Go::SolveCG Class Reference
['geometry' - a spline library]

Solve the equation system Ax=b where A is a symmetric positive definite matrix using the Conjugate Gradient Method. More...

#include <SolveCG.h>

List of all members.

Public Member Functions
	SolveCG ()
	Default constructor.
	~SolveCG ()
	Destructor.
void	attachMatrix (double *gmat, int nn)
	Attach the left side of the equation system to the current object and represent the matrix as a sparse matrix.
void	precondRILU (double relaxfac)
	Prepare for preconditioning.
int	solve (double *ex, double *eb, int nn)
	Solve the equation system by conjugate gradient method.
void	setTolerance (double tolerance=1.0e-6)
	Set numerical tolerance used by the solver.
void	setMaxIterations (int max_iterations)
	Set the maximal number of iterations to be used by the solver.

---

Detailed Description

Solve the equation system Ax=b where A is a symmetric positive definite matrix using the Conjugate Gradient Method.

Definition at line 59 of file SolveCG.h.

---

Member Function Documentation

void Go::SolveCG::attachMatrix	(	double *	gmat,
		int	nn
	)

Attach the left side of the equation system to the current object and represent the matrix as a sparse matrix.

No test is applied on whether the matrix really is symmetric and positive definite.

Parameters:

	gmat	the system matrix for the linear equations. Size is nn*nn, stored columnwise.
	nn	the number of unknowns in the system.

void Go::SolveCG::precondRILU

(

double

relaxfac

)

Prepare for preconditioning.

Parameters:

relaxfac

relaxation parameter

int Go::SolveCG::solve	(	double *	ex,
		double *	eb,
		int	nn
	)

Solve the equation system by conjugate gradient method.

Parameters:

	ex	the solution vector. The input should be the initial guess. Size is equal to nn.
	eb	the right side of the equation. Size is equal to nn.
	nn	the number of unknowns int the system.

Returns:

0: success, 1: iterationcount exceeded, < 0: error.

void Go::SolveCG::setTolerance

(

double

tolerance = 1.0e-6

)

[inline]

Set numerical tolerance used by the solver.

Parameters:

tolerance

numerical tolerance.

Definition at line 92 of file SolveCG.h.

void Go::SolveCG::setMaxIterations

(

int

max_iterations

)

[inline]

Set the maximal number of iterations to be used by the solver.

Parameters:

max_iterations

the maximal number of iterations.

Definition at line 97 of file SolveCG.h.

---

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

• SolveCG.h

---