SparseDirectMUMPS Class Reference

#include <sparse_direct.h>

Public Types
typedef types::global_dof_index	size_type

Public Member Functions
	SparseDirectMUMPS ()
	~SparseDirectMUMPS ()
	DeclException0 (ExcInitializeAlreadyCalled)
template<class Matrix >
void	initialize (const Matrix &matrix, const Vector< double > &vector)
template<class Matrix >
void	initialize (const Matrix &matrix)
void	solve (Vector< double > &vector)
void	vmult (Vector< double > &dst, const Vector< double > &src)

Private Member Functions
template<class Matrix >
void	initialize_matrix (const Matrix &matrix)
void	copy_solution (Vector< double > &vector)
void	copy_rhs_to_mumps (const Vector< double > &rhs)

Private Attributes
DMUMPS_STRUC_C	id
double *	a
std::vector< double >	rhs
int *	irn
int *	jcn
types::global_dof_index	n
types::global_dof_index	nz
bool	initialize_called

Detailed Description

This class provides an interface to the parallel sparse direct solver MUMPS. MUMPS is direct method based on a multifrontal approach, which performs a direct LU factorization. The matrix coming in may have either symmetric or nonsymmetric sparsity pattern.

Note

This class is useable if and only if a working installation of MUMPS exists on your system and was detected during configuration of deal.II.

Instantiations

There are instantiations of this class for SparseMatrix<double>, SparseMatrix<float>, BlockSparseMatrix<double>, and BlockSparseMatrix<float>.

Author

Markus Buerg, 2010

Definition at line 340 of file sparse_direct.h.

Member Typedef Documentation

typedef types::global_dof_index SparseDirectMUMPS::size_type

Declare type for container size.

Definition at line 382 of file sparse_direct.h.

Constructor & Destructor Documentation

SparseDirectMUMPS::SparseDirectMUMPS

(

)

Constructor

SparseDirectMUMPS::~SparseDirectMUMPS

(

)

Destructor

Member Function Documentation

template<class Matrix >

void SparseDirectMUMPS::initialize_matrix ( const Matrix &  matrix )

private

This function initializes a MUMPS instance and hands over the system's matrix matrix.

void SparseDirectMUMPS::copy_solution ( Vector< double > &  vector )

private

Copy the computed solution into the solution vector.

SparseDirectMUMPS::DeclException0

(

ExcInitializeAlreadyCalled

)

Exception

template<class Matrix >

void SparseDirectMUMPS::initialize	(	const Matrix &	matrix,
		const Vector< double > &	vector
	)

This function initializes a MUMPS instance and hands over the system's matrix matrix and right-hand side vector to the solver.

template<class Matrix >

void SparseDirectMUMPS::initialize

(

const Matrix &

matrix

)

This function initializes a MUMPS instance and computes the factorization of the system's matrix matrix.

void SparseDirectMUMPS::solve

(

Vector< double > &

vector

)

A function in which the linear system is solved and the solution vector is copied into the given vector. The right-hand side need to be supplied in initialize(matrix, vector);

void SparseDirectMUMPS::vmult	(	Vector< double > &	dst,
		const Vector< double > &	src
	)

A function in which the inverse of the matrix is applied to the input vector src and the solution is written into the output vector dst.

Member Data Documentation

bool SparseDirectMUMPS::initialize_called

private

Flags storing whether the function initialize () has already been called.

Definition at line 376 of file sparse_direct.h.

---

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

• include/deal.II/lac/sparse_direct.h