preconditioned_conjugate_gradient.c File Reference

#include "turtle.h"
#include "t_utilities.h"
#include "doublevector_utilities.h"
#include "preconditioned_conjugate_gradient.h"

Go to the source code of this file.

Macros
#define	MAX_VALUE_DIAG   1e-10
#define	MAX_REITERTION   1

Functions
int	get_diagonal (DOUBLEVECTOR *diagonal, t_Matrix_element Matrix)
long	jacobi_preconditioned_conjugate_gradient_search (long icnt, double epsilon, DOUBLEVECTOR *x, DOUBLEVECTOR *b, t_Matrix_element funz)

Variables
t_Matrix	Functional_M

Macro Definition Documentation

#define MAX_REITERTION   1

Definition at line 36 of file preconditioned_conjugate_gradient.c.

Referenced by jacobi_preconditioned_conjugate_gradient_search().

#define MAX_VALUE_DIAG   1e-10

Definition at line 35 of file preconditioned_conjugate_gradient.c.

Function Documentation

int get_diagonal	(	DOUBLEVECTOR *	diagonal,
		t_Matrix_element	Matrix
	)

Definition at line 41 of file preconditioned_conjugate_gradient.c.

References free_doublevector(), new_doublevector(), DOUBLEVECTOR::nh, and DOUBLEVECTOR::nl.

Referenced by jacobi_preconditioned_conjugate_gradient_search(), and time_loop().

long jacobi_preconditioned_conjugate_gradient_search	(	long	icnt,
		double	epsilon,
		DOUBLEVECTOR *	x,
		DOUBLEVECTOR *	b,
		t_Matrix_element	funz
	)

Parameters

icnt	- (long) number of reiterations
epsilon	- (double) required tollerance (2-order norm of the residuals)
x	- (DOUBLEVECTOR *) vector of the unknowns x in Ax=b
b	- (DOUBLEVECTOR *) vector of b in Ax=b
funz	- (t_Matrix_element) - (int) pointer to the application A (x and y doublevector y=A(param)x ) it return 0 in case of success, -1 otherwise.

algorithm proposed by Jonathan Richard Shewckuck in http://www.cs.cmu.edu/~jrs/jrspapers.html#cg and http://www.cs.cmu.edu/~quake-papers/painless-conjugate-gradient.pdf

Author

Emanuele Cordano

Date

June 2009

Returns

the number of reitarations

Definition at line 80 of file preconditioned_conjugate_gradient.c.

References free_doublevector(), get_diagonal(), max_doublevector(), MAX_REITERTION, new_doublevector(), DOUBLEVECTOR::nh, DOUBLEVECTOR::nl, and stop_execution().

Referenced by Newton_convergence().

Variable Documentation

t_Matrix Functional_M

Definition at line 39 of file preconditioned_conjugate_gradient.c.