[torqueusers] Scalability of Torque

Chi Chan chichan2008 at gmail.com
Wed Apr 29 09:15:30 MDT 2009


On Mon, Apr 27, 2009 at 1:24 PM, Chi Chan <chichan2008 at gmail.com> wrote:
> I believe node number is not a scalability anymore in modern versions of Torque, SGE or Platform LSF.

I mean a huge number of nodes is not a scalability _limit_ anymore.

Sorry for the typo, and thanks to those who pointed it it!!!

 --Chi


>
> It's just a few years ago that people complained about the scalability
> of OpenPBS on clusters of 100 nodes or more!!!
>
> --Chi
>
>
>
> On Thu, Apr 23, 2009 at 2:36 PM, Garrick Staples <garrick at usc.edu> wrote:
>> On Wed, Apr 22, 2009 at 10:24:58PM -0500, Chi Chan alleged:
>>> I got into a discussion this morning with a friend who mentioned that
>>> SGE scales to over 3900 nodes (the one at TACC - Texas Advanced
>>> Computing Center), and those are real nodes with one SGE's execd on
>>> each node. Platform LSF scales to clusters with similar number of
>>> nodes as well.
>>>
>>> I am wondering how many nodes the largest known Torque cluster has? I
>>> know that the #1 on the top500 is running torque, but in terms of the
>>> number of nodes, which one is the biggest?
>>
>> I currently have 2691 pbs_mom nodes and I don't see why it I couldn't double it
>> without changing anything.
>>
>> The largest are in private enterprise and won't show up on top500.  Sshhh.
>> It's a secret.
>>
>> AFAIK, Torque currently runs on clusters with tens of thousands of nodes.
>>
>> --
>> Garrick Staples, GNU/Linux HPCC SysAdmin
>> University of Southern California
>>
>> The pro-disease movement: http://www.jennymccarthybodycount.com/
>>
>>
>> _______________________________________________
>> torqueusers mailing list
>> torqueusers at supercluster.org
>> http://www.supercluster.org/mailman/listinfo/torqueusers
>>
>>
>


More information about the torqueusers mailing list