[BACK]Return to dlarrb.f CVS log [TXT] [DIR] Up to [local] / rpl / lapack / lapack

Diff for /rpl/lapack/lapack/dlarrb.f between versions 1.16 and 1.20

version 1.16, 2017/06/17 11:06:25 version 1.20, 2023/08/07 08:38:57
Line 91 Line 91
 *>          RTOL2 is DOUBLE PRECISION  *>          RTOL2 is DOUBLE PRECISION
 *>          Tolerance for the convergence of the bisection intervals.  *>          Tolerance for the convergence of the bisection intervals.
 *>          An interval [LEFT,RIGHT] has converged if  *>          An interval [LEFT,RIGHT] has converged if
 *>          RIGHT-LEFT.LT.MAX( RTOL1*GAP, RTOL2*MAX(|LEFT|,|RIGHT|) )  *>          RIGHT-LEFT < MAX( RTOL1*GAP, RTOL2*MAX(|LEFT|,|RIGHT|) )
 *>          where GAP is the (estimated) distance to the nearest  *>          where GAP is the (estimated) distance to the nearest
 *>          eigenvalue.  *>          eigenvalue.
 *> \endverbatim  *> \endverbatim
Line 107 Line 107
 *> \verbatim  *> \verbatim
 *>          W is DOUBLE PRECISION array, dimension (N)  *>          W is DOUBLE PRECISION array, dimension (N)
 *>          On input, W( IFIRST-OFFSET ) through W( ILAST-OFFSET ) are  *>          On input, W( IFIRST-OFFSET ) through W( ILAST-OFFSET ) are
 *>          estimates of the eigenvalues of L D L^T indexed IFIRST throug  *>          estimates of the eigenvalues of L D L^T indexed IFIRST through
 *>          ILAST.  *>          ILAST.
 *>          On output, these estimates are refined.  *>          On output, these estimates are refined.
 *> \endverbatim  *> \endverbatim
Line 117 Line 117
 *>          WGAP is DOUBLE PRECISION array, dimension (N-1)  *>          WGAP is DOUBLE PRECISION array, dimension (N-1)
 *>          On input, the (estimated) gaps between consecutive  *>          On input, the (estimated) gaps between consecutive
 *>          eigenvalues of L D L^T, i.e., WGAP(I-OFFSET) is the gap between  *>          eigenvalues of L D L^T, i.e., WGAP(I-OFFSET) is the gap between
 *>          eigenvalues I and I+1. Note that if IFIRST.EQ.ILAST  *>          eigenvalues I and I+1. Note that if IFIRST = ILAST
 *>          then WGAP(IFIRST-OFFSET) must be set to ZERO.  *>          then WGAP(IFIRST-OFFSET) must be set to ZERO.
 *>          On output, these gaps are refined.  *>          On output, these gaps are refined.
 *> \endverbatim  *> \endverbatim
Line 178 Line 178
 *> \author Univ. of Colorado Denver  *> \author Univ. of Colorado Denver
 *> \author NAG Ltd.  *> \author NAG Ltd.
 *  *
 *> \date December 2016  
 *  
 *> \ingroup OTHERauxiliary  *> \ingroup OTHERauxiliary
 *  *
 *> \par Contributors:  *> \par Contributors:
Line 196 Line 194
      $                   RTOL2, OFFSET, W, WGAP, WERR, WORK, IWORK,       $                   RTOL2, OFFSET, W, WGAP, WERR, WORK, IWORK,
      $                   PIVMIN, SPDIAM, TWIST, INFO )       $                   PIVMIN, SPDIAM, TWIST, INFO )
 *  *
 *  -- LAPACK auxiliary routine (version 3.7.0) --  *  -- LAPACK auxiliary routine --
 *  -- LAPACK is a software package provided by Univ. of Tennessee,    --  *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
 *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--  *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
 *     December 2016  
 *  *
 *     .. Scalar Arguments ..  *     .. Scalar Arguments ..
       INTEGER            IFIRST, ILAST, INFO, N, OFFSET, TWIST        INTEGER            IFIRST, ILAST, INFO, N, OFFSET, TWIST
Line 237 Line 234
 *  *
       INFO = 0        INFO = 0
 *  *
   *     Quick return if possible
   *
         IF( N.LE.0 ) THEN
            RETURN
         END IF
   *
       MAXITR = INT( ( LOG( SPDIAM+PIVMIN )-LOG( PIVMIN ) ) /        MAXITR = INT( ( LOG( SPDIAM+PIVMIN )-LOG( PIVMIN ) ) /
      $           LOG( TWO ) ) + 2       $           LOG( TWO ) ) + 2
       MNWDTH = TWO * PIVMIN        MNWDTH = TWO * PIVMIN
Removed from v.1.16  
changed lines
  Added in v.1.20

CVSweb interface <joel.bertrand@systella.fr>