LAPACK: TESTING/EIG/dckglm.f Source File

Go to the documentation of this file.
00001       SUBROUTINE DCKGLM( NN, MVAL, PVAL, NVAL, NMATS, ISEED, THRESH,
00002      $                   NMAX, A, AF, B, BF, X, WORK, RWORK, NIN, NOUT,
00003      $                   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, NMATS, NMAX, NN, NOUT
00011       DOUBLE PRECISION   THRESH
00012 *     ..
00013 *     .. Array Arguments ..
00014       INTEGER            ISEED( 4 ), MVAL( * ), NVAL( * ), PVAL( * )
00015       DOUBLE PRECISION   A( * ), AF( * ), B( * ), BF( * ), RWORK( * ),
00016      $                   WORK( * ), X( * )
00017 *     ..
00018 *
00019 *  Purpose
00020 *  =======
00021 *
00022 *  DCKGLM tests DGGGLM - subroutine for solving generalized linear
00023 *                        model problem.
00024 *
00025 *  Arguments
00026 *  =========
00027 *
00028 *  NN      (input) INTEGER
00029 *          The number of values of N, M and P contained in the vectors
00030 *          NVAL, MVAL and PVAL.
00031 *
00032 *  MVAL    (input) INTEGER array, dimension (NN)
00033 *          The values of the matrix column dimension M.
00034 *
00035 *  PVAL    (input) INTEGER array, dimension (NN)
00036 *          The values of the matrix column dimension P.
00037 *
00038 *  NVAL    (input) INTEGER array, dimension (NN)
00039 *          The values of the matrix row dimension N.
00040 *
00041 *  NMATS   (input) INTEGER
00042 *          The number of matrix types to be tested for each combination
00043 *          of matrix dimensions.  If NMATS >= NTYPES (the maximum
00044 *          number of matrix types), then all the different types are
00045 *          generated for testing.  If NMATS < NTYPES, another input line
00046 *          is read to get the numbers of the matrix types to be used.
00047 *
00048 *  ISEED   (input/output) INTEGER array, dimension (4)
00049 *          On entry, the seed of the random number generator.  The array
00050 *          elements should be between 0 and 4095, otherwise they will be
00051 *          reduced mod 4096, and ISEED(4) must be odd.
00052 *          On exit, the next seed in the random number sequence after
00053 *          all the test matrices have been generated.
00054 *
00055 *  THRESH  (input) DOUBLE PRECISION
00056 *          The threshold value for the test ratios.  A result is
00057 *          included in the output file if RESID >= THRESH.  To have
00058 *          every test ratio printed, use THRESH = 0.
00059 *
00060 *  NMAX    (input) INTEGER
00061 *          The maximum value permitted for M or N, used in dimensioning
00062 *          the work arrays.
00063 *
00064 *  A       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00065 *
00066 *  AF      (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00067 *
00068 *  B       (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00069 *
00070 *  BF      (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00071 *
00072 *  X       (workspace) DOUBLE PRECISION array, dimension (4*NMAX)
00073 *
00074 *  RWORK   (workspace) DOUBLE PRECISION array, dimension (NMAX)
00075 *
00076 *  WORK    (workspace) DOUBLE PRECISION array, dimension (NMAX*NMAX)
00077 *
00078 *  NIN     (input) INTEGER
00079 *          The unit number for input.
00080 *
00081 *  NOUT    (input) INTEGER
00082 *          The unit number for output.
00083 *
00084 *  INFO    (output) INTEGER
00085 *          = 0 :  successful exit
00086 *          > 0 :  If DLATMS returns an error code, the absolute value
00087 *                 of it is returned.
00088 *
00089 *  =====================================================================
00090 *
00091 *     .. Parameters ..
00092       INTEGER            NTYPES
00093       PARAMETER          ( NTYPES = 8 )
00094 *     ..
00095 *     .. Local Scalars ..
00096       LOGICAL            FIRSTT
00097       CHARACTER          DISTA, DISTB, TYPE
00098       CHARACTER*3        PATH
00099       INTEGER            I, IINFO, IK, IMAT, KLA, KLB, KUA, KUB, LDA,
00100      $                   LDB, LWORK, M, MODEA, MODEB, N, NFAIL, NRUN, P
00101       DOUBLE PRECISION   ANORM, BNORM, CNDNMA, CNDNMB, RESID
00102 *     ..
00103 *     .. Local Arrays ..
00104       LOGICAL            DOTYPE( NTYPES )
00105 *     ..
00106 *     .. External Functions ..
00107       DOUBLE PRECISION   DLARND
00108       EXTERNAL           DLARND
00109 *     ..
00110 *     .. External Subroutines ..
00111       EXTERNAL           ALAHDG, ALAREQ, ALASUM, DGLMTS, DLATB9, DLATMS
00112 *     ..
00113 *     .. Intrinsic Functions ..
00114       INTRINSIC          ABS
00115 *     ..
00116 *     .. Executable Statements ..
00117 *
00118 *     Initialize constants.
00119 *
00120       PATH( 1: 3 ) = 'GLM'
00121       INFO = 0
00122       NRUN = 0
00123       NFAIL = 0
00124       FIRSTT = .TRUE.
00125       CALL ALAREQ( PATH, NMATS, DOTYPE, NTYPES, NIN, NOUT )
00126       LDA = NMAX
00127       LDB = NMAX
00128       LWORK = NMAX*NMAX
00129 *
00130 *     Check for valid input values.
00131 *
00132       DO 10 IK = 1, NN
00133          M = MVAL( IK )
00134          P = PVAL( IK )
00135          N = NVAL( IK )
00136          IF( M.GT.N .OR. N.GT.M+P ) THEN
00137             IF( FIRSTT ) THEN
00138                WRITE( NOUT, FMT = * )
00139                FIRSTT = .FALSE.
00140             END IF
00141             WRITE( NOUT, FMT = 9997 )M, P, N
00142          END IF
00143    10 CONTINUE
00144       FIRSTT = .TRUE.
00145 *
00146 *     Do for each value of M in MVAL.
00147 *
00148       DO 40 IK = 1, NN
00149          M = MVAL( IK )
00150          P = PVAL( IK )
00151          N = NVAL( IK )
00152          IF( M.GT.N .OR. N.GT.M+P )
00153      $      GO TO 40
00154 *
00155          DO 30 IMAT = 1, NTYPES
00156 *
00157 *           Do the tests only if DOTYPE( IMAT ) is true.
00158 *
00159             IF( .NOT.DOTYPE( IMAT ) )
00160      $         GO TO 30
00161 *
00162 *           Set up parameters with DLATB9 and generate test
00163 *           matrices A and B with DLATMS.
00164 *
00165             CALL DLATB9( PATH, IMAT, M, P, N, TYPE, KLA, KUA, KLB, KUB,
00166      $                   ANORM, BNORM, MODEA, MODEB, CNDNMA, CNDNMB,
00167      $                   DISTA, DISTB )
00168 *
00169             CALL DLATMS( N, M, DISTA, ISEED, TYPE, RWORK, MODEA, CNDNMA,
00170      $                   ANORM, KLA, KUA, 'No packing', A, LDA, WORK,
00171      $                   IINFO )
00172             IF( IINFO.NE.0 ) THEN
00173                WRITE( NOUT, FMT = 9999 )IINFO
00174                INFO = ABS( IINFO )
00175                GO TO 30
00176             END IF
00177 *
00178             CALL DLATMS( N, P, DISTB, ISEED, TYPE, RWORK, MODEB, CNDNMB,
00179      $                   BNORM, KLB, KUB, 'No packing', B, LDB, WORK,
00180      $                   IINFO )
00181             IF( IINFO.NE.0 ) THEN
00182                WRITE( NOUT, FMT = 9999 )IINFO
00183                INFO = ABS( IINFO )
00184                GO TO 30
00185             END IF
00186 *
00187 *           Generate random left hand side vector of GLM
00188 *
00189             DO 20 I = 1, N
00190                X( I ) = DLARND( 2, ISEED )
00191    20       CONTINUE
00192 *
00193             CALL DGLMTS( N, M, P, A, AF, LDA, B, BF, LDB, X,
00194      $                   X( NMAX+1 ), X( 2*NMAX+1 ), X( 3*NMAX+1 ),
00195      $                   WORK, LWORK, RWORK, RESID )
00196 *
00197 *           Print information about the tests that did not
00198 *           pass the threshold.
00199 *
00200             IF( RESID.GE.THRESH ) THEN
00201                IF( NFAIL.EQ.0 .AND. FIRSTT ) THEN
00202                   FIRSTT = .FALSE.
00203                   CALL ALAHDG( NOUT, PATH )
00204                END IF
00205                WRITE( NOUT, FMT = 9998 )N, M, P, IMAT, 1, RESID
00206                NFAIL = NFAIL + 1
00207             END IF
00208             NRUN = NRUN + 1
00209 *
00210    30    CONTINUE
00211    40 CONTINUE
00212 *
00213 *     Print a summary of the results.
00214 *
00215       CALL ALASUM( PATH, NOUT, NFAIL, NRUN, 0 )
00216 *
00217  9999 FORMAT( ' DLATMS in DCKGLM INFO = ', I5 )
00218  9998 FORMAT( ' N=', I4, ' M=', I4, ', P=', I4, ', type ', I2,
00219      $      ', test ', I2, ', ratio=', G13.6 )
00220  9997 FORMAT( ' *** Invalid input  for GLM:  M = ', I6, ', P = ', I6,
00221      $      ', N = ', I6, ';', / '     must satisfy M <= N <= M+P  ',
00222      $      '(this set of values will be skipped)' )
00223       RETURN
00224 *
00225 *     End of DCKGLM
00226 *
00227       END