[Pw_forum] Convergence test, vc-relax calculation

Ari P Seitsonen Ari.P.Seitsonen at iki.fi
Sun Nov 26 14:33:56 CET 2017


Dear Mondol Anik Kumar,

  1) It depends on which quantities you mean hear: The ecutwfc (and 
ecutrho, if using Vanderbilt/USPP or PAW) are universal so long as you 
have not added or changed new pseudo potentials/PAW data sets; k points, 
one can in general estimate the spacing (in reci-procal space, naturally) 
from the bulk, unless the physics is radically different (for example one 
system metallic, the other one insulating)

  2) You can use "restart_mode = 'restart'" and you do not have to update 
the geometry or cell in the input file; I tend to do this anyway, in case 
I once want to do a "restart_mode = 'from_scratch'"

     Greetings from Paris,

        apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
     Ecole Normale Supérieure (ENS), Département de Chimie, Paris
     Mobile (F) : +33 789 37 24 25    (CH) : +41 79 71 90 935


On Sun, 26 Nov 2017, Anik Mondol wrote:

> Dear Users,I am new to quantum espresso. I appreciate the things that can be done by quantum espresso. I have  some specific queries. I will get
> to the points
> 
> My intention is to calculate bandstructure of GaN nanowire. I have carried convergence test ( ecutwfc, kpoint , lattice parameters( a,c)) for GaN
> supercell.
> Then I constructed a nanowire supercell by adding vacuum . Then I have performed a self-consistent and vc-relax calculation
> 
> Now my confusion is this :
> 
> 1. Did I need to carry out convergence test for the supercell as well ?
> 
> 2.After the vc -relax calculation, I want to do a final scf calculation do I need to update my input file with the optimized results  from the
> outputs of vc-relaxation or QE reads data from the saved directory ?
> 
> Regards
> 
> Mondol Anik Kumar
> Department of Electrical and Electronic Engineering
> Bangladesh University of Engineering and Technology 
>


More information about the users mailing list