solver_option2rheolef

Overview

This class implements a set of options for direct or iterative solvers, as implemented by the solver(2). An instance of this class can be supplied to the solver constructor.

Options

iterative

Indicate if we want to use an iterative solver or a a direct one. Default is direct for 1D or 2D problems and iterative for 3D ones. Direct solver is mumps when available, otherwise umfpack or builtin solver when none of these solvers are available at Rheolef configure time. Iterative solver is cg for symmetric positive definite systems and gmres otherwise.

Options for iterative solvers

tol

Tolerance for the stopping criterion. Default is the machine epsilon for the default Float type. The default Float type is double defined at Rheolef configure time and is double by default, when no special configure option is used.

max_iter

Maximum number of iteration when using iterative method.

absolute_stopping

Absolute or relative stopping criterion. With the absolute criterion, the algorithm stops when norm(A*x-b) < tol, otherwise it stops when norm(A*x-b) < tol*norm(b). Default is to use an absolute stopping criterion.

residue

On return, gives the obtained residue, optionally divided by the initial one when using a relatiive stopping. It is less or equal to tol when the iterative algorithm stops with succes.

n_iter

On return, gives the number of iterations performed. It is always less or equal to max_iter when the iterative algorithm stops with succes.

p_err

A pointer to the odiststeam standard error where residues are printed during iterations. When this pointer is zero, no errors are printed. Default is to print to derr.

label

When printing errors, each line is prefixed by [@var{label]}. When the label is empty, each iterative algorithm uses its default label, e.g. "cg" for the conjugate gradient cg. By default the label is empty and this option is used to superset the algorithm default.

krylov_dimension

The dimension of the Krylov space used by the gmres algorithm. Default is krylov_dimension=6.

Options for direct solvers

n_refinement

Number of iterative refinement, when using direct method (umfpack only support it). This option is only active when using iterative solvers.

compute_determinant

Compute also the determinant of the matrix. This option is only active when using direct solvers. Requires mumps or umfpack.

prefered_library

When both mumps and umfpack are available, then mumps is the default. This option allows one to force to choose umfpack, by setting the string to "umfpack". This option is only active when using direct solvers.

Options for developpers

verbose_level

Can be set to 0, 1, 2 or 3. The default is 0.

level_of_fill

Built an incomplete factorization with the prescribed level of fill [1:5].

do_check

Performs extra checks for debug.

force_seq

In distributed mode, restrict the linear system resolution to diagonal blocs per process. This option is only active when using the mumps direct solver.

ooc

Out-of-core limit (Mo/percent depending on compilation options). In development.

amalgamation

Level of amalgamation [10:70] for Kass. In development.

Implementation

class solver_option {
public:
  typedef std::size_t size_type;
  static const long int  decide = -1;
  mutable long int   iterative;
  Float              tol;
  size_type          max_iter;
  bool               absolute_stopping;
  mutable Float      residue;
  mutable size_type  n_iter;
  odiststream*       p_err;
  mutable std::string label;
  size_type          krylov_dimension;
  size_type          n_refinement;
  bool               compute_determinant;
  std::string        prefered_library;
  size_type          verbose_level;
  bool               do_check;
  bool               force_seq;
  size_type          level_of_fill;
  size_type          amalgamation;
  size_type          ooc;
// allocator and default values:
  solver_option()
   : iterative     (decide),
#if defined(_RHEOLEF_HAVE_FLOAT128)
     tol           (1e6*std::numeric_limits<Float>::epsilon()),
#else
     tol           (std::numeric_limits<Float>::epsilon()),
#endif
     max_iter      (100000),
     absolute_stopping (true),
     residue       (0),
     n_iter        (0),
     p_err         (&derr),
     label         (),
     krylov_dimension (6),
     n_refinement  (2),
     compute_determinant(false),
     prefered_library(),
     verbose_level (0),
     do_check      (false),
     force_seq     (false),
     level_of_fill (1),
     amalgamation  (10),
     ooc           (20000)
  {
  }
  solver_option (const solver_option&);
  solver_option& operator= (const solver_option&);
};