[gmx-users] Re: gmx-users Digest, Vol 81, Issue 140

Adwait Mevada adwait at imsc.res.in
Thu Jan 20 15:11:49 CET 2011


>    1. Re: avoiding bad contact in continutation run (Mark	Abraham)
> Adwait Mevada wrote:
>>
>>> On 01/20/11, Adwait Mevada  <adwait at imsc.res.in> wrote:
>>>>
>>>>  Dear gmx-users,
>>>>  I am simulating a dppc + chol system, the force field is martini
>>>> with gromacs 3.3.3 the system was initially run for 150 ns, but i
>>>> had to later extend the simulation, so using the confout.gro and
>>>> with mdp options:
>>>> gen_vel = no
>>>> unconstrained_start = yes
>>>>  I gave a re-run for another 850 ns, now due to some interaction  the
>>>> box explodes after 460ns of run.I found that the cause was due  to
>>>> some unwanted bad contact.
>>>> It is important that the system continue from the previous state,
>>>> but the extension is leading to this bad contact.
>>>>
>>>
>>> Actually, you didn't preserve your ensemble if you used the .gro  file
>>> for the coordinates. See
>>> http://www.gromacs.org/Documentation/How-tos/Doing_Restarts Actually
>>>  providing us with copies of the commands you're using is always a
>>> good idea when asking for help.
>>
>> 150ns.gro was the output of the previous simulation run i.e. confout.gro
>> mdp file for the previous simulation is taken as is except for change of
>> simulation steps and making unconstrained_start = yes
>> grompp -np 8 -f 850ns.mdp -p dppc_chol.top -c 150ns.gro -o
>> 512lip_256chol_850ns.tpr
>
> As Mark suspected, you have failed to preserve your ensemble, due to the fact
> that you're using a reduced precision (.gro) format for coordinates,  
>  and you are
> not supplying the .trr and .edr file from the previous run, so you're losing
> velocities, temperature and pressure coupling information, etc.
>
> -Justin
so i should be doing something like
tpbconv -s 512lip_150ns.tpr -f 512lip_150ns.trr -e 512lip_150ns.edr -o  
512lip_850ns.tpr -extend 850

mpiexe.openmpi -np 8 mdrun_mpi.openmpi -np 8 -s 512lip_850ns.tpr
  -o 512lip_850ns -e 512lip_850ns -g 512lip_850ns &

>> mpiexe.openmpi -np 8 mdrun_mpi.openmpi -np 8 -s 512lip_256chol_850ns.tpr
>> -o 512lip_256chol_850ns -e 512lip_256chol_850ns -g 512lip_256chol_850ns &
>>
>>> Mark
>>>
>>>>
>>>> My question is, is there a way to avoid this while retaining the
>>>> continuation of the previous simulation?
>>>>  Is using lower precision file a solution to this problem? if yes
>>>> how do i go about it.
>>>> -Adwait
>>>>
>>>> ----------------------------------------------------------------
>>>> This message was sent using IMP, the Internet Messaging Program.
>>>>
>>>> --


-Adwait

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.




More information about the gromacs.org_gmx-users mailing list