[Wien] Geometry minimizer via PORT (min_lapw)

[Laurence Marks]

MSR1a "seems" to take longer than PORT because it just keeps going.
However, for most systems and insulators in particular it is faster than
PORT.

Personally I do not think NEW1 is good (sorry Peter). For certain, there
are many cases where it will not work that well.

One comment: PORT will always work unless the model is badly constructed.
It is a conventional trust-region BFGS written by the top names in the
optimization field, and is about as good as one can have for a conventional
trust-region BFGS code. MSR1a uses a different algorithm.

N.B., if you have MSR1a in case.inm it does not matter if you use PORT or
NEW1, it is the MSR1a algorithm that is in fact being used.

On Wed, Apr 29, 2015 at 4:02 PM, Mohammed Abujafar <mabujafar at yahoo.com>
wrote:

>  Dear Prof. Laurence,
>  Thanks a lot for your nice clarification.Using MSR1a minimizer takes 25
> days to finish the minimization for 100 atoms supercell. I added DRIFT as
> you mentioned at the bottom of the case.inm file and it works fine even
> computationally is so expensive.I am trying now to use the port minimizer
> and  is working fine with MSR1a in case.inm and even after switching to
> MSR1 in case.inm for 40 atoms supercell .The job is still running and it
> seems to me the forces on the atoms are decreasing smoothly.Using  NEW1
> minimizer for optimizing the 100 atoms with MSR1a in case.inm was starting
> reasonably but after switching to MSR1 in case.inm it is running badly for
> ec ,cc and fc convergent.I got confused what to do!
>  With best regards
> Mohammed
>
>
>
>
>
>
>   On Wednesday, April 29, 2015 8:05 PM, Mohammed Abujafar <
> mabujafar at yahoo.com> wrote:
>
>
>   Dear WIEN2k Developers and users,
> Hi!
> I am using Geometry minimizer via PORT (min_lapw) for doing supercell
> calculations.I switched to  PORT in the file case.inM.While the job  is
> running ,I checked the case.inm and I found  that MSR1a has been inserted
> automatically in the file case.inm. I have changed it to MSR1.Is that Ok?
> Is it Ok to use PRATT in the case.inm for PORT minimization method? Thank
> you very much in advance.
> With best regards
> Mohammed
>
>
>
>


-- 
Professor Laurence Marks
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu
Corrosion in 4D: MURI4D.numis.northwestern.edu
Co-Editor, Acta Cryst A
"Research is to see what everybody else has seen, and to think what nobody
else has thought"
Albert Szent-Gyorgi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20150429/7766a9a3/attachment.html>