Class KrylovLinearSolver provides an abstract interface for the implementation of Krylov subspace solvers for linear problems of the form \(Ax=b\).

#include <ibtk/KrylovLinearSolver.h>

Inheritance diagram for IBTK::KrylovLinearSolver:

[legend]

Public Member Functions
	KrylovLinearSolver ()=default
	Default constructor. More...

	~KrylovLinearSolver ()=default
	Empty destructor. More...

void	setHierarchyMathOps (SAMRAI::tbox::Pointer< HierarchyMathOps > hier_math_ops) override
	Set the HierarchyMathOps object used by the solver. More...

General-purpose solver functionality.
const std::string &	getName () const
	Return the object name. More...

virtual bool	getIsInitialized () const
	Return whether the operator is initialized. More...

virtual bool	getHomogeneousBc () const
	Return whether the solver is using homogeneous boundary conditions. More...

virtual double	getSolutionTime () const
	Get the time at which the solution is being evaluated. More...

virtual std::pair< double, double >	getTimeInterval () const
	Get the current time interval. More...

virtual double	getDt () const
	Get the current time step size. More...

virtual SAMRAI::tbox::Pointer< HierarchyMathOps >	getHierarchyMathOps () const
	Get the HierarchyMathOps object used by the solver. More...

virtual bool	solveSystem (SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &x, SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &b)=0
	Solve the system of equations. More...

virtual void	initializeSolverState (const SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &x, const SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &b)
	Compute hierarchy dependent data required for solving \(F[x]=b\). More...

virtual void	deallocateSolverState ()
	Remove all hierarchy dependent data allocated by initializeSolverState(). More...

Functions to access solver parameters.
virtual void	setMaxIterations (int max_iterations)
	Set the maximum number of nonlinear iterations to use per solve. More...

virtual int	getMaxIterations () const
	Get the maximum number of nonlinear iterations to use per solve. More...

virtual void	setAbsoluteTolerance (double abs_residual_tol)
	Set the absolute residual tolerance for convergence. More...

virtual double	getAbsoluteTolerance () const
	Get the absolute residual tolerance for convergence. More...

virtual void	setRelativeTolerance (double rel_residual_tol)
	Set the relative residual tolerance for convergence. More...

virtual double	getRelativeTolerance () const
	Get the relative residual tolerance for convergence. More...

Logging functions.
std::string	d_object_name = "unitialized"

bool	d_is_initialized = false

bool	d_homogeneous_bc = false

double	d_solution_time = std::numeric_limits<double>::quiet_NaN()

double	d_current_time = std::numeric_limits<double>::quiet_NaN()

double	d_new_time = std::numeric_limits<double>::quiet_NaN()

double	d_rel_residual_tol = 0.0

double	d_abs_residual_tol = 0.0

int	d_max_iterations = 100

int	d_current_iterations = 0

double	d_current_residual_norm = std::numeric_limits<double>::quiet_NaN()

SAMRAI::tbox::Pointer< HierarchyMathOps >	d_hier_math_ops

bool	d_hier_math_ops_external = false

bool	d_enable_logging = false

virtual void	setLoggingEnabled (bool enable_logging=true)
	Enable or disable logging. More...

virtual bool	getLoggingEnabled () const
	Determine whether logging is enabled or disabled. More...

void	init (const std::string &object_name, bool homogeneous_bc)

virtual void	initSpecialized (const std::string &object_name, bool homogeneous_bc)

Krylov solver functionality.
SAMRAI::tbox::Pointer< LinearOperator >	d_A

SAMRAI::tbox::Pointer< LinearSolver >	d_pc_solver

SAMRAI::tbox::Pointer< SAMRAI::solv::SAMRAIVectorReal< NDIM, double > >	d_x

SAMRAI::tbox::Pointer< SAMRAI::solv::SAMRAIVectorReal< NDIM, double > >	d_b

virtual void	setOperator (SAMRAI::tbox::Pointer< LinearOperator > A)
	Set the linear operator used when solving \(Ax=b\). More...

virtual SAMRAI::tbox::Pointer< LinearOperator >	getOperator () const
	Retrieve the linear operator used when solving \(Ax=b\). More...

virtual void	setPreconditioner (SAMRAI::tbox::Pointer< LinearSolver > pc_solver=nullptr)
	Set the preconditioner used by the Krylov subspace method when solving \(Ax=b\). More...

virtual SAMRAI::tbox::Pointer< LinearSolver >	getPreconditioner () const
	Retrieve the preconditioner used by the Krylov subspace method when solving \(Ax=b\). More...

	KrylovLinearSolver (const KrylovLinearSolver &from)=delete
	Copy constructor. More...

KrylovLinearSolver &	operator= (const KrylovLinearSolver &that)=delete
	Assignment operator. More...

Functions to access solver parameters.
virtual void	setInitialGuessNonzero (bool initial_guess_nonzero=true)
	Set whether the initial guess is non-zero. More...

virtual bool	getInitialGuessNonzero () const
	Get whether the initial guess is non-zero. More...

Logging functions.
bool	d_initial_guess_nonzero = true

bool	d_nullspace_contains_constant_vec = false

std::vector< SAMRAI::tbox::Pointer< SAMRAI::solv::SAMRAIVectorReal< NDIM, double > > >	d_nullspace_basis_vecs

virtual void	printClassData (std::ostream &stream) override
	Print class data to stream. More...

Constructor & Destructor Documentation

◆ KrylovLinearSolver() [1/2]

IBTK::KrylovLinearSolver::KrylovLinearSolver ( )

default

◆ ~KrylovLinearSolver()

IBTK::KrylovLinearSolver::~KrylovLinearSolver ( )

default

◆ KrylovLinearSolver() [2/2]

IBTK::KrylovLinearSolver::KrylovLinearSolver ( const KrylovLinearSolver & from )

privatedelete

Note: This constructor is not implemented and should not be used.

Parameters

from	The value to copy to this object.

Member Function Documentation

◆ setHierarchyMathOps()

void IBTK::KrylovLinearSolver::setHierarchyMathOps ( SAMRAI::tbox::Pointer< HierarchyMathOps > hier_math_ops )

overridevirtual

Reimplemented from IBTK::GeneralSolver.

◆ setHomogeneousBc()

void IBTK::KrylovLinearSolver::setHomogeneousBc ( bool homogeneous_bc )

overridevirtual

Reimplemented from IBTK::GeneralSolver.

◆ setSolutionTime()

void IBTK::KrylovLinearSolver::setSolutionTime ( double solution_time )

overridevirtual

Reimplemented from IBTK::GeneralSolver.

◆ setTimeInterval()

void IBTK::KrylovLinearSolver::setTimeInterval	(	double	current_time,
		double	new_time
	)

overridevirtual

Reimplemented from IBTK::GeneralSolver.

◆ setOperator()

virtual void IBTK::KrylovLinearSolver::setOperator ( SAMRAI::tbox::Pointer< LinearOperator > A )

virtual

Reimplemented in IBTK::PETScKrylovLinearSolver.

◆ getOperator()

virtual SAMRAI::tbox::Pointer<LinearOperator> IBTK::KrylovLinearSolver::getOperator ( ) const

virtual

◆ setPreconditioner()

virtual void IBTK::KrylovLinearSolver::setPreconditioner ( SAMRAI::tbox::Pointer< LinearSolver > pc_solver = nullptr )

virtual

Note: If the preconditioner is nullptr, no preconditioning is performed.

Reimplemented in IBTK::PETScKrylovLinearSolver.

◆ getPreconditioner()

virtual SAMRAI::tbox::Pointer<LinearSolver> IBTK::KrylovLinearSolver::getPreconditioner ( ) const

virtual

◆ operator=()

KrylovLinearSolver& IBTK::KrylovLinearSolver::operator= ( const KrylovLinearSolver & that )

privatedelete

Note: This operator is not implemented and should not be used.

Parameters

that	The value to assign to this object.

Returns: A reference to this object.

◆ setNullSpace()

virtual void IBTK::LinearSolver::setNullSpace	(	bool	nullspace_contains_constant_vec,
		const std::vector< SAMRAI::tbox::Pointer< SAMRAI::solv::SAMRAIVectorReal< NDIM, double > > > &	nullspace_basis_vecs = `std::vector< SAMRAI::tbox::Pointer< SAMRAI::solv::SAMRAIVectorReal< NDIM, double > > >()`
	)

virtualinherited

Implementations can require the nullspace basis vectors to be orthogonal but should not assume the basis vectors to be orthonormal. If the basis vectors are not orthonormal, the solver may normalize them in place.

Reimplemented in IBTK::PETScKrylovLinearSolver, and IBTK::PETScLevelSolver.

◆ getNullSpaceContainsConstantVector()

virtual bool IBTK::LinearSolver::getNullSpaceContainsConstantVector ( ) const

virtualinherited

◆ getNullSpaceBasisVectors()

virtual const std::vector<SAMRAI::tbox::Pointer<SAMRAI::solv::SAMRAIVectorReal<NDIM, double> > >& IBTK::LinearSolver::getNullSpaceBasisVectors ( ) const

virtualinherited

◆ setInitialGuessNonzero()

virtual void IBTK::LinearSolver::setInitialGuessNonzero ( bool initial_guess_nonzero = true )

virtualinherited

Reimplemented in IBTK::BGaussSeidelPreconditioner, IBTK::FACPreconditioner, IBTK::BJacobiPreconditioner, IBTK::PETScPCLSWrapper, IBAMR::StaggeredStokesBlockFactorizationPreconditioner, IBAMR::StaggeredStokesProjectionPreconditioner, and IBAMR::VCStaggeredStokesProjectionPreconditioner.

◆ getInitialGuessNonzero()

virtual bool IBTK::LinearSolver::getInitialGuessNonzero ( ) const

virtualinherited

Reimplemented in IBTK::PETScPCLSWrapper.

◆ printClassData()

virtual void IBTK::LinearSolver::printClassData ( std::ostream & stream )

overridevirtualinherited

Reimplemented from IBTK::GeneralSolver.

◆ getName()

const std::string& IBTK::GeneralSolver::getName ( ) const

inherited

◆ getIsInitialized()

virtual bool IBTK::GeneralSolver::getIsInitialized ( ) const

virtualinherited

◆ getHomogeneousBc()

virtual bool IBTK::GeneralSolver::getHomogeneousBc ( ) const

virtualinherited

◆ getSolutionTime()

virtual double IBTK::GeneralSolver::getSolutionTime ( ) const

virtualinherited

◆ getTimeInterval()

virtual std::pair<double, double> IBTK::GeneralSolver::getTimeInterval ( ) const

virtualinherited

◆ getDt()

virtual double IBTK::GeneralSolver::getDt ( ) const

virtualinherited

◆ getHierarchyMathOps()

virtual SAMRAI::tbox::Pointer<HierarchyMathOps> IBTK::GeneralSolver::getHierarchyMathOps ( ) const

virtualinherited

◆ solveSystem()

virtual bool IBTK::GeneralSolver::solveSystem	(	SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &	x,
		SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &	b
	)

pure virtualinherited

Before calling solveSystem(), the form of the solution x and right-hand-side b vectors must be set properly by the user on all patch interiors on the specified range of levels in the patch hierarchy. The user is responsible for all data management for the quantities associated with the solution and right-hand-side vectors. In particular, patch data in these vectors must be allocated prior to calling this method.

Parameters

x	solution vector
b	right-hand-side vector

Conditions on Parameters:

vectors x and b must have same patch hierarchy
vectors x and b must have same structure, depth, etc.

Note: Subclasses must be implemented so that the vector arguments for solveSystem() need not match those for initializeSolverState(). However, they are allowed to require a certain degree of similarity, including:

hierarchy configuration (hierarchy pointer and range of levels)
number, type and alignment of vector component data
ghost cell widths of data in the solution x and right-hand-side b vectors

Note: Subclasses are required to be implemented so that the solver does not need to be initialized prior to calling solveSystem(); however, see initializeSolverState() and deallocateSolverState() for opportunities to save overhead when performing multiple consecutive solves.

See also: initializeSolverState; deallocateSolverState

Returns: true if the solver converged to the specified tolerances, false otherwise

Implemented in IBTK::PETScKrylovLinearSolver, IBTK::PETScNewtonKrylovSolver, IBTK::CCPoissonHypreLevelSolver, IBTK::SCPoissonHypreLevelSolver, IBTK::BGaussSeidelPreconditioner, IBTK::PETScLevelSolver, IBTK::FACPreconditioner, IBTK::BJacobiPreconditioner, IBAMR::StaggeredStokesBlockFactorizationPreconditioner, IBTK::PETScPCLSWrapper, IBAMR::StaggeredStokesProjectionPreconditioner, IBAMR::VCStaggeredStokesProjectionPreconditioner, IBAMR::CIBStaggeredStokesSolver, and IBAMR::IBImplicitStaggeredHierarchyIntegrator::IBImplicitStaggeredStokesSolver.

◆ initializeSolverState()

virtual void IBTK::GeneralSolver::initializeSolverState	(	const SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &	x,
		const SAMRAI::solv::SAMRAIVectorReal< NDIM, double > &	b
	)

virtualinherited

In a typical implementation, the solveSystem() method will compute some required hierarchy dependent data before the solve, and then remove that data after the solve. For multiple solves that use the same hierarchy configuration, it is more efficient to:

initialize the hierarchy-dependent data required by the solver via initializeSolverState(),
solve the system one or more times via solveSystem(), and
remove the hierarchy-dependent data via deallocateSolverState().

Note that it is generally necessary to reinitialize the solver state when the hierarchy configuration changes.

Parameters

x	solution vector
b	right-hand-side vector

Conditions on Parameters:

vectors x and b must have same patch hierarchy
vectors x and b must have same structure, depth, etc.

Note: Subclasses must be implemented so that the vector arguments for solveSystem() need not match those for initializeSolverState(). However, they are allowed to require a certain degree of similarity, including:

hierarchy configuration (hierarchy pointer and range of levels)
number, type and alignment of vector component data
ghost cell widths of data in the solution x and right-hand-side b vectors

Note: Subclasses are required to be implemented so that it is safe to call initializeSolverState() when the solver state is already initialized. In this case, the solver state should be first deallocated and then reinitialized.; Subclasses are required to be implemented so that when any operator objects have been registered with the solver via setOperator() or setJacobian(), they are also initialized by initializeSolverState().

See also: deallocateSolverState

Reimplemented in IBTK::PETScKrylovLinearSolver, IBTK::PETScNewtonKrylovSolver, IBTK::CCPoissonHypreLevelSolver, IBTK::SCPoissonHypreLevelSolver, IBTK::BGaussSeidelPreconditioner, IBTK::PETScLevelSolver, IBTK::FACPreconditioner, IBTK::BJacobiPreconditioner, IBAMR::StaggeredStokesBlockPreconditioner, IBTK::PETScPCLSWrapper, IBAMR::StaggeredStokesBlockFactorizationPreconditioner, IBAMR::StaggeredStokesProjectionPreconditioner, IBAMR::VCStaggeredStokesProjectionPreconditioner, and IBAMR::CIBStaggeredStokesSolver.

◆ deallocateSolverState()

virtual void IBTK::GeneralSolver::deallocateSolverState ( )

virtualinherited

Note: Subclasses are required to be implemented so that it is safe to call deallocateSolverState() when the solver state is already deallocated.; Subclasses are required to be implemented so that when any operator objects have been registered with the solver via setOperator() or setJacobian(), they are also deallocated by deallocateSolverState().

See also: initializeSolverState

Reimplemented in IBTK::PETScKrylovLinearSolver, IBTK::PETScNewtonKrylovSolver, IBTK::CCPoissonHypreLevelSolver, IBTK::SCPoissonHypreLevelSolver, IBTK::BGaussSeidelPreconditioner, IBTK::PETScLevelSolver, IBTK::FACPreconditioner, IBTK::BJacobiPreconditioner, IBAMR::StaggeredStokesBlockPreconditioner, IBTK::PETScPCLSWrapper, IBAMR::StaggeredStokesBlockFactorizationPreconditioner, IBAMR::StaggeredStokesProjectionPreconditioner, IBAMR::VCStaggeredStokesProjectionPreconditioner, and IBAMR::CIBStaggeredStokesSolver.

◆ setMaxIterations()

virtual void IBTK::GeneralSolver::setMaxIterations ( int max_iterations )

virtualinherited

Reimplemented in IBTK::BGaussSeidelPreconditioner, IBTK::FACPreconditioner, IBTK::BJacobiPreconditioner, IBTK::PETScPCLSWrapper, IBAMR::StaggeredStokesBlockFactorizationPreconditioner, IBAMR::StaggeredStokesProjectionPreconditioner, and IBAMR::VCStaggeredStokesProjectionPreconditioner.

◆ getMaxIterations()

virtual int IBTK::GeneralSolver::getMaxIterations ( ) const

virtualinherited

Reimplemented in IBTK::PETScPCLSWrapper.

◆ setAbsoluteTolerance()

virtual void IBTK::GeneralSolver::setAbsoluteTolerance ( double abs_residual_tol )

virtualinherited

Reimplemented in IBTK::PETScPCLSWrapper.

◆ getAbsoluteTolerance()

virtual double IBTK::GeneralSolver::getAbsoluteTolerance ( ) const

virtualinherited

Reimplemented in IBTK::PETScPCLSWrapper.

◆ setRelativeTolerance()

virtual void IBTK::GeneralSolver::setRelativeTolerance ( double rel_residual_tol )

virtualinherited

Reimplemented in IBTK::PETScPCLSWrapper.

◆ getRelativeTolerance()

virtual double IBTK::GeneralSolver::getRelativeTolerance ( ) const

virtualinherited

Reimplemented in IBTK::PETScPCLSWrapper.

◆ getNumIterations()

virtual int IBTK::GeneralSolver::getNumIterations ( ) const

virtualinherited

Reimplemented in IBTK::BGaussSeidelPreconditioner, IBTK::PETScPCLSWrapper, and IBTK::BJacobiPreconditioner.

◆ getResidualNorm()

virtual double IBTK::GeneralSolver::getResidualNorm ( ) const

virtualinherited

Reimplemented in IBTK::BGaussSeidelPreconditioner, IBTK::PETScPCLSWrapper, and IBTK::BJacobiPreconditioner.

◆ setLoggingEnabled()

virtual void IBTK::GeneralSolver::setLoggingEnabled ( bool enable_logging = true )

virtualinherited

◆ getLoggingEnabled()

virtual bool IBTK::GeneralSolver::getLoggingEnabled ( ) const

virtualinherited

◆ init()

void IBTK::GeneralSolver::init	(	const std::string &	object_name,
		bool	homogeneous_bc
	)

protectedinherited

◆ initSpecialized()

virtual void IBTK::GeneralSolver::initSpecialized	(	const std::string &	object_name,
		bool	homogeneous_bc
	)

protectedvirtualinherited

Reimplemented in IBTK::PoissonSolver.

Member Data Documentation

◆ d_A

SAMRAI::tbox::Pointer<LinearOperator> IBTK::KrylovLinearSolver::d_A

protected

◆ d_pc_solver

SAMRAI::tbox::Pointer<LinearSolver> IBTK::KrylovLinearSolver::d_pc_solver

protected

◆ d_x

SAMRAI::tbox::Pointer<SAMRAI::solv::SAMRAIVectorReal<NDIM, double> > IBTK::KrylovLinearSolver::d_x

protected

◆ d_b

SAMRAI::tbox::Pointer<SAMRAI::solv::SAMRAIVectorReal<NDIM, double> > IBTK::KrylovLinearSolver::d_b

protected

◆ d_initial_guess_nonzero

bool IBTK::LinearSolver::d_initial_guess_nonzero = true

protectedinherited

◆ d_nullspace_contains_constant_vec

bool IBTK::LinearSolver::d_nullspace_contains_constant_vec = false

protectedinherited

◆ d_nullspace_basis_vecs

std::vector<SAMRAI::tbox::Pointer<SAMRAI::solv::SAMRAIVectorReal<NDIM, double> > > IBTK::LinearSolver::d_nullspace_basis_vecs

protectedinherited

◆ d_object_name

std::string IBTK::GeneralSolver::d_object_name = "unitialized"

protectedinherited

◆ d_is_initialized

bool IBTK::GeneralSolver::d_is_initialized = false

protectedinherited

◆ d_homogeneous_bc

bool IBTK::GeneralSolver::d_homogeneous_bc = false

protectedinherited

◆ d_solution_time

double IBTK::GeneralSolver::d_solution_time = std::numeric_limits<double>::quiet_NaN()

protectedinherited

◆ d_current_time

double IBTK::GeneralSolver::d_current_time = std::numeric_limits<double>::quiet_NaN()

protectedinherited

◆ d_new_time

double IBTK::GeneralSolver::d_new_time = std::numeric_limits<double>::quiet_NaN()

protectedinherited

◆ d_rel_residual_tol

double IBTK::GeneralSolver::d_rel_residual_tol = 0.0

protectedinherited

◆ d_abs_residual_tol

double IBTK::GeneralSolver::d_abs_residual_tol = 0.0

protectedinherited

◆ d_max_iterations

int IBTK::GeneralSolver::d_max_iterations = 100

protectedinherited

◆ d_current_iterations

int IBTK::GeneralSolver::d_current_iterations = 0

protectedinherited

◆ d_current_residual_norm

double IBTK::GeneralSolver::d_current_residual_norm = std::numeric_limits<double>::quiet_NaN()

protectedinherited

◆ d_hier_math_ops

SAMRAI::tbox::Pointer<HierarchyMathOps> IBTK::GeneralSolver::d_hier_math_ops

protectedinherited

◆ d_hier_math_ops_external

bool IBTK::GeneralSolver::d_hier_math_ops_external = false

protectedinherited

◆ d_enable_logging

bool IBTK::GeneralSolver::d_enable_logging = false

protectedinherited

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

ibtk/include/ibtk/KrylovLinearSolver.h

Public Member Functions

General-purpose solver functionality.

Functions to access solver parameters.

Logging functions.

Krylov solver functionality.

Functions to access solver parameters.

Logging functions.

Constructor & Destructor Documentation

◆ KrylovLinearSolver() [1/2]

◆ ~KrylovLinearSolver()

◆ KrylovLinearSolver() [2/2]

Member Function Documentation

◆ setHierarchyMathOps()

◆ setHomogeneousBc()

◆ setSolutionTime()

◆ setTimeInterval()

◆ setOperator()

◆ getOperator()

◆ setPreconditioner()

◆ getPreconditioner()

◆ operator=()

◆ setNullSpace()

◆ getNullSpaceContainsConstantVector()

◆ getNullSpaceBasisVectors()

◆ setInitialGuessNonzero()

◆ getInitialGuessNonzero()

◆ printClassData()

◆ getName()

◆ getIsInitialized()

◆ getHomogeneousBc()

◆ getSolutionTime()

◆ getTimeInterval()

◆ getDt()

◆ getHierarchyMathOps()

◆ solveSystem()

◆ initializeSolverState()

◆ deallocateSolverState()

◆ setMaxIterations()

◆ getMaxIterations()

◆ setAbsoluteTolerance()

◆ getAbsoluteTolerance()

◆ setRelativeTolerance()

◆ getRelativeTolerance()

◆ getNumIterations()

◆ getResidualNorm()

◆ setLoggingEnabled()

◆ getLoggingEnabled()

◆ init()

◆ initSpecialized()

Member Data Documentation

◆ d_A

◆ d_pc_solver

◆ d_x

◆ d_b

◆ d_initial_guess_nonzero

◆ d_nullspace_contains_constant_vec

◆ d_nullspace_basis_vecs

◆ d_object_name

◆ d_is_initialized

◆ d_homogeneous_bc

◆ d_solution_time

◆ d_current_time

◆ d_new_time

◆ d_rel_residual_tol

◆ d_abs_residual_tol

◆ d_max_iterations

◆ d_current_iterations

◆ d_current_residual_norm

◆ d_hier_math_ops

◆ d_hier_math_ops_external

◆ d_enable_logging