IterativeLinSolver< T > Class Template Reference

The iterative linear solver base class template. More...

#include <LinSolver.hpp>

Inheritance diagram for IterativeLinSolver< T >:

[legend]Collaboration diagram for IterativeLinSolver< T >:

[legend]List of all members.

Public Member Functions
	IterativeLinSolver (double the_reduction, unsigned the_max_iter, double the_abs_prec)
	The explicit constructor.
	IterativeLinSolver (SGParam *list)
	Constructor for reading the halting criteria from SG parameters.
Protected Attributes
double	reduction
unsigned	max_iter
double	abs_prec

---

Detailed Description

template<typename T>
class IterativeLinSolver< T >

The iterative linear solver base class template.

This is the base class template for generic iterative solvers for large sparse linear systems. The only difference from LinSolver<T> is that this class has fields for storing the iteration parameters like the max. number of the iterations and the desired reduction. These parameters can be read form SG parameters using standard methods of this class.

Note:

SGDisplay is a virtual base class and its constructor should be called in the derived class.

---

Constructor & Destructor Documentation

template<typename T>

IterativeLinSolver< T >::IterativeLinSolver	(	double	the_reduction,
		unsigned	the_max_iter,
		double	the_abs_prec
	)			[inline]

The explicit constructor.

Parameters:

[in]	the_reduction	The desired reduction (0 to do 'max_iter' iter.)
[in]	the_max_iter	The maximum number of iterations
[in]	the_abs_prec	The minimum norm of the residual

template<typename T>

IterativeLinSolver< T >::IterativeLinSolver

(

SGParam *

list

)

[inline]

Constructor for reading the halting criteria from SG parameters.

The parameter names are:

• LinReduction - minimal reduction factor for the defect,
• LinMaxIter - maximum number of iterations to do,
• LinAbsPrec - minimum norm of the residual to achieve.

Parameters:

[in]

list

The list of the SG parameters

---

Member Data Documentation

template<typename T>

double IterativeLinSolver< T >::reduction [protected]

the desired reduction of the residual (0 to do 'max_iter' iter.)

template<typename T>

unsigned IterativeLinSolver< T >::max_iter [protected]

the maximum number of iterations to compute

template<typename T>

double IterativeLinSolver< T >::abs_prec [protected]

the minimum norm of the residual

---

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

• LinSolver.hpp
• LinSolver.cc

---