[gmx-users] number of DD cells

[Justin A. Lemkul]

Moeed wrote:
> Hello,
> 
> I am attempting to increase the density using NPT. As I increase the 
> pressure to compress the system after some steps simulation crashes. I 
> thought maybe its becasue I am compressing too fast but even when I take 
> a stepwise approach to compress gradually the same error comes up. The 
> density at the point of crash is about 30 SI while I need 650 which is 
> the actual density. Topology file is generated successfully and also I 
> did EM before MD.
> 
> Initial box size is 30 nm. As suggested in the link below I reduced the 
> size to 20 nm (there are 2500 atoms in the system ) but the same error 
> appears.
> http://lists.gromacs.org/pipermail/gmx-users/2009-April/041488.html
> 
> I also tried less -np from 7 to 4 but no success.
> 
> Can you please tell me what this exclamation mark mean which I get at 
> times in parallel trials.(vol 0.99 !)
> 

Domain decomposition relies on dividing your system up into reliably-sized 
domains in order to communicate force, coordinate, energy, etc. information.  If 
something happens to skew or annihilate these domains, the algorithm crashes. 
The exclamation marks indicate (in a general sense) that the DD algorithm has 
detected something dangerously wrong.

If you're trying to make drastic changes to the size or dimensions of your 
system, DD is a bad choice and you'll likely have to use particle decomposition 
(mdrun -pd).

<snip>

> fourierspacing      =  0.3                 

You're setting yourself up for inaccurate PME calculations with this.  Use 0.1 - 
0.12 unless you've got some compelling reason to decrease your accuracy.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================