NAG FL Interface
f11duf (complex_​gen_​solve_​bdilu)

Settings help

FL Name Style:

FL Specification Language:

1 Purpose

f11duf solves a complex sparse non-Hermitian system of linear equations, represented in coordinate storage format, using a restarted generalized minimal residual (RGMRES), conjugate gradient squared (CGS), stabilized bi-conjugate gradient (Bi-CGSTAB), or transpose-free quasi-minimal residual (TFQMR) method, with block Jacobi or additive Schwarz preconditioning.

2 Specification

Fortran Interface
Subroutine f11duf ( method, n, nnz, a, la, irow, icol, nb, istb, indb, lindb, ipivp, ipivq, istr, idiag, b, m, tol, maxitn, x, rnorm, itn, work, lwork, ifail)
Integer, Intent (In) :: n, nnz, la, irow(la), icol(la), nb, istb(nb+1), indb(lindb), lindb, istr(lindb+1), idiag(lindb), m, maxitn, lwork
Integer, Intent (Inout) :: ipivp(lindb), ipivq(lindb), ifail
Integer, Intent (Out) :: itn
Real (Kind=nag_wp), Intent (In) :: tol
Real (Kind=nag_wp), Intent (Out) :: rnorm
Complex (Kind=nag_wp), Intent (In) :: a(la), b(n)
Complex (Kind=nag_wp), Intent (Inout) :: x(n)
Complex (Kind=nag_wp), Intent (Out) :: work(lwork)
Character (*), Intent (In) :: method
C Header Interface
#include <nag.h>
void  f11duf_ (const char *method, const Integer *n, const Integer *nnz, const Complex a[], const Integer *la, const Integer irow[], const Integer icol[], const Integer *nb, const Integer istb[], const Integer indb[], const Integer *lindb, Integer ipivp[], Integer ipivq[], const Integer istr[], const Integer idiag[], const Complex b[], const Integer *m, const double *tol, const Integer *maxitn, Complex x[], double *rnorm, Integer *itn, Complex work[], const Integer *lwork, Integer *ifail, const Charlen length_method)
The routine may be called by the names f11duf or nagf_sparse_complex_gen_solve_bdilu.

3 Description

f11duf solves a complex sparse non-Hermitian linear system of equations
using a preconditioned RGMRES (see Saad and Schultz (1986)), CGS (see Sonneveld (1989)), Bi-CGSTAB() (see Van der Vorst (1989) and Sleijpen and Fokkema (1993)), or TFQMR (see Freund and Nachtigal (1991) and Freund (1993)) method.
f11duf uses the incomplete (possibly overlapping) block LU factorization determined by f11dtf as the preconditioning matrix. A call to f11duf must always be preceded by a call to f11dtf. Alternative preconditioners for the same storage scheme are available by calling f11dqf or f11dsf.
The matrix A, and the preconditioning matrix M, are represented in coordinate storage (CS) format (see Section 2.1.1 in the F11 Chapter Introduction) in the arrays a, irow and icol, as returned from f11dtf. The array a holds the nonzero entries in these matrices, while irow and icol hold the corresponding row and column indices.
f11duf is a Black Box routine which calls f11brf, f11bsf and f11btf. If you wish to use an alternative storage scheme, preconditioner, or termination criterion, or require additional diagnostic information, you should call these underlying routines directly.

4 References

Freund R W (1993) A transpose-free quasi-minimal residual algorithm for non-Hermitian linear systems SIAM J. Sci. Comput. 14 470–482
Freund R W and Nachtigal N (1991) QMR: a Quasi-Minimal Residual Method for Non-Hermitian Linear Systems Numer. Math. 60 315–339
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
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: method Character(*) Input
On entry: specifies the iterative method to be used.
Restarted generalized minimum residual method.
Conjugate gradient squared method.
Bi-conjugate gradient stabilized () method.
Transpose-free quasi-minimal residual method.
Constraint: method='RGMRES', 'CGS', 'BICGSTAB' or 'TFQMR'.
2: n Integer Input
3: nnz Integer Input
4: a(la) Complex (Kind=nag_wp) array Input
5: la Integer Input
6: irow(la) Integer array Input
7: icol(la) Integer array Input
8: nb Integer Input
9: istb(nb+1) Integer array Input
10: indb(lindb) Integer array Input
11: lindb Integer Input
12: ipivp(lindb) Integer array Input
13: ipivq(lindb) Integer array Input
14: istr(lindb+1) Integer array Input
15: idiag(lindb) Integer array Input
On entry: the values returned in arrays a, irow, icol, ipivp, ipivq and idiag by a previous call to f11dtf.
The arrays istb and indb together with the scalars n, nnz, la, nb and lindb must be the same values that were supplied in the preceding call to f11dtf.
16: b(n) Complex (Kind=nag_wp) array Input
On entry: the right-hand side vector b.
17: m Integer Input
On entry: if method='RGMRES', m is the dimension of the restart subspace.
If method='BICGSTAB', m is the order of the polynomial Bi-CGSTAB method.
Otherwise, m is not referenced.
  • if method='RGMRES', 0<mmin(n,50);
  • if method='BICGSTAB', 0<mmin(n,10).
18: tol Real (Kind=nag_wp) Input
On entry: the required tolerance. Let xk denote the approximate solution at iteration k, and rk the corresponding residual. The algorithm is considered to have converged at iteration k if
If tol0.0, τ=maxε,nε is used, where ε is the machine precision. Otherwise τ=max(tol,10ε,nε) is used.
Constraint: tol<1.0.
19: maxitn Integer Input
On entry: the maximum number of iterations allowed.
Constraint: maxitn1.
20: x(n) Complex (Kind=nag_wp) array Input/Output
On entry: an initial approximation to the solution vector x.
On exit: an improved approximation to the solution vector x.
21: rnorm Real (Kind=nag_wp) Output
On exit: the final value of the residual norm rk, where k is the output value of itn.
22: itn Integer Output
On exit: the number of iterations carried out.
23: work(lwork) Complex (Kind=nag_wp) array Workspace
24: lwork Integer Input
On entry: the dimension of the array work as declared in the (sub)program from which f11duf is called.
  • if method='RGMRES', lwork6×n+m×(m+n+5)+121;
  • if method='CGS', lwork10×n+120;
  • if method='BICGSTAB', lwork2×n×m+8×n+m×(m+2)+120;
  • if method='TFQMR', lwork13×n+120.
25: ifail Integer Input/Output
On entry: ifail must be set to 0, −1 or 1 to set behaviour on detection of an error; these values have no effect when no error is detected.
A value of 0 causes the printing of an error message and program execution will be halted; otherwise program execution continues. A value of −1 means that an error message is printed while a value of 1 means that it is not.
If halting is not appropriate, the value −1 or 1 is recommended. If message printing is undesirable, then the value 1 is recommended. Otherwise, the value 0 is recommended. When the value -1 or 1 is used it is essential to test the value of ifail on exit.
On exit: ifail=0 unless the routine detects an error or a warning has been flagged (see Section 6).

6 Error Indicators and Warnings

If on entry ifail=0 or −1, explanatory error messages are output on the current error message unit (as defined by x04aaf).
Errors or warnings detected by the routine:
On entry, for b=value, istb(b+1)=value and istb(b)=value.
Constraint: istb(b+1)>istb(b), for b=1,2,,nb.
On entry, indb(value)=value and n=value.
Constraint: 1indb(m)n, for m=1,2,,istb(nb+1)-1
On entry, istb(1)=value.
Constraint: istb(1)1.
On entry, la=value and nnz=value.
Constraint: la2×nnz.
On entry, lindb=value, istb(nb+1)-1=value and nb=value.
Constraint: lindbistb(nb+1)-1.
On entry, lwork=value.
Constraint: lworkvalue.
On entry, m=value and n=value.
Constraint: if method='RGMRES', 1mmin(n,value).
If method='BICGSTAB', 1mmin(n,value).
On entry, maxitn=value.
Constraint: maxitn1.
On entry, method=value.
Constraint: method='RGMRES', 'CGS' or 'BICGSTAB'.
On entry, n=value.
Constraint: n1.
On entry, nb=value and n=value.
Constraint: 1nbn.
On entry, nnz=value.
Constraint: nnz1.
On entry, nnz=value and n=value.
Constraint: nnzn2.
On entry, tol=value.
Constraint: tol<1.0.
On entry, element value of a was out of order.
Check that a, irow, icol, ipivp, ipivq, istr and idiag have not been corrupted between calls to f11dtf and f11duf.
On entry, icol(value)=value and n=value.
Constraint: 1icol(i)n, for i=1,2,,nnz.
Check that a, irow, icol, ipivp, ipivq, istr and idiag have not been corrupted between calls to f11dtf and f11duf.
On entry, irow(value)=value and n=value.
Constraint: 1irow(i)n, for i=1,2,,nnz.
Check that a, irow, icol, ipivp, ipivq, istr and idiag have not been corrupted between calls to f11dtf and f11duf.
On entry, location value of (irow,icol) was a duplicate.
Check that a, irow, icol, ipivp, ipivq, istr and idiag have not been corrupted between calls to f11dtf and f11duf.
The CS representation of the preconditioner is invalid.
Check that a, irow, icol, ipivp, ipivq, istr and idiag have not been corrupted between calls to f11dtf and f11duf.
The required accuracy could not be obtained. However, a reasonable accuracy may have been achieved. 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.
The solution has not converged after value iterations.
Algorithmic breakdown. A solution is returned, although it is possible that it is completely inaccurate.
An unexpected error has been triggered by this routine. Please contact NAG.
See Section 7 in the Introduction to the NAG Library FL Interface for further information.
Your licence key may have expired or may not have been installed correctly.
See Section 8 in the Introduction to the NAG Library FL Interface for further information.
Dynamic memory allocation failed.
See Section 9 in the Introduction to the NAG Library FL Interface for further information.

7 Accuracy

On successful termination, the final residual rk=b-Axk, where k=itn, satisfies the termination criterion
The value of the final residual norm is returned in rnorm.

8 Parallelism and Performance

Background information to multithreading can be found in the Multithreading documentation.
f11duf is threaded by NAG for parallel execution in multithreaded implementations of the NAG Library.
f11duf makes calls to BLAS and/or LAPACK routines, which may be threaded within the vendor library used by this implementation. Consult the documentation for the vendor library for further information.
Please consult the X06 Chapter Introduction for information on how to control and interrogate the OpenMP environment used within this routine. Please also consult the Users' Note for your implementation for any additional implementation-specific information.

9 Further Comments

The time taken by f11duf for each iteration is roughly proportional to the value of nnzc returned from the preceding call to f11dtf.
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−1A.

10 Example

This example program reads in a sparse matrix A and a vector b. It calls f11dtf, with the array lfill=0 and the array dtol=0.0, to compute an overlapping incomplete LU factorization. This is then used as an additive Schwarz preconditioner on a call to f11duf which uses the RGMRES method to solve Ax=b.

10.1 Program Text

Program Text (f11dufe.f90)

10.2 Program Data

Program Data (f11dufe.d)

10.3 Program Results

Program Results (f11dufe.r)