1ZHEEVX(1)             LAPACK driver routine (version 3.1)            ZHEEVX(1)
2
3
4

NAME

6       ZHEEVX  -  selected eigenvalues and, optionally, eigenvectors of a com‐
7       plex Hermitian matrix A
8

SYNOPSIS

10       SUBROUTINE ZHEEVX( JOBZ, RANGE, UPLO,  N,  A,  LDA,  VL,  VU,  IL,  IU,
11                          ABSTOL,  M,  W,  Z,  LDZ, WORK, LWORK, RWORK, IWORK,
12                          IFAIL, INFO )
13
14           CHARACTER      JOBZ, RANGE, UPLO
15
16           INTEGER        IL, INFO, IU, LDA, LDZ, LWORK, M, N
17
18           DOUBLE         PRECISION ABSTOL, VL, VU
19
20           INTEGER        IFAIL( * ), IWORK( * )
21
22           DOUBLE         PRECISION RWORK( * ), W( * )
23
24           COMPLEX*16     A( LDA, * ), WORK( * ), Z( LDZ, * )
25

PURPOSE

27       ZHEEVX computes selected eigenvalues and, optionally, eigenvectors of a
28       complex  Hermitian  matrix  A.   Eigenvalues  and  eigenvectors  can be
29       selected by specifying either a range of values or a range  of  indices
30       for the desired eigenvalues.
31
32

ARGUMENTS

34       JOBZ    (input) CHARACTER*1
35               = 'N':  Compute eigenvalues only;
36               = 'V':  Compute eigenvalues and eigenvectors.
37
38       RANGE   (input) CHARACTER*1
39               = 'A': all eigenvalues will be found.
40               =  'V':  all eigenvalues in the half-open interval (VL,VU] will
41               be found.  = 'I': the IL-th through IU-th eigenvalues  will  be
42               found.
43
44       UPLO    (input) CHARACTER*1
45               = 'U':  Upper triangle of A is stored;
46               = 'L':  Lower triangle of A is stored.
47
48       N       (input) INTEGER
49               The order of the matrix A.  N >= 0.
50
51       A       (input/output) COMPLEX*16 array, dimension (LDA, N)
52               On  entry,  the Hermitian matrix A.  If UPLO = 'U', the leading
53               N-by-N upper triangular part of A contains the upper triangular
54               part  of the matrix A.  If UPLO = 'L', the leading N-by-N lower
55               triangular part of A contains the lower triangular part of  the
56               matrix  A.   On  exit,  the lower triangle (if UPLO='L') or the
57               upper triangle (if UPLO='U') of A, including the  diagonal,  is
58               destroyed.
59
60       LDA     (input) INTEGER
61               The leading dimension of the array A.  LDA >= max(1,N).
62
63       VL      (input) DOUBLE PRECISION
64               VU       (input)  DOUBLE  PRECISION If RANGE='V', the lower and
65               upper bounds of the interval to be searched for eigenvalues. VL
66               < VU.  Not referenced if RANGE = 'A' or 'I'.
67
68       IL      (input) INTEGER
69               IU      (input) INTEGER If RANGE='I', the indices (in ascending
70               order) of the smallest and largest eigenvalues to be  returned.
71               1  <= IL <= IU <= N, if N > 0; IL = 1 and IU = 0 if N = 0.  Not
72               referenced if RANGE = 'A' or 'V'.
73
74       ABSTOL  (input) DOUBLE PRECISION
75               The absolute error tolerance for the eigenvalues.  An  approxi‐
76               mate  eigenvalue is accepted as converged when it is determined
77               to lie in an interval [a,b] of width less than or equal to
78
79               ABSTOL + EPS *   max( |a|,|b| ) ,
80
81               where EPS is the machine precision.  If ABSTOL is less than  or
82               equal  to zero, then  EPS*|T|  will be used in its place, where
83               |T| is the 1-norm of the tridiagonal matrix obtained by  reduc‐
84               ing A to tridiagonal form.
85
86               Eigenvalues will be computed most accurately when ABSTOL is set
87               to twice the underflow threshold 2*DLAMCH('S'), not  zero.   If
88               this  routine  returns with INFO>0, indicating that some eigen‐
89               vectors did not converge, try setting ABSTOL to 2*DLAMCH('S').
90
91               See "Computing Small Singular  Values  of  Bidiagonal  Matrices
92               with  Guaranteed  High Relative Accuracy," by Demmel and Kahan,
93               LAPACK Working Note #3.
94
95       M       (output) INTEGER
96               The total number of eigenvalues found.  0 <= M <= N.  If  RANGE
97               = 'A', M = N, and if RANGE = 'I', M = IU-IL+1.
98
99       W       (output) DOUBLE PRECISION array, dimension (N)
100               On  normal  exit, the first M elements contain the selected ei‐
101               genvalues in ascending order.
102
103       Z       (output) COMPLEX*16 array, dimension (LDZ, max(1,M))
104               If JOBZ = 'V', then if INFO = 0, the first M columns of Z  con‐
105               tain the orthonormal eigenvectors of the matrix A corresponding
106               to the selected eigenvalues, with the i-th column of Z  holding
107               the  eigenvector associated with W(i).  If an eigenvector fails
108               to converge, then that column of Z contains the latest approxi‐
109               mation  to the eigenvector, and the index of the eigenvector is
110               returned in IFAIL.  If JOBZ = 'N', then Z  is  not  referenced.
111               Note:  the  user must ensure that at least max(1,M) columns are
112               supplied in the array Z; if RANGE = 'V', the exact value  of  M
113               is not known in advance and an upper bound must be used.
114
115       LDZ     (input) INTEGER
116               The  leading dimension of the array Z.  LDZ >= 1, and if JOBZ =
117               'V', LDZ >= max(1,N).
118
119       WORK    (workspace/output) COMPLEX*16 array, dimension (MAX(1,LWORK))
120               On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
121
122       LWORK   (input) INTEGER
123               The length of the array WORK.  LWORK >= 1, when N <= 1;  other‐
124               wise  2*N.  For optimal efficiency, LWORK >= (NB+1)*N, where NB
125               is the max of the  blocksize  for  ZHETRD  and  for  ZUNMTR  as
126               returned by ILAENV.
127
128               If  LWORK  = -1, then a workspace query is assumed; the routine
129               only calculates the optimal size of  the  WORK  array,  returns
130               this  value  as the first entry of the WORK array, and no error
131               message related to LWORK is issued by XERBLA.
132
133       RWORK   (workspace) DOUBLE PRECISION array, dimension (7*N)
134
135       IWORK   (workspace) INTEGER array, dimension (5*N)
136
137       IFAIL   (output) INTEGER array, dimension (N)
138               If JOBZ = 'V', then if INFO = 0, the first M elements of  IFAIL
139               are  zero.  If INFO > 0, then IFAIL contains the indices of the
140               eigenvectors that failed to converge.   If  JOBZ  =  'N',  then
141               IFAIL is not referenced.
142
143       INFO    (output) INTEGER
144               = 0:  successful exit
145               < 0:  if INFO = -i, the i-th argument had an illegal value
146               >  0:   if  INFO  =  i, then i eigenvectors failed to converge.
147               Their indices are stored in array IFAIL.
148
149
150
151 LAPACK driver routine (version 3.N1o)vember 2006                       ZHEEVX(1)
Impressum