[CPMD-list] Crash when trying HF exchange

[Juerg Hutter]

Hi

the programming errors could lead to crashes (see other
mail) and if not are producing slightly wrong results.
The errors were in the cases I look at very small.
However, I would recommend that you test this for your
special case with the new release.

Juerg

----------------------------------------------------------
Juerg Hutter                   Phone : ++41 1 635 4491
Physical Chemistry Institute   FAX   : ++41 1 635 6838
University of Zurich           E-mail: hutter at pci.unizh.ch
Winterthurerstrasse 190
CH-8057 Zurich, Switzerland
----------------------------------------------------------


On Wed, 4 Feb 2004, Antonio Rodriguez Fortea wrote:

>
> Dear Prof. Hutter,
>
> Does it mean that the results obtained with CPMD v3.7.2 using B3LYP are
> wrong?
>
> Thank you. Best regards,
>
> Antonio Rodriguez-Fortea
>
> On Wed, 2004-02-04 at 16:58, Juerg Hutter wrote:
> > Dear Christian
> >
> > the upcomming new release of CPMD will have a much
> > improved and corrected version of HF exchange.
> > However, some problems regarding periodic systems
> > will not be solved.
> > The code in V3.7.2 is not working correctly.
> >
> > regards
> >
> > Juerg
> >
> > ----------------------------------------------------------
> > Juerg Hutter                   Phone : ++41 1 635 4491
> > Physical Chemistry Institute   FAX   : ++41 1 635 6838
> > University of Zurich           E-mail: hutter at pci.unizh.ch
> > Winterthurerstrasse 190
> > CH-8057 Zurich, Switzerland
> > ----------------------------------------------------------
> >
> >
> > On Wed, 4 Feb 2004, Christian Tuma wrote:
> >
> > > Dear CPMD experts,
> > >
> > > in order to see how long computations including Hartree-Fock exchange really
> > > take compared to non-hybrid density functionals I tried single-point calculations
> > > on an isolated system (water molecule).
> > > Using CPMD 3.7 on an IBM p690 or 3.7.2 on a Linux PC-cluster the job always
> > > crashes when using the (not officially documented) keyword B3LYP or the
> > > (officially documented) keyword HARTREE-FOCK. The coredump occurs during
> > > the first do-loop in the subroutine HFXAA (file hfx.F, lines 223-226),
> > > just at the beginning of the scf iterations.
> > >
> > > This is the input I used:
> > >
> > > ! H2O
> > > &CPMD
> > >  OPTIMIZE WAVEFUNCTION
> > > &END
> > > &SYSTEM
> > >  CELL
> > >   16.0  1.0  1.0  0.0  0.0  0.0
> > >  SYMMETRY
> > >   0
> > >  CUTOFF
> > >   70.0
> > > &END
> > > &ATOMS
> > > *TM/O_TM_PBE_pade_115 KLEINMAN-BYLANDER
> > >  LMAX=P
> > >  1
> > >     9.896855  10.112455   9.898462
> > > *TM/H_TM_PBE_pade_070 KLEINMAN-BYLANDER
> > >  LMAX=S
> > >  2
> > >    11.530911  10.209722  10.742015
> > >     9.519948   8.310182   9.869858
> > > &END
> > > &DFT
> > >  FUNCTIONAL B3LYP
> > > &END
> > >
> > >
> > > Did I make a mistake in my input?
> > >
> > > Did anybody ever successfully run a B3LYP calculation? If so, under which
> > > conditions?
> > >
> > > Thank you. With best regards,
> > >
> > > Christian Tuma.
> > >
> > >
> > > --
> > > Christian Tuma             Humboldt-Universitaet Berlin
> > > ct at chemie.hu-berlin.de     Arbeitsgruppe Quantenchemie (Prof. Sauer)
> > > phone: +49-30-20937140     Brook-Taylor-Str. 2, 12489 Berlin, GERMANY
> > > fax:   +49-30-20937136     http://www.chemie.hu-berlin.de/ag_sauer
> > > _______________________________________________
> > > CPMD-list mailing list
> > > CPMD-list at cpmd.org
> > > http://www.cpmd.org/mailman/listinfo/cpmd-list
> > >
> > _______________________________________________
> > CPMD-list mailing list
> > CPMD-list at cpmd.org
> > http://www.cpmd.org/mailman/listinfo/cpmd-list
> --
> Antonio Rodriguez Fortea
> Physical Chemistry Department    Phone: +41 (0)91 913 8809
> ETH Zurich/USI Campus              Fax: +41 (0)91 913 8817
> Via Giuseppe Buffi, 13          e-mail: arodrig at phys.chem.ethz.ch
> CH-6904 Lugano (Switzerland)
>
>