[Pw_forum] ecut vs energy

[Gabriele Sclauzero]

Dear Amin,

    I think that your convergence looks "slow" to you because your PP for H is norm-conserving: you'll need about 85 Ry to get a 1mRy accuracy, but that's not much surprise with a hard NCPP. I believe that by replacing the H PP with a corresponding USPP you could get convergence at a much lower energy. If instead you want to keep the NC PP for H, I think there's no need to use ecutrho=10*ecutwfc, since this is required only for the PP of B, which should converge at much lower ecutwfc (you could check this). Let's suppose that you need 35 Ry for B (hence you need to take an ecutrho of about 350 Ry), while 85 Ry are needed for H (therefore ecutrho=4*85=340 ~ 350): I believe that you can keep the ecutrho as low as 350 Ry (or 4*ecutwfc for higher ecutwfc) without any problem.


HTH

GS

Il giorno 14/giu/2011, alle ore 20.08, Amin Torabi ha scritto:

> Dear experts!
> 
> Could you please have a look at my input file, attached to this email.
> 
> My main problem is that I expect the total energy to converge at lower ecutwfc energy. ( or I am expecting too much?)
> 
> ecut    energy
> 25     -38.36026321
> 30     -38.44145201
> 35     -38.49537648
> 40     -38.53312799
> 45     -38.55904483
> 50     -38.57650475
> 55     -38.58840904
> 60     -38.59670631
> 65     -38.60253103
> 70     -38.60653038
> 75     -38.60918176
> 80     -38.61090436
> 85     -38.61200805
> 90     -38.61269007
> 95     -38.61308582
> 100    -38.61329742
> 
> 
> 
> Notes: 
> 
> 1. Do I have to be at the minimum structure to check for convergence of ecut vs energy, or I can check the convergency from a non-optimized structure? 
> 
> 2. I used USPP for one of my atoms and NCPP for the other atom, and set ecutrho = 10*ecutwfc...
> 
> 
> 
> Many thanks in advance!
> 
> <input.rtf>_______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum


§ Gabriele Sclauzero, EPFL SB ITP CSEA
   PH H2 462, Station 3, CH-1015 Lausanne

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20110615/fa740df0/attachment.html>