[CP2K:3193] Re: problem converging molecule in doublet state

[hut... at pci.uzh.ch]

Hi

if you have problems converging OT you should use safe options.
 
      &OUTER_SCF ON
        EPS_SCF ...
      &END
      &OT ON
        Minimizer CG
        LINESEARCH 3PNT
        PRECONDITIONER FULL_ALL
      &END OT

Outer SCF helps to get a good preconditioner. Keep the number of SCF
cycle in the inner loop small (10-20 max).

regards

Juerg

--------------------------------------------------------------
Juerg Hutter                         Phone : ++41 44 635 4491
Physical Chemistry Institute   FAX   : ++41 44 635 6838
University of Zurich               E-mail:  hut... at pci.uzh.ch
Winterthurerstrasse 190
CH-8057 Zurich, Switzerland
---------------------------------------------------------------

-----cp... at googlegroups.com wrote: -----
To: cp2k <cp... at googlegroups.com>
From: garold 
Sent by: cp... at googlegroups.com
Date: 04/09/2011 01:33AM
Subject: [CP2K:3193] Re: problem converging molecule in doublet state

Hi Sun,

Thank you for the suggestions.  I had tried with larger MAX_SCF but I
see I was not going far enough.  Now, after changing also some other
settings (based on sample inputs in ~/cp2k/tests/benchmarks*), I am
getting efficient convergence for B3LYP/6-31+G*.  Energies agree well
with molpro.  BLYP I cannot get to converge at all with this basis set
(but fortunately I don't need this).

As for use of smoothing, I believe this is not recommended when using
GAPW.  (If I do use smoothing, convergence is reached in fewer
iterations but with errors of few tenths of hartree in the energies.)

I  am still curious about why I cannot use OT for my system....

Garold


On Apr 7, 8:37 pm, "Ross, Sun" <sunshou... at yahoo.com.cn> wrote:
> Hi Garold,
>
> Firstly, MAX_SCF 20 in your input file. I think 20 steps can not
> guarantee any system to converge at such small value.
>
> Secondly,  you can add XC_SMOOTH in the XC domain, which may lead to
> the convergence faster.
>       &XC_GRID
>          XC_SMOOTH_RHO NN10
>          XC_DERIV SPLINE2_SMOOTH
>       &END XC_GRID

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To post to this group, send email to cp... at googlegroups.com.
To unsubscribe from this group, send email to cp2k+uns... at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/cp2k?hl=en.