LAPACK 3.12.0
LAPACK: Linear Algebra PACKage
Loading...
Searching...
No Matches

◆ chbevd_2stage()

subroutine chbevd_2stage ( character  jobz,
character  uplo,
integer  n,
integer  kd,
complex, dimension( ldab, * )  ab,
integer  ldab,
real, dimension( * )  w,
complex, dimension( ldz, * )  z,
integer  ldz,
complex, dimension( * )  work,
integer  lwork,
real, dimension( * )  rwork,
integer  lrwork,
integer, dimension( * )  iwork,
integer  liwork,
integer  info 
)

CHBEVD_2STAGE computes the eigenvalues and, optionally, the left and/or right eigenvectors for OTHER matrices

Download CHBEVD_2STAGE + dependencies [TGZ] [ZIP] [TXT]

Purpose:
 CHBEVD_2STAGE computes all the eigenvalues and, optionally, eigenvectors of
 a complex Hermitian band matrix A using the 2stage technique for
 the reduction to tridiagonal.  If eigenvectors are desired, it
 uses a divide and conquer algorithm.
Parameters
[in]JOBZ
          JOBZ is CHARACTER*1
          = 'N':  Compute eigenvalues only;
          = 'V':  Compute eigenvalues and eigenvectors.
                  Not available in this release.
[in]UPLO
          UPLO is CHARACTER*1
          = 'U':  Upper triangle of A is stored;
          = 'L':  Lower triangle of A is stored.
[in]N
          N is INTEGER
          The order of the matrix A.  N >= 0.
[in]KD
          KD is INTEGER
          The number of superdiagonals of the matrix A if UPLO = 'U',
          or the number of subdiagonals if UPLO = 'L'.  KD >= 0.
[in,out]AB
          AB is COMPLEX array, dimension (LDAB, N)
          On entry, the upper or lower triangle of the Hermitian band
          matrix A, stored in the first KD+1 rows of the array.  The
          j-th column of A is stored in the j-th column of the array AB
          as follows:
          if UPLO = 'U', AB(kd+1+i-j,j) = A(i,j) for max(1,j-kd)<=i<=j;
          if UPLO = 'L', AB(1+i-j,j)    = A(i,j) for j<=i<=min(n,j+kd).

          On exit, AB is overwritten by values generated during the
          reduction to tridiagonal form.  If UPLO = 'U', the first
          superdiagonal and the diagonal of the tridiagonal matrix T
          are returned in rows KD and KD+1 of AB, and if UPLO = 'L',
          the diagonal and first subdiagonal of T are returned in the
          first two rows of AB.
[in]LDAB
          LDAB is INTEGER
          The leading dimension of the array AB.  LDAB >= KD + 1.
[out]W
          W is REAL array, dimension (N)
          If INFO = 0, the eigenvalues in ascending order.
[out]Z
          Z is COMPLEX array, dimension (LDZ, N)
          If JOBZ = 'V', then if INFO = 0, Z contains the orthonormal
          eigenvectors of the matrix A, with the i-th column of Z
          holding the eigenvector associated with W(i).
          If JOBZ = 'N', then Z is not referenced.
[in]LDZ
          LDZ is INTEGER
          The leading dimension of the array Z.  LDZ >= 1, and if
          JOBZ = 'V', LDZ >= max(1,N).
[out]WORK
          WORK is COMPLEX array, dimension (MAX(1,LWORK))
          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
[in]LWORK
          LWORK is INTEGER
          The length of the array WORK. LWORK >= 1, when N <= 1;
          otherwise
          If JOBZ = 'N' and N > 1, LWORK must be queried.
                                   LWORK = MAX(1, dimension) where
                                   dimension = (2KD+1)*N + KD*NTHREADS
                                   where KD is the size of the band.
                                   NTHREADS is the number of threads used when
                                   openMP compilation is enabled, otherwise =1.
          If JOBZ = 'V' and N > 1, LWORK must be queried. Not yet available.

          If LWORK = -1, then a workspace query is assumed; the routine
          only calculates the optimal sizes of the WORK, RWORK and
          IWORK arrays, returns these values as the first entries of
          the WORK, RWORK and IWORK arrays, and no error message
          related to LWORK or LRWORK or LIWORK is issued by XERBLA.
[out]RWORK
          RWORK is REAL array,
                                         dimension (LRWORK)
          On exit, if INFO = 0, RWORK(1) returns the optimal LRWORK.
[in]LRWORK
          LRWORK is INTEGER
          The dimension of array RWORK.
          If N <= 1,               LRWORK must be at least 1.
          If JOBZ = 'N' and N > 1, LRWORK must be at least N.
          If JOBZ = 'V' and N > 1, LRWORK must be at least
                        1 + 5*N + 2*N**2.

          If LRWORK = -1, then a workspace query is assumed; the
          routine only calculates the optimal sizes of the WORK, RWORK
          and IWORK arrays, returns these values as the first entries
          of the WORK, RWORK and IWORK arrays, and no error message
          related to LWORK or LRWORK or LIWORK is issued by XERBLA.
[out]IWORK
          IWORK is INTEGER array, dimension (MAX(1,LIWORK))
          On exit, if INFO = 0, IWORK(1) returns the optimal LIWORK.
[in]LIWORK
          LIWORK is INTEGER
          The dimension of array IWORK.
          If JOBZ = 'N' or N <= 1, LIWORK must be at least 1.
          If JOBZ = 'V' and N > 1, LIWORK must be at least 3 + 5*N .

          If LIWORK = -1, then a workspace query is assumed; the
          routine only calculates the optimal sizes of the WORK, RWORK
          and IWORK arrays, returns these values as the first entries
          of the WORK, RWORK and IWORK arrays, and no error message
          related to LWORK or LRWORK or LIWORK is issued by XERBLA.
[out]INFO
          INFO is INTEGER
          = 0:  successful exit.
          < 0:  if INFO = -i, the i-th argument had an illegal value.
          > 0:  if INFO = i, the algorithm failed to converge; i
                off-diagonal elements of an intermediate tridiagonal
                form did not converge to zero.
Author
Univ. of Tennessee
Univ. of California Berkeley
Univ. of Colorado Denver
NAG Ltd.
Further Details:
  All details about the 2stage techniques are available in:

  Azzam Haidar, Hatem Ltaief, and Jack Dongarra.
  Parallel reduction to condensed forms for symmetric eigenvalue problems
  using aggregated fine-grained and memory-aware kernels. In Proceedings
  of 2011 International Conference for High Performance Computing,
  Networking, Storage and Analysis (SC '11), New York, NY, USA,
  Article 8 , 11 pages.
  http://doi.acm.org/10.1145/2063384.2063394

  A. Haidar, J. Kurzak, P. Luszczek, 2013.
  An improved parallel singular value algorithm and its implementation
  for multicore hardware, In Proceedings of 2013 International Conference
  for High Performance Computing, Networking, Storage and Analysis (SC '13).
  Denver, Colorado, USA, 2013.
  Article 90, 12 pages.
  http://doi.acm.org/10.1145/2503210.2503292

  A. Haidar, R. Solca, S. Tomov, T. Schulthess and J. Dongarra.
  A novel hybrid CPU-GPU generalized eigensolver for electronic structure
  calculations based on fine-grained memory aware tasks.
  International Journal of High Performance Computing Applications.
  Volume 28 Issue 2, Pages 196-209, May 2014.
  http://hpc.sagepub.com/content/28/2/196

Definition at line 251 of file chbevd_2stage.f.

254*
255 IMPLICIT NONE
256*
257* -- LAPACK driver routine --
258* -- LAPACK is a software package provided by Univ. of Tennessee, --
259* -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
260*
261* .. Scalar Arguments ..
262 CHARACTER JOBZ, UPLO
263 INTEGER INFO, KD, LDAB, LDZ, LIWORK, LRWORK, LWORK, N
264* ..
265* .. Array Arguments ..
266 INTEGER IWORK( * )
267 REAL RWORK( * ), W( * )
268 COMPLEX AB( LDAB, * ), WORK( * ), Z( LDZ, * )
269* ..
270*
271* =====================================================================
272*
273* .. Parameters ..
274 REAL ZERO, ONE
275 parameter( zero = 0.0e0, one = 1.0e0 )
276 COMPLEX CZERO, CONE
277 parameter( czero = ( 0.0e0, 0.0e0 ),
278 $ cone = ( 1.0e0, 0.0e0 ) )
279* ..
280* .. Local Scalars ..
281 LOGICAL LOWER, LQUERY, WANTZ
282 INTEGER IINFO, IMAX, INDE, INDWK2, INDRWK, ISCALE,
283 $ LLWORK, INDWK, LHTRD, LWTRD, IB, INDHOUS,
284 $ LIWMIN, LLRWK, LLWK2, LRWMIN, LWMIN
285 REAL ANRM, BIGNUM, EPS, RMAX, RMIN, SAFMIN, SIGMA,
286 $ SMLNUM
287* ..
288* .. External Functions ..
289 LOGICAL LSAME
290 INTEGER ILAENV2STAGE
291 REAL SLAMCH, CLANHB
292 EXTERNAL lsame, slamch, clanhb, ilaenv2stage
293* ..
294* .. External Subroutines ..
295 EXTERNAL sscal, ssterf, xerbla, cgemm, clacpy,
297* ..
298* .. Intrinsic Functions ..
299 INTRINSIC real, sqrt
300* ..
301* .. Executable Statements ..
302*
303* Test the input parameters.
304*
305 wantz = lsame( jobz, 'V' )
306 lower = lsame( uplo, 'L' )
307 lquery = ( lwork.EQ.-1 .OR. liwork.EQ.-1 .OR. lrwork.EQ.-1 )
308*
309 info = 0
310 IF( n.LE.1 ) THEN
311 lwmin = 1
312 lrwmin = 1
313 liwmin = 1
314 ELSE
315 ib = ilaenv2stage( 2, 'CHETRD_HB2ST', jobz, n, kd, -1, -1 )
316 lhtrd = ilaenv2stage( 3, 'CHETRD_HB2ST', jobz, n, kd, ib, -1 )
317 lwtrd = ilaenv2stage( 4, 'CHETRD_HB2ST', jobz, n, kd, ib, -1 )
318 IF( wantz ) THEN
319 lwmin = 2*n**2
320 lrwmin = 1 + 5*n + 2*n**2
321 liwmin = 3 + 5*n
322 ELSE
323 lwmin = max( n, lhtrd + lwtrd )
324 lrwmin = n
325 liwmin = 1
326 END IF
327 END IF
328 IF( .NOT.( lsame( jobz, 'N' ) ) ) THEN
329 info = -1
330 ELSE IF( .NOT.( lower .OR. lsame( uplo, 'U' ) ) ) THEN
331 info = -2
332 ELSE IF( n.LT.0 ) THEN
333 info = -3
334 ELSE IF( kd.LT.0 ) THEN
335 info = -4
336 ELSE IF( ldab.LT.kd+1 ) THEN
337 info = -6
338 ELSE IF( ldz.LT.1 .OR. ( wantz .AND. ldz.LT.n ) ) THEN
339 info = -9
340 END IF
341*
342 IF( info.EQ.0 ) THEN
343 work( 1 ) = lwmin
344 rwork( 1 ) = lrwmin
345 iwork( 1 ) = liwmin
346*
347 IF( lwork.LT.lwmin .AND. .NOT.lquery ) THEN
348 info = -11
349 ELSE IF( lrwork.LT.lrwmin .AND. .NOT.lquery ) THEN
350 info = -13
351 ELSE IF( liwork.LT.liwmin .AND. .NOT.lquery ) THEN
352 info = -15
353 END IF
354 END IF
355*
356 IF( info.NE.0 ) THEN
357 CALL xerbla( 'CHBEVD_2STAGE', -info )
358 RETURN
359 ELSE IF( lquery ) THEN
360 RETURN
361 END IF
362*
363* Quick return if possible
364*
365 IF( n.EQ.0 )
366 $ RETURN
367*
368 IF( n.EQ.1 ) THEN
369 w( 1 ) = real( ab( 1, 1 ) )
370 IF( wantz )
371 $ z( 1, 1 ) = cone
372 RETURN
373 END IF
374*
375* Get machine constants.
376*
377 safmin = slamch( 'Safe minimum' )
378 eps = slamch( 'Precision' )
379 smlnum = safmin / eps
380 bignum = one / smlnum
381 rmin = sqrt( smlnum )
382 rmax = sqrt( bignum )
383*
384* Scale matrix to allowable range, if necessary.
385*
386 anrm = clanhb( 'M', uplo, n, kd, ab, ldab, rwork )
387 iscale = 0
388 IF( anrm.GT.zero .AND. anrm.LT.rmin ) THEN
389 iscale = 1
390 sigma = rmin / anrm
391 ELSE IF( anrm.GT.rmax ) THEN
392 iscale = 1
393 sigma = rmax / anrm
394 END IF
395 IF( iscale.EQ.1 ) THEN
396 IF( lower ) THEN
397 CALL clascl( 'B', kd, kd, one, sigma, n, n, ab, ldab, info )
398 ELSE
399 CALL clascl( 'Q', kd, kd, one, sigma, n, n, ab, ldab, info )
400 END IF
401 END IF
402*
403* Call CHBTRD_HB2ST to reduce Hermitian band matrix to tridiagonal form.
404*
405 inde = 1
406 indrwk = inde + n
407 llrwk = lrwork - indrwk + 1
408 indhous = 1
409 indwk = indhous + lhtrd
410 llwork = lwork - indwk + 1
411 indwk2 = indwk + n*n
412 llwk2 = lwork - indwk2 + 1
413*
414 CALL chetrd_hb2st( "N", jobz, uplo, n, kd, ab, ldab, w,
415 $ rwork( inde ), work( indhous ), lhtrd,
416 $ work( indwk ), llwork, iinfo )
417*
418* For eigenvalues only, call SSTERF. For eigenvectors, call CSTEDC.
419*
420 IF( .NOT.wantz ) THEN
421 CALL ssterf( n, w, rwork( inde ), info )
422 ELSE
423 CALL cstedc( 'I', n, w, rwork( inde ), work, n, work( indwk2 ),
424 $ llwk2, rwork( indrwk ), llrwk, iwork, liwork,
425 $ info )
426 CALL cgemm( 'N', 'N', n, n, n, cone, z, ldz, work, n, czero,
427 $ work( indwk2 ), n )
428 CALL clacpy( 'A', n, n, work( indwk2 ), n, z, ldz )
429 END IF
430*
431* If matrix was scaled, then rescale eigenvalues appropriately.
432*
433 IF( iscale.EQ.1 ) THEN
434 IF( info.EQ.0 ) THEN
435 imax = n
436 ELSE
437 imax = info - 1
438 END IF
439 CALL sscal( imax, one / sigma, w, 1 )
440 END IF
441*
442 work( 1 ) = lwmin
443 rwork( 1 ) = lrwmin
444 iwork( 1 ) = liwmin
445 RETURN
446*
447* End of CHBEVD_2STAGE
448*
subroutine xerbla(srname, info)
Definition cblat2.f:3285
subroutine cgemm(transa, transb, m, n, k, alpha, a, lda, b, ldb, beta, c, ldc)
CGEMM
Definition cgemm.f:188
subroutine chetrd_hb2st(stage1, vect, uplo, n, kd, ab, ldab, d, e, hous, lhous, work, lwork, info)
CHETRD_HB2ST reduces a complex Hermitian band matrix A to real symmetric tridiagonal form T
integer function ilaenv2stage(ispec, name, opts, n1, n2, n3, n4)
ILAENV2STAGE
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
real function slamch(cmach)
SLAMCH
Definition slamch.f:68
real function clanhb(norm, uplo, n, k, ab, ldab, work)
CLANHB returns the value of the 1-norm, or the Frobenius norm, or the infinity norm,...
Definition clanhb.f:132
subroutine clascl(type, kl, ku, cfrom, cto, m, n, a, lda, info)
CLASCL multiplies a general rectangular matrix by a real scalar defined as cto/cfrom.
Definition clascl.f:143
logical function lsame(ca, cb)
LSAME
Definition lsame.f:48
subroutine sscal(n, sa, sx, incx)
SSCAL
Definition sscal.f:79
subroutine cstedc(compz, n, d, e, z, ldz, work, lwork, rwork, lrwork, iwork, liwork, info)
CSTEDC
Definition cstedc.f:206
subroutine ssterf(n, d, e, info)
SSTERF
Definition ssterf.f:86
Here is the call graph for this function:
Here is the caller graph for this function: