Force10 touts supercomputing success
By Bryan Betts
,
TechWorld
, 06/28/2008
- Share/Email
- Tweet This
- Print
10Gig Ethernet specialist Force10 Networks has claimed that its switch/routers are now integral to more than a dozen of the
world's top 40 supercomputers. They include the U.S. Los Alamos National Laboratory's Roadrunner, which currently ranks as
the world's most powerful supercomputer and was the first to reach a peak performance of 1 Petaflop.
Force10 added that its networking hardware is also a common thread that links the fastest Cray, Dell, IBM and Sun-based machines
on the Top500 ranking of supercomputers with the fastest systems in Germany, Korea, Russia and the U.S.
The report highlights the critical importance of always-on networking technology in supercomputer work, both in labs and in
commercial applications, said Sachi Sambandan, the company's VP of engineering.
"As supercomputers grow more powerful, it is increasingly important to minimize potential bottlenecks," he said. "Force10
products are designed to deliver reliability without compromising throughput - that's essential for supercomputers like Roadrunner
that are reaching Petaflop performance levels."
"If you have a powerful supercomputer, you also need to connect it to the outside world," agreed Klaus Wolkersdorfer, head
of the high-performance computing department at Germany's Forschungszentrum Jülich, whose quarter-Petaflop Jugene system is
ranked the sixth fastest in the world.
Jugene comprises 16 IBM BlueGene/P racks with a total of 65,536 processors and 152 I/O nodes. These are connected to each
other, and to the research facility's networks and other supercomputers, by a Force10 TeraScale E1200 switch.
"We needed over 200 10Gigabit ports, and only the Force10 system was capable of this," added Wolkersdorfer.
Force10 said that alongside Los Alamos and Forschungszentrum Jülich, its supercomputing users include:
-- Moscow State University, which is home to Russia's leading supercomputer, SKIF-GRID. It uses the Force10 C300 resilient
switch, as well as the company's S50 and S2410 switches;
-- The Tachyon supercomputer at the KISTI Supercomputing Center in Korea;
-- The Texas Advanced Computing Center's Sun-based Ranger;
-- Oak Ridge National Laboratory in the US, with its Cray supercomputer called Jaguar; and
-- Abe, at the U.S. National Center for Supercomputing Applications, which is number 23 on the list and is the most powerful
Dell supercomputer.
Partner Content
Simplify Your Branch Infrastructure
Learn how to simplify your branch infrastructure while dramatically increasing app performance with Citrix Branch Repeater.
Download the Free Info Kit
Next-Gen Load Balancing
Free Guide: "Next Gen Load Balancing: 8 Things You Need to Handle Today's Network Traffic" shows you the functionality needed in your next load balancer.
Download the Free Guide
Accelerate Your Web Apps by up to 5x
Free Guide: "The Secret to Getting Maximum Speed from your Web Applications."' Learn how you can deliver Web apps up to 5x faster.
Download the Free Guide
Comments (1)
BSBy Anonymous on July 5, 2008, 4:55 amThere is no way their store and forward switches (or s2410 - fulcrum trash) can deliver that performance. With latency in excess of 30us... This article is crafty...
Reply | Read entire comment
View all comments