[Pw_forum] Iron cluster relaxation

[Cyrille Barreteau]

In fact I have looked in more details at the results of pw.x calculations
and there is a funny behaviour. The restarted calculations converges towards
a totally different scf solution with a total magnetization  1mu_B larger
than the previous "crashed" calculation. I guess there are several magnetic
solutions and I "jumped" from one solution to the other.. Funny..

  cyrille


> Dear Kostya and Axel
>
> I have restarted my job with the last coordinates and
> the geometry optimization converged in just one scf cycle!!
>
> ==========================================================
>     BFGS Geometry Optimization
>
>     bfgs converged in   1 scf cycles and   0 bfgs steps
>
>     End of BFGS Geometry Optimization
> =============================================
> So I guess the previous "crashed run" did converge correctely
> toward the right optimal atomic configuration!!
>
> thanks
>
>    cyrille
>
>> On Fri, 10 Nov 2006, Konstantin Kudin wrote:
>>
>> kostya,
>>
>> KK>  It is not clear why the job should have a hard crash like this 
>> instead
>> KK> of either terminating gracefully, recovering from the condition 
>> in some
>> KK> way, or at least printing the very thing that you say, i.e. that
>> KK> "optimization terminated due to noise in the forces, increase 
>> cutoff",
>> KK> or something else similar.
>>
>> i agree. after this issue has come up quite a few times in the last 
>> months, it may be worth putting in the effort
>> and changing the code, so that it a) exits gracefully,
>> b) has a few more options on how to handle the resets,
>> e.g. force a minimum number of steps between two resets,
>> or have a small(!) randomization of coordinates.
>>
>> that just leaves the usual problem: who is going to do it?
>> anybody in the market for making a lot of people happy?
>>
>> KK>  Actually, in my experience just restarting the job with the last
>> KK> coordinates has a very good chance to advance further, and even 
>> find a
>> KK> lower energy structure since the BFGS will be started from scratch.
>>
>> hmmm. if you see the latter, then perhaps the reset is not
>> a complete reset and there may be a bug (unless you have
>> a 'randomization' of the coordinates due to limited
>> accuracy when reading in the so far optimized coordinates).
>>
>> cheers,
>>   axel.
>>
>> KK> KK>  Kostya
>> KK> KK> KK> KK> --- Axel Kohlmeyer <akohlmey at cmm.chem.upenn.edu> wrote:
>> KK> KK> > On Fri, 10 Nov 2006, Cyrille Barreteau wrote:
>> KK> > KK> > CB> Dear Pwscf users
>> KK> > KK> > dear cyrille,
>> KK> > KK> > did you check the mailing list archives?
>> KK> > KK> > IIRC, this error seems to happen if you try to
>> KK> > optimize the geometry too tightly for the
>> KK> > given convergence of the wavefunction or
>> KK> > when you suffer from 'ripples' due to a
>> KK> > too small density cutoff.
>> KK> > KK> > salut,
>> KK> >   axel.
>> KK> > KK> > CB> I am trying to relax an iron cluster (cubocathedron) and
>> KK> > CB> pw.x systematicaly ends with the following error:
>> KK> > CB> KK> > CB> KK> >
>> KK> 
>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
>>
>> KK> > CB>      from bfgs : error #         1
>> KK> > CB>      bfgs history already reset at previous step
>> KK> > CB> KK> >
>> KK> 
>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
>>
>> KK> > CB> KK> > CB>      stopping ...
>> KK> > CB> 2
>> KK> > CB> KK> > CB> KK> > CB> You can find the input and output files 
>> on the following web
>> KK> > site:
>> KK> > CB> KK> > CB>
>> KK> >
>> KK> 
>> http://www-drecam.cea.fr/Images/Pisp/cbarreteau/temporaire/cuboctahedron_relax.in 
>>
>> KK> > CB>
>> KK> >
>> KK> 
>> http://www-drecam.cea.fr/Images/Pisp/cbarreteau/temporaire/cuboctahedron_relax.out 
>>
>> KK> > CB> KK> > CB> KK> > CB> thanks in advance for any constructive 
>> comment :-)
>> KK> > CB> KK> > CB>    cyrille
>> KK> > CB> KK> > CB> KK> > KK> > -- KK> >
>> KK> 
>> =======================================================================
>> KK> > Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu  KK> > 
>> http://www.cmm.upenn.edu
>> KK> >    Center for Molecular Modeling   --   University of Pennsylvania
>> KK> > Department of Chemistry, 231 S.34th Street, Philadelphia, PA
>> KK> > 19104-6323
>> KK> > tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel:
>> KK> > 1-215-898-5425
>> KK> >
>> KK> 
>> =======================================================================
>> KK> > If you make something idiot-proof, the universe creates a better
>> KK> > idiot.
>> KK> > KK> > _______________________________________________
>> KK> > Pw_forum mailing list
>> KK> > Pw_forum at pwscf.org
>> KK> > http://www.democritos.it/mailman/listinfo/pw_forum
>> KK> > KK> KK> KK> KK>  KK> 
>> ____________________________________________________________________________________ 
>>
>> KK> Cheap talk?
>> KK> Check out Yahoo! Messenger's low PC-to-Phone call rates.
>> KK> http://voice.yahoo.com
>> KK> _______________________________________________
>> KK> Pw_forum mailing list
>> KK> Pw_forum at pwscf.org
>> KK> http://www.democritos.it/mailman/listinfo/pw_forum
>> KK>
>>  
>>
>
>


-- 
==================================================================

Cyrille Barreteau

CEA Saclay DSM/DRECAM/SPCSI

Batiment 462

91191 Gif sur Yvette Cedex FRANCE
~~~~~~~~~~~~~~~~~~~~~~~~
phone : +33 (0)1 69 08 29 51
fax : +33 (0)1 69 08 84 46
email : cyrille.barreteau at cea.fr

~~~~~~~~~~~~~~~~~~~~~~~~

http://www-drecam.cea.fr/Images/Pisp/cbarreteau/cbarreteau_fr.html

==================================================================