Store and solve a local sparse linear problem. More...

#include <Ifpack2_SparseContainer_decl.hpp>

Inheritance diagram for Ifpack2::SparseContainer< MatrixType, InverseType >:

Public Member Functions
Constructor and destructor
	SparseContainer (const Teuchos::RCP< const row_matrix_type > &matrix, const Teuchos::Array< Teuchos::Array< LO > > &partitions, const Teuchos::RCP< const import_type > &importer, bool pointIndexed)
	Constructor. More...

virtual	~SparseContainer ()
	Destructor (declared virtual for memory safety of derived classes). More...

Get and set methods
virtual void	setParameters (const Teuchos::ParameterList &List)
	Set all necessary parameters. More...

Mathematical functions
virtual void	initialize ()
	Do all set-up operations that only require matrix structure. More...

virtual void	compute ()
	Initialize and compute all blocks. More...

void	clearBlocks ()

virtual void	apply (ConstHostView X, HostView Y, int blockIndex, Teuchos::ETransp mode=Teuchos::NO_TRANS, SC alpha=Teuchos::ScalarTraits< SC >::one(), SC beta=Teuchos::ScalarTraits< SC >::zero()) const
	Compute `Y := alpha * M^{-1} X + beta*Y`. More...

virtual void	weightedApply (ConstHostView X, HostView Y, ConstHostView W, int blockIndex, Teuchos::ETransp mode=Teuchos::NO_TRANS, SC alpha=Teuchos::ScalarTraits< SC >::one(), SC beta=Teuchos::ScalarTraits< SC >::zero()) const
	Compute `Y := alpha * diag(D) * M^{-1} (diag(D) * X) + beta*Y`. More...

Miscellaneous methods
virtual std::ostream &	print (std::ostream &os) const
	Print information about this object to the given output stream. More...

Implementation of Teuchos::Describable
virtual std::string	description () const
	A one-line description of this object. More...

virtual void	describe (Teuchos::FancyOStream &out, const Teuchos::EVerbosityLevel verbLevel=Teuchos::Describable::verbLevel_default) const
	Print the object with some verbosity level to the given FancyOStream. More...

Public Member Functions inherited from Ifpack2::ContainerImpl< MatrixType, InverseType::scalar_type >
virtual	~ContainerImpl ()
	Destructor. More...

virtual void	applyInverseJacobi (const mv_type &, mv_type &, SC dampingFactor, bool, int) const
	Compute `Y := (1 - a) Y + a D^{-1} (X - R*Y)`. More...

void	applyMV (const mv_type &X, mv_type &Y) const
	Wrapper for apply with MVs, used in unit tests (never called by BlockRelaxation) More...

void	weightedApplyMV (const mv_type &X, mv_type &Y, vector_type &W) const
	Wrapper for weightedApply with MVs, used in unit tests (never called by BlockRelaxation) More...

Public Member Functions inherited from Ifpack2::Container< MatrixType >
	Container (const Teuchos::RCP< const row_matrix_type > &matrix, const Teuchos::Array< Teuchos::Array< LO > > &partitions, bool pointIndexed)
	Constructor. More...

virtual	~Container ()
	Destructor. More...

Teuchos::ArrayView< const LO >	getBlockRows (int blockIndex) const
	Local indices of the rows of the input matrix that belong to this block. More...

void	setBlockSizes (const Teuchos::Array< Teuchos::Array< LO > > &partitions)
	Initialize arrays with information about block sizes. More...

bool	isInitialized () const
	Whether the container has been successfully initialized. More...

bool	isComputed () const
	Whether the container has been successfully computed. More...

Static Public Member Functions
static std::string	getName ()
	Get the name of this container type for Details::constructContainer() More...

Static Public Member Functions inherited from Ifpack2::ContainerImpl< MatrixType, InverseType::scalar_type >
static std::string	getName ()

Static Public Member Functions inherited from Ifpack2::Container< MatrixType >
static std::string	getName ()

Additional Inherited Members
Protected Types inherited from Ifpack2::ContainerImpl< MatrixType, InverseType::scalar_type >
using	local_scalar_type = InverseType::scalar_type

using	SC = typename Container< MatrixType >::scalar_type

using	LO = typename Container< MatrixType >::local_ordinal_type

using	GO = typename Container< MatrixType >::global_ordinal_type

using	NO = typename Container< MatrixType >::node_type

using	StridedRowView = Details::StridedRowView< SC, LO, GO, NO >

using	LSC = InverseType::scalar_type
	The internal representation of LocalScalarType in Kokkos::View. More...

using	LISC = typename Kokkos::ArithTraits< LSC >::val_type

using	local_mv_type = Tpetra::MultiVector< LSC, LO, GO, NO >

using	HostViewLocal = typename local_mv_type::dual_view_type::t_host

using	HostSubviewLocal = Kokkos::View< LISC **, Kokkos::LayoutStride, typename HostViewLocal::memory_space >

using	ConstHostSubviewLocal = Kokkos::View< const LISC **, Kokkos::LayoutStride, typename HostViewLocal::memory_space >

Protected Types inherited from Ifpack2::Container< MatrixType >
using	ISC = typename Kokkos::ArithTraits< SC >::val_type
	Internal representation of Scalar in Kokkos::View. More...

using	HostView = typename mv_type::dual_view_type::t_host

Protected Member Functions inherited from Ifpack2::ContainerImpl< MatrixType, InverseType::scalar_type >
LO	translateRowToCol (LO row)

Details::StridedRowView< SC, LO, GO, NO >	getInputRowView (LO row) const
	View a row of the input matrix. More...

void	DoGSBlock (ConstHostView X, HostView Y, HostView Y2, HostView Resid, SC dampingFactor, LO i) const
	Do one step of Gauss-Seidel on block i (used by DoGaussSeidel and DoSGS) More...

virtual void	solveBlock (ConstHostSubviewLocal X, HostSubviewLocal Y, int blockIndex, Teuchos::ETransp mode, const LSC alpha, const LSC beta) const

Protected Attributes inherited from Ifpack2::ContainerImpl< MatrixType, InverseType::scalar_type >
HostViewLocal	X_local_
	Scratch vectors used in apply(). More...

HostViewLocal	weightedApplyScratch_

std::vector< HostSubviewLocal >	X_localBlocks_
	Views for holding pieces of X corresponding to each block. More...

std::vector< HostSubviewLocal >	Y_localBlocks_
	Views for holding pieces of Y corresponding to each block. More...

Protected Attributes inherited from Ifpack2::Container< MatrixType >
Teuchos::RCP< const row_matrix_type >	inputMatrix_
	The input matrix to the constructor. More...

Teuchos::RCP< const crs_matrix_type >	inputCrsMatrix_
	The input matrix, dynamic cast to CrsMatrix. May be null. More...

Teuchos::RCP< const block_crs_matrix_type >	inputBlockMatrix_
	The input matrix, dynamic cast to BlockCrsMatrix. May be null. More...

int	numBlocks_
	The number of blocks (partitions) in the container. More...

Teuchos::Array< LO >	blockRows_
	Local indices of the rows of the input matrix that belong to this block. More...

Teuchos::Array< LO >	blockSizes_
	Number of rows in each block. More...

Teuchos::Array< LO >	blockOffsets_
	Starting index in blockRows_ of local row indices for each block. More...

Teuchos::RCP< vector_type >	Diag_
	Diagonal elements. More...

bool	IsParallel_
	Whether the problem is distributed across multiple MPI processes. More...

LO	NumLocalRows_
	Number of local rows in input matrix. More...

GO	NumGlobalRows_
	Number of global rows in input matrix. More...

GO	NumGlobalNonzeros_
	Number of nonzeros in input matrix. More...

bool	hasBlockCrs_
	Whether the input matrix is a BlockCRS matrix. More...

int	bcrsBlockSize_
	If hasBlockCrs_, the number of DOFs per vertex. Otherwise 1. More...

bool	pointIndexed_
	(If hasBlockCrs_) Whether the blocks are described using sub-block row indices instead of full block rows. More...

LO	scalarsPerRow_

bool	IsInitialized_
	If `true`, the container has been successfully initialized. More...

bool	IsComputed_
	If `true`, the container has been successfully computed. More...

Detailed Description

template<typename MatrixType, typename InverseType>
class Ifpack2::SparseContainer< MatrixType, InverseType >

Store and solve a local sparse linear problem.

Template Parameters

A	specialization of Tpetra::RowMatrix.

Please refer to the documentation of the Container interface. Currently, Containers are used by BlockRelaxation. Block relaxations need to be able to do two things:

Store the diagonal blocks
Solve linear systems with each diagonal block

These correspond to the two template parameters:

MatrixType, which stores a sparse matrix
InverseType, which solves linear systems with that matrix

This class stores each block as a sparse matrix. Using a sparse matrix for each block is a good idea when the blocks are large and sparse. For small and / or dense blocks, it would probably be better to use an implementation of Container that stores the blocks densely, like DenseContainer. You may also want to consider BandedContainer.

The InverseType template parameter represents the class to use for solving linear systems with a block. In SparseContainer, this template parameter must be a specialization of Preconditioner. Specifically, InverseType must implement the following methods:

A constructor that takes an RCP<const MatrixType>
setParameters(Teuchos::ParameterList&)
initialize()
compute()
apply (const mv_type& X, mv_type& Y, ...), where mv_type is the appropriate specialization of Tpetra::MultiVector

We also assume that InverseType has the following typedefs:

scalar_type
local_ordinal_type
global_ordinal_type
node_type

MatrixType and InverseType may store values of different types, and may have different template parameters (e.g., local or global ordinal types). You may mix and match so long as implicit conversions are available. The most obvious use case for this are:

MatrixType::global_ordinal_type=long long and InverseType::global_ordinal_type=short
MatrixType::scalar_type=float and InverseType::scalar_type=double

SparseContainer currently assumes the following about the column and row Maps of the input matrix:

On all processes, the column and row Maps begin with the same set of on-process entries, in the same order. That is, on-process row and column indices are the same.
On all processes, all off-process indices in the column Map of the input matrix occur after that initial set.

These assumptions may be violated if the input matrix is a Tpetra::CrsMatrix that was constructed with a user-provided column Map. The assumptions are not mathematically necessary and could be relaxed at any time. Implementers who wish to do so will need to modify the extract() method, so that it translates explicitly between local row and column indices, instead of just assuming that they are the same.