[CPMD-list] Re: CPMD-list Digest, Vol 2, Issue 6

[Axel Kohlmeyer]

>>> "JD" == John Kofi Dogbe <dogbe at unr.nevada.edu> writes:

JD> Dear CPMD Users,

JD>     I sent out an e-mail with a request on how to efficiently do slab
JD>     calculations using the CPMD. I did not get any feedback on this. My
JD>     request appeared in the CPMD-list Digest vol. 2, Issue 6. In any case,
JD>     I've reproduced below my previous e-mail. If someone  have had an
JD>     experience to share with me on this, it will be very much appreciated.

dear john,

there are several points you have to keep in mind, when asking for help
here (or on most other mailing lists).

support is done on a voluntary basis, so if the people that are
willing and able to answer your question are otherwise busy, you
will have to wait, or - as you did - ask again and hope for the best.

usually you will get a faster answer, if you ask more specific questions.

there are two very nice write-ups about how to report bug and/or 
problems and how to answer them at 
<http://freshmeat.net/articles/view/149/>
and
<http://freshmeat.net/articles/view/1082/>

i'd also like to take the opportunity to encourage more people to
respond to questions here and contribute to the CPMD documentation. 
most of the time, you can already help people a lot by answering 
part of the question. furthermore, once you have - for example - figured
out something, that was not properly documented in the manual, please 
consider writing it down and sending it to the list of one of the cpmd 
developers for inclusion into the manual. again it does not need to
be perfect or complete (or in latex).


ok, now on to your questions.

JD>     Meanwhile, I've started a slab calculation which is taking unusually
JD>     very long compared to those I've done previously. It's been a week now!

so what is the difference between the runs? there are some options that
will increase the amount of cpu time needed significantly...

JD>     By the way, is it possible that CPMD could re-start (NOT using the
JD>     RESTART keyword) itself within the same job?

could you elaborate a bit more, what you mean by that and what you 
want to achieve.

JD>     See below my previous e-mail. Thank you.
JD>     John.
JD>     **************begin previous e-mail*******************
JD>     Dear All,
JD>     I need to do Si slab calculations and I'm wondering if those who have
JD>     experience in this could give me some pointers on how to do this
JD>     efficiently with CPMD. I've been able to obtain results (geometry
JD>     optimizations) using Si clusters. These results compare very well with
JD>     previous results from the Gaussian package.
JD>     Since I'm fairly new to CPMD, I'll appreciate it if anyone can give me
JD>     hints on how to set up the slab calculations. Specifically, I need
JD>     keywords or combinations thereof which could help me do this. I'm
JD>     currently experimenting with different keywords using the manual.
JD>     A sample input file with comments could do as well.

frankly, your request is too general so that one either can give you 
a fairly generals answer or would have to give you a full lecture on 
surface/slab calculations.

basically you'd just take a proper slice from a bulk crystal
configuration and enlarge your super cell in z-direction. due
to the fact, that with plane waves even calculating the vaccuum
will cost considerable amounts of cpu time, you select the vacuum 
area that the periodic images don't interact significantly.
alternatively you can use a poisson solver (hockney, tuckerman or
mortensen, activated by using SYMMETRY 0, in case of a slab
geometry you also have to specify the SURFACE keyword) to reduce that 
effect for smaller boxes (be careful, there are minimum distances 
of the atoms from the border required for the poisson solvers to work 
properly) at the expense of some cpu time. 

this again demonstrates the need to have a real cpmd tutorial.
until then you may want to check out the material from the CECAM
tutorial on juerg hutters homepage. 
<http://pciwww.unizh.ch/pci/hutter/information/information.html>

there also should be some questions and anwers regarding slab
calculations in the mailing list archives.

hope this helps,
     axel.


JD>     All suggestions will be appreciated. Thanks.
JD>     John
JD>     *************end previous e-mail**********************
JD>     --
JD> **The first is not necessarily the Leader**

JD>     	"If I have spoken evil, bear witness of the evil:
JD>     		but if well, why smitest thou me?"
JD>     				-- Jesus Christ (John, 18:23)
JD> _______________________________________________
JD> CPMD-list mailing list
JD> CPMD-list at cpmd.org
JD> http://cpmd.org/mailman/listinfo/cpmd-list



--

=======================================================================
Axel Kohlmeyer       e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.