSolver< T >::m_normType

public member data: norm to be used when evaluation convergence. 0 is max norm, 1 is L(1), 2 is L(2) and so on.