NERSCPowering Scientific Discovery for 50 Years

Lawrence Berkeley National Lab Leads International Team to Win High-Performance Computing Bandwidth Challenge

Held at SC2002 conference, team wins for the third year in a row at speeds five times greater than last year's challenge

November 25, 2002

BALTIMORE, Md. — An international team led by scientists at the U.S. Department of Energy's Lawrence Berkeley National Laboratory achieved their third consecutive victory in the High-Performance Bandwidth Challenge at the SC2002 Conference on high-performance networking and computing.

The team won top honors for the Highest Performing Application, moving data at a peak speed of 16.8 gigabits, or16.8 billion bits of data, per second, which was more than five times higher than the team's record-setting win at the SC2001 conference.

The team used clusters of computers at seven sites in the United States, the Netherlands, Poland and the Czech Republic. Entitled "Wide Area Distributed Simulations using Cactus, Globus and Visapult," the winning application modeled gravitational waves generated during the collision of black holes.

"It was great to win it again for the third straight year for a bandwidth challenge hat trick" said Berkeley Lab's John Shalf, leader of the team. "This year we proved that Moore's Law is too slow for networking."

Participating sites in the winning effort were the Parallel Distributed Systems Facility at the National Energy Research Scientific Computing (NERSC) Center at LBNL; and clusters at the SC2002 conference in Baltimore, Argonne National Laboratory, the National Center for Supercomputing Applications, the Pittsburgh Supercomputing Center, the University of Amsterdam, and the Masaryk University in the Czech Republic. Support was provided by the Albert Einstein Institute in Germany, the Poznan Supercomputing and Networking Center in Poland, DOE's Energy Sciences Network (ESnet), Sandia National Laboratories, SysKonnect, Hewlett-Packard and Force10 Networks Inc.

"We're happy to have helped Berkeley win again, especially since they were our first customer," said Rob Quiros, marketing director for Force10 Networks, makers of the E-Series high performance 10 gigabit Ethernet switch used by the team. "This successful collaboration again proves that Force10 continues to define high-performance networking for demanding application by delivering true 10 Gig E performance."

Shalf said the team used a "Global Grid Testbed Collaboration" to win. This same collaboration won two out of three awards in the SC2002 High Performance Computing Challenge.

"As we build a more global infrastructure, researchers will be able to choose from resources around the world to increase their throughput," Shalf said.

The team ran the Visapult volume rendering application (http://www-vis.lbl.gov/Research/visapult2/) at SC2002 to create visualizations from the simulations being run on the participating clusters. The OC-192 and OC-48 lines that fed the Baltimore convention center were aggregated into three 10 gigabit Ethernet links to a Force10 Networks switch that fed the HP/Compaq Linux cluster in the LBNL booth on the showfloor. The team used a cluster of Alpha-based computers loaned by Hewlett-Packard and SysKonnect network interface cards to put together the winning effort.

According to Wes Bethel, the head of LBNL's Visualization Group and developer of Visapult, the effort demonstrated how distributed resources can be used effectively.

Bethel said that improvements in Visapult, along with the evolving networking and Grid infrastructure of hardware, software and middleware helped push the team's data transfer to such a high rate. Members of the team tested the basic setup last July to demonstrate the feasibility of 10 gigabit Ethernet capability. Details on that demo can be found at http://www.lbl.gov/CS/Archive/10gig.html.

The LBNL-led team won the first ever Bandwidth Challenge at SC2000, moving data at an average of 596 megabits per second over 60 minutes and hitting a peak of 1.48 gigabits per second over a five-second period. At the SC2001 conference, the team took the top prize by achieving a sustained network performance level of 3.3 gigabits per second.

First held in 2000, the High-Performance Bandwidth Challenge encourages teams of researchers from around the world to use, if not swamp, the conference network to demonstrate applications using huge amounts of data. The challenge is sponsored by Qwest Communications and provides cash prizes for the winning teams in three categories.


About NERSC and Berkeley Lab
The National Energy Research Scientific Computing Center (NERSC) is a U.S. Department of Energy Office of Science User Facility that serves as the primary high performance computing center for scientific research sponsored by the Office of Science. Located at Lawrence Berkeley National Laboratory, NERSC serves almost 10,000 scientists at national laboratories and universities researching a wide range of problems in climate, fusion energy, materials science, physics, chemistry, computational biology, and other disciplines. Berkeley Lab is a DOE national laboratory located in Berkeley, California. It conducts unclassified scientific research and is managed by the University of California for the U.S. Department of Energy. »Learn more about computing sciences at Berkeley Lab.