[StarCluster] running jobs on more than 16 cores

[Ron Chen]

If you specify "cc2.8xlarge", then you get two Intel Xeon E5-2670 (each with 8 cores, and with HT enabled).

>From the Open Grid Scheduler/Grid Engine point of view, you have 32 job slots. From user program point of view, it is a 32-"core"  (virtual core) machine.

Many biotech code benefits from HT, as the working set is usually large and thus the extra HT thread can help hide the cache miss idle cycles.

  -Ron

************************************************************************
Open Grid Scheduler - the official open source Grid Engine: http://gridscheduler.sourceforge.net/



________________________________
From: Sagar Chhangawala <sagar.cornell at gmail.com>
To: starcluster at mit.edu 
Sent: Wednesday, February 6, 2013 12:13 PM
Subject: [StarCluster] running jobs on more than 16 cores


Hi,

I am a bioinformatician and I work on a lot of projects which require more than 16 cores. I see in amazon instance classes that the maximum number of cores in 1 instance is 16 cores. I would like to run jobs on more than just 16 cores. My tools are not MPI compatible. Other than splitting my jobs onto multiple instances, is there any other way to get more than 16 cores on an ec2 machine? 


Thank you.
_______________________________________________
StarCluster mailing list
StarCluster at mit.edu
http://mailman.mit.edu/mailman/listinfo/starcluster