[gmx-users] Fwd: Help for REMD

Mark Abraham Mark.Abraham at anu.edu.au
Mon Feb 1 15:45:01 CET 2010


Please keep GROMACS correspondence on the mailing list. Nobody here is a 
private tutor unless you're paying them.

-------- Original Message --------
Subject: Help for REMD
Date: Mon, 01 Feb 2010 14:58:08 +0200 (EET)
From: Murat Erman OGUZ <erman at hacettepe.edu.tr>
To: mark.abraham at anu.edu.au

Dear Mr.Mark Abraham
I am Erman Oguz from Turkey Hacettepe University.
I have a problem about Replica Excahange Method on Gromacs.
ı hope you can help me.

I am trying to perform replica exchange MD on a 'protein in
water' system. I have to perform the REMD simulation with 64 different
temperatures.As per advise on wiki,I equilibrated the system at
respective temperatures (total of 64 equilibration simulations in 200 ps).
After this I generated "tpr files" from the equilibrated structure.
when I submit final job for REMD with following command-line

command line: mpirun -np 64 mdrun -multi 64 -replex 1000 -s *_.tpr .....

it gives error that :

t=29.802 ps Water molecule starting at atom 5902 can not be settled.
check for bad contacs and/or reduce time step.

what can I do and how can I do... please help me...


You should start by searching on the GROMACS webpage errors page. The 
reasons for this kind of error are clearly described there, and are 
likely to have nothing to with REMD as such. If they were, you'll have 
to provide much more data for anyone to help.

You should try your .tpr files in a non-replica simulation and see which 
are broken and thus what to do about it. Probably you've not 
equilibrated well enough, or have a broken system, or have a temperature 
range that is too great for your integration to be stable at high T.

Mark



More information about the gromacs.org_gmx-users mailing list