[Crib-list] SPEAKER: JUSTIN RILEY (MIT) -- Computational Research in Boston Seminar -- Friday, September 4, 2009

[Shirley Entzminger]

		COMPUTATIONAL RESEARCH in BOSTON SEMINAR


DATE:		FRIDAY, SEPTEMBER 4, 2009
TIME:		12:30 PM
LOCATION:	Building 32, Room 144  (Stata Center)

Pizza and beverages will be provided at 12:15 pm outside Room 32-144.


TITLE:		StarCluster - High Performance 
		Computing on Amazon's Elastic Compute Cloud


SPEAKER:	JUSTIN RILEY  (Massachusetts Institute of Technology)


ABSTRACT:

StarCluster is a utility for creating and managing general purpose 
computing clusters hosted on Amazon's Elastic Compute Cloud (EC2).  
StarCluster makes it easy to create a traditional computing cluster used 
in research labs or for general purpose parallel and distributed computing 
applications in the cloud.  All that's needed to get started with your own 
computing cluster on EC2 is an Amazon AWS account and StarCluster.

Once you've created an AWS account and have StarCluster installed you 
create a simple configuration file and execute a single command to launch 
the cluster.  StarCluster then completely automates the process of 
creating a high performance computing cluster on Amazon EC2.  This 
involves requesting virtual machines from Amazon, configuring the cluster 
with Sun Grid Engine, OpenMPI, passwordless SSH, NFS shared /home and /opt 
directories, and ~140GB of /scratch space.

Topics:
------
-	Software Tools for Academics and Researchers (STAR) group at MIT
-	Quick intro to Amazon's Web Services 
-	Motivations behind StarCluster (education/research)
-	StarCluster Features
-	Case study (Materials Science)
-	Live Demo of installing StarCluster and launching a small 4 node 
-	OpenMPI/Sun Grid Engine cluster on EC2

Website:  http://web.mit.edu/starcluster

------------------------------------------------------------------------------

Massachusetts Institute of Technology
Cambridge, MA  02139

For more information about the CRiB Seminars, please visit

		http://math.mit.edu/crib

*******************************************************************************