The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,

1) Instead of "pulayit" use "npulayit" which I think is the correct form
2) use "iprcel 45" for instance
3) if your geometry is too far from equilibrium, instead of CG for the
ion movement prefer a MD-viscous damping for some iterations.

regards

PMA

On 9/14/07, david.hendry@mandmd.com.au <david.hendry@mandmd.com.au> wrote:
> Dear Abinitians,
>
> I'm trying to relax a single unit cell of CaTiO3 with two Ca removed and 
> one Sm added.  The geometry does not converge fast enough: after 200 nsteps 
> the potential residual is still above 1.  Each ntime step therefore only 
> marginaly reduces the atomic forces.  The input file is added below.  I'm 
> using Troullier-Martins norm-conserving psps.  Might there be some way to 
> allow this to work more efficiently?
>
> Regards,
>
> David.
>
>
> #Definition of lattice parameters
> #********************************
>   acell  10.1629 14.4451 10.2329
>
> #Definition of atomic positions and types
> #****************************************
>    natom       19
>    ntypat      4
>    znucl       22 20 62 8
>    typat       1 1 1 1 2 3 3 4 4 4 4 4 4 4 4 4 4 4 4
>    xred        0      0     0.5
>                0.5    0     0
>                0      0.5   0.5
>                0.5    0.5   0
>                0.9937 0.25  0.03339
>                0.5063 0.75  0.53339
>                0.4937 0.25  0.46661
>                0.0707 0.25  0.4842
>                0.4293 0.75  0.9842
>                0.9293 0.75  0.5158
>                0.5707 0.25  0.0158
>                0.711  0.037 0.2884
>                0.789  0.963 0.7884
>                0.289  0.537 0.7116
>                0.211  0.463 0.2116
>                0.289  0.963 0.7116
>                0.211  0.037 0.2116
>                0.711  0.463 0.2884
>                0.789  0.537 0.7884
>
>
>
> #Echange-correlation functional
> #******************************
>    ixc         1
>
> #Plane wave basis and k-point grid
> #*********************************
>    #pawecutdg   60
>    ecut        42
>    ngkpt       7 5 7
>    nshiftk     1
>    shiftk      0.5 0.5 0.5
>    nband       53
>    ngfft       60  90  60
>
> #Relax ions and unit cell
> #************************
>     #chkprim    0
>     kptopt     1
>     iscf       7
>     pulayit    20
>     tolvrs     1.0d-18
>     tolmxf     1.0d-8
>     nstep      200
>     optcell    0
>     ionmov     3
>     ntime      200
>     dilatmx    1.1
>     ecutsm     0.5
>     prtwf      0
>     prtden     0
>


-- 
Pierre-Matthieu Anglade