some issues with ab initio MD with xTB

[amin....@gmail.com]

Hi everyone,

I am trying to do some ab initio MD simulation with xTB to evaluate the interatomic interactions.

For a cubic cell with periodic boundary condition, for the same number of particles, as I reduce the number density (for the same number of particles only increasing the box size), the simulation time exponentially increases. Does any one have any idea?

Best regards,

Amin

[hut...@chem.uzh.ch]

Hi

without additional information you will not get any helpful replies.

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Institut für Chemie C FAX : ++41 44 635 6838
Universität Zürich E-mail: hut...@chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----
To: "cp2k" <cp...@googlegroups.com>
From: "amin....@gmail.com"
Sent by: cp...@googlegroups.com
Date: 05/26/2021 09:38AM
Subject: [CP2K:15464] some issues with ab initio MD with xTB

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/c48868f0-04ce-49bc-8ebf-7f0029c7b95cn%40googlegroups.com.

[amin....@gmail.com]

Thanks so much prof. Hutter for your reply.

Attached, I provide an example. In the zip file there are two input scripts main_1.inp and main_2.inp which both run MD simulation with interactions evaluated by xTB for a simulation box containing 8 CO2 molecules.

Everything is the same for the two scripts except the size of the simulation box which in main_2 is ca 2.5 times larger in each direction. For these two scripts, each MD step took on my system with 32 cores roughly 0.64 and 11 s, respectively which seems kind of unusual to me.

I am wondering first of all, is it normal that for larger cells which contain exact same particles and default cut-off computations become so slower and are there some ways to avoid it, e.g. if someone need to study a very dilute gas state?

Thanks in advance and best regards,

Amin

[hut...@chem.uzh.ch]

Hi

the increase in time is expected as your system becomes dominated
by the Ewald sum which scales like a^3 log(a) where a is the
size of your box. So this is definitely not an exponential increase.

If you only want to calculate a set of isolated molecules in a
large box, I would not use Ewald.

Are you sure your geometry is correct? SCF is not converging for me.

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Institut für Chemie C FAX : ++41 44 635 6838
Universität Zürich E-mail: hut...@chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----
To: "cp2k" <cp...@googlegroups.com>
From: "amin....@gmail.com"
Sent by: cp...@googlegroups.com

Date: 05/28/2021 11:08AM
Subject: Re: [CP2K:15476] some issues with ab initio MD with xTB

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/91275009-cf18-4e50-9b05-b86c412fc127n%40googlegroups.com.


[attachment "cp2k.zip" removed by Jürg Hutter/at/UZH]

[amin....@gmail.com]

many thanks prof. Hutter, that solved the issue.

About the initial structure, yes you are right. It was actually the initial structure for a box of water which I just manually changed the element tags to CO2 to see are there element related issues or not. thanks for noticing that.

Best regards,

Amin