jacsolve.h File Reference

#include "geom.h"
#include "getmatrix.h"
#include <vector>

Go to the source code of this file.

Classes
class	Multigrid
	The Multigrid hierarchy including meshes, vectors and matrices, prolongations is stored. More...

Functions
void	JacobiSolve (CRS_Matrix const &SK, std::vector< double > const &f, std::vector< double > &u)
void	JacobiSmoother (Matrix const &SK, std::vector< double > const &f, std::vector< double > &u, std::vector< double > &r, int nsmooth=1, double const omega=1.0, bool zero=false)
void	DiagPrecond (Matrix const &SK, std::vector< double > const &r, std::vector< double > &w, double const omega=1.0)
	Simple diagonale preconditioning.

Function Documentation

DiagPrecond()

void DiagPrecond	(	Matrix const &	SK,
		std::vector< double > const &	r,
		std::vector< double > &	w,
		double const	omega = 1.0
	)

Simple diagonale preconditioning.

The residuum r scaled by the inverse diagonal of matríx SK results in the correction w.

\( w := \omega D^{-1}*r \)

Parameters

[in]	SK	matrix
[in]	r	distributed local vector storing the residuum
[out]	w	accumulated local vector storing the correction.
[in]	omega	relaxation parameter.

Definition at line 89 of file jacsolve.cpp.

JacobiSmoother()

void JacobiSmoother	(	Matrix const &	SK,
		std::vector< double > const &	f,
		std::vector< double > &	u,
		std::vector< double > &	r,
		int	nsmooth = 1,
		double const	omega = 1.0,
		bool	zero = false
	)

Solves linear system of equations K u = f via the Jacobi iteration. We use a distributed symmetric CSR matrix SK and initial guess of the solution is set to 0.

In each smoothing step: \( \widehat{u} := u + \omega D^{-1}\left({f-K*u}\right) \)

Parameters

[in]	SK	CSR matrix
[in]	f	distributed local vector storing the right hand side
[out]	u	accumulated local vector storing the solution.
[in,out]	r	auxiliary local vector.
[in]	nsmooth	number of smoothing steps.
[in]	omega	relaxation parameter.
[in]	zero	initial solution u is assumed to be zero.

Definition at line 65 of file jacsolve.cpp.

JacobiSolve()

void JacobiSolve	(	CRS_Matrix const &	SK,
		std::vector< double > const &	f,
		std::vector< double > &	u
	)

Solves linear system of equations K u = f via the Jacobi iteration. We use a distributed symmetric CSR matrix SK and initial guess of the solution is set to 0.

Parameters

[in]	SK	CSR matrix
[in]	f	distributed local vector storing the right hand side
[out]	u	accumulated local vector storing the solution.