 LAPACK  3.10.1 LAPACK: Linear Algebra PACKage

## ◆ dckglm()

 subroutine dckglm ( integer NN, integer, dimension( * ) MVAL, integer, dimension( * ) PVAL, integer, dimension( * ) NVAL, integer NMATS, integer, dimension( 4 ) ISEED, double precision THRESH, integer NMAX, double precision, dimension( * ) A, double precision, dimension( * ) AF, double precision, dimension( * ) B, double precision, dimension( * ) BF, double precision, dimension( * ) X, double precision, dimension( * ) WORK, double precision, dimension( * ) RWORK, integer NIN, integer NOUT, integer INFO )

DCKGLM

Purpose:
``` DCKGLM tests DGGGLM - subroutine for solving generalized linear
model problem.```
Parameters
 [in] NN ``` NN is INTEGER The number of values of N, M and P contained in the vectors NVAL, MVAL and PVAL.``` [in] MVAL ``` MVAL is INTEGER array, dimension (NN) The values of the matrix column dimension M.``` [in] PVAL ``` PVAL is INTEGER array, dimension (NN) The values of the matrix column dimension P.``` [in] NVAL ``` NVAL is INTEGER array, dimension (NN) The values of the matrix row dimension N.``` [in] NMATS ``` NMATS is INTEGER The number of matrix types to be tested for each combination of matrix dimensions. If NMATS >= NTYPES (the maximum number of matrix types), then all the different types are generated for testing. If NMATS < NTYPES, another input line is read to get the numbers of the matrix types to be used.``` [in,out] ISEED ``` ISEED is INTEGER array, dimension (4) On entry, the seed of the random number generator. The array elements should be between 0 and 4095, otherwise they will be reduced mod 4096, and ISEED(4) must be odd. On exit, the next seed in the random number sequence after all the test matrices have been generated.``` [in] THRESH ``` THRESH is DOUBLE PRECISION The threshold value for the test ratios. A result is included in the output file if RESID >= THRESH. To have every test ratio printed, use THRESH = 0.``` [in] NMAX ``` NMAX is INTEGER The maximum value permitted for M or N, used in dimensioning the work arrays.``` [out] A ` A is DOUBLE PRECISION array, dimension (NMAX*NMAX)` [out] AF ` AF is DOUBLE PRECISION array, dimension (NMAX*NMAX)` [out] B ` B is DOUBLE PRECISION array, dimension (NMAX*NMAX)` [out] BF ` BF is DOUBLE PRECISION array, dimension (NMAX*NMAX)` [out] X ` X is DOUBLE PRECISION array, dimension (4*NMAX)` [out] RWORK ` RWORK is DOUBLE PRECISION array, dimension (NMAX)` [out] WORK ` WORK is DOUBLE PRECISION array, dimension (NMAX*NMAX)` [in] NIN ``` NIN is INTEGER The unit number for input.``` [in] NOUT ``` NOUT is INTEGER The unit number for output.``` [out] INFO ``` INFO is INTEGER = 0 : successful exit > 0 : If DLATMS returns an error code, the absolute value of it is returned.```

Definition at line 164 of file dckglm.f.

167 *
168 * -- LAPACK test routine --
169 * -- LAPACK is a software package provided by Univ. of Tennessee, --
170 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
171 *
172 * .. Scalar Arguments ..
173  INTEGER INFO, NIN, NMATS, NMAX, NN, NOUT
174  DOUBLE PRECISION THRESH
175 * ..
176 * .. Array Arguments ..
177  INTEGER ISEED( 4 ), MVAL( * ), NVAL( * ), PVAL( * )
178  DOUBLE PRECISION A( * ), AF( * ), B( * ), BF( * ), RWORK( * ),
179  \$ WORK( * ), X( * )
180 * ..
181 *
182 * =====================================================================
183 *
184 * .. Parameters ..
185  INTEGER NTYPES
186  parameter( ntypes = 8 )
187 * ..
188 * .. Local Scalars ..
189  LOGICAL FIRSTT
190  CHARACTER DISTA, DISTB, TYPE
191  CHARACTER*3 PATH
192  INTEGER I, IINFO, IK, IMAT, KLA, KLB, KUA, KUB, LDA,
193  \$ LDB, LWORK, M, MODEA, MODEB, N, NFAIL, NRUN, P
194  DOUBLE PRECISION ANORM, BNORM, CNDNMA, CNDNMB, RESID
195 * ..
196 * .. Local Arrays ..
197  LOGICAL DOTYPE( NTYPES )
198 * ..
199 * .. External Functions ..
200  DOUBLE PRECISION DLARND
201  EXTERNAL dlarnd
202 * ..
203 * .. External Subroutines ..
204  EXTERNAL alahdg, alareq, alasum, dglmts, dlatb9, dlatms
205 * ..
206 * .. Intrinsic Functions ..
207  INTRINSIC abs
208 * ..
209 * .. Executable Statements ..
210 *
211 * Initialize constants.
212 *
213  path( 1: 3 ) = 'GLM'
214  info = 0
215  nrun = 0
216  nfail = 0
217  firstt = .true.
218  CALL alareq( path, nmats, dotype, ntypes, nin, nout )
219  lda = nmax
220  ldb = nmax
221  lwork = nmax*nmax
222 *
223 * Check for valid input values.
224 *
225  DO 10 ik = 1, nn
226  m = mval( ik )
227  p = pval( ik )
228  n = nval( ik )
229  IF( m.GT.n .OR. n.GT.m+p ) THEN
230  IF( firstt ) THEN
231  WRITE( nout, fmt = * )
232  firstt = .false.
233  END IF
234  WRITE( nout, fmt = 9997 )m, p, n
235  END IF
236  10 CONTINUE
237  firstt = .true.
238 *
239 * Do for each value of M in MVAL.
240 *
241  DO 40 ik = 1, nn
242  m = mval( ik )
243  p = pval( ik )
244  n = nval( ik )
245  IF( m.GT.n .OR. n.GT.m+p )
246  \$ GO TO 40
247 *
248  DO 30 imat = 1, ntypes
249 *
250 * Do the tests only if DOTYPE( IMAT ) is true.
251 *
252  IF( .NOT.dotype( imat ) )
253  \$ GO TO 30
254 *
255 * Set up parameters with DLATB9 and generate test
256 * matrices A and B with DLATMS.
257 *
258  CALL dlatb9( path, imat, m, p, n, TYPE, KLA, KUA, KLB, KUB,
259  \$ ANORM, BNORM, MODEA, MODEB, CNDNMA, CNDNMB,
260  \$ DISTA, DISTB )
261 *
262  CALL dlatms( n, m, dista, iseed, TYPE, RWORK, MODEA, CNDNMA,
263  \$ ANORM, KLA, KUA, 'No packing', A, LDA, WORK,
264  \$ IINFO )
265  IF( iinfo.NE.0 ) THEN
266  WRITE( nout, fmt = 9999 )iinfo
267  info = abs( iinfo )
268  GO TO 30
269  END IF
270 *
271  CALL dlatms( n, p, distb, iseed, TYPE, RWORK, MODEB, CNDNMB,
272  \$ BNORM, KLB, KUB, 'No packing', B, LDB, WORK,
273  \$ IINFO )
274  IF( iinfo.NE.0 ) THEN
275  WRITE( nout, fmt = 9999 )iinfo
276  info = abs( iinfo )
277  GO TO 30
278  END IF
279 *
280 * Generate random left hand side vector of GLM
281 *
282  DO 20 i = 1, n
283  x( i ) = dlarnd( 2, iseed )
284  20 CONTINUE
285 *
286  CALL dglmts( n, m, p, a, af, lda, b, bf, ldb, x,
287  \$ x( nmax+1 ), x( 2*nmax+1 ), x( 3*nmax+1 ),
288  \$ work, lwork, rwork, resid )
289 *
290 * Print information about the tests that did not
291 * pass the threshold.
292 *
293  IF( resid.GE.thresh ) THEN
294  IF( nfail.EQ.0 .AND. firstt ) THEN
295  firstt = .false.
296  CALL alahdg( nout, path )
297  END IF
298  WRITE( nout, fmt = 9998 )n, m, p, imat, 1, resid
299  nfail = nfail + 1
300  END IF
301  nrun = nrun + 1
302 *
303  30 CONTINUE
304  40 CONTINUE
305 *
306 * Print a summary of the results.
307 *
308  CALL alasum( path, nout, nfail, nrun, 0 )
309 *
310  9999 FORMAT( ' DLATMS in DCKGLM INFO = ', i5 )
311  9998 FORMAT( ' N=', i4, ' M=', i4, ', P=', i4, ', type ', i2,
312  \$ ', test ', i2, ', ratio=', g13.6 )
313  9997 FORMAT( ' *** Invalid input for GLM: M = ', i6, ', P = ', i6,
314  \$ ', N = ', i6, ';', / ' must satisfy M <= N <= M+P ',
315  \$ '(this set of values will be skipped)' )
316  RETURN
317 *
318 * End of DCKGLM
319 *
subroutine alareq(PATH, NMATS, DOTYPE, NTYPES, NIN, NOUT)
ALAREQ
Definition: alareq.f:90
subroutine alahdg(IOUNIT, PATH)
ALAHDG
Definition: alahdg.f:62
subroutine alasum(TYPE, NOUT, NFAIL, NRUN, NERRS)
ALASUM
Definition: alasum.f:73
subroutine dlatb9(PATH, IMAT, M, P, N, TYPE, KLA, KUA, KLB, KUB, ANORM, BNORM, MODEA, MODEB, CNDNMA, CNDNMB, DISTA, DISTB)
DLATB9
Definition: dlatb9.f:170
subroutine dglmts(N, M, P, A, AF, LDA, B, BF, LDB, D, DF, X, U, WORK, LWORK, RWORK, RESULT)
DGLMTS
Definition: dglmts.f:146
double precision function dlarnd(IDIST, ISEED)
DLARND
Definition: dlarnd.f:73
subroutine dlatms(M, N, DIST, ISEED, SYM, D, MODE, COND, DMAX, KL, KU, PACK, A, LDA, WORK, INFO)
DLATMS
Definition: dlatms.f:321
Here is the call graph for this function:
Here is the caller graph for this function: