hide long namesshow long names
hide short namesshow short names
Integer type:  int32  int64  nag_int  show int32  show int32  show int64  show int64  show nag_int  show nag_int

PDF version (NAG web site, 64-bit version, 64-bit version)
Chapter Contents
Chapter Introduction
NAG Toolbox

NAG Toolbox: nag_sparse_real_symm_solve_ichol (f11jc)

Purpose

nag_sparse_real_symm_solve_ichol (f11jc) solves a real sparse symmetric system of linear equations, represented in symmetric coordinate storage format, using a conjugate gradient or Lanczos method, with incomplete Cholesky preconditioning.

Syntax

[x, rnorm, itn, ifail] = f11jc(method, nnz, a, irow, icol, ipiv, istr, b, tol, maxitn, x, 'n', n, 'la', la)
[x, rnorm, itn, ifail] = nag_sparse_real_symm_solve_ichol(method, nnz, a, irow, icol, ipiv, istr, b, tol, maxitn, x, 'n', n, 'la', la)

Description

nag_sparse_real_symm_solve_ichol (f11jc) solves a real sparse symmetric linear system of equations
Ax = b,
Ax=b,
using a preconditioned conjugate gradient method (see Meijerink and Van der Vorst (1977)), or a preconditioned Lanczos method based on the algorithm SYMMLQ (see Paige and Saunders (1975)). The conjugate gradient method is more efficient if AA is positive definite, but may fail to converge for indefinite matrices. In this case the Lanczos method should be used instead. For further details see Barrett et al. (1994).
nag_sparse_real_symm_solve_ichol (f11jc) uses the incomplete Cholesky factorization determined by nag_sparse_real_symm_precon_ichol (f11ja) as the preconditioning matrix. A call to nag_sparse_real_symm_solve_ichol (f11jc) must always be preceded by a call to nag_sparse_real_symm_precon_ichol (f11ja). Alternative preconditioners for the same storage scheme are available by calling nag_sparse_real_symm_solve_jacssor (f11je).
The matrix AA, and the preconditioning matrix MM, are represented in symmetric coordinate storage (SCS) format (see Section [Symmetric coordinate storage (SCS) format] in the F11 Chapter Introduction) in the arrays a, irow and icol, as returned from nag_sparse_real_symm_precon_ichol (f11ja). The array a holds the nonzero entries in the lower triangular parts of these matrices, while irow and icol hold the corresponding row and column indices.

References

Barrett R, Berry M, Chan T F, Demmel J, Donato J, Dongarra J, Eijkhout V, Pozo R, Romine C and Van der Vorst H (1994) Templates for the Solution of Linear Systems: Building Blocks for Iterative Methods SIAM, Philadelphia
Meijerink J and Van der Vorst H (1977) An iterative solution method for linear systems of which the coefficient matrix is a symmetric M-matrix Math. Comput. 31 148–162
Paige C C and Saunders M A (1975) Solution of sparse indefinite systems of linear equations SIAM J. Numer. Anal. 12 617–629
Salvini S A and Shaw G J (1995) An evaluation of new NAG Library solvers for large sparse symmetric linear systems NAG Technical Report TR1/95

Parameters

Compulsory Input Parameters

1:     method – string
Specifies the iterative method to be used.
method = 'CG'method='CG'
Conjugate gradient method.
method = 'SYMMLQ'method='SYMMLQ'
Lanczos method (SYMMLQ).
Constraint: method = 'CG'method='CG' or 'SYMMLQ''SYMMLQ'.
2:     nnz – int64int32nag_int scalar
The number of nonzero elements in the lower triangular part of the matrix AA. This must be the same value as was supplied in the preceding call to nag_sparse_real_symm_precon_ichol (f11ja).
Constraint: 1nnzn × (n + 1) / 21nnzn×(n+1)/2.
3:     a(la) – double array
la, the dimension of the array, must satisfy the constraint la2 × nnzla2×nnz.
The values returned in the array a by a previous call to nag_sparse_real_gen_precon_ilu (f11da).
4:     irow(la) – int64int32nag_int array
5:     icol(la) – int64int32nag_int array
6:     ipiv(n) – int64int32nag_int array
7:     istr(n + 1n+1) – int64int32nag_int array
la, the dimension of the array, must satisfy the constraint la2 × nnzla2×nnz.
The values returned in arrays irow, icol, ipiv and istr by a previous call to nag_sparse_real_symm_precon_ichol (f11ja).
8:     b(n) – double array
n, the dimension of the array, must satisfy the constraint n1n1.
The right-hand side vector bb.
9:     tol – double scalar
The required tolerance. Let xkxk denote the approximate solution at iteration kk, and rkrk the corresponding residual. The algorithm is considered to have converged at iteration kk if
rkτ × (b + Axk).
rkτ×(b+Axk).
If tol0.0tol0.0, τ = max (sqrt(ε),sqrt(n)ε)τ=max(ε,nε) is used, where εε is the machine precision. Otherwise τ = max (tol,10ε,sqrt(n)ε)τ=max(tol,10ε,nε) is used.
Constraint: tol < 1.0tol<1.0.
10:   maxitn – int64int32nag_int scalar
The maximum number of iterations allowed.
Constraint: maxitn1maxitn1.
11:   x(n) – double array
n, the dimension of the array, must satisfy the constraint n1n1.
An initial approximation to the solution vector xx.

Optional Input Parameters

1:     n – int64int32nag_int scalar
Default: The dimension of the arrays ipiv, b, x. (An error is raised if these dimensions are not equal.)
nn, the order of the matrix AA. This must be the same value as was supplied in the preceding call to nag_sparse_real_symm_precon_ichol (f11ja).
Constraint: n1n1.
2:     la – int64int32nag_int scalar
Default: The dimension of the arrays a, irow, icol. (An error is raised if these dimensions are not equal.)
The dimension of the arrays a, irow and icol as declared in the (sub)program from which nag_sparse_real_symm_solve_ichol (f11jc) is called. This must be the same value as was supplied in the preceding call to nag_sparse_real_symm_precon_ichol (f11ja).
Constraint: la2 × nnzla2×nnz.

Input Parameters Omitted from the MATLAB Interface

work lwork

Output Parameters

1:     x(n) – double array
An improved approximation to the solution vector xx.
2:     rnorm – double scalar
The final value of the residual norm rkrk, where kk is the output value of itn.
3:     itn – int64int32nag_int scalar
The number of iterations carried out.
4:     ifail – int64int32nag_int scalar
ifail = 0ifail=0 unless the function detects an error (see [Error Indicators and Warnings]).

Error Indicators and Warnings

Errors or warnings detected by the function:

Cases prefixed with W are classified as warnings and do not generate an error of type NAG:error_n. See nag_issue_warnings.

  ifail = 1ifail=1
On entry,method'CG'method'CG' or 'SYMMLQ''SYMMLQ',
orn < 1n<1,
ornnz < 1nnz<1,
ornnz > n × (n + 1) / 2nnz>n×(n+1)/2,
orla too small,
ortol1.0tol1.0,
ormaxitn < 1maxitn<1,
orlwork too small.
  ifail = 2ifail=2
On entry, the SCS representation of AA is invalid. Further details are given in the error message. Check that the call to nag_sparse_real_symm_solve_ichol (f11jc) has been preceded by a valid call to nag_sparse_real_symm_precon_ichol (f11ja), and that the arrays a, irow, and icol have not been corrupted between the two calls.
  ifail = 3ifail=3
On entry, the SCS representation of the preconditioning matrix MM is invalid. Further details are given in the error message. Check that the call to nag_sparse_real_symm_solve_ichol (f11jc) has been preceded by a valid call to nag_sparse_real_symm_precon_ichol (f11ja) and that the arrays a, irow, icol, ipiv and istr have not been corrupted between the two calls.
W ifail = 4ifail=4
The required accuracy could not be obtained. However, a reasonable accuracy has been obtained and further iterations could not improve the result.
  ifail = 5ifail=5
Required accuracy not obtained in maxitn iterations.
  ifail = 6ifail=6
The preconditioner appears not to be positive definite.
  ifail = 7ifail=7
The matrix of the coefficients appears not to be positive definite (conjugate gradient method only).
  ifail = 8ifail=8 (nag_sparse_real_symm_basic_setup (f11gd), nag_sparse_real_symm_basic_solver (f11ge) or nag_sparse_real_symm_basic_diag (f11gf))
A serious error has occurred in an internal call to one of the specified functions. Check all function calls and array sizes. Seek expert help.

Accuracy

On successful termination, the final residual rk = bAxkrk=b-Axk, where k = itnk=itn, satisfies the termination criterion
rkτ × (b + Axk).
rkτ×(b+Axk).
The value of the final residual norm is returned in rnorm.

Further Comments

The time taken by nag_sparse_real_symm_solve_ichol (f11jc) for each iteration is roughly proportional to the value of nnzc returned from the preceding call to nag_sparse_real_symm_precon_ichol (f11ja). One iteration with the Lanczos method (SYMMLQ) requires a slightly larger number of operations than one iteration with the conjugate gradient method.
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 matrix of the coefficients A = M1AA-=M-1A.
Some illustrations of the application of nag_sparse_real_symm_solve_ichol (f11jc) to linear systems arising from the discretization of two-dimensional elliptic partial differential equations, and to random-valued randomly structured symmetric positive definite linear systems, can be found in Salvini and Shaw (1995).

Example

function nag_sparse_real_symm_solve_ichol_example
method = 'CG';
nz = int64(16);
lfill = int64(1);
dtol = 0;
mic = 'N';
dscale = 0;
pstrat = 'M';
a = zeros(3*nz, 1);
a(1:16) = [4; 1; 5; 2; 2; 3; -1; 1; 4; 1; -2; 3; 2; -1; -2; 5];
irow = zeros(3*nz, 1, 'int64');
irow(1:16) = [int64(1); 2; 2; 3; 4; 4; 5; 5; 5; 6; 6; 6; 7; 7; 7; 7];
icol = zeros(3*nz, 1, 'int64');
icol(1:16) = [int64(1); 1; 2; 3; 2; 4; 1; 4; 5; 2; 5; 6; 1; 2; 3; 7];
ipiv = [int64(0);0;0;0;0;0;0];
b = [15; 18; -8; 21; 11; 10; 29];
tol = 1e-06;
maxitn = int64(100);
x = [0; 0; 0; 0; 0; 0; 0];
% Calculate incomplete Cholesky factorization
[a, irow, icol, ipiv, istr, nnzc, npivm, ifail] = ...
    nag_sparse_real_symm_precon_ichol(nz, a, irow, icol, lfill, dtol, mic, ...
                                      dscale, ipiv);

% Solve Ax = b
[xOut, rnorm, itn, ifail] = ...
    nag_sparse_real_symm_solve_ichol(method, nz, a, irow, icol, ipiv, istr, ...
                                     b, tol, maxitn, x);

fprintf('\nConverged in %d iterations\n', itn);
fprintf('Final residual norm = %16.3e\n', rnorm);
disp(xOut);
 

Converged in 1 iterations
Final residual norm =        7.105e-15
    1.0000
    2.0000
    3.0000
    4.0000
    5.0000
    6.0000
    7.0000


function f11jc_example
method = 'CG';
nz = int64(16);
lfill = int64(1);
dtol = 0;
mic = 'N';
dscale = 0;
pstrat = 'M';
a = zeros(3*nz, 1);
a(1:16) = [4; 1; 5; 2; 2; 3; -1; 1; 4; 1; -2; 3; 2; -1; -2; 5];
irow = zeros(3*nz, 1, 'int64');
irow(1:16) = [int64(1); 2; 2; 3; 4; 4; 5; 5; 5; 6; 6; 6; 7; 7; 7; 7];
icol = zeros(3*nz, 1, 'int64');
icol(1:16) = [int64(1); 1; 2; 3; 2; 4; 1; 4; 5; 2; 5; 6; 1; 2; 3; 7];
ipiv = [int64(0);0;0;0;0;0;0];
b = [15; 18; -8; 21; 11; 10; 29];
tol = 1e-06;
maxitn = int64(100);
x = [0; 0; 0; 0; 0; 0; 0];
% Calculate incomplete Cholesky factorization
[a, irow, icol, ipiv, istr, nnzc, npivm, ifail] = ...
    f11ja(nz, a, irow, icol, lfill, dtol, mic, dscale, ipiv);

% Solve Ax = b
[xOut, rnorm, itn, ifail] = ...
    f11jc(method, nz, a, irow, icol, ipiv, istr, b, tol, maxitn, x);

fprintf('\nConverged in %d iterations\n', itn);
fprintf('Final residual norm = %16.3e\n', rnorm);
disp(xOut);
 

Converged in 1 iterations
Final residual norm =        7.105e-15
    1.0000
    2.0000
    3.0000
    4.0000
    5.0000
    6.0000
    7.0000



PDF version (NAG web site, 64-bit version, 64-bit version)
Chapter Contents
Chapter Introduction
NAG Toolbox

© The Numerical Algorithms Group Ltd, Oxford, UK. 2009–2013