LocalOpt
 All Classes Files Functions Variables Friends
Support Class Reference

Supports for linear system solve and linear algebra. More...

#include <Support.h>

Public Member Functions

int solvesystem (const int max_iterations, const Graph *Mat, double *solution, const double *rhs, const double precision)
 The method used to solve systems. More...
 
int solveCG (const int max_iterations, const Graph *Mat, double *solution, const double *rhs, const double precision)
 The Conjugate Gradient method. Mat is assumed to be symmetric-positive-definite.
 
int solveCGNR (const int max_iterations, const Graph *Mat, double *solution, const double *rhs, const double precision)
 The Least-Squares Conjugate Gradient method. Mat is assumed to be symmetric and non-singular.
 
int solveMinRes (const int max_iterations, const Graph *Mat, double *solution, const double *rhs, const double precision, const bool show)
 The Minimal Residual method. Mat is assumed to be symmetric and non-singular.
 
void sum (double *lhs, const double *rhs, int dimension)
 
void subtract (double *lhs, const double *rhs, int dimension)
 
double norm (const double *x, int dimension)
 
void copyvec (double *lhs, const double *rhs, int dimension)
 
void mult_scalar (double *x, int dimension, const double scal)
 
double inner_prod (const double *x, const double *y, int dimension)
 
template<class ArrayType >
void print_array (ArrayType x, int length)
 

Detailed Description

Supports for linear system solve and linear algebra.

Member Function Documentation

int Support::solvesystem ( const int  max_iterations,
const Graph *  Mat,
double *  solution,
const double *  rhs,
const double  precision 
)

The method used to solve systems.

Where the type of system solve is determined, e.g., CG or MINRES or CGNR ...etc


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