[gmx-users] Gromacs 4.0.7 failing in parallel mode

[Mark Abraham]

----- Original Message -----
From: Jyoti Mahalik <jmahalik at polysci.umass.edu>
Date: Tuesday, September 28, 2010 9:19
Subject: [gmx-users] Gromacs 4.0.7 failing in parallel mode
To: gmx-users at gromacs.org

> Hi,
>  I am simulating two protein units in water and it works 
> fine in the single processor but when I use the parallel version 
> of the GROMACS, it fails.
> 
> My script:
> grompp_mpi_d -f md.mdp -c hba_b4md.gro -r hba_b4md.gro -p 
> hba.top -o hba_md.tpr
> mpirun -np $NPROCS mdrun_mpi_d -np $NPROCS -ddorder pp_pme -s 
> hba_md.tpr -o hba_md.trr -c hba_pmd.gro -g md.log -e md. edr
> 
> Error Messages:
> Step 0, time 0 (ps)  LINCS WARNING
> relative constraint deviation after LINCS:
> rms nan, max inf (between atoms 1 and 2)
> bonds that rotated more than 30 degrees:
>  atom 1 atom 2  angle  previous, current, 
> constraint length
>       1      
> 2   90.0    
> 0.1000      inf      0.1000
>     712    714   
> 90.0    0.1250      
> inf      0.1250
>     712    713   
> 90.0    0.1250      
> inf      0.1250
>     710    711   
> 90.0    0.1000      
> inf      0.1000
>     709    710   
> 90.0    0.1430      
> inf      0.1430
> 
> 
> Program mdrun_mpi_d, VERSION 4.0.7
> Source code file: constr.c, line: 136
> 
> Fatal error:
> Too many LINCS warnings (1286)
> If you know what you are doing you can adjust the lincs warning 
> threshold in your mdp file
> or set the environment variable GMX_MAXCONSTRWARN to -1,
> but normally it is better to fix the problem

See http://www.gromacs.org/Documentation/Errors#LINCS.2fSETTLE.2fSHAKE_warnings

Probably your single processor version was also fairly unhappy, but was lucky enough to succeed. Reconsider your system preparation workflow in the light of the "blowing up" page on the above link.

Mark

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100928/3043556f/attachment.html>