GmpArmadilloMatrix Class Reference

A class for wrapping up an Armadillo sparse matrix sharing the memory with a GeMA sparse matrix. This setup is better than simply using
directly an Armadillo sparse matrix since creating a GeMA sparse matrix is far more efficient than creating an Armadillo one. More...

#include <gmpArmadilloMatrix.h>

Inheritance diagram for GmpArmadilloMatrix:

Collaboration diagram for GmpArmadilloMatrix:

Public Member Functions
	GmpArmadilloMatrix (int nlin, const GmpArmadilloMatrix *sharedLayout, GmSparseMatrixOptions options)
	Constructor for the square matrix used by the Armadillo solver. If sharedLayout is different from NULL, this matrix layout will be given by it.
virtual	~GmpArmadilloMatrix ()
	Destructor.
const arma::sp_mat &	armaMat () const
	Returns the Armadillo matrix wrapped by this class.
virtual void	clear (bool keepSparseLayout)
virtual void	clearLineAndColumnSet (const QVector< int > &indexList, bool setDiagonal, bool keepSparseLayout)
virtual void	ensureDiagonal ()
Public Member Functions inherited from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >
	GmSparseMatrix (int n, const GmSparseMatrix< IndexType, T > *sharedLayout, GmSparseMatrixOptions options, bool arma)
virtual bool	supportsBatchInsert () const
virtual bool	supportsSparseLayouts () const
virtual bool	supportsRandomSet () const
virtual ParallelAddMode	supportedParallelAddMode () const
virtual bool	beginBatchInsert (size_t expectedEntries=0)
virtual bool	endBatchInsert (bool discardData=false)
virtual GmSparseMatrixLayoutBuilder *	layoutBuilder () const
virtual bool	emptyLayout () const
virtual bool	symmetric () const
virtual void	setSymmetric (bool sym)
virtual bool	isSymmetric (double relTol=GM_DOUBLECMP_RELTOL, double absTol=GM_DOUBLECMP_ABSTOL) const
virtual int	nlin () const
virtual int	ncol () const
virtual int	layoutSize () const
virtual double	at (int lin, int col) const
virtual bool	inLayout (int lin, int col) const
virtual void	set (int lin, int col, double value)
virtual void	set (GmSolverMatrix *A, GmSolverMatrix *B, double c, bool sameStructure)
virtual void	add (int lin, int col, double value)
virtual void	matAdd (const GmVector &a, const GmVector &b, double zeroTol=0.0)
virtual void	mul (const GmVector &a, GmVector &b) const
virtual void	mulAdd (const GmVector &a, GmVector &b) const
virtual void	mulSub (const GmVector &a, GmVector &b) const
virtual void	columnMulAdd (int col, GmVector &f, double v, const bool *skipRows=NULL) const
virtual size_t	usedMemory () const
virtual bool	hasNanInf () const
virtual bool	saveMatrixCoordinatesToFile (FILE *f) const
Public Member Functions inherited from GmSolverMatrix
double	operator() (int lin, int col) const
void	print (const GmLogCategory &logger, GmLogLevel level, int fieldWidth=0, char format='g', int precision=-1) const

Protected Member Functions
virtual void	matrixLayoutReady ()
	Allocates space for matrix values and build the Armadillo wrapper over matrix vectors.
virtual GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT > *	detachAndClear ()
Protected Member Functions inherited from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >
void	addBatch (int lin, int col, double value)
void	addToPtrIndexFromThread (int tid, int ptr, int index, double value)
GmSparseMatrixLayoutBuilder *	createLayoutBuilder (GmSparseMatrixLayoutMode mode, GmCSxSparseMatrixLayout< IndexType, T > *layout)

Private Attributes
arma::sp_mat	_armaMat
	An Armadillo matrix created sharing vectors with the base class memory.

Additional Inherited Members
Public Types inherited from GmSolverMatrix
enum	ParallelAddMode
Protected Attributes inherited from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >
GmSparseMatrixOptions	_opt
bool	_batch
int	_n
QAtomicInt	_sym
double *	_values
const GmCSxSparseMatrixLayout< IndexType, T > *	_layout
GmCSxSparseMatrixLayout< IndexType, T > *	_masterLayout
GmSparseMatrixLayoutBuilder *	_layoutBuilder
GmSparseMatrixTripletBuffer< T > *	_batchData

Detailed Description

A class for wrapping up an Armadillo sparse matrix sharing the memory with a GeMA sparse matrix. This setup is better than simply using
directly an Armadillo sparse matrix since creating a GeMA sparse matrix is far more efficient than creating an Armadillo one.

Constructor & Destructor Documentation

GmpArmadilloMatrix()

GmpArmadilloMatrix::GmpArmadilloMatrix	(	int	nlin,
		const GmpArmadilloMatrix *	sharedLayout,
		GmSparseMatrixOptions	options )

Constructor for the square matrix used by the Armadillo solver. If sharedLayout is different from NULL, this matrix layout will be given by it.

IMPORTANT: When sharing a layout with another matrix, this matrix will become invalid if the matrix owning the layout is destroyed.

Member Function Documentation

clear()

void GmpArmadilloMatrix::clear ( bool keepSparseLayout )

virtual

Reimplemented from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >.

clearLineAndColumnSet()

void GmpArmadilloMatrix::clearLineAndColumnSet ( const QVector< int > & indexList, bool setDiagonal, bool keepSparseLayout )

virtual

Reimplemented from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >.

detachAndClear()

GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT > * GmpArmadilloMatrix::detachAndClear ( )

protectedvirtual

Reimplemented from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >.

ensureDiagonal()

void GmpArmadilloMatrix::ensureDiagonal ( )

virtual

Reimplemented from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >.

matrixLayoutReady()

void GmpArmadilloMatrix::matrixLayoutReady ( )

protectedvirtual

Allocates space for matrix values and build the Armadillo wrapper over matrix vectors.

Reimplemented from GmSparseMatrix< arma::uword, GM_CSC_SPARSE_FORMAT >.

---

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

• inc/gmpArmadilloMatrix.h
• src/gmpArmadilloMatrix.cpp