LAPACK: TESTING/EIG/dckgsv.f Source File

Go to the documentation of this file.
00001       SUBROUTINE DCKGSV( NM, MVAL, PVAL, NVAL, NMATS, ISEED, THRESH,
00002      $                   NMAX, A, AF, B, BF, U, V, Q, ALPHA, BETA, R,
00003      $                   IWORK, WORK, RWORK, NIN, NOUT, INFO )
00004 *
00005 *  -- LAPACK test routine (version 3.1) --
00006 *     Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
00007 *     November 2006
00008 *
00009 *     .. Scalar Arguments ..
00010       INTEGER            INFO, NIN, NM, NMATS, NMAX, NOUT
00011       DOUBLE PRECISION   THRESH
00012 *     ..
00013 *     .. Array Arguments ..
00014       INTEGER            ISEED( 4 ), IWORK( * ), MVAL( * ), NVAL( * ),
00015      $                   PVAL( * )
00016       DOUBLE PRECISION   A( * ), AF( * ), ALPHA( * ), B( * ), BETA( * ),
00017      $                   BF( * ), Q( * ), R( * ), RWORK( * ), U( * ),
00018      $                   V( * ), WORK( * )
00019 *     ..
00020 *
00021 *  Purpose
00022 *  =======
00023 *
00024 *  DCKGSV tests DGGSVD:
00025 *         the GSVD for M-by-N matrix A and P-by-N matrix B.
00026 *
00027 *  Arguments
00028 *  =========
00029 *
00030 *  NM      (input) INTEGER
00031 *          The number of values of M contained in the vector MVAL.
00032 *
00033 *  MVAL    (input) INTEGER array, dimension (NM)
00034 *          The values of the matrix row dimension M.
00035 *
00036 *  PVAL    (input) INTEGER array, dimension (NP)
00037 *          The values of the matrix row dimension P.
00038 *
00039 *  NVAL    (input) INTEGER array, dimension (NN)
00040 *          The values of the matrix column dimension N.
00041 *
00042 *  NMATS   (input) INTEGER
00043 *          The number of matrix types to be tested for each combination
00044 *          of matrix dimensions.  If NMATS >= NTYPES (the maximum
00045 *          number of matrix types), then all the different types are
00046 *          generated for testing.  If NMATS < NTYPES, another input line
00047 *          is read to get the numbers of the matrix types to be used.
00048 *
00049 *  ISEED   (input/output) INTEGER array, dimension (4)
00050 *          On entry, the seed of the random number generator.  The array
00051 *          elements should be between 0 and 4095, otherwise they will be
00052 *          reduced mod 4096, and ISEED(4) must be odd.
00053 *          On exit, the next seed in the random number sequence after
00054 *          all the test matrices have been generated.
00055 *
00056 *  THRESH  (input) DOUBLE PRECISION
00057 *          The threshold value for the test ratios.  A result is
00058 *          included in the output file if RESULT >= THRESH.  To have
00059 *          every test ratio printed, use THRESH = 0.
00060 *
00061 *  NMAX    (input) INTEGER
00062 *          The maximum value permitted for M or N, used in dimensioning
00063 *          the work arrays.
00064 *
00065 *  A       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00066 *
00067 *  AF      (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00068 *
00069 *  B       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00070 *
00071 *  BF      (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00072 *
00073 *  U       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00074 *
00075 *  V       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00076 *
00077 *  Q       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00078 *
00079 *  ALPHA   (workspace) DOUBLE PRECISION array, dimension (NMAX)
00080 *
00081 *  BETA    (workspace) DOUBLE PRECISION array, dimension (NMAX)
00082 *
00083 *  R       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00084 *
00085 *  IWORK   (workspace) INTEGER array, dimension (NMAX)
00086 *
00087 *  WORK    (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00088 *
00089 *  RWORK   (workspace) DOUBLE PRECISION array, dimension (NMAX)
00090 *
00091 *  NIN     (input) INTEGER
00092 *          The unit number for input.
00093 *
00094 *  NOUT    (input) INTEGER
00095 *          The unit number for output.
00096 *
00097 *  INFO    (output) INTEGER
00098 *          = 0 :  successful exit
00099 *          > 0 :  If DLATMS returns an error code, the absolute value
00100 *                 of it is returned.
00101 *
00102 *  =====================================================================
00103 *
00104 *     .. Parameters ..
00105       INTEGER            NTESTS
00106       PARAMETER          ( NTESTS = 7 )
00107       INTEGER            NTYPES
00108       PARAMETER          ( NTYPES = 8 )
00109 *     ..
00110 *     .. Local Scalars ..
00111       LOGICAL            FIRSTT
00112       CHARACTER          DISTA, DISTB, TYPE
00113       CHARACTER*3        PATH
00114       INTEGER            I, IINFO, IM, IMAT, KLA, KLB, KUA, KUB, LDA,
00115      $                   LDB, LDQ, LDR, LDU, LDV, LWORK, M, MODEA,
00116      $                   MODEB, N, NFAIL, NRUN, NT, P
00117       DOUBLE PRECISION   ANORM, BNORM, CNDNMA, CNDNMB
00118 *     ..
00119 *     .. Local Arrays ..
00120       LOGICAL            DOTYPE( NTYPES )
00121       DOUBLE PRECISION   RESULT( NTESTS )
00122 *     ..
00123 *     .. External Subroutines ..
00124       EXTERNAL           ALAHDG, ALAREQ, ALASUM, DGSVTS, DLATB9, DLATMS
00125 *     ..
00126 *     .. Intrinsic Functions ..
00127       INTRINSIC          ABS
00128 *     ..
00129 *     .. Executable Statements ..
00130 *
00131 *     Initialize constants and the random number seed.
00132 *
00133       PATH( 1: 3 ) = 'GSV'
00134       INFO = 0
00135       NRUN = 0
00136       NFAIL = 0
00137       FIRSTT = .TRUE.
00138       CALL ALAREQ( PATH, NMATS, DOTYPE, NTYPES, NIN, NOUT )
00139       LDA = NMAX
00140       LDB = NMAX
00141       LDU = NMAX
00142       LDV = NMAX
00143       LDQ = NMAX
00144       LDR = NMAX
00145       LWORK = NMAX*NMAX
00146 *
00147 *     Do for each value of M in MVAL.
00148 *
00149       DO 30 IM = 1, NM
00150          M = MVAL( IM )
00151          P = PVAL( IM )
00152          N = NVAL( IM )
00153 *
00154          DO 20 IMAT = 1, NTYPES
00155 *
00156 *           Do the tests only if DOTYPE( IMAT ) is true.
00157 *
00158             IF( .NOT.DOTYPE( IMAT ) )
00159      $         GO TO 20
00160 *
00161 *           Set up parameters with DLATB9 and generate test
00162 *           matrices A and B with DLATMS.
00163 *
00164             CALL DLATB9( PATH, IMAT, M, P, N, TYPE, KLA, KUA, KLB, KUB,
00165      $                   ANORM, BNORM, MODEA, MODEB, CNDNMA, CNDNMB,
00166      $                   DISTA, DISTB )
00167 *
00168 *           Generate M by N matrix A
00169 *
00170             CALL DLATMS( M, N, DISTA, ISEED, TYPE, RWORK, MODEA, CNDNMA,
00171      $                   ANORM, KLA, KUA, 'No packing', A, LDA, WORK,
00172      $                   IINFO )
00173             IF( IINFO.NE.0 ) THEN
00174                WRITE( NOUT, FMT = 9999 )IINFO
00175                INFO = ABS( IINFO )
00176                GO TO 20
00177             END IF
00178 *
00179             CALL DLATMS( P, N, DISTB, ISEED, TYPE, RWORK, MODEB, CNDNMB,
00180      $                   BNORM, KLB, KUB, 'No packing', B, LDB, WORK,
00181      $                   IINFO )
00182             IF( IINFO.NE.0 ) THEN
00183                WRITE( NOUT, FMT = 9999 )IINFO
00184                INFO = ABS( IINFO )
00185                GO TO 20
00186             END IF
00187 *
00188             NT = 6
00189 *
00190             CALL DGSVTS( M, P, N, A, AF, LDA, B, BF, LDB, U, LDU, V,
00191      $                   LDV, Q, LDQ, ALPHA, BETA, R, LDR, IWORK, WORK,
00192      $                   LWORK, RWORK, RESULT )
00193 *
00194 *           Print information about the tests that did not
00195 *           pass the threshold.
00196 *
00197             DO 10 I = 1, NT
00198                IF( RESULT( I ).GE.THRESH ) THEN
00199                   IF( NFAIL.EQ.0 .AND. FIRSTT ) THEN
00200                      FIRSTT = .FALSE.
00201                      CALL ALAHDG( NOUT, PATH )
00202                   END IF
00203                   WRITE( NOUT, FMT = 9998 )M, P, N, IMAT, I,
00204      $               RESULT( I )
00205                   NFAIL = NFAIL + 1
00206                END IF
00207    10       CONTINUE
00208             NRUN = NRUN + NT
00209    20    CONTINUE
00210    30 CONTINUE
00211 *
00212 *     Print a summary of the results.
00213 *
00214       CALL ALASUM( PATH, NOUT, NFAIL, NRUN, 0 )
00215 *
00216  9999 FORMAT( ' DLATMS in DCKGSV   INFO = ', I5 )
00217  9998 FORMAT( ' M=', I4, ' P=', I4, ', N=', I4, ', type ', I2,
00218      $      ', test ', I2, ', ratio=', G13.6 )
00219       RETURN
00220 *
00221 *     End of DCKGSV
00222 *
00223       END