[P]arallel [Hi]gh-order [Li]brary for [P]DEs  Latest
Parallel High-Order Library for PDEs through hp-adaptive Discontinuous Galerkin methods
PHiLiP::Parameters::LinearSolverParam Class Reference

Parameters related to the linear solver. More...

#include <parameters_linear_solver.h>

Collaboration diagram for PHiLiP::Parameters::LinearSolverParam:

Public Types

enum  LinearSolverEnum { direct, gmres }
 Types of linear solvers available. More...
 

Public Member Functions

void parse_parameters (dealii::ParameterHandler &prm)
 Parses input file and sets the variables.
 

Static Public Member Functions

static void declare_parameters (dealii::ParameterHandler &prm)
 Declares the possible variables and sets the defaults.
 

Public Attributes

OutputEnum linear_solver_output
 Can either be verbose or quiet. More...
 
LinearSolverEnum linear_solver_type
 direct or gmres.
 
double ilut_drop
 Threshold to drop terms close to zero.
 
double ilut_rtol
 Multiplies diagonal by ilut_rtol for more diagonal dominance.
 
double ilut_atol
 Add ilu_rtol to diagonal for more diagonal dominance.
 
int ilut_fill
 ILU fill-in.
 
double linear_residual
 Tolerance for linear residual.
 
int max_iterations
 Maximum number of linear iteration.
 
int restart_number
 Number of iterations before restarting GMRES.
 
double newton_residual
 Tolerance for Newton iteration residual (for Jacobian-free Newton-Krylov)
 
int newton_max_iterations
 Maximum number of Newton iterations (for Jacobian-free Newton-Krylov)
 
double perturbation_magnitude
 Small perturbation magnitude for Jacobian-free methods.
 

Detailed Description

Parameters related to the linear solver.

Definition at line 13 of file parameters_linear_solver.h.

Member Enumeration Documentation

◆ LinearSolverEnum

Types of linear solvers available.

Enumerator
gmres 

LU.

GMRES.

Definition at line 17 of file parameters_linear_solver.h.

Member Data Documentation

◆ linear_solver_output

OutputEnum PHiLiP::Parameters::LinearSolverParam::linear_solver_output

Can either be verbose or quiet.

Verbose will print the full dense matrix. Will not work for large matricesquiet or verbose.

Definition at line 25 of file parameters_linear_solver.h.


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