[gmx-users] Segmentation fault in mdrun!!!!!!!!!!!!

[Paul Barrett]

I had a similar problem last week.i.e. a segmentation fault in mdrun. It
arose because I had several parameters set very wrongly leading to
enormous values for the x y and z coordinates for my atoms (and I mean
huge: ~ solar system radius !). I guess the byte size for those values
caused the array writing to spill outside of its allocated area.

so: look back over your files, esp any called Step 0 or step -1 and see
if anything untoward is going on with the values that are being
generated.

Paul Barrett
barrett at biop.ox.ac.uk



>Some time ago I had a similar problem. It turned out to be unrelated to 
>Gromacs. In fact, I had installed additional software on my machine that for 
>some reason (that I did not further investigate in) interacted with gromacs 
>(maybe setting "ulimit" or providing a command with a name is also used in 
>gromacs). 
>
>The fix was to clean up my .bashrc (or .profile/.cshrc) and keep only 
>gromacs-related stuff. 
>
>Kind regards, Karsten.
>
>
>On Sunday 23 February 2003 21:15, David wrote:
>> On Sun, 2003-02-23 at 05:38, nanyu101 wrote:
>> > Dear gmx-users,
>> >  I don't know what happened to my system.I just try to minimize my
>> > bilayer(POPC128a.PDB) downloaded from Dr.Tieleman't web.I think I'v
>> > succeeded in energy minimization and tried to run some dynamics
>> > simulation.But when I run, mdrun -s popcmd.tpr -o popcmd.trr -c
>> > popcmd.gro
>> >
>> > The system told as follows. The file popcmd.tpr comes from my energy
>> > minization's results.
>> >
>> > Back Off! I just backed up ener.edr to ./#ener.edr.1#
>> > starting mdrun 'popc'
>> > 20000 steps,     20.0 ps.
>> >
>> >
>> > Back Off! I just backed up popcem.trr to ./#popcem.trr.1#
>> > Segmentation fault
>> >
>> >
>> > Would you please tell me what happened?
>>
>> What happened is a memory violation in the program. Writing an array
>> outside its bounds. This can be due to two things: bugs, or large
>> forces. Since you have reported problems equilibrating your system the
>> latter seems most likely.
>
>_______________________________________________
>gmx-users mailing list
>gmx-users at gromacs.org
>http://www.gromacs.org/mailman/listinfo/gmx-users
>Please don't post (un)subscribe requests to the list. Use the 
>www interface or send it to gmx-users-request at gromacs.org.
>--------------------------------------------------------------------------------
>Return-Path: <gmx-users-admin at gromacs.org>
>Received: from oxmail.ox.ac.uk by nutmeg.biop.ox.ac.uk (MX V4.2 AXP) with SMTP;
>          Mon, 24 Feb 2003 09:05:54 +0000
>Received: from hawk.theophys.kth.se ([130.237.25.19]) by oxmail.ox.ac.uk with
>          smtp (Exim 4.10) id 18nEYc-0007DV-04 for barrett at biop.ox.ac.uk; Mon,
>          24 Feb 2003 09:05:54 +0000
>Received: (qmail 23385 invoked from network); 24 Feb 2003 09:06:03 -0000
>Received: from localhost (HELO hawk.theophys.kth.se) (127.0.0.1) by 0 with
>          SMTP; 24 Feb 2003 09:06:03 -0000
>Received: (qmail 23342 invoked from network); 24 Feb 2003 09:05:12 -0000
>Received: from igs.cnrs-mrs.fr (193.50.234.245) by 0 with SMTP; 24 Feb 2003
>          09:05:12 -0000
>Received: from newton (newton.cnrs-mrs.fr [193.50.234.237]) by igs.cnrs-mrs.fr
>          (8.8.8/jtpda-5.3) with ESMTP id JAA01216 for <gmx-users at gromacs.org>;
>          Mon, 24 Feb 2003 09:39:26 +0100 (MET)
>Message-ID: <200302240940.12084.Karsten.Suhre at igs.cnrs-mrs.fr>