LAPACK  3.10.0
LAPACK: Linear Algebra PACKage

◆ cunhr_col01()

subroutine cunhr_col01 ( integer  M,
integer  N,
integer  MB1,
integer  NB1,
integer  NB2,
real, dimension(6)  RESULT 
)

CUNHR_COL01

Purpose:
 CUNHR_COL01 tests CUNGTSQR and CUNHR_COL using CLATSQR, CGEMQRT.
 Therefore, CLATSQR (part of CGEQR), CGEMQRT (part of CGEMQR)
 have to be tested before this test.
Parameters
[in]M
          M is INTEGER
          Number of rows in test matrix.
[in]N
          N is INTEGER
          Number of columns in test matrix.
[in]MB1
          MB1 is INTEGER
          Number of row in row block in an input test matrix.
[in]NB1
          NB1 is INTEGER
          Number of columns in column block an input test matrix.
[in]NB2
          NB2 is INTEGER
          Number of columns in column block in an output test matrix.
[out]RESULT
          RESULT is REAL array, dimension (6)
          Results of each of the six tests below.

            A is a m-by-n test input matrix to be factored.
            so that A = Q_gr * ( R )
                               ( 0 ),

            Q_qr is an implicit m-by-m unitary Q matrix, the result
            of factorization in blocked WY-representation,
            stored in CGEQRT output format.

            R is a n-by-n upper-triangular matrix,

            0 is a (m-n)-by-n zero matrix,

            Q is an explicit m-by-m unitary matrix Q = Q_gr * I

            C is an m-by-n random matrix,

            D is an n-by-m random matrix.

          The six tests are:

          RESULT(1) = |R - (Q**H) * A| / ( eps * m * |A| )
            is equivalent to test for | A - Q * R | / (eps * m * |A|),

          RESULT(2) = |I - (Q**H) * Q| / ( eps * m ),

          RESULT(3) = | Q_qr * C - Q * C | / (eps * m * |C|),

          RESULT(4) = | (Q_gr**H) * C - (Q**H) * C | / (eps * m * |C|)

          RESULT(5) = | D * Q_qr - D * Q | / (eps * m * |D|)

          RESULT(6) = | D * (Q_qr**H) - D * (Q**H) | / (eps * m * |D|),

          where:
            Q_qr * C, (Q_gr**H) * C, D * Q_qr, D * (Q_qr**H) are
            computed using CGEMQRT,

            Q * C, (Q**H) * C, D * Q, D * (Q**H)  are
            computed using CGEMM.
Author
Univ. of Tennessee
Univ. of California Berkeley
Univ. of Colorado Denver
NAG Ltd.

Definition at line 118 of file cunhr_col01.f.

119  IMPLICIT NONE
120 *
121 * -- LAPACK test routine --
122 * -- LAPACK is a software package provided by Univ. of Tennessee, --
123 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
124 *
125 * .. Scalar Arguments ..
126  INTEGER M, N, MB1, NB1, NB2
127 * .. Return values ..
128  REAL RESULT(6)
129 *
130 * =====================================================================
131 *
132 * ..
133 * .. Local allocatable arrays
134  COMPLEX , ALLOCATABLE :: A(:,:), AF(:,:), Q(:,:), R(:,:),
135  $ WORK( : ), T1(:,:), T2(:,:), DIAG(:),
136  $ C(:,:), CF(:,:), D(:,:), DF(:,:)
137  REAL , ALLOCATABLE :: RWORK(:)
138 *
139 * .. Parameters ..
140  REAL ZERO
141  parameter( zero = 0.0e+0 )
142  COMPLEX CONE, CZERO
143  parameter( cone = ( 1.0e+0, 0.0e+0 ),
144  $ czero = ( 0.0e+0, 0.0e+0 ) )
145 * ..
146 * .. Local Scalars ..
147  LOGICAL TESTZEROS
148  INTEGER INFO, I, J, K, L, LWORK, NB1_UB, NB2_UB, NRB
149  REAL ANORM, EPS, RESID, CNORM, DNORM
150 * ..
151 * .. Local Arrays ..
152  INTEGER ISEED( 4 )
153  COMPLEX WORKQUERY( 1 )
154 * ..
155 * .. External Functions ..
156  REAL SLAMCH, CLANGE, CLANSY
157  EXTERNAL slamch, clange, clansy
158 * ..
159 * .. External Subroutines ..
160  EXTERNAL clacpy, clarnv, claset, clatsqr, cunhr_col,
162 * ..
163 * .. Intrinsic Functions ..
164  INTRINSIC ceiling, real, max, min
165 * ..
166 * .. Scalars in Common ..
167  CHARACTER(LEN=32) SRNAMT
168 * ..
169 * .. Common blocks ..
170  COMMON / srmnamc / srnamt
171 * ..
172 * .. Data statements ..
173  DATA iseed / 1988, 1989, 1990, 1991 /
174 *
175 * TEST MATRICES WITH HALF OF MATRIX BEING ZEROS
176 *
177  testzeros = .false.
178 *
179  eps = slamch( 'Epsilon' )
180  k = min( m, n )
181  l = max( m, n, 1)
182 *
183 * Dynamically allocate local arrays
184 *
185  ALLOCATE ( a(m,n), af(m,n), q(l,l), r(m,l), rwork(l),
186  $ c(m,n), cf(m,n),
187  $ d(n,m), df(n,m) )
188 *
189 * Put random numbers into A and copy to AF
190 *
191  DO j = 1, n
192  CALL clarnv( 2, iseed, m, a( 1, j ) )
193  END DO
194  IF( testzeros ) THEN
195  IF( m.GE.4 ) THEN
196  DO j = 1, n
197  CALL clarnv( 2, iseed, m/2, a( m/4, j ) )
198  END DO
199  END IF
200  END IF
201  CALL clacpy( 'Full', m, n, a, m, af, m )
202 *
203 * Number of row blocks in CLATSQR
204 *
205  nrb = max( 1, ceiling( real( m - n ) / real( mb1 - n ) ) )
206 *
207  ALLOCATE ( t1( nb1, n * nrb ) )
208  ALLOCATE ( t2( nb2, n ) )
209  ALLOCATE ( diag( n ) )
210 *
211 * Begin determine LWORK for the array WORK and allocate memory.
212 *
213 * CLATSQR requires NB1 to be bounded by N.
214 *
215  nb1_ub = min( nb1, n)
216 *
217 * CGEMQRT requires NB2 to be bounded by N.
218 *
219  nb2_ub = min( nb2, n)
220 *
221  CALL clatsqr( m, n, mb1, nb1_ub, af, m, t1, nb1,
222  $ workquery, -1, info )
223  lwork = int( workquery( 1 ) )
224  CALL cungtsqr( m, n, mb1, nb1, af, m, t1, nb1, workquery, -1,
225  $ info )
226 
227  lwork = max( lwork, int( workquery( 1 ) ) )
228 *
229 * In CGEMQRT, WORK is N*NB2_UB if SIDE = 'L',
230 * or M*NB2_UB if SIDE = 'R'.
231 *
232  lwork = max( lwork, nb2_ub * n, nb2_ub * m )
233 *
234  ALLOCATE ( work( lwork ) )
235 *
236 * End allocate memory for WORK.
237 *
238 *
239 * Begin Householder reconstruction routines
240 *
241 * Factor the matrix A in the array AF.
242 *
243  srnamt = 'CLATSQR'
244  CALL clatsqr( m, n, mb1, nb1_ub, af, m, t1, nb1, work, lwork,
245  $ info )
246 *
247 * Copy the factor R into the array R.
248 *
249  srnamt = 'CLACPY'
250  CALL clacpy( 'U', n, n, af, m, r, m )
251 *
252 * Reconstruct the orthogonal matrix Q.
253 *
254  srnamt = 'CUNGTSQR'
255  CALL cungtsqr( m, n, mb1, nb1, af, m, t1, nb1, work, lwork,
256  $ info )
257 *
258 * Perform the Householder reconstruction, the result is stored
259 * the arrays AF and T2.
260 *
261  srnamt = 'CUNHR_COL'
262  CALL cunhr_col( m, n, nb2, af, m, t2, nb2, diag, info )
263 *
264 * Compute the factor R_hr corresponding to the Householder
265 * reconstructed Q_hr and place it in the upper triangle of AF to
266 * match the Q storage format in CGEQRT. R_hr = R_tsqr * S,
267 * this means changing the sign of I-th row of the matrix R_tsqr
268 * according to sign of of I-th diagonal element DIAG(I) of the
269 * matrix S.
270 *
271  srnamt = 'CLACPY'
272  CALL clacpy( 'U', n, n, r, m, af, m )
273 *
274  DO i = 1, n
275  IF( diag( i ).EQ.-cone ) THEN
276  CALL cscal( n+1-i, -cone, af( i, i ), m )
277  END IF
278  END DO
279 *
280 * End Householder reconstruction routines.
281 *
282 *
283 * Generate the m-by-m matrix Q
284 *
285  CALL claset( 'Full', m, m, czero, cone, q, m )
286 *
287  srnamt = 'CGEMQRT'
288  CALL cgemqrt( 'L', 'N', m, m, k, nb2_ub, af, m, t2, nb2, q, m,
289  $ work, info )
290 *
291 * Copy R
292 *
293  CALL claset( 'Full', m, n, czero, czero, r, m )
294 *
295  CALL clacpy( 'Upper', m, n, af, m, r, m )
296 *
297 * TEST 1
298 * Compute |R - (Q**H)*A| / ( eps * m * |A| ) and store in RESULT(1)
299 *
300  CALL cgemm( 'C', 'N', m, n, m, -cone, q, m, a, m, cone, r, m )
301 *
302  anorm = clange( '1', m, n, a, m, rwork )
303  resid = clange( '1', m, n, r, m, rwork )
304  IF( anorm.GT.zero ) THEN
305  result( 1 ) = resid / ( eps * max( 1, m ) * anorm )
306  ELSE
307  result( 1 ) = zero
308  END IF
309 *
310 * TEST 2
311 * Compute |I - (Q**H)*Q| / ( eps * m ) and store in RESULT(2)
312 *
313  CALL claset( 'Full', m, m, czero, cone, r, m )
314  CALL cherk( 'U', 'C', m, m, -cone, q, m, cone, r, m )
315  resid = clansy( '1', 'Upper', m, r, m, rwork )
316  result( 2 ) = resid / ( eps * max( 1, m ) )
317 *
318 * Generate random m-by-n matrix C
319 *
320  DO j = 1, n
321  CALL clarnv( 2, iseed, m, c( 1, j ) )
322  END DO
323  cnorm = clange( '1', m, n, c, m, rwork )
324  CALL clacpy( 'Full', m, n, c, m, cf, m )
325 *
326 * Apply Q to C as Q*C = CF
327 *
328  srnamt = 'CGEMQRT'
329  CALL cgemqrt( 'L', 'N', m, n, k, nb2_ub, af, m, t2, nb2, cf, m,
330  $ work, info )
331 *
332 * TEST 3
333 * Compute |CF - Q*C| / ( eps * m * |C| )
334 *
335  CALL cgemm( 'N', 'N', m, n, m, -cone, q, m, c, m, cone, cf, m )
336  resid = clange( '1', m, n, cf, m, rwork )
337  IF( cnorm.GT.zero ) THEN
338  result( 3 ) = resid / ( eps * max( 1, m ) * cnorm )
339  ELSE
340  result( 3 ) = zero
341  END IF
342 *
343 * Copy C into CF again
344 *
345  CALL clacpy( 'Full', m, n, c, m, cf, m )
346 *
347 * Apply Q to C as (Q**H)*C = CF
348 *
349  srnamt = 'CGEMQRT'
350  CALL cgemqrt( 'L', 'C', m, n, k, nb2_ub, af, m, t2, nb2, cf, m,
351  $ work, info )
352 *
353 * TEST 4
354 * Compute |CF - (Q**H)*C| / ( eps * m * |C|)
355 *
356  CALL cgemm( 'C', 'N', m, n, m, -cone, q, m, c, m, cone, cf, m )
357  resid = clange( '1', m, n, cf, m, rwork )
358  IF( cnorm.GT.zero ) THEN
359  result( 4 ) = resid / ( eps * max( 1, m ) * cnorm )
360  ELSE
361  result( 4 ) = zero
362  END IF
363 *
364 * Generate random n-by-m matrix D and a copy DF
365 *
366  DO j = 1, m
367  CALL clarnv( 2, iseed, n, d( 1, j ) )
368  END DO
369  dnorm = clange( '1', n, m, d, n, rwork )
370  CALL clacpy( 'Full', n, m, d, n, df, n )
371 *
372 * Apply Q to D as D*Q = DF
373 *
374  srnamt = 'CGEMQRT'
375  CALL cgemqrt( 'R', 'N', n, m, k, nb2_ub, af, m, t2, nb2, df, n,
376  $ work, info )
377 *
378 * TEST 5
379 * Compute |DF - D*Q| / ( eps * m * |D| )
380 *
381  CALL cgemm( 'N', 'N', n, m, m, -cone, d, n, q, m, cone, df, n )
382  resid = clange( '1', n, m, df, n, rwork )
383  IF( dnorm.GT.zero ) THEN
384  result( 5 ) = resid / ( eps * max( 1, m ) * dnorm )
385  ELSE
386  result( 5 ) = zero
387  END IF
388 *
389 * Copy D into DF again
390 *
391  CALL clacpy( 'Full', n, m, d, n, df, n )
392 *
393 * Apply Q to D as D*QT = DF
394 *
395  srnamt = 'CGEMQRT'
396  CALL cgemqrt( 'R', 'C', n, m, k, nb2_ub, af, m, t2, nb2, df, n,
397  $ work, info )
398 *
399 * TEST 6
400 * Compute |DF - D*(Q**H)| / ( eps * m * |D| )
401 *
402  CALL cgemm( 'N', 'C', n, m, m, -cone, d, n, q, m, cone, df, n )
403  resid = clange( '1', n, m, df, n, rwork )
404  IF( dnorm.GT.zero ) THEN
405  result( 6 ) = resid / ( eps * max( 1, m ) * dnorm )
406  ELSE
407  result( 6 ) = zero
408  END IF
409 *
410 * Deallocate all arrays
411 *
412  DEALLOCATE ( a, af, q, r, rwork, work, t1, t2, diag,
413  $ c, d, cf, df )
414 *
415  RETURN
416 *
417 * End of CUNHR_COL01
418 *
subroutine clatsqr(M, N, MB, NB, A, LDA, T, LDT, WORK, LWORK, INFO)
CLATSQR
Definition: clatsqr.f:166
subroutine cscal(N, CA, CX, INCX)
CSCAL
Definition: cscal.f:78
subroutine cgemm(TRANSA, TRANSB, M, N, K, ALPHA, A, LDA, B, LDB, BETA, C, LDC)
CGEMM
Definition: cgemm.f:187
subroutine cherk(UPLO, TRANS, N, K, ALPHA, A, LDA, BETA, C, LDC)
CHERK
Definition: cherk.f:173
real function clange(NORM, M, N, A, LDA, WORK)
CLANGE returns the value of the 1-norm, Frobenius norm, infinity-norm, or the largest absolute value ...
Definition: clange.f:115
subroutine cgemqrt(SIDE, TRANS, M, N, K, NB, V, LDV, T, LDT, C, LDC, WORK, INFO)
CGEMQRT
Definition: cgemqrt.f:168
subroutine claset(UPLO, M, N, ALPHA, BETA, A, LDA)
CLASET initializes the off-diagonal elements and the diagonal elements of a matrix to given values.
Definition: claset.f:106
subroutine clarnv(IDIST, ISEED, N, X)
CLARNV returns a vector of random numbers from a uniform or normal distribution.
Definition: clarnv.f:99
subroutine clacpy(UPLO, M, N, A, LDA, B, LDB)
CLACPY copies all or part of one two-dimensional array to another.
Definition: clacpy.f:103
subroutine cunhr_col(M, N, NB, A, LDA, T, LDT, D, INFO)
CUNHR_COL
Definition: cunhr_col.f:259
subroutine cungtsqr(M, N, MB, NB, A, LDA, T, LDT, WORK, LWORK, INFO)
CUNGTSQR
Definition: cungtsqr.f:175
real function clansy(NORM, UPLO, N, A, LDA, WORK)
CLANSY returns the value of the 1-norm, or the Frobenius norm, or the infinity norm,...
Definition: clansy.f:123
real function slamch(CMACH)
SLAMCH
Definition: slamch.f:68
Here is the call graph for this function:
Here is the caller graph for this function: