AMBER Archive (2005)

Subject: RE: AMBER: runtime error in sander?

From: Dave S Walker (dswalker_at_darkwing.uoregon.edu)
Date: Mon Jun 20 2005 - 13:02:16 CDT

Hi Ross,

Your suggestions took me a little while to address. I have no quotas
installed on the cpus and there is ample disk space. I am running on 32bit
cpus and had previously been producing mdvel and mdcrd files ranging
anywhere from 5GB to >10GB without problems, but running sander in shorter
increments would have been easy enough to do. I recently ran into this
problem again while running an equilibration, during which time I wasn't
producing any mdcrd or mdvel files at all.

On the possibility this may be useful to know: my home directory is nfs
mounted on the cpus and I typically passwordless-ssh into them to run
sander. All i/o files are contained in my home directory except for
mdcrd/mdvel files which are stored elsewhere because of their size. When
this truncation error happenned, about 650ps into a 1ns simulation, .dip
and .crd files appeared to have been written and the next simulation did
appear to have started anyway.

Any additional insight into this problem would also be greatly
appreciated. Thanks again for your time.

dsw

On Tue, 31 May 2005, Ross Walker wrote:

> Hi Dave,
>
> > Hello all. I wrote a simple script to govern the simulation of about a
> > dozen different systems; most of them ran fine but for a few of them I
> > started getting this type of message:
> >
> > endfile: truncation failed in endfile
> > apparent state: unit 6 named outfiles/ctcdcemd2.o
> > lately writing sequential formatted external IO
>
> A couple of things to check...
>
> 1) Have you gone over your quota and/or filled your disk?
>
> 2) Are you running on a 32bit machine? If yes then has any file, output,
> mdcrd, mdvel, mden etc grown to larger than 2GB?
>
> Either of these could explain the problem you are seeing. You may find you
> need to run your jobs in shorter increments and then move the files off onto
> another system in order to avoid these problems.
>
> If the above two options check out then please get back to me.
>
> All the best
> Ross
>
> /\
> \/
> |\oss Walker
>
> | Department of Molecular Biology TPC15 |
> | The Scripps Research Institute |
> | Tel:- +1 858 784 8889 | EMail:- ross_at_rosswalker.co.uk |
> | http://www.rosswalker.co.uk/ | PGP Key available on request |
>
> Note: Electronic Mail is not secure, has no guarantee of delivery, may not
> be read every day, and should not be used for urgent or sensitive issues.
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu