PardisoSolver Class Reference

creates solver class using pardiso for factorising, solving and selectively inverting linear system. More...

#include <PardisoSolver.h>

Inheritance diagram for PardisoSolver:

Public Member Functions
	PardisoSolver (int &MPI_rank)
	constructor. initialises parameters, check pardiso license.
void	symbolic_factorization (SpMat &Q, int &init)
	performs the symbolic factorisation.
void	factorize (SpMat &Q, double &log_det, double &t_priorLatChol)
	numerical factorisation.
void	factorize_w_constr (SpMat &Q, const MatrixXd &D, double &log_det, MatrixXd &V)
void	factorize_solve (SpMat &Q, Vect &rhs, Vect &sol, double &log_det, double &t_condLatChol, double &t_condLatSolve)
	factorises and solves matrix in one call (to reuse pardiso objects)
void	factorize_solve_w_constr (SpMat &Q, Vect &rhs, const MatrixXd &Dxy, double &log_det, Vect &sol, MatrixXd &V)
void	selected_inversion (SpMat &Q, Vect &inv_diag)
	selected inversion of the diagonal elements of Q.
void	selected_inversion_w_constr (SpMat &Q, const MatrixXd &D, Vect &inv_diag, MatrixXd &V)
void	compute_inverse_pardiso (MatrixXd &H, MatrixXd &C)
	inversion of the entire matrix (only meant for small matrices) by means of using identity right-hand side.
	~PardisoSolver ()
	class destructor. Frees memory allocated by pardiso.

Private Attributes
int	MPI_rank
int	n
long unsigned int	nnz
SpMat	Q
int *	ia
int *	ja
double *	a
double *	b
double *	x
void *	pt [64]
int	iparm [64]
double	dparm [64]
int	maxfct
int	mnum
int	phase
int	error
int	msglvl
int	solver
int	num_procs
int	i
int	k
long unsigned int	l
double	ddum
int	idum
int	mtype
int	init

Additional Inherited Members
Protected Attributes inherited from Solver
int	init
double	log_det
SpMat	Q
Vect	rhs
Vect	sol

Detailed Description

creates solver class using pardiso for factorising, solving and selectively inverting linear system.

divided into set up, symbolic factorisation, numerical factorisation, numerical factorisation & solve and selected inversion (of the diagonal elements)

Member Function Documentation

compute_inverse_pardiso()

void PardisoSolver::compute_inverse_pardiso	(	MatrixXd &	H,
		MatrixXd &	C
	)

inversion of the entire matrix (only meant for small matrices) by means of using identity right-hand side.

Parameters

[in]	H	dense matrix.
[in,out]	C	inverse of H.

factorize()

void PardisoSolver::factorize ( SpMat &  Q, double &  log_det, double &  t_priorLatChol  )

virtual

numerical factorisation.

Parameters

[in]	Q	precision matrix to be factorised.
[in,out]	log_det	computes log determinant of Q.

Implements Solver.

factorize_solve()

void PardisoSolver::factorize_solve ( SpMat &  Q, Vect &  rhs, Vect &  sol, double &  log_det, double &  t_condLatChol, double &  t_condLatSolve  )

virtual

factorises and solves matrix in one call (to reuse pardiso objects)

Parameters

[in]	Q	precision matrix.
[in]	rhs	right-hand side of the system.
[in,out]	sol	solution of the system.
[in,out]	log_det	log determinant of Q.

Implements Solver.

factorize_solve_w_constr()

void PardisoSolver::factorize_solve_w_constr ( SpMat &  Q, Vect &  rhs, const MatrixXd &  Dxy, double &  log_det, Vect &  sol, MatrixXd &  V  )

virtual

Implements Solver.

factorize_w_constr()

void PardisoSolver::factorize_w_constr ( SpMat &  Q, const MatrixXd &  D, double &  log_det, MatrixXd &  V  )

virtual

Implements Solver.

selected_inversion()

void PardisoSolver::selected_inversion ( SpMat &  Q, Vect &  inv_diag  )

virtual

selected inversion of the diagonal elements of Q.

Parameters

[in]	Q	precision matrix.
[in,out]	inv_diag	inverse diagonal to hold the solution vector.

Implements Solver.

selected_inversion_w_constr()

void PardisoSolver::selected_inversion_w_constr ( SpMat &  Q, const MatrixXd &  D, Vect &  inv_diag, MatrixXd &  V  )

virtual

Implements Solver.

symbolic_factorization()

void PardisoSolver::symbolic_factorization ( SpMat &  Q, int &  init  )

virtual

performs the symbolic factorisation.

Parameters

[in]	Q	precision matrix to be factorised.
[in,out]	init	integer value indicating status of symbolic factorisation. Changed to one at the end of the function.

For each PardisoSolver object symbolic factorisation only needs to be performed once as sparsity patters if assumed to remain the same.

Implements Solver.

Member Data Documentation

a

double* PardisoSolver::a

private

CSR format. values.

b

double* PardisoSolver::b

private

right-hand side.

ddum

double PardisoSolver::ddum

private

Double dummy

ia

int* PardisoSolver::ia

private

CSR format. row indices.

idum

int PardisoSolver::idum

private

Integer dummy.

init

int PardisoSolver::init

private

flag that indicates if symbolic factorisation already performed.

ja

int* PardisoSolver::ja

private

CSR format. col pointers.

MPI_rank

int PardisoSolver::MPI_rank

private

pass on current MPI rank

mtype

int PardisoSolver::mtype

private

matrix type

n

int PardisoSolver::n

private

size of the matrix

nnz

long unsigned int PardisoSolver::nnz

private

number of nonzeros

num_procs

int PardisoSolver::num_procs

private

Number of processors.

pt

void* PardisoSolver::pt[64]

private

Internal solver memory pointer pt

Q

SpMat PardisoSolver::Q

private

sparse precision matrix Q. Eigen format.

x

double* PardisoSolver::x

private

placeholder for solution.

---

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

• INLA_main/PardisoSolver.h
• INLA_main/PardisoSolver.cpp