12/13/00 -- Latest NAS Technical Reports Online
Take a look four recent NAS Technical Reports detailing the division's research activities.

RISC Performance Improvements To the LAURA Code by Robert J. Bergeron, presents performance data resulting from simple algorithm modifications to an efficiently vectorized NAS workload code.

Preproduction Performance of the SGI Origins on the NAS Workload, also by Bergeron, reports floating-point and related performance measurements for SGI Origin machines executing the NAS workload over a one-year preproduction period.

On Animating 2D Velocity Fields, by David Kao and Alex Pang, looks at several techniques for animating velocity fields and proposes two new alternatives.

The Design of a Templated C++ Small Vector Class for Numerical Computing, by Patrick J. Moran, describes the design and implementation of a templated C++ class for vector length and component type, and reviews some essential template features.

You can find other NAS Technical Reports for 2000, 1999, and 1998. For prior years, go to the NAS Technical Report Archive.

12/5/00 -- NAS Scientist Named to Editorial Post
NAS Systems Division senior scientist Cun-Zheng Ning has been named associate editor of the Journal of Quantum Electronics (JQE), the monthly journal of the IEEE-LEOS (Lasers and Electro-Optics Society). The IEEE-LEOS Board of Governors unanimously approved the appointment at a recent meeting.

During his three-year tenure, which begins in January 2001, Ning's duties will include identifying reviewers and making decisions regarding acceptance of submitted papers. He has published extensively in the fields of optoelectronics, lasers, and nonlinear physics, and has served as conference chair or member of technical committees in the field of optoelectronics at many international conferences.

Ning is task lead for device modeling projects within the NAS division. He obtained his Ph.D. in theoretical physics at the University of Stuttgart, Germany. He was a research assistant professor at University of Arizona before joining NAS. Ning is employed by Computer Sciences Corp.

IEEE-LEOS is considered to be the one of the most influential organizations worldwide in the fields of lasers, optoelectronics, and optics sciences and engineering. Since inception in the early '60s, JQE has consistently received one of the highest citation ratings of any IEEE journal.

For more information on IEEE-LEOS, see http://www.ileos.org/. For the current JQE table of contents, see http://www.ieee.org/organizations/pubs/pub_preview/qe_toc.html.

11/30/00 -- New Version of PEGASUS Code Released
A public beta release of the PEGASUS code (version 5.1) in now available. PEGASUS is a key component in a series of tool used in computational fluid dynamics (CFD) analysis. Version 5.1 significantly reduces user effort and allows for easier operation of the code.

The PEGASUS code is responsible for connecting overset volume grids and preparing them for the OVERFLOW flow solver. The new version dramatically reduces the time required to perform the overset connection. The result is a process which is automatic for many flow analysis problems.

The PEGASUS code, developed at NASA Ames Research Center in collaboration with Microcraft Corp., is used by both government and industry users. The code is used in conjunction with the Chimera Grid Tools (used for grid generation) and OVERFLOW. A parallel version of PEGASUS is in the works.

Get more information on PEGASUS, including a software request form and user's manual. Contact NAS researcher Stuart Rogers at rogers@nas.nasa.gov for additional details.

11/21/00 -- New NASA Grant Reports Online
Investigators working on two related research projects funded by NASA grants have recently published several papers. The papers are available online in various formats.

Scalable Numerical Algorithms on Non-heterogeneous Computing Environments with Dynamic Adaptive Routing and Load Balancing. Principal Investigator: Jinchao Xu, Pennsylvania State University. The research team is investigating scalable numerical algorithms, primarily multigrid and domain decomposition algorithms.

Scalable Numerical Algorithms on Non-Heterogeneous Computing Environments with Dynamic Routing and Load Balancing. Principal Investigator: Tony F. Chan, University of California, Los Angeles. In this related project, particular attention is given to the design and implementation of scalable algorithms in the presence of dynamic adaptive routing and load balancing.

For more information on NAS-related research grants, contact Timothy Barth at barth@nas.nasa.gov.

11/14/00 --New NAS Research Branch Chief Appointed
NAS Systems Division Chief Bill Feiereisen announced recently that Dr. Bryan Biegel has been appointed chief of the division's Research Branch. The branch consists of some 50 programmers and scientists whose work encompasses computer science research and development for aerospace applications, NASA's Information Power Grid (IPG), and nanotechnology, among other areas.

"The challenges and possibilities of the Research Branch Chief position are compelling to me," Biegel said. His first order of business is to ensure that the branch technical staff can work with minimal bureaucratic "red tape," and that the staff's talents and expertise are fully utilized and acknowledged. "I'm very excited about helping everyone in the branch to work happily and efficiently, so that we will in turn help NAS continue to do great things in high performance computing."

Biegel has worked in the NAS organization for more than three years as a researcher in the semiconductor device modeling group. Before accepting the branch chief position, he was employed by Computer Sciences Corp. under a NASA contract. His most recent research projects have focused quantum effects in ultra-small conventional electronic devices, thermal and radiation effects in space electronics, and electronic device modeling in the IPG environment.

Prior to joining the NAS division in 1997, Biegel was a research assistant at Stanford University, where he developed SQUADS, the Stanford Quantum Device Simulator, an automated, extensible software package for accurately modeling one-dimensional quantum electronic devices. Biegel holds both a bachelor of science degree and master's degree in electrical engineering from the University of Oklahoma, and earned his Ph.D. in electrical engineering from Stanford.

For more information on NAS research projects, contact Biegel at biegel@nas.nasa.gov.

11/3/00 -- Scientists Show Latest Technologies at SC2000
NAS Systems Division researchers will be on hand to demonstrate molecular simulations and innovative tools at the SC2000 High Performance Computing and Networking Conference in Dallas, November 4-10.

One demonstration will show how NASA Ames scientists have used supercomputer simulations to help improve the DeBakey heart assist device, leading to human trials with patients awaiting heart transplants. Ames researchers join colleagues from four other NASA centers for the annual show. For more information on Ames demonstrations, contact Jill Dunbar at jdunbar@mail.arc.nasa.gov.

10/31/00 -- New Website Highlights NASA Images, Research
A new NASA website showcases the agency's latest innovations in supercomputing and networking technologies. Five NASA centers will demonstrate work featured on the website at the SC2000 Conference, November 4-10 in Dallas.

NASA's SC2000 website, produced by team members in the NAS Systems Division, contains extraordinary high-resolution images to download, plus descriptions of all demonstrations to be shown in the NASA booth.

A key demonstration will show how scientists in the NAS division combined computational fluid dynamics techniques and high-speed computers to help improve the DeBakey heart assist pump. The pump extends the lives of patients awaiting heart transplants.

The website also includes a schedule of technical papers and panel discussions by NASA researchers at SC2000, a guide to finding demos in NASA's 50-by 50-foot booth, and a map of the conference exhibit floor.

For complete details, go to www.nas.nasa.gov/SC2000.

10/17/00 -- NASA IPG Team Reaches Another Major Milestone
On September 30, NASA's Information Power Grid (IPG) team demonstrated a complete version of a grid system, reaching another major project milestone. To demonstrate high-speed data access on the grid's infrastructure, the group employed the IPG Virtual Laboratory (ILab) program used for managing parameter studies. They also utilized the program manager Condor, which provides access to idle workstations, adding more computing power for executing jobs on the grid.

The IPG is designed to take a large collection of dispersed and heterogeneous resources -- computing systems, storage systems, and scientific instruments -- and define a standard set of services for accessing those resources for research. NASA's Ames, Glenn, and Langley Research Centers are collaborating to develop infrastructure for the grid.

ILab, a problem-solving tool designed by NAS researchers Maurice Yarrow and Karen McCann, manages parameter studies in the IPG environment. Running ILab on the grid dramatically reduces turnaround time for complex computations. The system was recently used to manage a parameter sweep study on the X-38 Crew Return Vehicle across a collection of resources using the new IPG infrastructure. The ILab X-38 test demonstrates uniform access to different IPG resources.

The IPG's uniform interface for batch queuing systems, the Globus middleware toolkit (developed at Argonne National Laboratory), was recently integrated into the Condor system manager. Condor utilizes Globus resources to provide more computing power by scavenging processor hours from a pool of 60 -100 Sun and SGI workstations. "I think the Condor project is part and parcel of getting the IPG infrastructure integrated into NAS," says IPG Project Manager Bill Johnston. NAS researcher Al Globus has been using Condor to run a collection of genetic algorithms, exploring different molecular structures. In roughly one year, Globus has accumulated nearly half a million CPU hours for this single problem using Condor.

Acknowledging the teamwork involved in meeting another major milestone, Leigh Ann Tanner, IPG deployment and integration project manager says, "There are lots of pieces to the IPG, and until you take a closer look, it seems like a trivial process to assemble the infrastructure. It takes a lot of work and effort. I would like to thank not only the IPG testbed group, but everyone in the NAS division -- as well as our outside collaborators -- for meeting this milestone." For more information on the IPG, contact Tanner at tanner@nas.nasa.gov.

10/3/00 -- IPG Workshop Builds Partnership
On September 19 and 20, scientists and engineers from U.S. government, industry, and academic organizations participated in the second Information Power Grid (IPG) Workshop at Ames Research Center, lead site for NASA's IPG project. Hosted by the NAS Systems Division, the technical workshop marks two years of progress on the IPG, designed to connect remotely located computational resources, data storage facilities, and scientific instruments for collaborative problem solving.

Workshop participants had the opportunity to review IPG development, demonstrate accomplishments, address issues, and share ideas with colleagues on the future of the grid. NAS Division Chief Bill Feiereisen welcomed the group, emphasizing the importance of collaboration to advancing grid technology. "We are a partnership, we are trying to build a grid together. This is a collaborative effort that crosses organizational boundaries," he said, adding, "This partnership will build the infrastructure upon which science and engineering computing of the future will be based."

Presentation topics ranged widely and included: infrastructure for monitoring and managing grids; user interfaces; security; job scheduling; debuggers for computational grid applications. NAS researcher M. Jahed Djomehri presented "Large-Scale OVERFLOW Applications and Performance Assessment," an investigation of a technique that will enable scientific computing of large-scale problems in a wide-area computational grid environment.

Argonne National Laboratory's Ian Foster, who co-authored "The Grid: Blueprint for a New Computing Infrastructure" was especially pleased with the meeting, "I really enjoyed the workshop. NASA is to be congratulated for putting together a meeting that not only showcased the progress made on the IPG to date but also provided a wonderful forum for the exchange of ideas within the burgeoning grid community."

Get presentation abstracts and information about the IPG Workshop. Online proceedings will be available the week of October 9. For more information on NASA's IPG project, contact Arsi Vaziri, deputy project manager, at vaziri@nas.nasa.gov.