1,024 Supeprcomputer Makes More Accurate Climate Assessments

NASA scientists can now evaluate the global impact of natural and
human-induced activities on climate and better predict probable
climate patterns in the future, thanks to the world’s first
1,024-processor supercomputer.

The newly installed 1,024-processor machine at NASA’s Ames Research
Center in California’s Silicon Valley, along with a 512-processor
supercomputer at NASA’s Goddard Space Flight Center, Greenbelt, Md.,
are producing a 10-fold improvement in Earth science applications.
Scientists say the performance gains achieved on these supercomputing
systems will allow the United States to develop objective policies
for large, future industrial activities, such as urban planning, and
for examining alternative plans for urban development. The
supercomputers — SGIª Originª 3800 machines — also can portray
current climate more quantitatively and simulate future global
warming scenarios.

“The new 1024-processor SGI Origin 3800 supercomputer at NASA Ames
will lead to faster and better development of climate models for the
Earth science community, government and industry,” said William
Feiereisen, chief of the NASA Advanced Supercomputing (NAS) division
at Ames. “We have improved our ability to merge observed data and
simulation by a factor of 10 with considerably greater increases in
the core climate solver. Such a substantial increase in performance
allows Earth scientists to complete climate simulations in days,
rather than months, leading to a better understanding of how human
activity has changed climate patterns.”

For the past three years, NASA Ames and SGI have been testing the
limits of single-system-image shared memory, in which all processors
share the supercomputer’s memory as if it were a single entity, to
improve performance significantly over other clustered architectures.
A series of joint research agreements between SGI and NASA Ames has
resulted in SGI expanding the original SGI Origin 2000 product
offering from 128 to 512 CPUs, and most recently from 512 to 1,024
CPUs for the SGI Origin 3000 product line.

“The new techniques have demonstrated a development path that will
allow us to move forward to100-times performance improvements over
the next few years,” said James Taft, co-director of the Advanced
Computing Technologies Group at Ames. “At these performance levels,
we can begin to execute climate simulations at truly high resolution,
while taking advantage of the huge data streams emerging from the
latest Earth resources satellites.”

To improve the prediction capabilities of the climate models, the
1,024-processor SGI Origin 3800 supercomputer at NASA Ames
assimilates thousands of gigabytes of observational data from the
whole Earth to create a database for verifying the physics of the
computer model. NASA Ames then backs up a few years and runs the
climate model to see how good its predictions are. The computer
models then can be adjusted to improve their accuracy for future
predictions.

A memorandum of agreement with NASA Ames placed a separate
512-processor SGI Origin 3800 supercomputer at NASA Goddard, which is
only the second site in the world to put an SGI Origin 3800 of this
type into production.

“This collaboration between Goddard and Ames to acquire the latest
supercomputing technology grants NASA scientists a significant new
capability for understanding the intricacies of our planet’s climate
system,” said Dr. Ghassem Asrar, Associate Administrator for the
Office of Earth Sciences, NASA Headquarters, Washington, D.C. “For
instance, the Goddard Institute for Space Studies has been able to
complete in two months research that would have taken six months on
their previous computing platform. This latest supercomputing
technology will grant NASA scientists a significant new capability
for understanding the intricacies of our planet’s climate system and
being able to simulate them,” Asrar added.

The primary user of the new SGI Origin 3800 supercomputer is
Goddard’s Data Assimilation Office (DAO), which is preparing for the
Aqua satellite by building NASA’s next-generation software for
incorporating observations into global climate models. Data
assimilation uses observations from satellites and other sources to
define the physical processes that make up weather and climate.

“With the SGI Origin 3800, NASA will more than double the amount of
data it ingests to 800,000 observations each day,” said Dr. Richard
B. Rood, DAO senior scientist and acting director of the NASA-NOAA
Joint Center for Satellite Data Assimilation. “We will also integrate
assimilation systems for several satellites so that, like the real
Earth, the impact of one type of data will be felt by another type of
data.”

The SGI Origin 3800’s processing power, along with the multi-level
parallelism (MLP) software developed by Taft, which takes advantage
of its unique memory design, will enable the DAO climate models to
run more than four times faster and at double the resolution. Climate
models divide the globe into a grid of stacked boxes, solving the
relevant equations in each box and then assembling the full results.
With a box only 1/2-degree wide (or 30 miles over the continental
United States), the model will more faithfully represent atmospheric
conditions worldwide for periods as long as 15 years.
“These advances will reduce uncertainties in the climate assessments
that are an essential ingredient of the U.S. Global Change Research
Program,” Rood noted.