Features:

INTERVIEW WITH BARBARA HORNER-MILLER, ARSC
By Tim Curns, Assistant Editor, HPCwire

HPCwire: What are the long term goals and objectives of the Arctic Region Supercomputing Center? What obstacles have stood in the way of these goals? What's new with the ARSC? Any major breaking news to announce?

BARBARA HORNER-MILLER: ARSC wants to continue to support research of national significance that requires the use of high performance computers while allowing development of computational models and research teams that tie together academic and federal agency interests, especially as they pertain to the Arctic.

We have been extremely fortunate in hiring excellent staff and bringing leading supercomputing technology to ARSC. ARSC is celebrating its tenth year with a major technology refresh. In late September, we completed the installation and acceptance of a 128-processor Cray X1. Currently, we are proceeding with the installation of a large IBM system consisting of P655+ and P690+ servers that will be tied together with early Federation technology. We expect this system to be ready for users after the beginning of the year.

In addition, we installed a Mechdyne Flying Flex immersive environment powered by an SGI Onyx system in the spring and are in the process of releasing a new storage system from Sun. As you can see, this has been a busy year and has provided the center with real expansion of the computational resources and growth opportunities for the staff.

Current scientific thrust areas for the center include tsunami research, ocean modeling, space physics and bioinformatics. Researchers at the University of Alaska are working with Alaska disaster planners in the areas of tsunami runup and inundation. Several ocean modeling efforts of interest are currently underway. In the SALMON project, ARSC and UAF oceanographers, Kate Hedstrom and Dave Musgrave, and their colleagues are combining observational data and computer modeling to create an ocean and ecosystem-coupled model on ARSC computers; this effort may someday allow marine biologists to estimate fluctuations in fish, bird and marine mammal populations in the Gulf of Alaska. Another project, sponsored by the Department of Defense and headed by Wieslaw Maslowski of the Naval Postgraduate School, suggests that major changes in ice/ocean circulation have occurred over the last two decades. You can learn more about these projects and other research underway on ARSC systems, in the latest issue of our magazine Challenges in Science and Engineering. Copies are available at our booth on the show floor.

Some very interesting work is being done by researchers in ARSC and the Geophysical Institute. This work demonstrates the synergy that ARSC feels it brings to the table as a combination university and DoD center. And bringing this up now is especially timely due to the high solar flare activity of the recent weeks. Several years ago, Dr. Sergei Maurits, then a Ph.D. student, and his advisor Dr. Brenton Watkins, began work on a Eulerian Polar Ionospheric model. The model was later parallelized for the ARSC T3E. This model is used for operational, real-time forecasting of space weather as well as for research objectives. About a year ago, the Department of Defense became interested in the model as it was higher resolution than other known ionospheric models. The 3-D, time-dependent model, combined with ray-tracing techniques, is being utilized to delineate the effects of solar activity on ionospheric radio wave propagation, taking into effect the perturbations introduced by solar activity, geomagnetic activity and polar ionosphere dynamics induced by solar wind variability.

HPC: How has the relationship between supercomputing and the Arctic region affected breakthroughs in the field?

BHM: The University of Alaska-Fairbanks is home to several world class research organizations including the Geophysical Institute, the International Arctic Research Center and the Institute of Marine Science. These institutes provide the research teams that tackle large problems like those described above. The Arctic is a bellwether area in that what will eventually happen to the climate and the environment throughout the world will happen first in the Arctic. This, coupled with the fact that most global models gloss over the Arctic, provides a lot of incentive for us to support researchers who are focusing on this region. These are the researchers who will bring about future breakthroughs in climate, ocean and atmospheric science.

HPC: What contributions has your research made to the country as a whole? Please comment on the impact your work has had on defense projects, etc.

BHM: At ARSC, we provide significant resources to the Department of Defense High Performance Computing Modernization Program. As an unclassified center, ARSC contributes greatly to the 6.1 and 6.2 research in the DoD. We provide significant resources each year to DoD Challenge projects that support specific goals within the DoD community. In addition to ocean modeling projects such as the work by Dr. Maslowski, we support DoD researchers who are designing submarines and developing new fuels for propulsion systems and vehicles in the U.S. Air Force space program. We also support researchers working in the areas of nanotechnology and bioinformatics.

HPC: Please comment on your involvement with CUG (Cray User Group). What kind of advancements in high performance computing can be partly or wholly attributed to ARSC's participation in CUG?

BHM: I have been involved with CUG at a leadership level for about 10 years. CUG is an independent corporation, not an appendage of the vendor. The organization has struggled for the last couple of years as has the industry but both now seem to be on firmer footing. I feel that I, along with other members of the CUG Board of Directors, helped to keep the CUG together during very stressful times. CUG in turn, has supported the industry.

Any user group has two main functions from the user's point of view: to give and take support from its fellow members and to influence the vendor to act in ways that benefit the users (and hopefully the vendor at the same time). I am really glad to say that CUG is still in a position to do both.

One obvious impact that CUG has had on the industry is that it serves as a role model for user groups. It is my understanding, based on comments from vendors and users alike, is that several other user group have patterned their organizations after CUG.

Over the years, CUG has had a major effect on the direction that Cray has taken in both hardware and software directions. I have been in meetings where the membership was able to change the direction that Cray set. This might have been as simple as raising the priority of a new feature or it might send them back to the drawing board for the implementation of a product. Cray's first design of an on-line documentation system was totally revamped after an especially vocal session within the user services special interest group. Similarly, the operating system special interest group ordered Cray priorities for new features and upgrades through the "wish list" developed by that group. This process was responsible for the addition of checkpoint/restart, hardware performance monitoring and more.

The bottom line is that an organization like CUG, working with a vendor, in this case Cray, can and has made a major impact on the usability of the vendor's products. This is not only of significant benefit to the company and the users, it raises the bar across the entire industry.

HPC: Finally, please feel free to comment on anything else you wish to include here.

BHM: I'd like to thank you for the opportunity to share my opinions with your readers and to say that we really enjoy your coverage of the supercomputing news and community. We're happy to see so much attention being given to supercomputing at the highest levels of the nation as evidenced by the multitude of current studies and reports. Several sessions as SC2003 will focus on the recent release of these studies and the reaction of different facets of the community to them. I hope that many of your readers are in Phoenix and have the opportunity to hear directly from some of the principals who have spent considerable time this year creating these reports of national and international interest.

Thanks again.