NAG Library Routine Document

f07mpf  (zhesvx)

 Contents

    1  Purpose
    7  Accuracy

1
Purpose

f07mpf (zhesvx) uses the diagonal pivoting factorization to compute the solution to a complex system of linear equations
AX=B ,  
where A is an n by n Hermitian matrix and X and B are n by r matrices. Error bounds on the solution and a condition estimate are also provided.

2
Specification

Fortran Interface
Subroutine f07mpf ( fact, uplo, n, nrhs, a, lda, af, ldaf, ipiv, b, ldb, x, ldx, rcond, ferr, berr, work, lwork, rwork, info)
Integer, Intent (In):: n, nrhs, lda, ldaf, ldb, ldx, lwork
Integer, Intent (Inout):: ipiv(*)
Integer, Intent (Out):: info
Real (Kind=nag_wp), Intent (Inout):: ferr(*), berr(*), rwork(*)
Real (Kind=nag_wp), Intent (Out):: rcond
Complex (Kind=nag_wp), Intent (In):: a(lda,*), b(ldb,*)
Complex (Kind=nag_wp), Intent (Inout):: af(ldaf,*), x(ldx,*)
Complex (Kind=nag_wp), Intent (Out):: work(max(1,lwork))
Character (1), Intent (In):: fact, uplo
C Header Interface
#include nagmk26.h
void  f07mpf_ ( const char *fact, const char *uplo, const Integer *n, const Integer *nrhs, const Complex a[], const Integer *lda, Complex af[], const Integer *ldaf, Integer ipiv[], const Complex b[], const Integer *ldb, Complex x[], const Integer *ldx, double *rcond, double ferr[], double berr[], Complex work[], const Integer *lwork, double rwork[], Integer *info, const Charlen length_fact, const Charlen length_uplo)
The routine may be called by its LAPACK name zhesvx.

3
Description

f07mpf (zhesvx) performs the following steps:
1. If fact='N', the diagonal pivoting method is used to factor A. The form of the factorization is A=UDUH if uplo='U' or A=LDLH if uplo='L', where U (or L) is a product of permutation and unit upper (lower) triangular matrices, and D is Hermitian and block diagonal with 1 by 1 and 2 by 2 diagonal blocks.
2. If some dii=0, so that D is exactly singular, then the routine returns with info=i. Otherwise, the factored form of A is used to estimate the condition number of the matrix A. If the reciprocal of the condition number is less than machine precision, info=n+1 is returned as a warning, but the routine still goes on to solve for X and compute error bounds as described below.
3. The system of equations is solved for X using the factored form of A.
4. Iterative refinement is applied to improve the computed solution matrix and to calculate error bounds and backward error estimates for it.

4
References

Anderson E, Bai Z, Bischof C, Blackford S, Demmel J, Dongarra J J, Du Croz J J, Greenbaum A, Hammarling S, McKenney A and Sorensen D (1999) LAPACK Users' Guide (3rd Edition) SIAM, Philadelphia http://www.netlib.org/lapack/lug
Golub G H and Van Loan C F (1996) Matrix Computations (3rd Edition) Johns Hopkins University Press, Baltimore
Higham N J (2002) Accuracy and Stability of Numerical Algorithms (2nd Edition) SIAM, Philadelphia

5
Arguments

1:     fact – Character(1)Input
On entry: specifies whether or not the factorized form of the matrix A has been supplied.
fact='F'
af and ipiv contain the factorized form of the matrix A. af and ipiv will not be modified.
fact='N'
The matrix A will be copied to af and factorized.
Constraint: fact='F' or 'N'.
2:     uplo – Character(1)Input
On entry: if uplo='U', the upper triangle of A is stored.
If uplo='L', the lower triangle of A is stored.
Constraint: uplo='U' or 'L'.
3:     n – IntegerInput
On entry: n, the number of linear equations, i.e., the order of the matrix A.
Constraint: n0.
4:     nrhs – IntegerInput
On entry: r, the number of right-hand sides, i.e., the number of columns of the matrix B.
Constraint: nrhs0.
5:     alda* – Complex (Kind=nag_wp) arrayInput
Note: the second dimension of the array a must be at least max1,n.
On entry: the n by n Hermitian matrix A.
  • If uplo='U', the upper triangular part of A must be stored and the elements of the array below the diagonal are not referenced.
  • If uplo='L', the lower triangular part of A must be stored and the elements of the array above the diagonal are not referenced.
6:     lda – IntegerInput
On entry: the first dimension of the array a as declared in the (sub)program from which f07mpf (zhesvx) is called.
Constraint: ldamax1,n.
7:     afldaf* – Complex (Kind=nag_wp) arrayInput/Output
Note: the second dimension of the array af must be at least max1,n.
On entry: if fact='F', af contains the block diagonal matrix D and the multipliers used to obtain the factor U or L from the factorization a=UDUH or a=LDLH as computed by f07mrf (zhetrf).
On exit: if fact='N', af returns the block diagonal matrix D and the multipliers used to obtain the factor U or L from the factorization a=UDUH or a=LDLH.
8:     ldaf – IntegerInput
On entry: the first dimension of the array af as declared in the (sub)program from which f07mpf (zhesvx) is called.
Constraint: ldafmax1,n.
9:     ipiv* – Integer arrayInput/Output
Note: the dimension of the array ipiv must be at least max1,n.
On entry: if fact='F', ipiv contains details of the interchanges and the block structure of D, as determined by f07mrf (zhetrf).
  • if ipivi=k>0, dii is a 1 by 1 pivot block and the ith row and column of A were interchanged with the kth row and column;
  • if uplo='U' and ipivi-1=ipivi=-l<0, di-1,i-1d-i,i-1 d-i,i-1dii  is a 2 by 2 pivot block and the i-1th row and column of A were interchanged with the lth row and column;
  • if uplo='L' and ipivi=ipivi+1=-m<0, diidi+1,idi+1,idi+1,i+1 is a 2 by 2 pivot block and the i+1th row and column of A were interchanged with the mth row and column.
On exit: if fact='N', ipiv contains details of the interchanges and the block structure of D, as determined by f07mrf (zhetrf), as described above.
10:   bldb* – Complex (Kind=nag_wp) arrayInput
Note: the second dimension of the array b must be at least max1,nrhs.
On entry: the n by r right-hand side matrix B.
11:   ldb – IntegerInput
On entry: the first dimension of the array b as declared in the (sub)program from which f07mpf (zhesvx) is called.
Constraint: ldbmax1,n.
12:   xldx* – Complex (Kind=nag_wp) arrayOutput
Note: the second dimension of the array x must be at least max1,nrhs.
On exit: if info=0 or n+1, the n by r solution matrix X.
13:   ldx – IntegerInput
On entry: the first dimension of the array x as declared in the (sub)program from which f07mpf (zhesvx) is called.
Constraint: ldxmax1,n.
14:   rcond – Real (Kind=nag_wp)Output
On exit: the estimate of the reciprocal condition number of the matrix A. If rcond=0.0, the matrix may be exactly singular. This condition is indicated by info>0andinfon. Otherwise, if rcond is less than the machine precision, the matrix is singular to working precision. This condition is indicated by info=n+1.
15:   ferr* – Real (Kind=nag_wp) arrayOutput
Note: the dimension of the array ferr must be at least max1,nrhs.
On exit: if info=0 or n+1, an estimate of the forward error bound for each computed solution vector, such that x^j-xj/xjferrj where x^j is the jth column of the computed solution returned in the array x and xj is the corresponding column of the exact solution X. The estimate is as reliable as the estimate for rcond, and is almost always a slight overestimate of the true error.
16:   berr* – Real (Kind=nag_wp) arrayOutput
Note: the dimension of the array berr must be at least max1,nrhs.
On exit: if info=0 or n+1, an estimate of the component-wise relative backward error of each computed solution vector x^j (i.e., the smallest relative change in any element of A or B that makes x^j an exact solution).
17:   workmax1,lwork – Complex (Kind=nag_wp) arrayWorkspace
On exit: if info=0, work1 returns the optimal lwork.
18:   lwork – IntegerInput
On entry: the dimension of the array work as declared in the (sub)program from which f07mpf (zhesvx) is called.
lworkmax1,2×n, and for best performance, when fact='N', lworkmax1,2×n,n×nb, where nb is the optimal block size for f07mrf (zhetrf).
If lwork=-1, a workspace query is assumed; the routine only calculates the optimal size of the work array, returns this value as the first entry of the work array, and no error message related to lwork is issued.
19:   rwork* – Real (Kind=nag_wp) arrayWorkspace
Note: the dimension of the array rwork must be at least max1,n.
20:   info – IntegerOutput
On exit: info=0 unless the routine detects an error (see Section 6).

6
Error Indicators and Warnings

info<0
If info=-i, argument i had an illegal value. An explanatory message is output, and execution of the program is terminated.
info>0andinfon
Element value of the diagonal is exactly zero. The factorization has been completed, but the factor D is exactly singular, so the solution and error bounds could not be computed. rcond=0.0 is returned.
info=n+1
D is nonsingular, but rcond is less than machine precision, meaning that the matrix is singular to working precision. Nevertheless, the solution and error bounds are computed because there are a number of situations where the computed solution can be more accurate than the value of rcond would suggest.

7
Accuracy

For each right-hand side vector b, the computed solution x^ is the exact solution of a perturbed system of equations A+Ex^=b, where
E1 = Oε A1 ,  
where ε is the machine precision. See Chapter 11 of Higham (2002) for further details.
If x^ is the true solution, then the computed solution x satisfies a forward error bound of the form
x-x^ x^ wc condA,x^,b  
where condA,x^,b = A-1 A x^ + b / x^ condA = A-1 A κ A. If x^  is the j th column of X , then wc  is returned in berrj  and a bound on x - x^ / x^  is returned in ferrj . See Section 4.4 of Anderson et al. (1999) for further details.

8
Parallelism and Performance

f07mpf (zhesvx) is threaded by NAG for parallel execution in multithreaded implementations of the NAG Library.
f07mpf (zhesvx) 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 factorization of A  requires approximately 43 n3  floating-point operations.
For each right-hand side, computation of the backward error involves a minimum of 16n2  floating-point operations. Each step of iterative refinement involves an additional 24n2  operations. At most five steps of iterative refinement are performed, but usually only one or two steps are required. Estimating the forward error involves solving a number of systems of equations of the form Ax=b ; the number is usually 4 or 5 and never more than 11. Each solution involves approximately 8n2  operations.
The real analogue of this routine is f07mbf (dsysvx). The complex symmetric analogue of this routine is f07npf (zsysvx).

10
Example

This example solves the equations
AX=B ,  
where A  is the Hermitian matrix
A = -1.84i+0.00 0.11-0.11i -1.78-1.18i 3.91-1.50i 0.11+0.11i -4.63i+0.00 -1.84+0.03i 2.21+0.21i -1.78+1.18i -1.84-0.03i -8.87i+0.00 1.58-0.90i 3.91+1.50i 2.21-0.21i 1.58+0.90i -1.36i+0.00  
and
B = 2.98-10.18i 28.68-39.89i -9.58+03.88i -24.79-08.40i -0.77-16.05i 4.23-70.02i 7.79+05.48i -35.39+18.01i .  
Error estimates for the solutions, and an estimate of the reciprocal of the condition number of the matrix A  are also output.

10.1
Program Text

Program Text (f07mpfe.f90)

10.2
Program Data

Program Data (f07mpfe.d)

10.3
Program Results

Program Results (f07mpfe.r)

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