VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Mon Sep 04 2006 - 14:00:38 CDT

On 9/4/06, Gordon Wells <gordon.wells_at_gmail.com> wrote:
> Hi Axel
>
> Thanks for your comments
>
> Some of the script is a leftover from a file initially for submitting namd
> jobs. Those parts shouldn't be causing any problems though.
>
> The jobs are being launched from a master server on the node selected by pbs,
> thats why the ssh. As far as I understand pbs just returns the resource info,
> it's then up to the rest of the script and the programs it calls to make use
> of that.

no. not on all the PBS servers that i have administrated or used
with one exception (cray xt3, but on that machine, you _cannot_ ssh
into the nodes at all, as it is more like a transputer type machine
in terms of batch processing).

you can easily test this by running a script containing only the
hostname command and see whether its output contains the
name of the server or that of a compute node.

> As for the loading order, the coordinates are indeed the first frame.

so then you should definitely load the .psf first.

> I don't get any errors using ">&". pre_rmsd_per_res prepares various files and
> variables before calling bigdcd. I can post it, but as far as I know it
> shouldn't be the cause of the problem. If I run the "ssh..." command (or the
> script) directly all works fine. The problem comes when calling from the
> script via qsub. Using sleep doesn't help either, it seems to interfere with
> the repeated frame loads from bigdcd and isn't executed after bigdcd.

hmmm... there could be a problem with getting a proper console, but it
is unlikely, as it would indicate that somebody has to have a serious chat
with your system administrator...

> I'd still like to know why vmd simply when returnig control, since the
> intended quit command which get's called at the end of bigdcd is never even
> reached?

it actually shouldn't. just to make sure, can you run the _identical_ command
line, that you run from qsub script, in a shell on the master without any
problem?

in order to debug this, you can also submit an interactive shell to the queue.

qsub -q md -l nodes=1 -I

(not that the last flag is a capital i)

there might just be an error in one of the scripts/procedures used.
you may want to look through the mailing list archives for more
suggstions. i remember that there were discussions about running
bigdcd through a batch system quite a while ago.

cheers,
   axel.

>
> regards
> gordon
>
> --
>
> If you want to know what God thinks of money, just look at the people he
> gave it to.
>
> -- Dorothy Parker
>
> Gordon Wells
> Bioinformatics and Computational Biology Unit
> Department of Biochemistry
> University of Pretoria
>
> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey_at_cmm.chem.upenn.edu   http://www.cmm.upenn.edu
  Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.