NAG Library Routine Document

Integer, Intent (In)	::	n, ka, kb, ldab, ldbb, ldx
Integer, Intent (Out)	::	info
Real (Kind=nag_wp), Intent (Out)	::	rwork(n)
Complex (Kind=nag_wp), Intent (In)	::	bb(ldbb,*)
Complex (Kind=nag_wp), Intent (Inout)	::	ab(ldab,), x(ldx,)
Complex (Kind=nag_wp), Intent (Out)	::	work(n)
Character (1), Intent (In)	::	vect, uplo

C Header Interface

#include <nagmk26.h>

void

f08usf_ (const char *vect, const char *uplo, const Integer *n, const Integer *ka, const Integer *kb, Complex ab[], const Integer *ldab, const Complex bb[], const Integer *ldbb, Complex x[], const Integer *ldx, Complex work[], double rwork[], Integer *info, const Charlen length_vect, const Charlen length_uplo)

The routine may be called by its LAPACK name zhbgst.

3

Description

To reduce the complex Hermitian-definite generalized eigenproblem

A z = λ B z

to the standard form

C y = λ y

, where

A

B

and

C

are banded, f08usf (zhbgst) must be preceded by a call to f08utf (zpbstf) which computes the split Cholesky factorization of the positive definite matrix

B

B = S^{H} S

. The split Cholesky factorization, compared with the ordinary Cholesky factorization, allows the work to be approximately halved.

This routine overwrites

A

with

C = X^{H} A X

, where

X = S^{- 1} Q

and

Q

is a unitary matrix chosen (implicitly) to preserve the bandwidth of

A

. The routine also has an option to allow the accumulation of

X

, and then, if

z

is an eigenvector of

C

X z

is an eigenvector of the original system.

4

References

Crawford C R (1973) Reduction of a band-symmetric generalized eigenvalue problem Comm. ACM 16 41–44

Kaufman L (1984) Banded eigenvalue solvers on vector machines ACM Trans. Math. Software 10 73–86

5

Arguments

1: $vect$ – Character(1)Input

On entry: indicates whether

X

is to be returned.

$vect ='N'$: $X$ is not returned.
$vect ='V'$: $X$ is returned.

Constraint:

vect ='N'

'V'

2: $uplo$ – Character(1)Input

On entry: indicates whether the upper or lower triangular part of

A

is stored.

$uplo ='U'$: The upper triangular part of $A$ is stored.
$uplo ='L'$: The lower triangular part of $A$ is stored.

Constraint:

uplo ='U'

'L'

3: $n$ – IntegerInput

On entry:

n

, the order of the matrices

A

and

B

Constraint:

n \geq 0

4: $ka$ – IntegerInput

On entry: if

uplo ='U'

, the number of superdiagonals,

k_{a}

, of the matrix

A

uplo ='L'

, the number of subdiagonals,

k_{a}

, of the matrix

A

Constraint:

ka \geq 0

5: $kb$ – IntegerInput

On entry: if

uplo ='U'

, the number of superdiagonals,

k_{b}

, of the matrix

B

uplo ='L'

, the number of subdiagonals,

k_{b}

, of the matrix

B

Constraint:

ka \geq kb \geq 0

6: $ab (ldab *)$ – Complex (Kind=nag_wp) arrayInput/Output

Note: the second dimension of the array ab must be at least

\max (1, n)

On entry: the upper or lower triangle of the

n

n

Hermitian band matrix

A

The matrix is stored in rows

1

k_{a} + 1

, more precisely,

if $uplo ='U'$ , the elements of the upper triangle of $A$ within the band must be stored with element $A_{i j}$ in $ab (k_{a} + 1 + i - j, j) for \max (1, j - k_{a}) \leq i \leq j$ ;
if $uplo ='L'$ , the elements of the lower triangle of $A$ within the band must be stored with element $A_{i j}$ in $ab (1 + i - j, j) for j \leq i \leq \min (n, j + k_{a}) .$

On exit: the upper or lower triangle of ab is overwritten by the corresponding upper or lower triangle of

C

as specified by uplo.

7: $ldab$ – IntegerInput

On entry: the first dimension of the array ab as declared in the (sub)program from which f08usf (zhbgst) is called.

Constraint:

ldab \geq ka + 1

8: $bb (ldbb *)$ – Complex (Kind=nag_wp) arrayInput

Note: the second dimension of the array bb must be at least

\max (1, n)

On entry: the banded split Cholesky factor of

B

as specified by uplo, n and kb and returned by f08utf (zpbstf).

9: $ldbb$ – IntegerInput

On entry: the first dimension of the array bb as declared in the (sub)program from which f08usf (zhbgst) is called.

Constraint:

ldbb \geq kb + 1

10: $x (ldx *)$ – Complex (Kind=nag_wp) arrayOutput

Note: the second dimension of the array x must be at least

\max (1, n)

vect ='V'

and at least

1

vect ='N'

On exit: the

n

n

matrix

X = S^{- 1} Q

, if

vect ='V'

vect ='N'

, x is not referenced.

11: $ldx$ – IntegerInput

On entry: the first dimension of the array x as declared in the (sub)program from which f08usf (zhbgst) is called.

Constraints:

if $vect ='V'$ , $ldx \geq \max (1, n)$ ;
if $vect ='N'$ , $ldx \geq 1$ .

12: $work (n)$ – Complex (Kind=nag_wp) arrayWorkspace

13: $rwork (n)$ – Real (Kind=nag_wp) arrayWorkspace

14: $info$ – 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

Forming the reduced matrix

C

is a stable procedure. However it involves implicit multiplication by

B^{- 1}

. When f08usf (zhbgst) is used as a step in the computation of eigenvalues and eigenvectors of the original problem, there may be a significant loss of accuracy if

B

is ill-conditioned with respect to inversion.

8

Parallelism and Performance

f08usf (zhbgst) 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 real floating-point operations is approximately

20 n^{2} k_{B}

, when

vect ='N'

, assuming

n ≫ k_{A}, k_{B}

; there are an additional

5 n^{3} (k_{B} / k_{A})

operations when

vect ='V'

The real analogue of this routine is f08uef (dsbgst).

10

Example

This example computes all the eigenvalues of

A z = λ B z

, where

A = (\begin{array}{r} - 1.13 + 0.00 i & 1.94 - 2.10 i & - 1.40 + 0.25 i & 0.00 + 0.00 i \\ 1.94 + 2.10 i & - 1.91 + 0.00 i & - 0.82 - 0.89 i & - 0.67 + 0.34 i \\ - 1.40 - 0.25 i & - 0.82 + 0.89 i & - 1.87 + 0.00 i & - 1.10 - 0.16 i \\ 0.00 + 0.00 i & - 0.67 - 0.34 i & - 1.10 + 0.16 i & 0.50 + 0.00 i \end{array})

and

B = (\begin{array}{r} 9.89 + 0.00 i & 1.08 - 1.73 i & 0.00 + 0.00 i & 0.00 + 0.00 i \\ 1.08 + 1.73 i & 1.69 + 0.00 i & - 0.04 + 0.29 i & 0.00 + 0.00 i \\ 0.00 + 0.00 i & - 0.04 - 0.29 i & 2.65 + 0.00 i & - 0.33 + 2.24 i \\ 0.00 + 0.00 i & 0.00 + 0.00 i & - 0.33 - 2.24 i & 2.17 + 0.00 i \end{array}) .

Here

A

is Hermitian,

B

is Hermitian positive definite, and

A

and

B

are treated as band matrices.

B

must first be factorized by f08utf (zpbstf). The program calls f08usf (zhbgst) to reduce the problem to the standard form

C y = λ y

, then f08hsf (zhbtrd) to reduce

C

to tridiagonal form, and f08jff (dsterf) to compute the eigenvalues.

NAG Library Routine Document

f08usf (zhbgst)

▸▿ Contents

1 Purpose

2 Specification

3 Description

4 References

5 Arguments

6 Error Indicators and Warnings

7 Accuracy

8 Parallelism and Performance

9 Further Comments

10 Example

10.1 Program Text

10.2 Program Data

10.3 Program Results

1

Purpose

2

Specification

3

Description

4

References

5

Arguments

6

Error Indicators and Warnings

7

Accuracy

8

Parallelism and Performance

9

Further Comments

10

Example

10.1

Program Text

10.2

Program Data

10.3

Program Results