IBM Continues Supercomputing Deal Spree

One week to the day after it agreed to provide
supercomputing capabilities for the European Centre for Medium-Range Weather Forecasts (ECMWF), IBM Corp.
Friday did the same at the University of Texas for Austin’s advanced computing center (TACC). Financial terms were not disclosed.

While Big Blue offered its services to the European organization to help its scientists improve their weather forecasting prowess,
the giant has dipped into the institution of higher learning well to help the UT Center for Space Research (CSR) prepare for a
future manned landing on Mars.

The CSR will employ four eServer p690s (code-named Regatta) and two IBM Linux Clusters to improve the precision of interplanetary
navigation and will gauge the Martian environment. In a public statement, IBM said “CSR will use the
system to handle the monster number-crunching assignments” endemic to space travel.

“The raw floating point performance coupled with the excellent bandwidth of the IBM systems gives our researchers vast improvements
in performance,” said Dr. Jay Boisseau, director of TACC. “For science and engineering applications, bandwidth is often the key to
numerically intensive simulations running more accurately.”

Specifically, TACC’s systems include an eServer Cluster 1600 Unix system, comprised of four Regattas, as well as two eServer Cluster
1300 Linux systems, the first of which will be rolled out with an Intel IA64 cluster with 40 Itanium 800 MHz processors with 80GB of
memory. The second cluster will be an Intel IA32, based on 64 Pentium III 1 GHz processors with 32 GB of memory.

In addition to supercomputing, which employs parallel processing, the systems will also be used
to test a new computing
grids, which requires the use of software that can divide and farm out pieces of a program to as many as several thousand computers.
In this case, TACC will link one of the Linux clusters to IA32 clusters already in place to test applications.

While IBM has long maintained that grid computing will drastically change computing by enabling heterogeneous systems to share
resources over the Web, TACC is focused on developing mature software to harness a grid’s power.

The systems will be available to UT in January 2002.