NAG Library Routine Document

f08cwf  (zungrq)

 Contents

    1  Purpose
    7  Accuracy

1
Purpose

f08cwf (zungrq) generates all or part of the complex n by n unitary matrix Q from an RQ factorization computed by f08cvf (zgerqf).

2
Specification

Fortran Interface
Subroutine f08cwf ( m, n, k, a, lda, tau, work, lwork, info)
Integer, Intent (In):: m, n, k, lda, lwork
Integer, Intent (Out):: info
Complex (Kind=nag_wp), Intent (In):: tau(*)
Complex (Kind=nag_wp), Intent (Inout):: a(lda,*)
Complex (Kind=nag_wp), Intent (Out):: work(max(1,lwork))
C Header Interface
#include nagmk26.h
void  f08cwf_ ( const Integer *m, const Integer *n, const Integer *k, Complex a[], const Integer *lda, const Complex tau[], Complex work[], const Integer *lwork, Integer *info)
The routine may be called by its LAPACK name zungrq.

3
Description

f08cwf (zungrq) is intended to be used following a call to f08cvf (zgerqf), which performs an RQ factorization of a complex matrix A and represents the unitary matrix Q as a product of k elementary reflectors of order n.
This routine may be used to generate Q explicitly as a square matrix, or to form only its trailing rows.
Usually Q is determined from the RQ factorization of a p by n matrix A with pn. The whole of Q may be computed by:
Call ZUNGRQ(n,n,p,a,lda,tau,work,lwork,info)
(note that the matrix A must have at least n rows), or its trailing p rows as:
Call ZUNGRQ(p,n,p,a,lda,tau,work,lwork,info)
The rows of Q returned by the last call form an orthonormal basis for the space spanned by the rows of A; thus f08cvf (zgerqf) followed by f08cwf (zungrq) can be used to orthogonalize the rows of A.
The information returned by f08cvf (zgerqf) also yields the RQ factorization of the trailing k rows of A, where k<p. The unitary matrix arising from this factorization can be computed by:
Call ZUNGRQ(n,n,k,a,lda,tau,work,lwork,info)
or its leading k columns by:
Call ZUNGRQ(k,n,k,a,lda,tau,work,lwork,info)

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

5
Arguments

1:     m – IntegerInput
On entry: m, the number of rows of the matrix Q.
Constraint: m0.
2:     n – IntegerInput
On entry: n, the number of columns of the matrix Q.
Constraint: nm.
3:     k – IntegerInput
On entry: k, the number of elementary reflectors whose product defines the matrix Q.
Constraint: mk0.
4:     alda* – Complex (Kind=nag_wp) arrayInput/Output
Note: the second dimension of the array a must be at least max1,n.
On entry: details of the vectors which define the elementary reflectors, as returned by f08cvf (zgerqf).
On exit: the m by n matrix Q.
5:     lda – IntegerInput
On entry: the first dimension of the array a as declared in the (sub)program from which f08cwf (zungrq) is called.
Constraint: ldamax1,m.
6:     tau* – Complex (Kind=nag_wp) arrayInput
Note: the dimension of the array tau must be at least max1,k.
On entry: taui must contain the scalar factor of the elementary reflector Hi, as returned by f08cvf (zgerqf).
7:     workmax1,lwork – Complex (Kind=nag_wp) arrayWorkspace
On exit: if info=0, the real part of work1 contains the minimum value of lwork required for optimal performance.
8:     lwork – IntegerInput
On entry: the dimension of the array work as declared in the (sub)program from which f08cwf (zungrq) is called.
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.
Suggested value: for optimal performance, lworkn×nb, where nb is the optimal block size.
Constraint: lworkmax1,m or lwork=-1.
9:     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.

7
Accuracy

The computed matrix Q differs from an exactly unitary matrix by a matrix E such that
E2 = Oε  
and ε is the machine precision.

8
Parallelism and Performance

f08cwf (zungrq) 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 total number of floating-point operations is approximately 16mnk-8m+nk2+163k3; when m=k this becomes 83m23n-m.
The real analogue of this routine is f08cjf (dorgrq).

10
Example

This example generates the first four rows of the matrix Q of the RQ factorization of A as returned by f08cvf (zgerqf), where
A = 0.96-0.81i -0.98+1.98i 0.62-0.46i -0.37+0.38i 0.83+0.51i 1.08-0.28i -0.03+0.96i -1.20+0.19i 1.01+0.02i 0.19-0.54i 0.20+0.01i 0.20-0.12i -0.91+2.06i -0.66+0.42i 0.63-0.17i -0.98-0.36i -0.17-0.46i -0.07+1.23i -0.05+0.41i -0.81+0.56i -1.11+0.60i 0.22-0.20i 1.47+1.59i 0.26+0.26i .  
Note that the block size (NB) of 64 assumed in this example is not realistic for such a small problem, but should be suitable for large problems.

10.1
Program Text

Program Text (f08cwfe.f90)

10.2
Program Data

Program Data (f08cwfe.d)

10.3
Program Results

Program Results (f08cwfe.r)

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