[torqueusers] Torque job options

Gus Correa gus at ldeo.columbia.edu
Mon Feb 18 11:23:01 MST 2013


Hi Kamran

You could add node properties to your server_priv/nodes file,
say:

suraci np=8 head
n001 np=12 compute
n002 np=12 compute

restart the pbs_server,

and ask for specific node properties in your job script:

#PBS -l nodes=1:head+1:compute

say, to get one node of each type (total 20 cores).

Is this what you want to do?

See this:
http://docs.adaptivecomputing.com/torque/help.htm#topics/1-installConfig/specifyNodeFeatures.htm
http://docs.adaptivecomputing.com/torque/help.htm#topics/2-jobs/requestingRes.htm


Or do you want to pack as much work on the available cores?

Then, see this:
http://docs.adaptivecomputing.com/torque/Content/topics/11-troubleshooting/faq.htm#qsubNotAllow

and this recent thread:

http://www.supercluster.org/pipermail/torqueusers/2013-February/015528.html

I hope this helps,
Gus Correa

On 02/18/2013 12:33 PM, Kamran Khan wrote:
> Hi Ricardo,
>
> Thanks for the speedy response.  That makes perfect sense to me.  I just
> wasn’t sure if I was able to make those specifications in the ‘#PBS –l’
> settings.
>
> CLOSED
>
> --
>
> Kamran Khan
>
> PSSC Labs
>
> Support Technician
>
> *From:*torqueusers-bounces at supercluster.org
> [mailto:torqueusers-bounces at supercluster.org] *On Behalf Of *Ricardo
> Román Brenes
> *Sent:* Monday, February 18, 2013 9:16 AM
> *To:* Torque Users Mailing List
> *Subject:* Re: [torqueusers] Torque job options
>
> if you want to use suraci, you cant ask for ppn=12 =(
>
> Sorry...
>
> you can ask for all nodes with 8 cores. and that would get you the 3
> machines reserved, there yo ucan filter by hostname or IP which ones
> have 12 cores and use the 12, and use just 8 in suraci in you code or
> bash script.
>
> On Mon, Feb 18, 2013 at 11:05 AM, Kamran Khan <kamran at pssclabs.com
> <mailto:kamran at pssclabs.com>> wrote:
>
> Hi Ricardo,
>
> Yeah, suraci is the pbsserver and has 4 cores not reserved for torque.
>
> So I can specify ‘#PBS -l nodes=n001+n002:ppn=12‘, but can I also
> specify suraci?
>
> That’s where I am getting a little confused because I have the 2 compute
> nodes with 12 available cores and the head node with only 8 available cores.
>
> Is it at all possible to tell specific systems to use a specific amount
> of cores, or would I just have to lower the amount of cores being used
> on each system to 8? Such as:
>
> PBS -l nodes=3:ppn=8
>
> Thanks in advance for your support.
>
> --
>
> Kamran Khan
>
> PSSC Labs
>
> Support Technician
>
> *From:*torqueusers-bounces at supercluster.org
> <mailto:torqueusers-bounces at supercluster.org>
> [mailto:torqueusers-bounces at supercluster.org
> <mailto:torqueusers-bounces at supercluster.org>] *On Behalf Of *Ricardo
> Román Brenes
> *Sent:* Friday, February 15, 2013 3:39 PM
> *To:* Torque Users Mailing List
> *Subject:* Re: [torqueusers] Torque job options
>
> hello! :-)
>
> you can specify specific hosts with "-l nodes=n001+n002".
>
> that option will request both big nodes :-)
>
> and on a side note if suraci has the pbsserver and has 8 cores, leave 1
> or 2 reserved not for torque :-P
>
> On Feb 15, 2013 5:07 PM, "Kamran Khan" <kamran at pssclabs.com
> <mailto:kamran at pssclabs.com>> wrote:
>
> Hi Everyone,
>
> This is my first message to the community, so HIIIII!!
>
> Below is my torquetest.sh script which I use to test the functionality
> of torque across my cluster.
>
> #!/bin/bash
>
> #PBS -l walltime=00:1:00
>
> #PBS -l nice=19
>
> #PBS -q default
>
> #PBS -l nodes=2:ppn=12
>
> mpiexec -machinefile /opt/machinelist -np 24 hostname
>
> echo "end"
>
> And here is a copy of my pbsnodes output:
>
> suraci
>
> state = free
>
> np = 8
>
> ntype = cluster
>
> status = rectime=1360968542,varattr=,jobs=,state=free,netload=2013931431
> <tel:2013931431>,gres=,loadave=1.00,ncpus=12,physmem=24562968kb,availmem=48500940kb,totmem=49728784kb,idletime=275,nusers=2,nsessions=2,sessions=18767
> 23996,uname=Linux icarus.beowulf6.actainc.com
> <http://icarus.beowulf6.actainc.com> 2.6.32-279.19.1.el6.x86_64 #1 SMP
> Sat Nov 24 14:35:28 EST 2012 x86_64,opsys=linux
>
> mom_service_port = 15002
>
> mom_manager_port = 15003
>
> n001
>
> state = free
>
> np = 12
>
> ntype = cluster
>
> status =
> rectime=1360968542,varattr=,jobs=,state=free,netload=865146013,gres=,loadave=0.00,ncpus=12,physmem=24562968kb,availmem=49216676kb,totmem=49728784kb,idletime=81117,nusers=0,nsessions=0,uname=Linux
> n001.beowulf6.actainc.com <http://n001.beowulf6.actainc.com>
> 2.6.32-279.19.1.el6.x86_64 #1 SMP Sat Nov 24 14:35:28 EST 2012
> x86_64,opsys=linux
>
> mom_service_port = 15002
>
> mom_manager_port = 15003
>
> n002
>
> state = free
>
> np = 12
>
> ntype = cluster
>
> status =
> rectime=1360968542,varattr=,jobs=,state=free,netload=647756235,gres=,loadave=1.00,ncpus=12,physmem=24562968kb,availmem=49226348kb,totmem=49728784kb,idletime=808775,nusers=0,nsessions=0,uname=Linux
> n002.beowulf6.actainc.com <http://n002.beowulf6.actainc.com>
> 2.6.32-279.19.1.el6.x86_64 #1 SMP Sat Nov 24 14:35:28 EST 2012
> x86_64,opsys=linux
>
> mom_service_port = 15002
>
> mom_manager_port = 15003
>
> As you can see, ‘suraci’ has 8 free processors, while n001 and n002 have
> 12 free processors. How do I establish these settings in my
> torquetest.sh? Since I have 2 nodes with 12 and 1 node with 8.
>
> Does this make sense?
>
> Thanks in advance.
>
> --
>
> Kamran Khan
>
> PSSC Labs
>
> Support Technician
>
>
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org <mailto:torqueusers at supercluster.org>
> http://www.supercluster.org/mailman/listinfo/torqueusers
>
>
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org <mailto:torqueusers at supercluster.org>
> http://www.supercluster.org/mailman/listinfo/torqueusers
>
>
>
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers



More information about the torqueusers mailing list