The ABINIT Users Mailing List ( CLOSED )

[Marc Torrent <marc.torrent@cea.fr>]

I tested your case with a v5.2.3 version of Abinit and did not have any
problem to converge (no divergence and no problem with dilatmx).
With the future v5.3 version, the run will be more efficient (to be
release in dec.).
Just a remark: you put pawecutdg=ecut in your input file. I may cause
inacurracy of calculations. I advice to put pawecutdg=50.
It increases the memory need but you can save memory putting the
pawmixdg variable to zero (mixing of densities will be done on a
"coarse" grid).

Marc Torrent


Yong Liu a écrit :
> Dear Torrent:
>     I'm so sorry. The Abinit version is 4.6.5 and the input files and 
> paw files are attached name as paw.zip. Since the parameters in the 
> input files working well with the norm-conserving files are not changed 
> except for ecut, pawcutdg and iscf, I'm afraid the crash results from 
> the paw files.
> 
> 
> */Marc Torrent <marc.torrent@cea.fr>/* 写道：
> 
>     Hum, probably a forces/stress convergence problem; check the value of
>     tolmxf/tolvrs keywords...
> 
>     But giving a right answer is difficult !
>     What about the Netiquette requirements (Abinit version, input files,
>     paw
>     files, ...) ?
> 
>     Marc Torrent
> 
>     occupierliu@yahoo.com.cn a écrit :
>      > Dear abiniters:
>      > I've create a series PAW potentials for PbTiO3 by Vanderbilt USPP
>     package following ABINIT tutorial. For geometry optimisation of
>     cubic PbTiO3, these work well, and the optimized lattice constant
>     a=7.4546 Bohr as well as Bulk modul B=199.2981 GPa which have a very
>     agreement with experimental data. But when I perform the geometry
>     optimisation of tetragonal PbTiO3 with these potentails, the abinit
>     program crashed due to the parameter dilatmx. Even if dilatmx was
>     increased from 1.05 to 1.20, the program still can't work well.
>     Since I have perform the geometry optimisation successfully using
>     the same input file with norm-conserving pseudopential downloaded
>     from the abinit website, I just wonder whether the PAW potential is
>     only suitable for the structral optimisation of the cubic crystal
>     rather than tetragonal crystal or other crystal with lower symmetry.
>      >
>      > With best regards,
>      >
>      > Liu
> 
> 
> 
> 
> Sincerely.
> Yong Liu
> 
> ------------------------------------------------------------------------
> 抢注雅虎免费邮箱-3.5G容量，20M附件！ <http://cn.mail.yahoo.com>