[Pw_forum] convergence issues

[Jonas Baltrusaitis]

WIth all being said, how do I test pseudopotentials, grid and cutoff for my system? My understanding is this:

select pseudopotential, use it on a real system, have grid fixed, calculate total energy at , say, 20 Ry and several ecutrho values, the increase cutoff value, calculate total energy at sevarl ecutrho  again etc. The take the lowest cutoff and ecutrho values (when it doesn;t change anymore) and increase grid while having them constant.

Am I thinking right? I am a cluster type of guy so it's somewhat out of my understanding

thanks

JOnas
University of Iowa


--- On Tue, 1/27/09, Gabriele Sclauzero <sclauzer at sissa.it> wrote:

> From: Gabriele Sclauzero <sclauzer at sissa.it>
> Subject: Re: [Pw_forum] convergence issues
> To: "PWSCF Forum" <pw_forum at pwscf.org>
> Date: Tuesday, January 27, 2009, 9:02 AM
> Hi Jonas
> 
> Jonas Baltrusaitis wrote:
> > 
> > INPUT
> > 
> >  &control
> >     calculation='relax',
> >     prefix='44tpcb_PBE_optimization',
> >    
> pseudo_dir='/share/apps/espresso-4.0.4/pseudo/',
> >     outdir='/home/jbaltrus/atom-lsda/',
> >     nstep=200
> >  /
> >  &system
> >    ibrav=12,
> >    nat=38,
> >    ntyp=4,
> >    a=11.367,
> >    b=9.749,
> >    c=14.273,
> >    cosab=-0.36922777,
> >    ecutwfc=25.0
> 
> As a first hint I would suggest to specify ecutrho, since
> you're using Ultrasoft PPs, and 
> test convergence of energy (or other quantities you're
> interested in) w.r.t. to this 
> parameter (I usually choose ecutrho ~ 10*ecutwfc).
> 
> If you want to do spin polarized calculations, you need to
> specify nspin=2 and a non-zero 
> starting magnetization for one species at least.
> 
> >  /
> >  &electrons
> >    mixing_beta=0.5D0,
> 
> Then maybe decrease the mixing parameter, or use another
> mixing (for instance, 
> mixing_mode='local-TF' works well for highly
> inhomogeneous systems).
> 
> 
> >    conv_thr = 1.D-8
> >  /
> >  &ions
> > ion_dynamics = ’none’
> 
> Are you doing a relaxation or what?
> 
> >  /
> > K_POINTS {automatic}
> > 4 4 4 0 0 0
> 
> Do you want to study an isolated molecule? If so you
> don't need a k-point sampling, simply 
> use
> 
>  > K_POINTS GAMMA
> 
> Regards
> 
> GS
> 
> 
> 
> > 
> > 
> > 
> >       
> > _______________________________________________
> > Pw_forum mailing list
> > Pw_forum at pwscf.org
> > http://www.democritos.it/mailman/listinfo/pw_forum
> 
> -- 
> 
> 
> o ------------------------------------------------ o
> | Gabriele Sclauzero, PhD Student                  |
> | c/o:   SISSA & CNR-INFM Democritos,              |
> |        via Beirut 2-4, 34014 Trieste (Italy)     |
> | email: sclauzer at sissa.it                         |
> | phone: +39 040 3787 511                          |
> | skype: gurlonotturno                             |
> o ------------------------------------------------ o
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum