[Pw_forum] scf_must_converge

[florence liu]

Dear Davide, Dear Ari and Dear Bramha, and Dear Sonu,


thankyou (again) for your responses.
my very initial problem has been solved, it is as Bramha pointed out, a
simple copy paste mistake. I forgot to change the calculation tag to
'relax', that why it did not make a second optimization step.

@ Davide,
you are right, my convergence criteria is chosen to be rather tight. I was
thinking, dass for the first couple of steps, it may be ok, if the scf does
not fully converge. as long as the scf converges for the last step(s) of
the optimization, it should be ok.

2012/11/2 Ari P Seitsonen <Ari.P.Seitsonen at iki.fi>

>
> Dear Florence,
>
>   I would try different values of 'mixing_beta' in the namelist
> 'electrons'; typically in such transition metal systems I use values
> 0.1-0.3.
>
>     Greetings from Sunny Zurich,
>
>        apsi
>
> PS I often use values 1e-8 or similar in metallic systems. Please notice
> that the suitable value for 'degauss' depends also on the algorithm to
> used: I use the value '0.0036749326', but with Fermi-Dirac broadening; a
> similar value with Gaussian broadening is indeed quite different
>
> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=***=-=*=-=*=-=*=-=*=-=*=-=*=-=*=**
> -=*=-=*=-=*=-=*=-
>   Ari P Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
>   Physikalisch-Chemisches Institut der Universitaet Zuerich
>   Tel: +41 44 63 54 497  /  Mobile: +41 79 71 90 935
>
>
>
> On Fri, 2 Nov 2012, Bramha Pandey wrote:
>
>
>> On Fri, Nov 2, 2012 at 3:09 PM, florence liu <florenceliu86 at gmail.com>
>> wrote:
>>       Dear Davide, Dear Bramha,
>>
>> thankyou for your answers.
>> i am only optimizing atomic positions of a Pt surface with vdw-DF
>> funtionals,.... so nothing too special, i guess.
>>
>> as you suggested, here is the complete input file:
>>
>>
>>  &control
>>     prefix='Ptsurface47',
>>     pseudo_dir= '(...)',
>>     outdir = '(...)',
>>     calculation='scf',
>>
>> As per your problem you want to optimize, so set 'relax' or 'vc-relax' in
>> calculation field.
>>           forc_conv_thr=0.000389,
>>           tefield=.TRUE.,
>>           dipfield=.TRUE.,
>>        /
>>        &system
>>           ibrav=  0,  nat=  16, ntyp= 1,
>>           ecutwfc = 47, occupations='smearing', smearing='gaussian',
>>           degauss= 0.00367487
>>
>> it seems very low value to me. set it as a 0.02 or 0.01. and apply what i
>> am
>> already said before.
>>           edir=3,
>>           emaxpos=0.6,
>>           eopreg=0.1,
>>           input_dft='sla+pw+rpb+vdw1'
>>        /
>>        &electrons
>>        electron_maxstep=100,
>>        conv_thr=7.35D-8,
>>        scf_must_converge=.false.
>>        /
>>  &ions
>>  /
>>  &cell
>>  /
>> CELL_PARAMETERS (angstrom)
>> 5.7210401300000000    0.0000000000000000    0.0000000000000000
>> 2.8605200700000000    4.9545660900000000    0.0000000000000000
>> 0.0000000000000000    0.0000000000000000   22.9590000000000000
>> ATOMIC_SPECIES
>>  Pt  196.966   Pt.revPBE-n-kjpaw.UPF
>> ATOMIC_POSITIONS (crystal)
>> Pt  0.0000000000000000  0.0000000000000000  0.3051881400000000 1 1 1
>> Pt  0.5000000000000000  0.0000000000000000  0.3051881400000000 1 1 1
>> Pt  0.0000000000000000  0.5000000000000000  0.3051881400000000 1 1 1
>> Pt  0.5000000000000000  0.5000000000000000  0.3051881400000000 1 1 1
>> Pt  0.3333333333333333  0.3333333333333333  0.2034587900000000 1 1 1
>> Pt  0.8333333333333333  0.3333333333333333  0.2034587900000000 1 1 1
>> Pt  0.3333333333333333  0.8333333333333333  0.2034587900000000 1 1 1
>> Pt  0.8333333333333333  0.8333333333333333  0.2034587900000000 1 1 1
>> Pt  0.1666666666666666  0.1666666666666666  0.1017293800000000 0 0 0
>> Pt  0.6666666666666666  0.1666666666666666  0.1017293800000000 0 0 0
>> Pt  0.1666666666666666  0.6666666666666666  0.1017293800000000 0 0 0
>> Pt  0.6666666666666666  0.6666666666666666  0.1017293800000000 0 0 0
>> Pt  0.0000000000000000  0.0000000000000000  0.0000000000000000 0 0 0
>> Pt  0.5000000000000000  0.0000000000000000  0.0000000000000000 0 0 0
>> Pt  0.0000000000000000  0.5000000000000000  0.0000000000000000 0 0 0
>> Pt  0.5000000000000000  0.5000000000000000  0.0000000000000000 0 0 0
>>
>> K_POINTS (automatic)
>> 5 5 1 0 0 0
>>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20121102/a1f76c1c/attachment.html>