Dune::PDELab::ISTLBackend_CG_AMG_SSOR< GO, s > Class Template Reference

Overlapping parallel conjugate gradient solver preconditioned with AMG smoothed by SSOR. More...

#include <dune/pdelab/backend/istl/ovlpistlsolverbackend.hh>

Inheritance diagram for Dune::PDELab::ISTLBackend_CG_AMG_SSOR< GO, s >:

Public Types
typedef Dune::Amg::Parameters	Parameters
	Parameters object to customize matrix hierachy building.

Public Member Functions
	ISTLBackend_CG_AMG_SSOR (const GFS &gfs_, unsigned maxiter_=5000, int verbose_=1, bool reuse_=false, bool usesuperlu_=true)
	Constructor.
void	setParameters (const Parameters &params_)
	set AMG parameters
const Parameters &	parameters () const
	Get the parameters describing the behaviuour of AMG.
void	setReuse (bool reuse_)
	Set whether the AMG should be reused again during call to apply().
bool	getReuse () const
	Return whether the AMG is reused during call to apply()
V::ElementType	norm (const V &v) const
	compute global norm of a vector
void	apply (M &A, V &z, V &r, typename Dune::template FieldTraits< typename V::ElementType >::real_type reduction)
	solve the given linear system
const ISTLAMGStatistics &	statistics () const
	Get statistics of the AMG solver (no of levels, timings).
const Dune::PDELab::LinearSolverResult< double > &	result () const
	Return access to result data.

Protected Attributes
Dune::PDELab::LinearSolverResult< double >	res

Detailed Description

template<class GO, int s = 96>
class Dune::PDELab::ISTLBackend_CG_AMG_SSOR< GO, s >

Overlapping parallel conjugate gradient solver preconditioned with AMG smoothed by SSOR.

Template Parameters

GO	The type of the grid operator (or the fakeGOTraits class for the old grid operator space).
s	The bits to use for the global index.

Member Typedef Documentation

Parameters

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

typedef Dune::Amg::Parameters Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::Parameters

inherited

Parameters object to customize matrix hierachy building.

Constructor & Destructor Documentation

ISTLBackend_CG_AMG_SSOR()

template<class GO , int s = 96>

Dune::PDELab::ISTLBackend_CG_AMG_SSOR< GO, s >::ISTLBackend_CG_AMG_SSOR ( const GFS &  gfs_, unsigned  maxiter_ = 5000, int  verbose_ = 1, bool  reuse_ = false, bool  usesuperlu_ = true  )

inline

Constructor.

Parameters

gfs_	The grid function space used.
maxiter_	The maximum number of iterations allowed.
verbose_	The verbosity level to use.
reuse_	Set true, if the Matrix to be used is always identical (AMG aggregation is then only performed once).
usesuperlu_	Set false, to suppress the no SuperLU warning

Member Function Documentation

apply()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

void Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::apply ( M &  A, V &  z, V &  r, typename Dune::template FieldTraits< typename V::ElementType >::real_type  reduction  )

inlineinherited

solve the given linear system

Parameters

[in]	A	the given matrix
[out]	z	the solution vector to be computed
[in]	r	right hand side
[in]	reduction	to be achieved

getReuse()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

bool Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::getReuse ( ) const

inlineinherited

Return whether the AMG is reused during call to apply()

norm()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

V::ElementType Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::norm ( const V &  v ) const

inlineinherited

compute global norm of a vector

Parameters

[in]

v

the given vector

parameters()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

const Parameters & Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::parameters ( ) const

inlineinherited

Get the parameters describing the behaviuour of AMG.

The returned object can be adjusted to ones needs and then can be reset using setParameters.

Returns

The object holding the parameters of AMG.

result()

const Dune::PDELab::LinearSolverResult< double > & Dune::PDELab::LinearResultStorage::result ( ) const

inlineinherited

Return access to result data.

setParameters()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

void Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::setParameters ( const Parameters &  params_ )

inlineinherited

set AMG parameters

Parameters

[in]

params_

a parameter object of Type Dune::Amg::Parameters

setReuse()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

void Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::setReuse ( bool  reuse_ )

inlineinherited

Set whether the AMG should be reused again during call to apply().

statistics()

template<class GO , int s, template< class, class, class, int > class Preconditioner, template< class > class Solver>

const ISTLAMGStatistics & Dune::PDELab::ISTLBackend_AMG< GO, s, Preconditioner, Solver >::statistics ( ) const

inlineinherited

Get statistics of the AMG solver (no of levels, timings).

Returns

statistis of the AMG solver.

Member Data Documentation

res

Dune::PDELab::LinearSolverResult<double> Dune::PDELab::LinearResultStorage::res

protectedinherited

---

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

• ovlpistlsolverbackend.hh