Ifpack_SparseContainer: a class for storing and solving linear systems using sparse matrices. More...

#include <Ifpack_SparseContainer.h>

Inheritance diagram for Ifpack_SparseContainer< T >:

Collaboration diagram for Ifpack_SparseContainer< T >:

Public Member Functions
	Ifpack_SparseContainer (const int NumRows, const int NumVectors=1)
	Constructor.

	Ifpack_SparseContainer (const Ifpack_SparseContainer< T > &rhs)
	Copy constructor.

virtual	~Ifpack_SparseContainer ()
	Destructor.

Ifpack_SparseContainer &	operator= (const Ifpack_SparseContainer< T > &rhs)
	Operator =.

virtual int	NumRows () const
	Returns the number of rows of the matrix and LHS/RHS.

virtual int	NumVectors () const
	Returns the number of vectors in LHS/RHS.

virtual int	SetNumVectors (const int NumVectors_in)
	Sets the number of vectors for LHS/RHS.

virtual double &	LHS (const int i, const int Vector=0)
	Returns the i-th component of the vector Vector of LHS.

virtual double &	RHS (const int i, const int Vector=0)
	Returns the i-th component of the vector Vector of RHS.

virtual int &	ID (const int i)
	Returns the ID associated to local row i.

virtual int	SetMatrixElement (const int row, const int col, const double value)
	Set the matrix element (row,col) to `value`.

virtual bool	IsInitialized () const
	Returns `true` is the container has been successfully initialized.

virtual bool	IsComputed () const
	Returns `true` is the container has been successfully computed.

virtual int	SetParameters (Teuchos::ParameterList &List)
	Sets all necessary parameters.

virtual const char *	Label () const
	Returns the label of this container.

Teuchos::RCP< const Epetra_Map >	Map () const
	Returns a pointer to the internally stored map.

Teuchos::RCP< const Epetra_MultiVector >	LHS () const
	Returns a pointer to the internally stored solution multi-vector.

Teuchos::RCP< const Epetra_MultiVector >	RHS () const
	Returns a pointer to the internally stored rhs multi-vector.

Teuchos::RCP< const Epetra_CrsMatrix >	Matrix () const
	Returns a pointer to the internally stored matrix.

const Epetra_IntSerialDenseVector &	ID () const
	Returns a pointer to the internally stored ID's.

Teuchos::RCP< const T >	Inverse () const
	Returns a pointer to the internally stored inverse operator.

virtual int	Initialize ()
	Initializes the container, by completing all the operations based on matrix structure.

virtual int	Compute (const Epetra_RowMatrix &Matrix_in)
	Finalizes the linear system matrix and prepares for the application of the inverse.

virtual int	Apply ()
	Apply the matrix to RHS, result is stored in LHS.

virtual int	ApplyInverse ()
	Apply the inverse of the matrix to RHS, result is stored in LHS.

virtual int	Destroy ()
	Destroys all data.

virtual double	InitializeFlops () const
	Returns the flops in Compute().

virtual double	ComputeFlops () const
	Returns the flops in Compute().

virtual double	ApplyFlops () const
	Returns the flops in Apply().

virtual double	ApplyInverseFlops () const
	Returns the flops in ApplyInverse().

virtual std::ostream &	Print (std::ostream &os) const
	Prints basic information on iostream. This function is used by operator<<.

Public Member Functions inherited from Ifpack_Container
virtual	~Ifpack_Container ()
	Destructor.

Detailed Description

template<typename T>
class Ifpack_SparseContainer< T >

Ifpack_SparseContainer: a class for storing and solving linear systems using sparse matrices.

To understand what an IFPACK container is, please refer to the documentation of the pure virtual class Ifpack_Container. Currently, containers are used by class Ifpack_BlockRelaxation.

Using block methods, one needs to store all diagonal blocks and to be also to apply the inverse of each diagonal block. Using class Ifpack_DenseContainer, one can store the blocks as sparse matrices (Epetra_CrsMatrix), which can be advantageous when the blocks are large. Otherwise, class Ifpack_DenseContainer is probably more appropriate.

Sparse containers are templated with a type T, which represent the class to use in the application of the inverse. (T is not used in Ifpack_DenseContainer). In SparseContainer, T must be an Ifpack_Preconditioner derived class. The container will allocate a T object, use SetParameters() and Compute(), then use T every time the linear system as to be solved (using the ApplyInverse() method of T).

Author: Marzio Sala, SNL 9214.

Date: Last modified on Nov-04.

Definition at line 99 of file Ifpack_SparseContainer.h.

Constructor & Destructor Documentation

◆ Ifpack_SparseContainer() [1/2]

template<typename T >

Ifpack_SparseContainer< T >::Ifpack_SparseContainer	(	const int	NumRows,
		const int	NumVectors = `1`
	)

Constructor.

Definition at line 319 of file Ifpack_SparseContainer.h.

◆ Ifpack_SparseContainer() [2/2]

template<typename T >

Ifpack_SparseContainer< T >::Ifpack_SparseContainer ( const Ifpack_SparseContainer< T > & rhs )

Copy constructor.

Definition at line 338 of file Ifpack_SparseContainer.h.

References Ifpack_SparseContainer< T >::LHS(), Ifpack_SparseContainer< T >::Map(), Ifpack_SparseContainer< T >::Matrix(), and Ifpack_SparseContainer< T >::RHS().

◆ ~Ifpack_SparseContainer()

template<typename T >

Ifpack_SparseContainer< T >::~Ifpack_SparseContainer ( )

virtual

Destructor.

Definition at line 367 of file Ifpack_SparseContainer.h.

Member Function Documentation

◆ Apply()

template<typename T >

int Ifpack_SparseContainer< T >::Apply ( )

virtual

Apply the matrix to RHS, result is stored in LHS.

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ Ifpack_SparseContainer() [1/2]

◆ Ifpack_SparseContainer() [2/2]

◆ ~Ifpack_SparseContainer()

Member Function Documentation

◆ Apply()

◆ ApplyFlops()

◆ ApplyInverse()

◆ ApplyInverseFlops()

◆ Compute()

◆ ComputeFlops()

◆ Destroy()

◆ ID() [1/2]

◆ ID() [2/2]

◆ Initialize()

◆ InitializeFlops()

◆ Inverse()

◆ IsComputed()

◆ IsInitialized()

◆ Label()

◆ LHS() [1/2]

◆ LHS() [2/2]

◆ Map()

◆ Matrix()

◆ NumRows()

◆ NumVectors()

◆ Print()

◆ RHS() [1/2]

◆ RHS() [2/2]

◆ SetMatrixElement()

◆ SetNumVectors()

◆ SetParameters()