A two-step version of an overlapping Schwarz preconditioner using one step ILU0 as. More...

#include <ParallelOverlappingILU0.hpp>

Inheritance diagram for Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >:

Classes
struct	CRS

Public Types
using	matrix_type = typename std::remove_const< Matrix >::type
	The matrix type the preconditioner is for.

using	domain_type = Domain
	The domain type of the preconditioner.

using	range_type = Range
	The range type of the preconditioner.

using	field_type = typename Domain::field_type
	The field type of the preconditioner.

using	block_type = typename matrix_type::block_type

using	size_type = typename matrix_type::size_type

Public Member Functions
virtual bool	hasPerfectUpdate () const override

Dune::SolverCategory::Category	category () const override

	ParallelOverlappingILU0 (const Matrix &A, const int n, const field_type w, MILU_VARIANT milu, bool redblack=false, bool reorder_sphere=true)
	Constructor.

	ParallelOverlappingILU0 (const Matrix &A, const ParallelInfo &comm, const int n, const field_type w, MILU_VARIANT milu, bool redblack=false, bool reorder_sphere=true)
	Constructor gets all parameters to operate the prec.

	ParallelOverlappingILU0 (const Matrix &A, const field_type w, MILU_VARIANT milu, bool redblack=false, bool reorder_sphere=true)
	Constructor.

	ParallelOverlappingILU0 (const Matrix &A, const ParallelInfo &comm, const field_type w, MILU_VARIANT milu, bool redblack=false, bool reorder_sphere=true)
	Constructor.

	ParallelOverlappingILU0 (const Matrix &A, const ParallelInfo &comm, const field_type w, MILU_VARIANT milu, size_type interiorSize, bool redblack=false, bool reorder_sphere=true)
	Constructor.

void	pre (Domain &, Range &) override
	Prepare the preconditioner.

void	apply (Domain &v, const Range &d) override
	Apply the preconditoner.

template<class V >
void	copyOwnerToAll (V &v) const

void	post (Range &) override
	Clean up.

void	update () override

Protected Member Functions
Range &	reorderD (const Range &d)
	Reorder D if needed and return a reference to it.

Domain &	reorderV (Domain &v)
	Reorder V if needed and return a reference to it.

void	reorderBack (const Range &reorderedV, Range &v)

Protected Attributes
std::unique_ptr< Matrix >	ILU_
	The ILU0 decomposition of the matrix.

CRS	lower_

CRS	upper_

std::vector< block_type >	inv_

std::vector< std::size_t >	ordering_
	the reordering of the unknowns

Range	reorderedD_
	The reordered right hand side.

Domain	reorderedV_
	The reordered left hand side.

const ParallelInfo *	comm_

const field_type	w_
	The relaxation factor to use.

const bool	relaxation_

size_type	interiorSize_

const Matrix *	A_

int	iluIteration_

MILU_VARIANT	milu_

bool	redBlack_

bool	reorderSphere_

Detailed Description

template<class Matrix, class Domain, class Range, class ParallelInfoT>
class Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >

A two-step version of an overlapping Schwarz preconditioner using one step ILU0 as.

This preconditioner differs from a ParallelRestrictedOverlappingSchwarz with Dune:SeqILU0 in the following way: During apply we make sure that the current residual is consistent (i.e. each process knows the same value for each index. Then we solve Ly = d for y and make y consistent again. Last we solve Ux = y and make sure that x is consistent. In contrast for ParallelRestrictedOverlappingSchwarz we solve (LU)x = d for x without forcing consistency between the two steps.

Template Parameters

Matrix	The type of the Matrix.
Domain	The type of the Vector representing the domain.
Range	The type of the Vector representing the range.
ParallelInfo	The type of the parallel information object used, e.g. Dune::OwnerOverlapCommunication

Constructor & Destructor Documentation

◆ ParallelOverlappingILU0() [1/5]

template<class Matrix , class Domain , class Range , class ParallelInfoT >

Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::ParallelOverlappingILU0	(	const Matrix &	A,
		const int	n,
		const field_type	w,
		MILU_VARIANT	milu,
		bool	redblack = `false`,
		bool	reorder_sphere = `true`
	)

Constructor.

Constructor gets all parameters to operate the prec.

Parameters

A	The matrix to operate on.
n	ILU fill in level (for testing). This does not work in parallel.
w	The relaxation factor.
milu	The modified ILU variant to use. 0 means traditional ILU.

See also: MILU_VARIANT.

Parameters

redblack	Whether to use a red-black ordering.
reorder_sphere	If true, we start the reordering at a root node. The vertices on each layer aound it (same distance) are ordered consecutivly. If false, we preserver the order of the vertices with the same color.

◆ ParallelOverlappingILU0() [2/5]

template<class Matrix , class Domain , class Range , class ParallelInfoT >

Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::ParallelOverlappingILU0	(	const Matrix &	A,
		const ParallelInfo &	comm,
		const int	n,
		const field_type	w,
		MILU_VARIANT	milu,
		bool	redblack = `false`,
		bool	reorder_sphere = `true`
	)

Constructor gets all parameters to operate the prec.

Parameters

A	The matrix to operate on.
comm	communication object, e.g. Dune::OwnerOverlapCopyCommunication
n	ILU fill in level (for testing). This does not work in parallel.
w	The relaxation factor.
milu	The modified ILU variant to use. 0 means traditional ILU.

See also: MILU_VARIANT.

Parameters

redblack	Whether to use a red-black ordering.
reorder_sphere	If true, we start the reordering at a root node. The vertices on each layer aound it (same distance) are ordered consecutivly. If false, we preserver the order of the vertices with the same color.

◆ ParallelOverlappingILU0() [3/5]

template<class Matrix , class Domain , class Range , class ParallelInfoT >

Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::ParallelOverlappingILU0	(	const Matrix &	A,
		const field_type	w,
		MILU_VARIANT	milu,
		bool	redblack = `false`,
		bool	reorder_sphere = `true`
	)

Constructor.

Constructor gets all parameters to operate the prec.

Parameters

A	The matrix to operate on.
w	The relaxation factor.
milu	The modified ILU variant to use. 0 means traditional ILU.

See also: MILU_VARIANT.

Parameters

redblack	Whether to use a red-black ordering.
reorder_sphere	If true, we start the reordering at a root node. The vertices on each layer aound it (same distance) are ordered consecutivly. If false, we preserver the order of the vertices with the same color.

◆ ParallelOverlappingILU0() [4/5]

template<class Matrix , class Domain , class Range , class ParallelInfoT >

Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::ParallelOverlappingILU0	(	const Matrix &	A,
		const ParallelInfo &	comm,
		const field_type	w,
		MILU_VARIANT	milu,
		bool	redblack = `false`,
		bool	reorder_sphere = `true`
	)

Constructor.

Constructor gets all parameters to operate the prec.

Parameters

A	The matrix to operate on.
comm	communication object, e.g. Dune::OwnerOverlapCopyCommunication
w	The relaxation factor.
milu	The modified ILU variant to use. 0 means traditional ILU.

See also: MILU_VARIANT.

Parameters

redblack	Whether to use a red-black ordering.
reorder_sphere	If true, we start the reordering at a root node. The vertices on each layer aound it (same distance) are ordered consecutivly. If false, we preserver the order of the vertices with the same color.

◆ ParallelOverlappingILU0() [5/5]

template<class Matrix , class Domain , class Range , class ParallelInfoT >

Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::ParallelOverlappingILU0	(	const Matrix &	A,
		const ParallelInfo &	comm,
		const field_type	w,
		MILU_VARIANT	milu,
		size_type	interiorSize,
		bool	redblack = `false`,
		bool	reorder_sphere = `true`
	)

Constructor.

Constructor gets all parameters to operate the prec.

Parameters

A	The matrix to operate on.
n	ILU fill in level (for testing). This does not work in parallel.
w	The relaxation factor.
milu	The modified ILU variant to use. 0 means traditional ILU.

See also: MILU_VARIANT.

Parameters

interiorSize	The number of interior/owner rows in the matrix.
redblack	Whether to use a red-black ordering.
reorder_sphere	If true, we start the reordering at a root node. The vertices on each layer aound it (same distance) are ordered consecutivly. If false, we preserver the order of the vertices with the same color.

Member Function Documentation

◆ apply()

template<class Matrix , class Domain , class Range , class ParallelInfoT >

void Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::apply	(	Domain &	v,
		const Range &	d
	)

override

Apply the preconditoner.

◆ hasPerfectUpdate()

template<class Matrix , class Domain , class Range , class ParallelInfoT >

virtual bool Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::hasPerfectUpdate ( ) const

inlineoverridevirtual

Implements Dune::PreconditionerWithUpdate< Domain, Range >.

◆ post()

template<class Matrix , class Domain , class Range , class ParallelInfoT >

void Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::post ( Range & )

inlineoverride

Clean up.

◆ pre()

template<class Matrix , class Domain , class Range , class ParallelInfoT >

void Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::pre	(	Domain &	,
		Range &
	)

inlineoverride

Prepare the preconditioner.

◆ update()

template<class Matrix , class Domain , class Range , class ParallelInfoT >

void Opm::ParallelOverlappingILU0< Matrix, Domain, Range, ParallelInfoT >::update ( )

overridevirtual

Implements Dune::PreconditionerWithUpdate< Domain, Range >.

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

opm/simulators/linalg/ParallelOverlappingILU0.hpp
opm/simulators/linalg/ParallelOverlappingILU0_impl.hpp

Classes

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ ParallelOverlappingILU0() [1/5]

◆ ParallelOverlappingILU0() [2/5]

◆ ParallelOverlappingILU0() [3/5]

◆ ParallelOverlappingILU0() [4/5]

◆ ParallelOverlappingILU0() [5/5]

Member Function Documentation

◆ apply()

◆ hasPerfectUpdate()

◆ post()

◆ pre()

◆ update()