[gmx-users] Problem when running in many nodes

[Justin A. Lemkul]

Tomy van Batis wrote:
> Dear all
> 
> The last days I switched from Gromacs 3.3.3 to Gromacs 4.5.3, but I 
> experience some difficulties when running in parallel.
> 
> Initially, my simulation box has dimensions :
> 
> * 3.30507   2.67145  41.15800*
> 
> and it consists of 25 polymers chains with 50 beads/chain
> 
> When I run NVT using 8 or 16 cores, everything is working fine, but when 
> I compress it to dimensions:
> 
> *3.30507   2.67145  10.64441*
> 
> and try again to run NVT in 8 cores, I have the following error message:
> 
> /There is no domain decomposition for 8 nodes that is compatible with 
> the given box and a minimum cell size of 2.5 nm
> Change the number of nodes or mdrun option -rcon or -dds or your LINCS 
> settings/
> 
> As I saw this is well known error, but I can't understand how to fix it :-)
> 

A nearly identical case was posted just a few days ago, and reasons for the 
error discussed:

http://lists.gromacs.org/pipermail/gmx-users/2011-March/059063.html

with a possible work-around:

http://lists.gromacs.org/pipermail/gmx-users/2011-March/059079.html

>  From what I understood, this is happening because_* I use constrains in 
> all the bonds of my system*_
> 

Constraints are but one of many considerations in establishing DD cells.

-Justin

> In my .log file I see the following:
> 
> /Initializing Domain Decomposition on 8 nodes
> Dynamic load balancing: auto
> Will sort the charge groups at every domain (re)decomposition
> Minimum cell size due to bonded interactions: 0.000 nm
> Maximum distance for 5 constraints, at 120 deg. angles, all-trans: 2.000 nm
> Estimated maximum distance required for P-LINCS: 2.000 nm
> This distance will limit the DD cell size, you can override this with -rcon
> Scaling the initial minimum size with 1/0.8 (option -dds) = 1.25
> Optimizing the DD grid for 8 cells with a minimum initial size of 2.500 nm
> The maximum allowed number of cells is: X 1 Y 1 Z 4/
> 
> , but I can't understand how to use -rcon or -dds in order to fix my problem
> 
> Any suggestions?
> 
> Thanks in advance
> Chrysostomos
> 
> 
> 
> 
> 
> 
> 

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

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

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