[Pw_forum] Error 61 in rdiaghg

[Axel Kohlmeyer]

On Thu, 5 Jul 2007, Marcos Verissimo Alves wrote:

MVA> Hi Eyvaz,
MVA> 
MVA> I am trying to run a test for NdS, which has the NaCl structure. I
MVA> generated a GGA pseudo for Nd based on an existing LDA one using
MVA> Vanderbilt's code, converted it to UPF (successfully according to
MVA> uspp2upf.x) and wanted to test it to see its performance in a simple
MVA> calculation. I am using ibrav=2, and one atom at 0.00 0.00 0.00 and
MVA> another at 0.50 0.50 0.50 - is this the correct setting for the NaCl
MVA> structure? Now I have tried to change the diagonalization method to
MVA> davidson, cg and diis, to no avail. The error is still there, and at the
MVA> very beginning of the scf cycle (the code stops at "Starting wfc are
MVA> atomic").

hi marcos,

that is frequently a sign of a bad pseudopotential (check in the psp 
generation log for negative density), but also some BLAS/LAPACK 
libraries are more sensitive than others. try switching between
ATLAS, MKL, ACML, GOTO-BLAS, and plain BLAS/LAPACK.

ciao,
   axel.

MVA> Marcos
MVA> 
MVA> > Hi Marcos,
MVA> >
MVA> > Similar problem  was discussed before in the forum
MVA> > many times.
MVA> > Briefly, the problem is due to a fact that an
MVA> > eigenvalue can not converge with a machine precision
MVA> > in this particular case.
MVA> > One reason is that a structure you used is not
MVA> > correct. So, have a look at your input file again.
MVA> > If it is not the case use "conjugated-gradient" method
MVA> > for diagonalization in &electrons.
MVA> >
MVA> > Bests,
MVA> > Eyvaz.
MVA> >
MVA> > --- Marcos Verissimo Alves <mverissi at ictp.it> wrote:
MVA> >
MVA> >> Hi all,
MVA> >>
MVA> >> I am trying to run a simple calculation (scf,
MVA> >> ibrav=2, 2 atoms in the unit
MVA> >> cell, gamma point) using espresso 3.0 but I'm
MVA> >> getting
MVA> >>
MVA> >> from rdiaghg : error #        61
MVA> >> info =/= 0
MVA> >>
MVA> >> How can I solve this? If necessary, I can post my
MVA> >> input and pseudo files.
MVA> >>
MVA> >> Thanks in advance,
MVA> >>
MVA> >> Marcos
MVA> >>
MVA> >> --
MVA> >> Dr. Marcos Verissimo Alves
MVA> >> Post-Doctoral Fellow
MVA> >> Condensed Matter and Statistical Physics Sector
MVA> >> International Centre for Theoretical Physics
MVA> >> Trieste, Italy
MVA> >>
MVA> >> --------
MVA> >>
MVA> >> I have become so addicted to vi that I try to exit
MVA> >> OpenOffice by typing :wq!
MVA> >>
MVA> >> _______________________________________________
MVA> >> Pw_forum mailing list
MVA> >> Pw_forum at pwscf.org
MVA> >> http://www.democritos.it/mailman/listinfo/pw_forum
MVA> >>
MVA> >
MVA> >
MVA> > -------------------------------------------------------------------
MVA> > Prof. Eyvaz Isaev,
MVA> > Theoretical Physics Department, Moscow State Institute of Steel & Alloys,
MVA> > Russia, and
MVA> > Condensed Matter Theory Group, Uppsala University, Sweden
MVA> > Eyvaz.Isaev at fysik.uu.se, eyvaz_isaev at yahoo.com
MVA> >
MVA> >
MVA> >
MVA> > ____________________________________________________________________________________
MVA> > Pinpoint customers who are looking for what you sell.
MVA> > http://searchmarketing.yahoo.com/
MVA> > _______________________________________________
MVA> > Pw_forum mailing list
MVA> > Pw_forum at pwscf.org
MVA> > http://www.democritos.it/mailman/listinfo/pw_forum
MVA> >
MVA> 
MVA> 
MVA> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.