HPCwire
 The global publication of record for High Performance Computing / May 14, 2004: Vol. 13, No. 19

  |  Table of Contents  |  

Cluster Computing:

FIRM OF 55 CREATES EXTREMELY COMPETITIVE SUPERCOMPUTER

California Digital, a small firm with just 55 employees, has created a supercomputer which can compete with some of the most powerful in the world.

The Silicon Valley based company has used 1,024 Itanium 2 servers and an unusually small group of employees to create the supercomputer, which allows for 19.94 trillion operations per second. Because the computer, dubbed Thunder, was created past the deadline for the Top 500 list of most powerful supercomputers, it did not show up on the list. Experts believe that it would have easily secured second place on the list, which is compiled by University of Mannheim, University of Tennessee and Lawrence Berkeley National Laboratory.

In a similar fashion to Virginia Tech's clustering endeavors, California Digital connected relatively inexpensive computers running Linux operating systems to summon the power of a traditional supercomputer, which would normally cost three times as much. Virginia Tech used Apple computers and IBM chips to create a cheap, but powerful cluster.

In addition, California Digital has supplied several Fortune 500 companies with large Linux clusters. Other small firms, like Linux Networx, which is building a 2,816-processor cluster and a 512-processor cluster for Los Alamos National Laboratory, as well as a 2,132 Intel Xeon processor for the US Army Research Laboratory, are taking notice of the opportunities in this field.

The National Center for Supercomputing Applications (NCSA) is looking at adopting another large cluster, in addition to the two current Linux clusters already installed.

The technological abilities, coupled with the more affordable prices, are beginning to attract companies who want to break away from the large, monolithic machines running proprietary products. Most necessary applications, aside from things like weather prediction typically performed by supercomputer's like NEC's Earth Simulator, can be run on two- and four- processor based Intel or AMD systems running Linux.



Often referred to as "cluster computing on steroids," Grid computing is for real. Web services, utility computing, .NET, CPU harvesting and distributed computing are just a few of the technologies that fall under the Grid computing umbrella. Gt04 -- a premiere enterprise Grid computing conference targeting industrial and commercial users -- will gather experts, and outline strategies and road maps for Grid deployment. For more information, visit http://www.gt04.com.



Top of Page

  |  Table of Contents  |