[gmx-users] Error in analyzing traj.trr file using g_clustsize, etc.

[Mark Abraham]

On 27/09/2011 10:42 PM, A Taqia wrote:
> Dear Gromacs experts,
>
> I ran a Gromacs simulation (with Gromacs 4.5.4) on a BlueGene/P system 
> and generated a trajectory file.  However, when I performed analysis 
> on this file (traj.trr) (for example using g_clustsize), the following 
> error was encountered:
> -------------------------------------------------------
> Program trjconv_d, VERSION 4.5.1
> Source code file: trnio.c, line: 68
>
> Fatal error:
> Float size 164738970. Maybe different CPU?
> For more information and tips for troubleshooting, please check the 
> GROMACS
> website at http://www.gromacs.org/Documentation/Errors
> -------------------------------------------------------
>
> I used the post processing tool (g_clustsize) from different versions 
> of Gromacs (4.5.4, 4.5.1, 4.0.x) on both the front end and back end 
> nodes of the BlueGene.  I even tried using g_clustsize built on a 
> Linux, but encountered the same error.  I was wondering if anyone can 
> shed me some lights on how to resolve this problem.  Your assistance 
> is greatly appreciated.

The simplest explanation for the symptoms is that your trajectory file 
is corrupt somehow - most likely because of something someone did to the 
file, or to the file system. The size of a .trr file is linear in the 
number of frames, so if you can run a short version of the simulation 
separately, you can estimate how big this file should have been. 
gmxcheck is available to probe the integrity of the file, but I expect 
it will also simply tell you something is catastrophically wrong.

Mark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110928/69ab6a00/attachment.html>