| Drucella Andersen
Headquarters, Washington, D.C. October 21, 1992
Paula Cleggett-Haleim
Headquarters, Washington, D.C.
RELEASE: 92-177
NASA has chosen 29 supercomputing research proposals
that will pave the way for a revolution in the way scientists
study the Earth and space.
The goal is to achieve computational capabilities far
beyond those of today's machines, enabling researchers to
produce realistic simulations of phenomena such as the
interaction of Earth's oceans, air and land masses and
reconstructions of the evolution of stars, galaxies and the
universe.
"This is the first time a broad spectrum of the
country's best and brightest people will engage in leading-
edge supercomputing research for Earth and space science
applications," said Lee Holcomb, Director for High
Performance Computing and Communications (HPCC) at NASA
Headquarters, Washington, D.C. "We deliberately selected a
diversity of institutions, applications and computing
platforms to get the best return on our investment."
Eight of the projects will involve teams of space and
computer scientists to create computer models and computing
techniques that can attack today's intensive, large-scale
problems in Earth and space science -- the so-called "Grand
Challenges." The other projects are focused efforts that
address specific issues in algorithms and computational
methods as they apply to Earth and space studies.
The proposals selected came from a wide variety of
sources. The investigators and collaborators represent 27
academic institutions, 9 federal laboratories and 3 private
firms.
The research will be done under the Earth and Space
Sciences project of the HPCC initiative, a leading-edge
effort to produce a 1000-fold increase in supercomputer speed
and a 100-fold improvement in communications capability by
1996.
The 8 large projects will begin early in calendar year
1993. The others will start in two phases: 10 early in 1993
and 11 in the last quarter of that year. Total funding for
the efforts is estimated at $10.5 million through fiscal year
1995.
"Combining computer science and technology with space
science gives us the opportunity to bring the two communities
together and for computer scientists to take an increased
role in helping us meet our research goals," said Joe
Alexander, Assistant Associate Administrator for Space
Science and Applications at NASA Headquarters.
WHY HIGH PERFORMANCE COMPUTING?
These projects will use "parallel processing" computers-
-machines that use up to thousands of processors to work
simultaneously on a problem. These processors are simpler
than those in traditional supercomputers but they allow
faster computation since there are many more of them.
The advances from these projects will be vital to Earth
and space scientists because current supercomputers are
neither fast nor powerful enough to digest the volume of
data expected in the 1990s or to model the full complexity of
many challenging scientific problems.
NASA's Goddard Space Flight Center, Greenbelt, Md.,
manages the HPCC Earth and Space Sciences project with
assistance from the Jet Propulsion Laboratory, Pasadena,
Calif., for NASA's Office of Aeronautics and Space Technology
and Office of Space Science and Applications.
The winning proposals broken down by state are:
Maryland, 6; California, 4; Illinois, 4; Texas, 4; Arizona,
2; and one each from the District of Columbia, Hawaii, Iowa,
Louisiana, Massachusetts, Michigan, New Mexico, Utah and
Washington.
|
| HQ 92-195/COMPUTER CLUSTER
Drucella Andersen
Headquarters, Washington, D.C. November 5, 1992
Mary Ann Peto
Lewis Research Center, Cleveland
RELEASE: 92-195
An array of powerful, interconnected computers will help NASA solve
some of the complex problems associated with designing aircraft propulsion
systems.
The Lewis Advanced Cluster Environment (LACE) system at NASA's Lewis
Research Center, Cleveland, is a network of 33 IBM RISC/6000, stand-alone
computer workstations that can be set up for both distributed and parallel
computing. LACE can perform up to 1 billion floating point operations per
second and has 3 billion bytes of memory, making it the world's most powerful
IBM workstation cluster.
"This is the first system of its kind for NASA," said Karen Pischel,
LACE Project Manager at Lewis. "By applying the power of a network of high-
performance computers, we are creating new, more efficient and more economical
ways to perform computational research."
LACE is a test facility for NASA's role in the multi-agency High
Performance Computing and Communications program. This high-priority national
effort fosters the development of advanced computing techniques by tackling
"grand challenges" that eventually will require one trillion floating point
operations per second.
One of the program's grand challenges involves the complete simulation
of a supersonic aircraft in flight. In a key LACE project, researchers are
modeling the performance of a supersonic throughflow fan engine -- an engine in
which the airflow remains supersonic throughout -- by distributing the various
engine components among the LACE computers.
|
| Drucella Andersen
Headquarters, Washington, D.C. November 16, 1992
Paula Cleggett-Haleim
Headquarters, Washington, D.C.
Franklin O'Donnell
Jet Propulsion Laboratory, Pasadena, Calif.
RELEASE: 92-205
NASA is collaborating with Cray Research, Inc., Eagan, Minn., to
conduct joint research and development activities using the company's most
powerful supercomputer.
The partnership is a NASA response to the multi-agency High Performance
Computing and Communications program, a bold national initiative to advance
U.S. capabilities in supercomputing.
The computer will be located at NASA's Jet Propulsion Laboratory (JPL),
Pasadena, Calif., with the California Institute of Technology, Pasadena, as a
collaborator in the project.
"The Cray T3D system, together with our expertise in parallel
processing, will allow us to tackle new computational problems in Earth and
space sciences," said Dr. Carl Kukkonen, Supercomputing Project Manager at JPL.
"More importantly, we will be able to feed back JPL's and Caltech's
experiences to Cray and thus, contribute to maintaining U.S. leadership in
supercomputing," Kukkonen added.
"This collaboration will help us address the grand challenges of Earth
and space sciences -- analyzing the enormous data sets from NASA's Earth and
planetary missions," said Joseph Bredekamp of the Office of Space Science and
Applications, NASA Headquarters. "We expect that 20 to 25 percent of our
scientific computing will be performed on parallel computers within 3 years."
NASA Applications
JPL will use the new Cray T3D system for applications that require
high-power computers -- turning planetary data from spacecraft into
three-dimensional animations; electromagnetic simulations for the design of
communications antennas; analyzing Earth satellite data; studying the dynamics
of chemical reactions and the flow of space plasmas; and computational fluid
dynamics.
"NASA is committed to being an early user of new parallel
supercomputers, and Cray will be an important player," said Lee Holcomb,
Director for the High Performance Computing and Communications program at NASA
Headquarters.
In the fall of 1993, JPL will receive a Cray T3D system, the company's
new "massively parallel" supercomputer. The innovative machine uses 256
processors to reach a peak speed of 38 billion floating point mathematical
operations per second, making it Cray's most powerful computer. The effort
involves JPL, Cray and the California Institute of Technology (Caltech),
Pasadena.
Center Of Excellence
Cray has designated JPL/Caltech as one of four "Cray Centers of
Excellence in Parallel Computing." Under this program, the company will locate
staff engineers at JPL to carry out joint research in parallel computing
techniques.
Cray President John Carlson said the company chose JPL as a site for
its Center of Excellence program "because of the great expertise developed at
JPL and the Caltech campus in parallel computation."
The project is funded by NASA's Office of Aeronautics and Office of
Space Science and Applications, Washington, D.C.
|