[CP2K-user] [CP2K:15476] some issues with ab initio MD with xTB

[hut... at 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... at chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp... at googlegroups.com wrote: -----
To: "cp2k" <cp... at googlegroups.com>
From: "am... at gmail.com" 
Sent by: cp... at googlegroups.com
Date: 05/28/2021 11:08AM
Subject: Re: [CP2K:15476] some issues with ab initio MD with xTB

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

On Thursday, May 27, 2021 at 4:46:35 PM UTC+2 jgh wrote:
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: h... at chem.uzh.ch 
Winterthurerstrasse 190 
CH-8057 Zürich, Switzerland 
--------------------------------------------------------------- 
 
-----c... at googlegroups.com wrote: ----- 
To: "cp2k" <c... at googlegroups.com> 
From: "am... at gmail.com"  
Sent by: c... at googlegroups.com 
Date: 05/26/2021 09:38AM 
Subject: [CP2K:15464] some issues with ab initio MD with xTB 
 
 
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 
   
  --  
 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 cp... at googlegroups.com. 
 To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/c48868f0-04ce-49bc-8ebf-7f0029c7b95cn%40googlegroups.com. 
  
  
  -- 
 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 cp... at googlegroups.com.
 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]