[Pw_forum] Fwd: Re: A long relaxation

[Pedro Augusto F. P. Moreira]

   Thanks, Guido.

Are atoms moving considerably, meanwhile?


Yes, they are.

do you really need so large ecutrho and small mixing_beta?


There is no necessity about a large ecutrho. Which value do you think is 
more suitable?

For mixing_beta, I tried the default value, but the program always 
stopped with some error, which I do not remember anymore. Then, I 
followed the suggestion of FAQs to try a value between 0.1 and 0.3..

Pedro Moreira
IFGW-Unicamp-Brazil

-------- Mensagem original --------
Assunto: 	Re: [Pw_forum] A long relaxation
Data: 	Fri, 18 May 2012 09:44:24 +0200
De: 	Guido Fratesi <fratesi at mater.unimib.it>
Para: 	pw_forum at pwscf.org



Dear Pedro,

if the energy is decreasing you are not trapped in a local minimum (not
yet, at least: maybe you are approaching it).

  From the extract you sent us, forces are rather increasing than
oscillating, and if the energy is instead decreasing that suggests me
that your system started close to some saddle point or energy maximum,
then after some relaxation steps it found some stronger energy gradient
and started following it.
Are atoms moving considerably, meanwhile?

PS do you really need so large ecutrho and small mixing_beta?

Best,
Guido

On 05/17/2012 08:07 PM, Pedro Augusto F. P. Moreira wrote:
>     Dear all,
>
>     I am doing a relaxation which is consuming a very long time. My input
>  is below, but I think it is OK, because there is nothing wrong with my
>  output and the system energies are decreasing .
>     However, I have noted that the total forces and the forces acting on
>  each atoms have oscillated a lot. I summarized below the total forces,
>  which are showed in the end of each SC calculation.
>     Is it right? Or could the program find a local minimum and cannot get
>  out of it (or anything similar)?
>
>     Thanks a lot,
>
>     Pedro
>
>  ####Total forces#####
>         Total force =     0.162129     Total SCF correction =     0.000046
>         Total force =     0.131410     Total SCF correction =     0.039799
>         Total force =     0.199885     Total SCF correction =     0.004809
>         Total force =     0.254130     Total SCF correction =     0.004143
>         Total force =     0.285538     Total SCF correction =     0.003092
>         Total force =     0.339246     Total SCF correction =     0.001453
>         Total force =     0.419218     Total SCF correction =     0.000630
>         Total force =     0.509140     Total SCF correction =     0.000243
>         Total force =     0.593476     Total SCF correction =     0.000133
>         Total force =     0.534968     Total SCF correction =     0.000206
>         Total force =     0.431853     Total SCF correction =     0.004582
>         Total force =     0.308932     Total SCF correction =     0.000404
>
>  #####Input########
>  &control
>        calculation='relax',
>        restart_mode='from_scratch',
>        pseudo_dir = '/quantum-espresso/espresso/pseudo/',
>        outdir='/quantum-espresso/espresso/tmp/',
>        prefix='ap5',
>        tprnfor = .true.,
>        nstep = 1000,
>  /
>  &system
>        ibrav = 4, celldm(1) = 17.8749, celldm(3) = 0.7241,
>        nat = 42, ntyp = 4,
>        ecutwfc = 52.0, ecutrho = 900.0,
>  /
>  &electrons
>        electron_maxstep = 500
>        mixing_beta = 0.1,
>  /
>  &ions
>        ion_dynamics='bfgs',
>  /
>  ATOMIC_SPECIES
>     Ca 40.0780 Ca.pz-n-vbc.UPF
>     P 30.9738 P.pz-van_ak.UPF
>     O 15.9994 O.pz-van_ak.UPF
>     F 18.99840 F.pz-van_asa.UPF
>  ATOMIC_POSITIONS angstrom
>  .
>  .
>  .
>  ##############################
>

-- 
Guido Fratesi

Dipartimento di Scienza dei Materiali
Universita` degli Studi di Milano-Bicocca
via Cozzi 53, 20125 Milano, Italy
_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org
http://www.democritos.it/mailman/listinfo/pw_forum

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20120518/e4f40270/attachment.html>