sdrvsx.f

Go to the documentation of this file.

*> \brief \b SDRVSX
*
*  =========== DOCUMENTATION ===========
*
* Online html documentation available at 
*            http://www.netlib.org/lapack/explore-html/ 
*
*  Definition:
*  ===========
*
*       SUBROUTINE SDRVSX( NSIZES, NN, NTYPES, DOTYPE, ISEED, THRESH,
*                          NIUNIT, NOUNIT, A, LDA, H, HT, WR, WI, WRT,
*                          WIT, WRTMP, WITMP, VS, LDVS, VS1, RESULT, WORK,
*                          LWORK, IWORK, BWORK, INFO )
* 
*       .. Scalar Arguments ..
*       INTEGER            INFO, LDA, LDVS, LWORK, NIUNIT, NOUNIT, NSIZES,
*      $                   NTYPES
*       REAL               THRESH
*       ..
*       .. Array Arguments ..
*       LOGICAL            BWORK( * ), DOTYPE( * )
*       INTEGER            ISEED( 4 ), IWORK( * ), NN( * )
*       REAL               A( LDA, * ), H( LDA, * ), HT( LDA, * ),
*      $                   RESULT( 17 ), VS( LDVS, * ), VS1( LDVS, * ),
*      $                   WI( * ), WIT( * ), WITMP( * ), WORK( * ),
*      $                   WR( * ), WRT( * ), WRTMP( * )
*       ..
*  
*
*> \par Purpose:
*  =============
*>
*> \verbatim
*>
*>    SDRVSX checks the nonsymmetric eigenvalue (Schur form) problem
*>    expert driver SGEESX.
*>
*>    SDRVSX uses both test matrices generated randomly depending on
*>    data supplied in the calling sequence, as well as on data
*>    read from an input file and including precomputed condition
*>    numbers to which it compares the ones it computes.
*>
*>    When SDRVSX is called, a number of matrix "sizes" ("n's") and a
*>    number of matrix "types" are specified.  For each size ("n")
*>    and each type of matrix, one matrix will be generated and used
*>    to test the nonsymmetric eigenroutines.  For each matrix, 15
*>    tests will be performed:
*>
*>    (1)     0 if T is in Schur form, 1/ulp otherwise
*>           (no sorting of eigenvalues)
*>
*>    (2)     | A - VS T VS' | / ( n |A| ulp )
*>
*>      Here VS is the matrix of Schur eigenvectors, and T is in Schur
*>      form  (no sorting of eigenvalues).
*>
*>    (3)     | I - VS VS' | / ( n ulp ) (no sorting of eigenvalues).
*>
*>    (4)     0     if WR+sqrt(-1)*WI are eigenvalues of T
*>            1/ulp otherwise
*>            (no sorting of eigenvalues)
*>
*>    (5)     0     if T(with VS) = T(without VS),
*>            1/ulp otherwise
*>            (no sorting of eigenvalues)
*>
*>    (6)     0     if eigenvalues(with VS) = eigenvalues(without VS),
*>            1/ulp otherwise
*>            (no sorting of eigenvalues)
*>
*>    (7)     0 if T is in Schur form, 1/ulp otherwise
*>            (with sorting of eigenvalues)
*>
*>    (8)     | A - VS T VS' | / ( n |A| ulp )
*>
*>      Here VS is the matrix of Schur eigenvectors, and T is in Schur
*>      form  (with sorting of eigenvalues).
*>
*>    (9)     | I - VS VS' | / ( n ulp ) (with sorting of eigenvalues).
*>
*>    (10)    0     if WR+sqrt(-1)*WI are eigenvalues of T
*>            1/ulp otherwise
*>            If workspace sufficient, also compare WR, WI with and
*>            without reciprocal condition numbers
*>            (with sorting of eigenvalues)
*>
*>    (11)    0     if T(with VS) = T(without VS),
*>            1/ulp otherwise
*>            If workspace sufficient, also compare T with and without
*>            reciprocal condition numbers
*>            (with sorting of eigenvalues)
*>
*>    (12)    0     if eigenvalues(with VS) = eigenvalues(without VS),
*>            1/ulp otherwise
*>            If workspace sufficient, also compare VS with and without
*>            reciprocal condition numbers
*>            (with sorting of eigenvalues)
*>
*>    (13)    if sorting worked and SDIM is the number of
*>            eigenvalues which were SELECTed
*>            If workspace sufficient, also compare SDIM with and
*>            without reciprocal condition numbers
*>
*>    (14)    if RCONDE the same no matter if VS and/or RCONDV computed
*>
*>    (15)    if RCONDV the same no matter if VS and/or RCONDE computed
*>
*>    The "sizes" are specified by an array NN(1:NSIZES); the value of
*>    each element NN(j) specifies one size.
*>    The "types" are specified by a logical array DOTYPE( 1:NTYPES );
*>    if DOTYPE(j) is .TRUE., then matrix type "j" will be generated.
*>    Currently, the list of possible types is:
*>
*>    (1)  The zero matrix.
*>    (2)  The identity matrix.
*>    (3)  A (transposed) Jordan block, with 1's on the diagonal.
*>
*>    (4)  A diagonal matrix with evenly spaced entries
*>         1, ..., ULP  and random signs.
*>         (ULP = (first number larger than 1) - 1 )
*>    (5)  A diagonal matrix with geometrically spaced entries
*>         1, ..., ULP  and random signs.
*>    (6)  A diagonal matrix with "clustered" entries 1, ULP, ..., ULP
*>         and random signs.
*>
*>    (7)  Same as (4), but multiplied by a constant near
*>         the overflow threshold
*>    (8)  Same as (4), but multiplied by a constant near
*>         the underflow threshold
*>
*>    (9)  A matrix of the form  U' T U, where U is orthogonal and
*>         T has evenly spaced entries 1, ..., ULP with random signs
*>         on the diagonal and random O(1) entries in the upper
*>         triangle.
*>
*>    (10) A matrix of the form  U' T U, where U is orthogonal and
*>         T has geometrically spaced entries 1, ..., ULP with random
*>         signs on the diagonal and random O(1) entries in the upper
*>         triangle.
*>
*>    (11) A matrix of the form  U' T U, where U is orthogonal and
*>         T has "clustered" entries 1, ULP,..., ULP with random
*>         signs on the diagonal and random O(1) entries in the upper
*>         triangle.
*>
*>    (12) A matrix of the form  U' T U, where U is orthogonal and
*>         T has real or complex conjugate paired eigenvalues randomly
*>         chosen from ( ULP, 1 ) and random O(1) entries in the upper
*>         triangle.
*>
*>    (13) A matrix of the form  X' T X, where X has condition
*>         SQRT( ULP ) and T has evenly spaced entries 1, ..., ULP
*>         with random signs on the diagonal and random O(1) entries
*>         in the upper triangle.
*>
*>    (14) A matrix of the form  X' T X, where X has condition
*>         SQRT( ULP ) and T has geometrically spaced entries
*>         1, ..., ULP with random signs on the diagonal and random
*>         O(1) entries in the upper triangle.
*>
*>    (15) A matrix of the form  X' T X, where X has condition
*>         SQRT( ULP ) and T has "clustered" entries 1, ULP,..., ULP
*>         with random signs on the diagonal and random O(1) entries
*>         in the upper triangle.
*>
*>    (16) A matrix of the form  X' T X, where X has condition
*>         SQRT( ULP ) and T has real or complex conjugate paired
*>         eigenvalues randomly chosen from ( ULP, 1 ) and random
*>         O(1) entries in the upper triangle.
*>
*>    (17) Same as (16), but multiplied by a constant
*>         near the overflow threshold
*>    (18) Same as (16), but multiplied by a constant
*>         near the underflow threshold
*>
*>    (19) Nonsymmetric matrix with random entries chosen from (-1,1).
*>         If N is at least 4, all entries in first two rows and last
*>         row, and first column and last two columns are zero.
*>    (20) Same as (19), but multiplied by a constant
*>         near the overflow threshold
*>    (21) Same as (19), but multiplied by a constant
*>         near the underflow threshold
*>
*>    In addition, an input file will be read from logical unit number
*>    NIUNIT. The file contains matrices along with precomputed
*>    eigenvalues and reciprocal condition numbers for the eigenvalue
*>    average and right invariant subspace. For these matrices, in
*>    addition to tests (1) to (15) we will compute the following two
*>    tests:
*>
*>   (16)  |RCONDE - RCDEIN| / cond(RCONDE)
*>
*>      RCONDE is the reciprocal average eigenvalue condition number
*>      computed by SGEESX and RCDEIN (the precomputed true value)
*>      is supplied as input.  cond(RCONDE) is the condition number
*>      of RCONDE, and takes errors in computing RCONDE into account,
*>      so that the resulting quantity should be O(ULP). cond(RCONDE)
*>      is essentially given by norm(A)/RCONDV.
*>
*>   (17)  |RCONDV - RCDVIN| / cond(RCONDV)
*>
*>      RCONDV is the reciprocal right invariant subspace condition
*>      number computed by SGEESX and RCDVIN (the precomputed true
*>      value) is supplied as input. cond(RCONDV) is the condition
*>      number of RCONDV, and takes errors in computing RCONDV into
*>      account, so that the resulting quantity should be O(ULP).
*>      cond(RCONDV) is essentially given by norm(A)/RCONDE.
*> \endverbatim
*
*  Arguments:
*  ==========
*
*> \param[in] NSIZES
*> \verbatim
*>          NSIZES is INTEGER
*>          The number of sizes of matrices to use.  NSIZES must be at
*>          least zero. If it is zero, no randomly generated matrices
*>          are tested, but any test matrices read from NIUNIT will be
*>          tested.
*> \endverbatim
*>
*> \param[in] NN
*> \verbatim
*>          NN is INTEGER array, dimension (NSIZES)
*>          An array containing the sizes to be used for the matrices.
*>          Zero values will be skipped.  The values must be at least
*>          zero.
*> \endverbatim
*>
*> \param[in] NTYPES
*> \verbatim
*>          NTYPES is INTEGER
*>          The number of elements in DOTYPE. NTYPES must be at least
*>          zero. If it is zero, no randomly generated test matrices
*>          are tested, but and test matrices read from NIUNIT will be
*>          tested. If it is MAXTYP+1 and NSIZES is 1, then an
*>          additional type, MAXTYP+1 is defined, which is to use
*>          whatever matrix is in A.  This is only useful if
*>          DOTYPE(1:MAXTYP) is .FALSE. and DOTYPE(MAXTYP+1) is .TRUE. .
*> \endverbatim
*>
*> \param[in] DOTYPE
*> \verbatim
*>          DOTYPE is LOGICAL array, dimension (NTYPES)
*>          If DOTYPE(j) is .TRUE., then for each size in NN a
*>          matrix of that size and of type j will be generated.
*>          If NTYPES is smaller than the maximum number of types
*>          defined (PARAMETER MAXTYP), then types NTYPES+1 through
*>          MAXTYP will not be generated.  If NTYPES is larger
*>          than MAXTYP, DOTYPE(MAXTYP+1) through DOTYPE(NTYPES)
*>          will be ignored.
*> \endverbatim
*>
*> \param[in,out] ISEED
*> \verbatim
*>          ISEED is INTEGER array, dimension (4)
*>          On entry ISEED specifies the seed of the random number
*>          generator. The array elements should be between 0 and 4095;
*>          if not they will be reduced mod 4096.  Also, ISEED(4) must
*>          be odd.  The random number generator uses a linear
*>          congruential sequence limited to small integers, and so
*>          should produce machine independent random numbers. The
*>          values of ISEED are changed on exit, and can be used in the
*>          next call to SDRVSX to continue the same random number
*>          sequence.
*> \endverbatim
*>
*> \param[in] THRESH
*> \verbatim
*>          THRESH is REAL
*>          A test will count as "failed" if the "error", computed as
*>          described above, exceeds THRESH.  Note that the error
*>          is scaled to be O(1), so THRESH should be a reasonably
*>          small multiple of 1, e.g., 10 or 100.  In particular,
*>          it should not depend on the precision (single vs. double)
*>          or the size of the matrix.  It must be at least zero.
*> \endverbatim
*>
*> \param[in] NIUNIT
*> \verbatim
*>          NIUNIT is INTEGER
*>          The FORTRAN unit number for reading in the data file of
*>          problems to solve.
*> \endverbatim
*>
*> \param[in] NOUNIT
*> \verbatim
*>          NOUNIT is INTEGER
*>          The FORTRAN unit number for printing out error messages
*>          (e.g., if a routine returns INFO not equal to 0.)
*> \endverbatim
*>
*> \param[out] A
*> \verbatim
*>          A is REAL array, dimension (LDA, max(NN))
*>          Used to hold the matrix whose eigenvalues are to be
*>          computed.  On exit, A contains the last matrix actually used.
*> \endverbatim
*>
*> \param[in] LDA
*> \verbatim
*>          LDA is INTEGER
*>          The leading dimension of A, and H. LDA must be at
*>          least 1 and at least max( NN ).
*> \endverbatim
*>
*> \param[out] H
*> \verbatim
*>          H is REAL array, dimension (LDA, max(NN))
*>          Another copy of the test matrix A, modified by SGEESX.
*> \endverbatim
*>
*> \param[out] HT
*> \verbatim
*>          HT is REAL array, dimension (LDA, max(NN))
*>          Yet another copy of the test matrix A, modified by SGEESX.
*> \endverbatim
*>
*> \param[out] WR
*> \verbatim
*>          WR is REAL array, dimension (max(NN))
*> \endverbatim
*>
*> \param[out] WI
*> \verbatim
*>          WI is REAL array, dimension (max(NN))
*>
*>          The real and imaginary parts of the eigenvalues of A.
*>          On exit, WR + WI*i are the eigenvalues of the matrix in A.
*> \endverbatim
*>
*> \param[out] WRT
*> \verbatim
*>          WRT is REAL array, dimension (max(NN))
*> \endverbatim
*>
*> \param[out] WIT
*> \verbatim
*>          WIT is REAL array, dimension (max(NN))
*>
*>          Like WR, WI, these arrays contain the eigenvalues of A,
*>          but those computed when SGEESX only computes a partial
*>          eigendecomposition, i.e. not Schur vectors
*> \endverbatim
*>
*> \param[out] WRTMP
*> \verbatim
*>          WRTMP is REAL array, dimension (max(NN))
*> \endverbatim
*>
*> \param[out] WITMP
*> \verbatim
*>          WITMP is REAL array, dimension (max(NN))
*>
*>          More temporary storage for eigenvalues.
*> \endverbatim
*>
*> \param[out] VS
*> \verbatim
*>          VS is REAL array, dimension (LDVS, max(NN))
*>          VS holds the computed Schur vectors.
*> \endverbatim
*>
*> \param[in] LDVS
*> \verbatim
*>          LDVS is INTEGER
*>          Leading dimension of VS. Must be at least max(1,max(NN)).
*> \endverbatim
*>
*> \param[out] VS1
*> \verbatim
*>          VS1 is REAL array, dimension (LDVS, max(NN))
*>          VS1 holds another copy of the computed Schur vectors.
*> \endverbatim
*>
*> \param[out] RESULT
*> \verbatim
*>          RESULT is REAL array, dimension (17)
*>          The values computed by the 17 tests described above.
*>          The values are currently limited to 1/ulp, to avoid overflow.
*> \endverbatim
*>
*> \param[out] WORK
*> \verbatim
*>          WORK is REAL array, dimension (LWORK)
*> \endverbatim
*>
*> \param[in] LWORK
*> \verbatim
*>          LWORK is INTEGER
*>          The number of entries in WORK.  This must be at least
*>          max(3*NN(j),2*NN(j)**2) for all j.
*> \endverbatim
*>
*> \param[out] IWORK
*> \verbatim
*>          IWORK is INTEGER array, dimension (max(NN)*max(NN))
*> \endverbatim
*>
*> \param[out] BWORK
*> \verbatim
*>          BWORK is LOGICAL array, dimension (max(NN))
*> \endverbatim
*>
*> \param[out] INFO
*> \verbatim
*>          INFO is INTEGER
*>          If 0,  successful exit.
*>            <0,  input parameter -INFO is incorrect
*>            >0,  SLATMR, SLATMS, SLATME or SGET24 returned an error
*>                 code and INFO is its absolute value
*>
*>-----------------------------------------------------------------------
*>
*>     Some Local Variables and Parameters:
*>     ---- ----- --------- --- ----------
*>     ZERO, ONE       Real 0 and 1.
*>     MAXTYP          The number of types defined.
*>     NMAX            Largest value in NN.
*>     NERRS           The number of tests which have exceeded THRESH
*>     COND, CONDS,
*>     IMODE           Values to be passed to the matrix generators.
*>     ANORM           Norm of A; passed to matrix generators.
*>
*>     OVFL, UNFL      Overflow and underflow thresholds.
*>     ULP, ULPINV     Finest relative precision and its inverse.
*>     RTULP, RTULPI   Square roots of the previous 4 values.
*>             The following four arrays decode JTYPE:
*>     KTYPE(j)        The general type (1-10) for type "j".
*>     KMODE(j)        The MODE value to be passed to the matrix
*>                     generator for type "j".
*>     KMAGN(j)        The order of magnitude ( O(1),
*>                     O(overflow^(1/2) ), O(underflow^(1/2) )
*>     KCONDS(j)       Selectw whether CONDS is to be 1 or
*>                     1/sqrt(ulp).  (0 means irrelevant.)
*> \endverbatim
*
*  Authors:
*  ========
*
*> \author Univ. of Tennessee 
*> \author Univ. of California Berkeley 
*> \author Univ. of Colorado Denver 
*> \author NAG Ltd. 
*
*> \date November 2011
*
*> \ingroup single_eig
*
*  =====================================================================
      SUBROUTINE sdrvsx( NSIZES, NN, NTYPES, DOTYPE, ISEED, THRESH,
     $                   niunit, nounit, a, lda, h, ht, wr, wi, wrt,
     $                   wit, wrtmp, witmp, vs, ldvs, vs1, result, work,
     $                   lwork, iwork, bwork, info )
*
*  -- LAPACK test routine (version 3.4.0) --
*  -- LAPACK is a software package provided by Univ. of Tennessee,    --
*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
*     November 2011
*
*     .. Scalar Arguments ..
      INTEGER            info, lda, ldvs, lwork, niunit, nounit, nsizes,
     $                   ntypes
      REAL               thresh
*     ..
*     .. Array Arguments ..
      LOGICAL            bwork( * ), dotype( * )
      INTEGER            iseed( 4 ), iwork( * ), nn( * )
      REAL               a( lda, * ), h( lda, * ), ht( lda, * ),
     $                   result( 17 ), vs( ldvs, * ), vs1( ldvs, * ),
     $                   wi( * ), wit( * ), witmp( * ), work( * ),
     $                   wr( * ), wrt( * ), wrtmp( * )
*     ..
*
*  =====================================================================
*
*     .. Parameters ..
      REAL               zero, one
      parameter( zero = 0.0e0, one = 1.0e0 )
      INTEGER            maxtyp
      parameter( maxtyp = 21 )
*     ..
*     .. Local Scalars ..
      LOGICAL            badnn
      CHARACTER*3        path
      INTEGER            i, iinfo, imode, itype, iwk, j, jcol, jsize,
     $                   jtype, mtypes, n, nerrs, nfail, nmax,
     $                   nnwork, nslct, ntest, ntestf, ntestt
      REAL               anorm, cond, conds, ovfl, rcdein, rcdvin,
     $                   rtulp, rtulpi, ulp, ulpinv, unfl
*     ..
*     .. Local Arrays ..
      CHARACTER          adumma( 1 )
      INTEGER            idumma( 1 ), ioldsd( 4 ), islct( 20 ),
     $                   kconds( maxtyp ), kmagn( maxtyp ),
     $                   kmode( maxtyp ), ktype( maxtyp )
*     ..
*     .. Arrays in Common ..
      LOGICAL            selval( 20 )
      REAL               selwi( 20 ), selwr( 20 )
*     ..
*     .. Scalars in Common ..
      INTEGER            seldim, selopt
*     ..
*     .. Common blocks ..
      common             / sslct / selopt, seldim, selval, selwr, selwi
*     ..
*     .. External Functions ..
      REAL               slamch
      EXTERNAL           slamch
*     ..
*     .. External Subroutines ..
      EXTERNAL           sget24, slabad, slasum, slatme, slatmr, slatms,
     $                   slaset, xerbla
*     ..
*     .. Intrinsic Functions ..
      INTRINSIC          abs, max, min, sqrt
*     ..
*     .. Data statements ..
      DATA               ktype / 1, 2, 3, 5*4, 4*6, 6*6, 3*9 /
      DATA               kmagn / 3*1, 1, 1, 1, 2, 3, 4*1, 1, 1, 1, 1, 2,
     $                   3, 1, 2, 3 /
      DATA               kmode / 3*0, 4, 3, 1, 4, 4, 4, 3, 1, 5, 4, 3,
     $                   1, 5, 5, 5, 4, 3, 1 /
      DATA               kconds / 3*0, 5*0, 4*1, 6*2, 3*0 /
*     ..
*     .. Executable Statements ..
*
      path( 1: 1 ) = 'Single precision'
      path( 2: 3 ) = 'SX'
*
*     Check for errors
*
      ntestt = 0
      ntestf = 0
      info = 0
*
*     Important constants
*
      badnn = .false.
*
*     12 is the largest dimension in the input file of precomputed
*     problems
*
      nmax = 12
      DO 10 j = 1, nsizes
         nmax = max( nmax, nn( j ) )
         IF( nn( j ).LT.0 )
     $      badnn = .true.
   10 continue
*
*     Check for errors
*
      IF( nsizes.LT.0 ) THEN
         info = -1
      ELSE IF( badnn ) THEN
         info = -2
      ELSE IF( ntypes.LT.0 ) THEN
         info = -3
      ELSE IF( thresh.LT.zero ) THEN
         info = -6
      ELSE IF( niunit.LE.0 ) THEN
         info = -7
      ELSE IF( nounit.LE.0 ) THEN
         info = -8
      ELSE IF( lda.LT.1 .OR. lda.LT.nmax ) THEN
         info = -10
      ELSE IF( ldvs.LT.1 .OR. ldvs.LT.nmax ) THEN
         info = -20
      ELSE IF( max( 3*nmax, 2*nmax**2 ).GT.lwork ) THEN
         info = -24
      END IF
*
      IF( info.NE.0 ) THEN
         CALL xerbla( 'SDRVSX', -info )
         return
      END IF
*
*     If nothing to do check on NIUNIT
*
      IF( nsizes.EQ.0 .OR. ntypes.EQ.0 )
     $   go to 150
*
*     More Important constants
*
      unfl = slamch( 'Safe minimum' )
      ovfl = one / unfl
      CALL slabad( unfl, ovfl )
      ulp = slamch( 'Precision' )
      ulpinv = one / ulp
      rtulp = sqrt( ulp )
      rtulpi = one / rtulp
*
*     Loop over sizes, types
*
      nerrs = 0
*
      DO 140 jsize = 1, nsizes
         n = nn( jsize )
         IF( nsizes.NE.1 ) THEN
            mtypes = min( maxtyp, ntypes )
         ELSE
            mtypes = min( maxtyp+1, ntypes )
         END IF
*
         DO 130 jtype = 1, mtypes
            IF( .NOT.dotype( jtype ) )
     $         go to 130
*
*           Save ISEED in case of an error.
*
            DO 20 j = 1, 4
               ioldsd( j ) = iseed( j )
   20       continue
*
*           Compute "A"
*
*           Control parameters:
*
*           KMAGN  KCONDS  KMODE        KTYPE
*       =1  O(1)   1       clustered 1  zero
*       =2  large  large   clustered 2  identity
*       =3  small          exponential  Jordan
*       =4                 arithmetic   diagonal, (w/ eigenvalues)
*       =5                 random log   symmetric, w/ eigenvalues
*       =6                 random       general, w/ eigenvalues
*       =7                              random diagonal
*       =8                              random symmetric
*       =9                              random general
*       =10                             random triangular
*
            IF( mtypes.GT.maxtyp )
     $         go to 90
*
            itype = ktype( jtype )
            imode = kmode( jtype )
*
*           Compute norm
*
            go to( 30, 40, 50 )kmagn( jtype )
*
   30       continue
            anorm = one
            go to 60
*
   40       continue
            anorm = ovfl*ulp
            go to 60
*
   50       continue
            anorm = unfl*ulpinv
            go to 60
*
   60       continue
*
            CALL slaset( 'Full', lda, n, zero, zero, a, lda )
            iinfo = 0
            cond = ulpinv
*
*           Special Matrices -- Identity & Jordan block
*
*              Zero
*
            IF( itype.EQ.1 ) THEN
               iinfo = 0
*
            ELSE IF( itype.EQ.2 ) THEN
*
*              Identity
*
               DO 70 jcol = 1, n
                  a( jcol, jcol ) = anorm
   70          continue
*
            ELSE IF( itype.EQ.3 ) THEN
*
*              Jordan Block
*
               DO 80 jcol = 1, n
                  a( jcol, jcol ) = anorm
                  IF( jcol.GT.1 )
     $               a( jcol, jcol-1 ) = one
   80          continue
*
            ELSE IF( itype.EQ.4 ) THEN
*
*              Diagonal Matrix, [Eigen]values Specified
*
               CALL slatms( n, n, 'S', iseed, 'S', work, imode, cond,
     $                      anorm, 0, 0, 'N', a, lda, work( n+1 ),
     $                      iinfo )
*
            ELSE IF( itype.EQ.5 ) THEN
*
*              Symmetric, eigenvalues specified
*
               CALL slatms( n, n, 'S', iseed, 'S', work, imode, cond,
     $                      anorm, n, n, 'N', a, lda, work( n+1 ),
     $                      iinfo )
*
            ELSE IF( itype.EQ.6 ) THEN
*
*              General, eigenvalues specified
*
               IF( kconds( jtype ).EQ.1 ) THEN
                  conds = one
               ELSE IF( kconds( jtype ).EQ.2 ) THEN
                  conds = rtulpi
               ELSE
                  conds = zero
               END IF
*
               adumma( 1 ) = ' '
               CALL slatme( n, 'S', iseed, work, imode, cond, one,
     $                      adumma, 'T', 'T', 'T', work( n+1 ), 4,
     $                      conds, n, n, anorm, a, lda, work( 2*n+1 ),
     $                      iinfo )
*
            ELSE IF( itype.EQ.7 ) THEN
*
*              Diagonal, random eigenvalues
*
               CALL slatmr( n, n, 'S', iseed, 'S', work, 6, one, one,
     $                      'T', 'N', work( n+1 ), 1, one,
     $                      work( 2*n+1 ), 1, one, 'N', idumma, 0, 0,
     $                      zero, anorm, 'NO', a, lda, iwork, iinfo )
*
            ELSE IF( itype.EQ.8 ) THEN
*
*              Symmetric, random eigenvalues
*
               CALL slatmr( n, n, 'S', iseed, 'S', work, 6, one, one,
     $                      'T', 'N', work( n+1 ), 1, one,
     $                      work( 2*n+1 ), 1, one, 'N', idumma, n, n,
     $                      zero, anorm, 'NO', a, lda, iwork, iinfo )
*
            ELSE IF( itype.EQ.9 ) THEN
*
*              General, random eigenvalues
*
               CALL slatmr( n, n, 'S', iseed, 'N', work, 6, one, one,
     $                      'T', 'N', work( n+1 ), 1, one,
     $                      work( 2*n+1 ), 1, one, 'N', idumma, n, n,
     $                      zero, anorm, 'NO', a, lda, iwork, iinfo )
               IF( n.GE.4 ) THEN
                  CALL slaset( 'Full', 2, n, zero, zero, a, lda )
                  CALL slaset( 'Full', n-3, 1, zero, zero, a( 3, 1 ),
     $                         lda )
                  CALL slaset( 'Full', n-3, 2, zero, zero, a( 3, n-1 ),
     $                         lda )
                  CALL slaset( 'Full', 1, n, zero, zero, a( n, 1 ),
     $                         lda )
               END IF
*
            ELSE IF( itype.EQ.10 ) THEN
*
*              Triangular, random eigenvalues
*
               CALL slatmr( n, n, 'S', iseed, 'N', work, 6, one, one,
     $                      'T', 'N', work( n+1 ), 1, one,
     $                      work( 2*n+1 ), 1, one, 'N', idumma, n, 0,
     $                      zero, anorm, 'NO', a, lda, iwork, iinfo )
*
            ELSE
*
               iinfo = 1
            END IF
*
            IF( iinfo.NE.0 ) THEN
               WRITE( nounit, fmt = 9991 )'Generator', iinfo, n, jtype,
     $            ioldsd
               info = abs( iinfo )
               return
            END IF
*
   90       continue
*
*           Test for minimal and generous workspace
*
            DO 120 iwk = 1, 2
               IF( iwk.EQ.1 ) THEN
                  nnwork = 3*n
               ELSE
                  nnwork = max( 3*n, 2*n*n )
               END IF
               nnwork = max( nnwork, 1 )
*
               CALL sget24( .false., jtype, thresh, ioldsd, nounit, n,
     $                      a, lda, h, ht, wr, wi, wrt, wit, wrtmp,
     $                      witmp, vs, ldvs, vs1, rcdein, rcdvin, nslct,
     $                      islct, result, work, nnwork, iwork, bwork,
     $                      info )
*
*              Check for RESULT(j) > THRESH
*
               ntest = 0
               nfail = 0
               DO 100 j = 1, 15
                  IF( result( j ).GE.zero )
     $               ntest = ntest + 1
                  IF( result( j ).GE.thresh )
     $               nfail = nfail + 1
  100          continue
*
               IF( nfail.GT.0 )
     $            ntestf = ntestf + 1
               IF( ntestf.EQ.1 ) THEN
                  WRITE( nounit, fmt = 9999 )path
                  WRITE( nounit, fmt = 9998 )
                  WRITE( nounit, fmt = 9997 )
                  WRITE( nounit, fmt = 9996 )
                  WRITE( nounit, fmt = 9995 )thresh
                  WRITE( nounit, fmt = 9994 )
                  ntestf = 2
               END IF
*
               DO 110 j = 1, 15
                  IF( result( j ).GE.thresh ) THEN
                     WRITE( nounit, fmt = 9993 )n, iwk, ioldsd, jtype,
     $                  j, result( j )
                  END IF
  110          continue
*
               nerrs = nerrs + nfail
               ntestt = ntestt + ntest
*
  120       continue
  130    continue
  140 continue
*
  150 continue
*
*     Read in data from file to check accuracy of condition estimation
*     Read input data until N=0
*
      jtype = 0
  160 continue
      READ( niunit, fmt = *, END = 200 )n, nslct
      IF( n.EQ.0 )
     $   go to 200
      jtype = jtype + 1
      iseed( 1 ) = jtype
      IF( nslct.GT.0 )
     $   READ( niunit, fmt = * )( islct( i ), i = 1, nslct )
      DO 170 i = 1, n
         READ( niunit, fmt = * )( a( i, j ), j = 1, n )
  170 continue
      READ( niunit, fmt = * )rcdein, rcdvin
*
      CALL sget24( .true., 22, thresh, iseed, nounit, n, a, lda, h, ht,
     $             wr, wi, wrt, wit, wrtmp, witmp, vs, ldvs, vs1,
     $             rcdein, rcdvin, nslct, islct, result, work, lwork,
     $             iwork, bwork, info )
*
*     Check for RESULT(j) > THRESH
*
      ntest = 0
      nfail = 0
      DO 180 j = 1, 17
         IF( result( j ).GE.zero )
     $      ntest = ntest + 1
         IF( result( j ).GE.thresh )
     $      nfail = nfail + 1
  180 continue
*
      IF( nfail.GT.0 )
     $   ntestf = ntestf + 1
      IF( ntestf.EQ.1 ) THEN
         WRITE( nounit, fmt = 9999 )path
         WRITE( nounit, fmt = 9998 )
         WRITE( nounit, fmt = 9997 )
         WRITE( nounit, fmt = 9996 )
         WRITE( nounit, fmt = 9995 )thresh
         WRITE( nounit, fmt = 9994 )
         ntestf = 2
      END IF
      DO 190 j = 1, 17
         IF( result( j ).GE.thresh ) THEN
            WRITE( nounit, fmt = 9992 )n, jtype, j, result( j )
         END IF
  190 continue
*
      nerrs = nerrs + nfail
      ntestt = ntestt + ntest
      go to 160
  200 continue
*
*     Summary
*
      CALL slasum( path, nounit, nerrs, ntestt )
*
 9999 format( / 1x, a3, ' -- Real Schur Form Decomposition Expert ',
     $      'Driver', / ' Matrix types (see SDRVSX for details):' )
*
 9998 format( / ' Special Matrices:', / '  1=Zero matrix.             ',
     $      '           ', '  5=Diagonal: geometr. spaced entries.',
     $      / '  2=Identity matrix.                    ', '  6=Diagona',
     $      'l: clustered entries.', / '  3=Transposed Jordan block.  ',
     $      '          ', '  7=Diagonal: large, evenly spaced.', / '  ',
     $      '4=Diagonal: evenly spaced entries.    ', '  8=Diagonal: s',
     $      'mall, evenly spaced.' )
 9997 format( ' Dense, Non-Symmetric Matrices:', / '  9=Well-cond., ev',
     $      'enly spaced eigenvals.', ' 14=Ill-cond., geomet. spaced e',
     $      'igenals.', / ' 10=Well-cond., geom. spaced eigenvals. ',
     $      ' 15=Ill-conditioned, clustered e.vals.', / ' 11=Well-cond',
     $      'itioned, clustered e.vals. ', ' 16=Ill-cond., random comp',
     $      'lex ', / ' 12=Well-cond., random complex ', '         ',
     $      ' 17=Ill-cond., large rand. complx ', / ' 13=Ill-condi',
     $      'tioned, evenly spaced.     ', ' 18=Ill-cond., small rand.',
     $      ' complx ' )
 9996 format( ' 19=Matrix with random O(1) entries.    ', ' 21=Matrix ',
     $      'with small random entries.', / ' 20=Matrix with large ran',
     $      'dom entries.   ', / )
 9995 format( ' Tests performed with test threshold =', f8.2,
     $      / ' ( A denotes A on input and T denotes A on output)',
     $      / / ' 1 = 0 if T in Schur form (no sort), ',
     $      '  1/ulp otherwise', /
     $      ' 2 = | A - VS T transpose(VS) | / ( n |A| ulp ) (no sort)',
     $      / ' 3 = | I - VS transpose(VS) | / ( n ulp ) (no sort) ', /
     $      ' 4 = 0 if WR+sqrt(-1)*WI are eigenvalues of T (no sort),',
     $      '  1/ulp otherwise', /
     $      ' 5 = 0 if T same no matter if VS computed (no sort),',
     $      '  1/ulp otherwise', /
     $      ' 6 = 0 if WR, WI same no matter if VS computed (no sort)',
     $      ',  1/ulp otherwise' )
 9994 format( ' 7 = 0 if T in Schur form (sort), ', '  1/ulp otherwise',
     $      / ' 8 = | A - VS T transpose(VS) | / ( n |A| ulp ) (sort)',
     $      / ' 9 = | I - VS transpose(VS) | / ( n ulp ) (sort) ',
     $      / ' 10 = 0 if WR+sqrt(-1)*WI are eigenvalues of T (sort),',
     $      '  1/ulp otherwise', /
     $      ' 11 = 0 if T same no matter what else computed (sort),',
     $      '  1/ulp otherwise', /
     $      ' 12 = 0 if WR, WI same no matter what else computed ',
     $      '(sort), 1/ulp otherwise', /
     $      ' 13 = 0 if sorting succesful, 1/ulp otherwise',
     $      / ' 14 = 0 if RCONDE same no matter what else computed,',
     $      ' 1/ulp otherwise', /
     $      ' 15 = 0 if RCONDv same no matter what else computed,',
     $      ' 1/ulp otherwise', /
     $      ' 16 = | RCONDE - RCONDE(precomputed) | / cond(RCONDE),',
     $      / ' 17 = | RCONDV - RCONDV(precomputed) | / cond(RCONDV),' )
 9993 format( ' N=', i5, ', IWK=', i2, ', seed=', 4( i4, ',' ),
     $      ' type ', i2, ', test(', i2, ')=', g10.3 )
 9992 format( ' N=', i5, ', input example =', i3, ',  test(', i2, ')=',
     $      g10.3 )
 9991 format( ' SDRVSX: ', a, ' returned INFO=', i6, '.', / 9x, 'N=',
     $      i6, ', JTYPE=', i6, ', ISEED=(', 3( i5, ',' ), i5, ')' )
*
      return
*
*     End of SDRVSX
*
      END