3-D finite difference-type linear system solver using Multigrid Preconditioned conjugate gradient (MGPCG). More...

#include <jet/fdm_mgpcg_solver3.h>

Inheritance diagram for jet::FdmMgpcgSolver3:

Public Member Functions
	FdmMgpcgSolver3 (unsigned int numberOfCgIter, size_t maxNumberOfLevels, unsigned int numberOfRestrictionIter=5, unsigned int numberOfCorrectionIter=5, unsigned int numberOfCoarsestIter=20, unsigned int numberOfFinalIter=20, double maxTolerance=1e-9, double sorFactor=1.5, bool useRedBlackOrdering=false)

bool	solve (FdmMgLinearSystem3 *system) override
	Solves the given linear system. More...

unsigned int	maxNumberOfIterations () const
	Returns the max number of Jacobi iterations. More...

unsigned int	lastNumberOfIterations () const
	Returns the last number of Jacobi iterations the solver made. More...

double	tolerance () const
	Returns the max residual tolerance for the Jacobi method. More...

double	lastResidual () const
	Returns the last residual after the Jacobi iterations. More...

Public Member Functions inherited from jet::FdmMgSolver3
	FdmMgSolver3 ()=default

virtual	~FdmMgSolver3 ()=default

	FdmMgSolver3 (size_t maxNumberOfLevels, unsigned int numberOfRestrictionIter=5, unsigned int numberOfCorrectionIter=5, unsigned int numberOfCoarsestIter=20, unsigned int numberOfFinalIter=20, double maxTolerance=1e-9, double sorFactor=1.5, bool useRedBlackOrdering=false)
	Constructs the solver with given parameters. More...

const MgParameters< FdmBlas3 > &	params () const
	Returns the Multigrid parameters. More...

double	sorFactor () const
	Returns the SOR (Successive Over Relaxation) factor. More...

bool	useRedBlackOrdering () const
	Returns true if red-black ordering is enabled. More...

bool	solve (FdmLinearSystem3 *system) final
	No-op. Multigrid-type solvers do not solve FdmLinearSystem3. More...

Public Member Functions inherited from jet::FdmLinearSystemSolver3
	FdmLinearSystemSolver3 ()=default

virtual	~FdmLinearSystemSolver3 ()=default

virtual bool	solveCompressed (FdmCompressedLinearSystem3 *)
	Solves the given compressed linear system. More...

Detailed Description

3-D finite difference-type linear system solver using Multigrid Preconditioned conjugate gradient (MGPCG).

See also: McAdams, Aleka, Eftychios Sifakis, and Joseph Teran. "A parallel multigrid Poisson solver for fluids simulation on large grids." Proceedings of the 2010 ACM SIGGRAPH/Eurographics Symposium on Computer Animation. Eurographics Association, 2010.

Constructor & Destructor Documentation

◆ FdmMgpcgSolver3()

jet::FdmMgpcgSolver3::FdmMgpcgSolver3	(	unsigned int	numberOfCgIter,
		size_t	maxNumberOfLevels,
		unsigned int	numberOfRestrictionIter = `5`,
		unsigned int	numberOfCorrectionIter = `5`,
		unsigned int	numberOfCoarsestIter = `20`,
		unsigned int	numberOfFinalIter = `20`,
		double	maxTolerance = `1e-9`,
		double	sorFactor = `1.5`,
		bool	useRedBlackOrdering = `false`
	)

Constructs the solver with given parameters.

Parameters

numberOfCgIter	- Number of CG iterations.
maxNumberOfLevels	- Number of maximum MG levels.
numberOfRestrictionIter	- Number of restriction iterations.
numberOfCorrectionIter	- Number of correction iterations.
numberOfCoarsestIter	- Number of iterations at the coarsest grid.
numberOfFinalIter	- Number of final iterations.
maxTolerance	- Number of max residual tolerance.

Member Function Documentation

◆ lastNumberOfIterations()

unsigned int jet::FdmMgpcgSolver3::lastNumberOfIterations ( ) const

Returns the last number of Jacobi iterations the solver made.

◆ lastResidual()

double jet::FdmMgpcgSolver3::lastResidual ( ) const

Returns the last residual after the Jacobi iterations.

◆ maxNumberOfIterations()

unsigned int jet::FdmMgpcgSolver3::maxNumberOfIterations ( ) const

Returns the max number of Jacobi iterations.

◆ solve()

bool jet::FdmMgpcgSolver3::solve ( FdmMgLinearSystem3 * system )

overridevirtual

Solves the given linear system.

Reimplemented from jet::FdmMgSolver3.

◆ tolerance()

double jet::FdmMgpcgSolver3::tolerance ( ) const

Returns the max residual tolerance for the Jacobi method.

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

jet/fdm_mgpcg_solver3.h

Public Member Functions