[Wien] Improving k-mesh/RKMAX etc

[Laurence Marks]

This type of question has come up many times, so let me expand/clarify what
Peter said.

In general, you can optimize positions and converge the density using a
smaller number of k-points and also a smaller RKMAX. You then save this
density and improve both the k-points and RKMAX, and sometimes also the
RMTs (using reduce_rmt_lapw with the "-a" option). In most cases the forces
only change by a few mRyd/au. While the energy and how much of the density
is within the RMTs will change, normally it will converge in these
"improved" runs in a small (5-10) number of iterations. This is much less
than the 20-50 that are normally required to converge from atomic densities
[1].

The reason that Peter recommended saving the prior run is just in case
something goes wrong. (Even experienced useres sometimes mess up!)

[1] If you need more than 50 iterations to converge just the densities
(fixed positions), you may have something wrong with your model.

On Mon, Aug 9, 2021 at 8:22 AM Peter Blaha <pblaha at theochem.tuwien.ac.at>
wrote:

> What do you mean by "can I wait for scf convergence with 4 k-points in a
> new folder ????
>
> Even in a new folder:   NEVER do an   init_lapw
> again, but always use the previous saved files.
>
> mkdir new
> cd new
> cp ../relaxed_4k-points.* .
> restore relaxed_4k-points
>
> x kgen    with better k-mesh
>
> runsp_lapw ...
>
>
> -------- Weitergeleitete Nachricht --------
> Betreff:        Query about denser k-mesh for calculating more accurate
> total
> energy/unit cell
> Datum:  Mon, 9 Aug 2021 17:58:13 +0530
> Von:    shamik chakrabarti <shamik15041981 at gmail.com>
> An:     pblaha at theochem.tuwien.ac.at
>
>
>
> Dear Prof. Blaha,
>                                As has been discussed in the wien2k
> mailing list I am following your advice regarding the structure
> optimization & steps to be followed after runsp_lapw -fc 1.0 -ec 0.0001
> -cc 0.0001 -min.
>
>    save_lapw relaxed_4k-points   (if not already done)
> x kgen   with as many k-points as you want.
> runsp ...   (first without -min, because most likely the forces will not
> change with the bigger k-mesh).  --
>
> Here I have a query. Can I wait till the convergence is achieved with 4
> k points (during total energy calculation in a new folder)& then (after
> convergence with 4 k points) increase the k-points to 14 k points?
>
> with regards,
>
> Dr. Shamik Chakrabarti
> Research Fellow
> Department of Physics
> Indian Institute of Technology Patna
> Bihta-801103
> Patna
> Bihar, India
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
>
> https://urldefense.com/v3/__http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien__;!!Dq0X2DkFhyF93HkjWTBQKhk!A0fR70GC0UYtbhJl4TwLdfZeFN8NsDYDt3pQIReHtL5-hz2vOOUTdWINozbuynIocXtQsQ$
> SEARCH the MAILING-LIST at:
> https://urldefense.com/v3/__http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!A0fR70GC0UYtbhJl4TwLdfZeFN8NsDYDt3pQIReHtL5-hz2vOOUTdWINozbuynIYZHVwHQ$
>


-- 
Professor Laurence Marks
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu
"Research is to see what everybody else has seen, and to think what nobody
else has thought" Albert Szent-Györgyi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20210809/24442064/attachment.htm>