nag_sparse_nsym_fac_sol (f11dcc) (PDF version)
f11 Chapter Contents
f11 Chapter Introduction
NAG C Library Manual

NAG Library Function Document

nag_sparse_nsym_fac_sol (f11dcc)

+ Contents

    1  Purpose
    7  Accuracy

1  Purpose

nag_sparse_nsym_fac_sol (f11dcc) solves a real sparse nonsymmetric system of linear equations, represented in coordinate storage format, using a restarted generalized minimal residual (RGMRES), conjugate gradient squared (CGS), or stabilized bi-conjugate gradient (Bi-CGSTAB) method, with incomplete LU  preconditioning.

2  Specification

#include <nag.h>
#include <nagf11.h>
void  nag_sparse_nsym_fac_sol (Nag_SparseNsym_Method method, Integer n, Integer nnz, const double a[], Integer la, const Integer irow[], const Integer icol[], const Integer ipivp[], const Integer ipivq[], const Integer istr[], const Integer idiag[], const double b[], Integer m, double tol, Integer maxitn, double x[], double *rnorm, Integer *itn, Nag_Sparse_Comm *comm, NagError *fail)

3  Description

nag_sparse_nsym_fac_sol (f11dcc) solves a real sparse nonsymmetric linear system of equations:
Ax = b ,
using a preconditioned RGMRES (see Saad and Schultz (1986)), CGS (see Sonneveld (1989)), or Bi-CGSTAB  method (see Van der Vorst (1989), Sleijpen and Fokkema (1993)).
nag_sparse_nsym_fac_sol (f11dcc) uses the incomplete LU  factorization determined by nag_sparse_nsym_fac (f11dac) as the preconditioning matrix. A call to nag_sparse_nsym_fac_sol (f11dcc) must always be preceded by a call to nag_sparse_nsym_fac (f11dac). Alternative preconditioners for the same storage scheme are available by calling nag_sparse_nsym_sol (f11dec).
The matrix A , and the preconditioning matrix M , are represented in coordinate storage (CS) format (see the f11 Chapter Introduction) in the arrays a, irow and icol, as returned from nag_sparse_nsym_fac (f11dac). The array a holds the nonzero entries in these matrices, while irow and icol hold the corresponding row and column indices.

4  References

Saad Y and Schultz M (1986) GMRES: a generalized minimal residual algorithm for solving nonsymmetric linear systems SIAM J. Sci. Statist. Comput. 7 856–869
Salvini S A and Shaw G J (1996) An evaluation of new NAG Library solvers for large sparse unsymmetric linear systems NAG Technical Report TR2/96
Sleijpen G L G and Fokkema D R (1993) BiCGSTAB for linear equations involving matrices with complex spectrum ETNA 1 11–32
Sonneveld P (1989) CGS, a fast Lanczos-type solver for nonsymmetric linear systems SIAM J. Sci. Statist. Comput. 10 36–52
Van der Vorst H (1989) Bi-CGSTAB, a fast and smoothly converging variant of Bi-CG for the solution of nonsymmetric linear systems SIAM J. Sci. Statist. Comput. 13 631–644

5  Arguments

1:     methodNag_SparseNsym_MethodInput
On entry: specifies the iterative method to be used.
method=Nag_SparseNsym_RGMRES
The restarted generalized minimum residual method is used.
method=Nag_SparseNsym_CGS
The conjugate gradient squared method is used.
method=Nag_SparseNsym_BiCGSTAB
Then the bi-conjugate gradient stabilised  method is used.
Constraint: method=Nag_SparseNsym_RGMRES, Nag_SparseNsym_CGS or Nag_SparseNsym_BiCGSTAB.
2:     nIntegerInput
On entry: the order of the matrix A . This must be the same value as was supplied in the preceding call to nag_sparse_nsym_fac (f11dac).
Constraint: n1 .
3:     nnzIntegerInput
On entry: the number of nonzero-elements in the matrix A . This must be the same value as was supplied in the preceding call to nag_sparse_nsym_fac (f11dac).
Constraint: 1 nnz n 2 .
4:     a[la]const doubleInput
On entry: the values returned in the array a by a previous call to nag_sparse_nsym_fac (f11dac).
5:     laIntegerInput
On entry: this must be the same value as returned by a previous call to nag_sparse_nsym_fac (f11dac).
Constraint: la 2 × nnz .
6:     irow[la]const IntegerInput
7:     icol[la]const IntegerInput
8:     ipivp[n]const IntegerInput
9:     ipivq[n]const IntegerInput
10:   istr[n+1]const IntegerInput
11:   idiag[n]const IntegerInput
On entry: the values returned in the arrays irow, icol, ipivp, ipivq, istr and idiag by a previous call to nag_sparse_nsym_fac (f11dac).
12:   b[n]const doubleInput
On entry: the right-hand side vector b .
13:   mIntegerInput
On entry: if method=Nag_SparseNsym_RGMRES, m is the dimension of the restart subspace.
If method=Nag_SparseNsym_BiCGSTAB, m is the order  of the polynomial Bi-CGSTAB method otherwise, m is not referenced.
Constraints:
  • if method=Nag_SparseNsym_RGMRES, 0 < m minn,50 ;
  • if method=Nag_SparseNsym_BiCGSTAB, 0 < m minn,10 .
14:   toldoubleInput
On entry: the required tolerance. Let x k  denote the approximate solution at iteration k , and r k  the corresponding residual. The algorithm is considered to have converged at iteration k  if:
r k τ × b + A x k .
If tol0.0 , τ = max ε , n ,ε  is used, where ε  is the machine precision. Otherwise τ = maxtol,10ε, n ,ε  is used.
Constraint: tol<1.0 .
15:   maxitnIntegerInput
On entry: the maximum number of iterations allowed.
Constraint: maxitn1 .
16:   x[n]doubleInput/Output
On entry: an initial approximation to the solution vector x .
On exit: an improved approximation to the solution vector x .
17:   rnormdouble *Output
On exit: the final value of the residual norm r k , where k  is the output value of itn.
18:   itnInteger *Output
On exit: the number of iterations carried out.
19:   commNag_Sparse_Comm *Input/Output
On entry/exit: a pointer to a structure of type Nag_Sparse_Comm whose members are used by the iterative solver.
20:   failNagError *Input/Output
The NAG error argument (see Section 3.6 in the Essential Introduction).

6  Error Indicators and Warnings

NE_2_INT_ARG_LT
On entry, la=value  while nnz=value . These arguments must satisfy la 2 × nnz .
NE_ACC_LIMIT
The required accuracy could not be obtained. However, a reasonable accuracy has been obtained and further iterations cannot improve the result.
You should check the output value of rnorm for acceptability. This error code usually implies that your problem has been fully and satisfactorily solved to within, or close to, the accuracy available on your system. Further iterations are unlikely to improve on this situation.
NE_ALLOC_FAIL
Dynamic memory allocation failed.
NE_BAD_PARAM
On entry, argument method had an illegal value.
NE_INT_2
On entry, m=value , minn,10 = value.
Constraint: 0 < m minn,10  when method=Nag_SparseNsym_BiCGSTAB.
On entry, m=value , minn,50 = value.
Constraint: 0 < m minn,50  when method=Nag_SparseNsym_RGMRES.
On entry, nnz=value , n=value .
Constraint: 1 nnz n 2 .
NE_INT_ARG_LT
On entry, maxitn=value.
Constraint: maxitn1.
On entry, n=value.
Constraint: n1.
NE_INVALID_CS
On entry, the CS representation of A  is invalid. Check that the call to nag_sparse_nsym_fac_sol (f11dcc) has been preceded by a valid call to nag_sparse_nsym_fac (f11dac), and that the arrays a, irow and icol have not been corrupted between the two calls.
NE_INVALID_CS_PRECOND
On entry, the CS representation of the preconditioning matrix M is invalid. Check that the call to nag_sparse_nsym_fac_sol (f11dcc) has been preceded by a valid call to nag_sparse_nsym_fac (f11dac), and that the arrays a, irow, icol, ipivp, ipivq, istr and idiag have not been corrupted between the two calls.
NE_NOT_REQ_ACC
The required accuracy has not been obtained in maxitn iterations.
NE_REAL_ARG_GE
On entry, tol must not be greater than or equal to 1.0: tol=value .

7  Accuracy

On successful termination, the final residual r k = b - A x k , where k=itn , satisfies the termination criterion
r k τ × b + A x k .
The value of the final residual norm is returned in rnorm.

8  Further Comments

The time taken by nag_sparse_nsym_fac_sol (f11dcc) for each iteration is roughly proportional to the value of nnzc returned from the preceding call to nag_sparse_nsym_fac (f11dac).
The number of iterations required to achieve a prescribed accuracy cannot be easily determined a priori, as it can depend dramatically on the conditioning and spectrum of the preconditioned coefficient matrix, A - = M -1 A .
Some illustrations of the application of nag_sparse_nsym_fac_sol (f11dcc) to linear systems arising from the discretization of two-dimensional elliptic partial differential equations, and to random-valued randomly structured linear systems, can be found in Salvini and Shaw (1996).

9  Example

This example program solves a sparse linear system of equations using the CGS method, with incomplete LU  preconditioning.

9.1  Program Text

Program Text (f11dcce.c)

9.2  Program Data

Program Data (f11dcce.d)

9.3  Program Results

Program Results (f11dcce.r)


nag_sparse_nsym_fac_sol (f11dcc) (PDF version)
f11 Chapter Contents
f11 Chapter Introduction
NAG C Library Manual

© The Numerical Algorithms Group Ltd, Oxford, UK. 2012