[Pw_forum] Geometry_Optimization_ERROR

[Ari P Seitsonen]

Dear Amir M. Mofrad, [ next time please with affiliation ]

   Since no qualified expert has answered so far, I give some quick 
comments:

  - Do you know what you are doing with the variable 'nr1'/'nr2'/'nr3'?

  - I found only 107 atoms in your input file, 'nat = 108'

  - I would actually find it easier to use 'ibrav = 4' and 'a = 13.6750', 
'c = 14.76700'; now you do not give many digits in the second, y 
coordinate of the second lattice vector, so a hexagonal cell is maybe not 
found; the alternative would be to more digits [13.6750/2*sqrt(3) = ...]

  - It is not a good idea to shift the k points laterally away from the 
Gamma point, unless there is a good reason (please see either my previous 
comments on the Forum or the appendix of my PhD thesis; the latter shows 
in practise how the Monkhort-Pack points, if not including the Gamma 
point, have a lower symmetry than the hexagonal cell and more k points 
need to be generated)

  - I do not know if the pseudo potential that you want to use have been 
improved in the later versions of PSLibrary; personally I would try to use 
the latest versions, if there is no specific reason for employing the 
elder ones

  - I would also check if you can symmetrise the coordinates; 'pw.x' gives 
coordinates that resemble having some "nice numbers" inside. But for 
'pw.x' to find symmetries the coordinates have to be very close to being 
symmetric, some five digits or so (I know of some other programs where one 
can ask the code to symmetrise the coordinates if found to be withing some 
given epsilon); symmetries would further reduce the number of k points 
that are generated with your input

  - ... and I would begin with less k points than the 4x4x4, as your cell 
is a bit bigger (lattice vectors ~10 Å) and your base material probably 
has a band gap (metallic systems usually require more k points); once you 
have the relaxation converging you might shift to more k points

   I do not know if these are of any help; Good Luck in any case. :)

     Greetings from Montrouge,

        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 Fri, 12 Feb 2016, Mofrad, Amir Mehdi (MU-Student) wrote:

> 
> Dear all QE experts and developers,
> 
> 
> I want to do a geometry optimization on a zeolite but I get the following
> error:
> 
>  task #         0
>      from sym_rho_init_shell : error #         2
>      lone vector
> Here is my input file in case you need it.
> 
> 
> &CONTROL
>  calculation = 'relax' ,
>  restart_mode = 'from_scratch' ,
>  wf_collect = .true. ,
>  outdir = './scratch' ,
>  wfcdir = './scratch' ,
>  pseudo_dir = '/global/espresso/pseudo' ,
>  prefix = 'CHA' ,
>  verbosity = 'high' ,
>  etot_conv_thr = 1e-5 ,
>  forc_conv_thr = 1e-4 ,
>  nstep = 50 ,
>  tstress = .true. ,
>  tprnfor = .true. ,
> 
>  /
> 
> &SYSTEM
>  ibrav = 0,
>  nat = 108,
>  ntyp = 2,
>  nr1= 32,
>  nr2=32,
>  nr3=32,
>  ecutwfc = 30 ,
>  ecutrho = 300 ,
> 
> /
> &ELECTRONS
>   electron_maxstep = 100,
>   conv_thr = 3e-8 ,
>   mixing_mode = 'plain' ,
>   mixing_beta = 0.7 ,
>   diagonalization = 'david' ,
> 
> /
> 
> &IONS
>                 ion_dynamics = 'bfgs' ,
>             trust_radius_ini = 0.5 ,
>  /
> CELL_PARAMETERS angstrom
>  13.6750   0.000000   0.000000
> -6.83750  11.84290   0.000000
> -0.00000  -0.00000   14.76700
> 
> ATOMIC_SPECIES
>     Si   28.08600  Si.pbe-n-rrkjus_psl.0.1.UPF
>     O   15.99940  O.pbe-n-rrkjus_psl.0.1.UPF
> ATOMIC_POSITIONS angstrom
>  O  11.66477   1.16060   1.81191
>  O  11.23606   3.67248   2.46166
>  O   0.00068   2.84822   1.94186
> ...
> Si   1.54596   5.21048  11.39668
> 
> K_POINTS automatic
> 4 4 4 1 1 1
> 
> Any help would be thoroughly appreciated.
> 
> 
> Best,
> 
> 
> Amir M. Mofrad   
>