The ABINIT Users Mailing List ( CLOSED )

[Masayoshi Mikami <mmikami@rc.m-kagaku.co.jp>]

Dear Ghada,

I do not know what your system is like,
but I guess that you may have many possibilities:

- increase dilatmx, but it may be tested last, before that ...
- please be aware that your system is metallic.
  So that occopt needs to be set  3 ... 7. (the default is 1 !)
  You will need set tsmear as well. Please try our Tutorials/Test*
  in the tar.gz package to learn the details.
  In passing, diemac may be set to the default.
- nstep (10) may be much more increased ...
  you might notice some warnings about SCF convergence ?
- "toldfe 1.0d-6" may not be good. Try other tol* parameter,
  e.g. tolvrs (1.0d-12 - 1.0d-10 ?)
- Your pseudopotential may be still improved ? (Which one)
  please be aware that ecut may depend on your pseudos.
- ... (or something else ?) If your system is just a model,
  the initial guess for the geometry may be far from
  a (possible) plausible structure ...

In any case, you might want to try the Tutorial/Test*.
Such a "detour" (you might think so ?) may be a shortcut
to your goal...

In passing, you might want to have more appropriate
subject, if you could take a look at our netiquette:
http://www.abinit.org/community/?text=netiquette

Good luck,
Masayoshi (now in Yokohama... ;-)

On 2005/09/05, at 1:46, ghada ameereh wrote:

> Dear sir :
> i attend to make convergence tests but i faced a
> problem.
> input file:
>
> #Fe4Si3Ge Crystal :
>
> strprecon 0.1
> ionmov 3
> ntime 10
> dilatmx 1.1
> ecutsm 0.5
> optcell 1
>
>
> kptopt 1
> nshiftk 1
> shiftk 0.5 0.5 0.5
>
>
>
>
> ngkpt 6 6 6
>
> rprim 1 0 0
>       0 1 0
>       0 0 1
>
> acell 3*9.5
>
> ixc 1
> ntypat 3
> znucl 26 14 32
> natom 8
> typat 1 1 1 1 2 2 2 3
> xred  0.3525 0.3525 0.3525
>     -0.3525 0.8525 0.1475
>      0.1475 -0.3525 0.8525
>      0.8525 0.1475 -0.3525
>     -0.3525 -0.3525 -0.3525
>      0.3525 0.1475 0.8525
>      0.8525 0.3525 0.1475
>      0.1475 0.8525 0.3525
>
> ecut 30
> nstep 10
> toldfe 1.0d-6
> diemac 12.0
>
>
> The log file give me: chkdilatmx: ERROR -
>   The new primitive vectors rprimd (an evolving
> quantity)
>   are too large with respect to the old rprimd and the
> accompanying dilatmx :
>   this large change of unit cell parameters is not
> allowed by the present value of dilatmx.
>   You need at least dilatmx=    1.281035E+00
>   Action : increase the input variable dilatmx.
>
>  leave_new : decision taken to exit ...
> FORTRAN STOP  1
>
> i will appreciate your advice.
>
> --- verstraete@pcpm.ucl.ac.be wrote:
>
> > Your dilatmx is huge. Go see the definition and use
> > <= 1.1
> >
> > http://www.abinit.org/Infos_v4.6/varrlx.html#dilatmx
> >
> > For the rest, send your output/log etc... please
> > read the netiquette
> >
> > http://www.abinit.org/community/?text=netiquette
> >
> > Matthieu