NAG Library Routine Document

1Purpose

f07quf (zspcon) estimates the condition number of a complex symmetric matrix $A$, where $A$ has been factorized by f07qrf (zsptrf), using packed storage.

2Specification

Fortran Interface
 Subroutine f07quf ( uplo, n, ap, ipiv, work, info)
 Integer, Intent (In) :: n, ipiv(*) Integer, Intent (Out) :: info Real (Kind=nag_wp), Intent (In) :: anorm Real (Kind=nag_wp), Intent (Out) :: rcond Complex (Kind=nag_wp), Intent (In) :: ap(*) Complex (Kind=nag_wp), Intent (Out) :: work(2*n) Character (1), Intent (In) :: uplo
#include nagmk26.h
 void f07quf_ ( const char *uplo, const Integer *n, const Complex ap[], const Integer ipiv[], const double *anorm, double *rcond, Complex work[], Integer *info, const Charlen length_uplo)
The routine may be called by its LAPACK name zspcon.

3Description

f07quf (zspcon) estimates the condition number (in the $1$-norm) of a complex symmetric matrix $A$:
 $κ1A=A1A-11 .$
Since $A$ is symmetric, ${\kappa }_{1}\left(A\right)={\kappa }_{\infty }\left(A\right)={‖A‖}_{\infty }{‖{A}^{-1}‖}_{\infty }$.
Because ${\kappa }_{1}\left(A\right)$ is infinite if $A$ is singular, the routine actually returns an estimate of the reciprocal of ${\kappa }_{1}\left(A\right)$.
The routine should be preceded by a call to f06ugf to compute ${‖A‖}_{1}$ and a call to f07qrf (zsptrf) to compute the Bunch–Kaufman factorization of $A$. The routine then uses Higham's implementation of Hager's method (see Higham (1988)) to estimate ${‖{A}^{-1}‖}_{1}$.

4References

Higham N J (1988) FORTRAN codes for estimating the one-norm of a real or complex matrix, with applications to condition estimation ACM Trans. Math. Software 14 381–396

5Arguments

1:     $\mathbf{uplo}$ – Character(1)Input
On entry: specifies how $A$ has been factorized.
${\mathbf{uplo}}=\text{'U'}$
$A=PUD{U}^{\mathrm{T}}{P}^{\mathrm{T}}$, where $U$ is upper triangular.
${\mathbf{uplo}}=\text{'L'}$
$A=PLD{L}^{\mathrm{T}}{P}^{\mathrm{T}}$, where $L$ is lower triangular.
Constraint: ${\mathbf{uplo}}=\text{'U'}$ or $\text{'L'}$.
2:     $\mathbf{n}$ – IntegerInput
On entry: $n$, the order of the matrix $A$.
Constraint: ${\mathbf{n}}\ge 0$.
3:     $\mathbf{ap}\left(*\right)$ – Complex (Kind=nag_wp) arrayInput
Note: the dimension of the array ap must be at least $\mathrm{max}\phantom{\rule{0.125em}{0ex}}\left(1,{\mathbf{n}}×\left({\mathbf{n}}+1\right)/2\right)$.
On entry: the factorization of $A$ stored in packed form, as returned by f07qrf (zsptrf).
4:     $\mathbf{ipiv}\left(*\right)$ – Integer arrayInput
Note: the dimension of the array ipiv must be at least $\mathrm{max}\phantom{\rule{0.125em}{0ex}}\left(1,{\mathbf{n}}\right)$.
On entry: details of the interchanges and the block structure of $D$, as returned by f07qrf (zsptrf).
5:     $\mathbf{anorm}$ – Real (Kind=nag_wp)Input
On entry: the $1$-norm of the original matrix $A$, which may be computed by calling f06ugf with its argument ${\mathbf{norm}}=\text{'1'}$. anorm must be computed either before calling f07qrf (zsptrf) or else from a copy of the original matrix $A$.
Constraint: ${\mathbf{anorm}}\ge 0.0$.
6:     $\mathbf{rcond}$ – Real (Kind=nag_wp)Output
On exit: an estimate of the reciprocal of the condition number of $A$. rcond is set to zero if exact singularity is detected or the estimate underflows. If rcond is less than machine precision, $A$ is singular to working precision.
7:     $\mathbf{work}\left(2×{\mathbf{n}}\right)$ – Complex (Kind=nag_wp) arrayWorkspace
8:     $\mathbf{info}$ – IntegerOutput
On exit: ${\mathbf{info}}=0$ unless the routine detects an error (see Section 6).

6Error Indicators and Warnings

${\mathbf{info}}<0$
If ${\mathbf{info}}=-i$, argument $i$ had an illegal value. An explanatory message is output, and execution of the program is terminated.

7Accuracy

The computed estimate rcond is never less than the true value $\rho$, and in practice is nearly always less than $10\rho$, although examples can be constructed where rcond is much larger.

8Parallelism and Performance

f07quf (zspcon) 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.

A call to f07quf (zspcon) involves solving a number of systems of linear equations of the form $Ax=b$; the number is usually $5$ and never more than $11$. Each solution involves approximately $8{n}^{2}$ real floating-point operations but takes considerably longer than a call to f07qsf (zsptrs) with one right-hand side, because extra care is taken to avoid overflow when $A$ is approximately singular.
The real analogue of this routine is f07pgf (dspcon).

10Example

This example estimates the condition number in the $1$-norm (or $\infty$-norm) of the matrix $A$, where
 $A= -0.39-0.71i 5.14-0.64i -7.86-2.96i 3.80+0.92i 5.14-0.64i 8.86+1.81i -3.52+0.58i 5.32-1.59i -7.86-2.96i -3.52+0.58i -2.83-0.03i -1.54-2.86i 3.80+0.92i 5.32-1.59i -1.54-2.86i -0.56+0.12i .$
Here $A$ is symmetric, stored in packed form, and must first be factorized by f07qrf (zsptrf). The true condition number in the $1$-norm is $32.92$.

10.1Program Text

Program Text (f07qufe.f90)

10.2Program Data

Program Data (f07qufe.d)

10.3Program Results

Program Results (f07qufe.r)

© The Numerical Algorithms Group Ltd, Oxford, UK. 2017