Posting this in case someone searches for it later.
I found that using the batch command (
http://unixhelp.ed.ac.uk/CGI/man-cgi?batch+1) worked great. obviously lots
of different ways to use it, but simply using batch (which only runs a job
in the queue when the machine load drops below a certain level) will run
jobs sequentially.
for example
$ echo "mpirun -np 4 $AMBERHOME/bin/pmemd.MPI -O -i mdin -o mdout -p prmtop
-c inpcrd" | batch
in each directory of the bechmarks, allowed me to stack all the reference
calculations to be run. Once one is done, the other begins. the man page
gives easy and simple instructions, but for my purposes, this works just
fine. (I am the only user right now)
on a side note, pbs was a pain to get setup on a single machine, and since
it has to run a server, it is using resources... does it make difference...
i'm not going to investigate.
Thanks all for the advice and support.
Jonathan
On Fri, Jun 1, 2012 at 4:56 PM, Jonathan Gough
<jonathan.d.gough.gmail.com>wrote:
> Thanks everyone.
>
> I did some reading and I believe that torque can serve and run on a single
> machine. I'll give that a try...
>
>
> On Fri, Jun 1, 2012 at 4:02 PM, Anselm Horn <
> Anselm.Horn.biochem.uni-erlangen.de> wrote:
>
>> Dear Jonathan,
>>
>> >> is there a batch queue app that people use to run amber calculations
>> on a
>> >> single machine?
>> >>
>> >> I am familiar with pbs/torque/moab for clusters. But does anyone use
>> or
>> >> has anyone built one for a single machine and use it to run amber
>> >> calculations?
>> >>
>> >> (I know someone built a simple one to run to run gamess calcs
>> (gamessQ).
>>
>> we used OpenPBS for running jobs through a queue in serial mode on a
>> server machine.
>>
>> The setup was rather easy at that time, because OpenPBS was part of our
>> Linux distribution, at that time. We installed the complete package
>> (server, mom, scheduler) and had all services started at boot time.
>> Then, we used the machine for running local Amber jobs.
>>
>> The pbs/torque/moab way should thus be feasible.
>>
>> Regards,
>>
>> Anselm
>>
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Jun 05 2012 - 15:00:02 PDT
