Features:

REED: 'THE FUTURE CAN BE BRIGHT, GRAB YOUR SHADES'
by Tim Curns, Editor

Dr. Daniel A. Reed, Vice-Chancellor for Information Technology and Chief Information Officer for the University of North Carolina at Chapel Hill, has been a busy man this past year. Between serving as director of the Renaissance Computing Institute (RENCI) and chairing the computational science subcommittee of the President's Information Technology Advisory Committee (PITAC), Dr. Reed managed to squeeze in some time to speak with HPCwire about the current state of HPC and what will be happening at Supercomputing 2004 this year.

HPCwire: Dan, last year, you were named as one of HPCwire's Top People to Watch in 2004! What kinds of things have happened over the past year? What specific advances have you had a hand in?

Daniel Reed: I've been focusing on two major activities this year: (a) bootstrapping the new Renaissance Computing Institute (RENCI – see http://www.renci.org) and (b) serving as chair of the computational science subcommittee of the President's Information Technology Advisory Committee (PITAC). Each has been an exciting, full time job!

The Renaissance Computing Institute (RENCI) is a major new collaboration that combines the strengths of three universities (UNC, Duke, and NCSU) with the social, business and research opportunities of the Research Triangle and the State of North Carolina. Within that framework, we have launched a wide range of new projects and collaborations. For example, one NIH-funded biomedical collaboration, as part of the NIH Roadmap, is examining the genetic basis of disease, with a goal of coupling public health, clinical and model organism data via federated data models and access tools. Another NSF-funded, jointly involving UNC, Duke and NCSU, is creating the National Evolutionary Synthesis Center (NESC), a collaboratory for coordinating the diverse data and models associated with evolutionary biology research. A third example is our participation in the SURA Coastal Ocean Observing Program (SCOOP), which is creating a regional Grid of distributed sensors, data archives and models for ocean analysis.

With respect to PITAC, we have been engaged in fact finding, as a prelude to a formal response to the computational science charge (see http://www.itrd.gov/pitac/20040609_compsci_charge.pdf for the questions posed by the White House). At SC04, we will be holding a Birds of a Feather town hall meeting to solicit further community input to the subcommittee. I encourage SC04 attendees to participate in the town hall and respond to the subcommittee's preliminary findings on the importance of computational science, the need for sustained investment and the opportunities for greater coordination.

For both of these causes, I have been traveling extensively to carry the message to companies, universities, and agencies. It seems as if I've been in every major airport in North America this year!

HPCwire: Understandable! So in your most modest opinion, has your placement on our list been well deserved?

DR: I will leave that for others to decide. I'm focused on trying to make a difference, both in support for computing and its broad applications to critical problems.

HPCwire: Sounds like a busy year! Anything that has slowed you down?

DR: If anything, I have even busier than I was in Illinois [at NCSA]. I've been teaching myself a bit of biology, to allow me to talk to biomedical researchers about their computing needs and opportunities. That self- education has been both exhilarating and humbling – I've learned both the depth of my biological ignorance and gained insights into some of the scintillating discoveries in flight.

HPCwire: Please describe Renaissance Computing. What does this mean? How did Donna Cox, last year's scheduled SC keynote, coin the phrase? Why is it so important now?

DR: As new discoveries increasingly lie at the interstices of traditional disciplines, computing is also the enabler for a scholarship in the arts, humanities, creative practice and public policy. Computing has also become deeply embedded in the fabric of everyday life, with profound social implications. The term Renaissance Computing, coined by my friend and colleague Donna Cox, is intended to both evoke and capture the breadth of such intellectual activities enriching and empowering human potential, as well as creating intellectual communities that span the sciences and engineering, the arts, the humanities and commerce. Donna has been using the term for many years as part of her own vision for the future.

HPCwire: Are other institutions advancing Renaissance computing endeavors? Where do you see these initiatives being most prominent?

DR: Many groups are rallying behind this banner. The Humanities, Arts, Science and Technology Advanced Collaboratory (HASTAC) is one notable example. Emerging activities in bioethics, cybersecurity and privacy, and complex modeling are others. Anywhere people are grappling with complex problems, Renaissance computing is there.

HPCwire: You mentioned the Renaissance Computing Institute in North Carolina earlier. What else is happening there?

DR: We are aggressively hiring staff and researchers, so if someone is looking for an opportunity, we're interested in talking.

At a high level, RENCI is exploring the application of computing technologies to a broad range of disciplines, ranging from art to zoology. RENCI's mission is to explore the "out of the box" research opportunities that are the hallmark of true innovation. Leveraging the biological research strengths of UNC and the Research Triangle, RENCI targets a broad range of activities in agriculture, biology, bioinformatics, biomedicine and public health. However, its scientific focus spans the entire range of scientific and engineering research. In addition, the Institute will support new applications of computing technology to the arts and humanities, with a goal of enabling new creative expression (e.g., via digital media, music and the arts) as well as historical scholarship, curation and social debate.

HPCwire: Please comment on recent TeraGrid developments and your involvement in these. What happens next?

DR: I continue to be involved in TeraGrid activities, and I am evangelizing for broader application engagement and bringing the biomedical computing community to the TeraGrid. I view the TeraGrid as the vanguard of a new backbone infrastructure, around which other resources and infrastructure will accrete, just as ARPAnet and NSFNET anchored the Internet. Grids allow us to co-locate virtually the data archives, instruments, computing systems and people who cannot otherwise be co-located due to economics or geography.

HPCwire: Speaking of geography, now that you're in North Carolina, how has your life changed, both professionally and personally this past year?

DR: I'm a lot closer to the beach, though I haven't been able to spend much time there! In addition, being near a larger airport has made travel much simpler, especially reaching Washington.

HPCwire: Regarding Washington, you chaired the High-end Computing Revitalization Task Force (HECRTF) workshop and were a member of the President's Information Technology Advisory Committee (PITAC). How has Congress advanced High-End computing initiatives this year? What changes do you think still need to be made, or at least need to be concentrated on further? In other words, how are things progressing in this area?

DR: I have testified to Congress twice this year on high-end revitalization, once to the House Science Committee and a second time to the Senate Energy Natural Resources Committee. The HECRTF authorization bill was passed by both houses of Congress. In both hearings, and in my 2003 testimony to the House, I urged that we pursue a balanced and sustained investment program that supports exploration of new architectures, programming models and tools, algorithms and applications, as well as procurements and deployments in support of scientific opportunities and national needs.

I believe that "staying the course" is one of our biggest challenges. We tend to focus on initiatives that we believe can solve problems in a couple of years, yet many of the core problems of high-end computing (e.g., usability, software development complexity and achieved performance) are complex and will require many years of coordinated effort to solve. PITAC is also examining these same issues.

HPCwire: Do you think the apparent polarization of our country during the recent elections will have any effect on these initiatives?

DR: Support for scientific discovery has always had bipartisan support. Hence, I think as a community we must focus on articulating the clear and compelling benefits from investment in computational science in ways that are intelligible and understandable by decision makers and the public.

HPCwire: It's clear that the U.S. is trying to stay ahead of competition in regards to HPC. How have the recent BlueGene/L and Columbia announcements affected US supercomputing initiatives to stay ahead of competition?

DR: The IBM BlueGene/L and SGI/NASA Columbia news is indicative of the rapid ferment in the field. We all know that the Top500 numbers provide "bragging rights," but that the Linpack benchmark is not always representative of achieved performance on scientific applications.

The real measure of progress is our ability to solve important problems more rapidly and effectively. As our problems become more complex, we must focus on enabling multidisciplinary teams to work together more effectively and productively. Integration of instrumentation and sensor data, as well as a true assessment of the time and cost to solution, including the human factors, must rise to the forefront. We can no longer focus just on the cost of the hardware or its peak performance.

HPCwire: How is high-performance computing different from other scientific mediums?

DR: One aspect of computing distinguishes it from other scientific instruments – its universality as an intellectual amplifier. Powerful new telescopes advance astronomy, but not materials science. Powerful new particle accelerators advance high energy physics, but not genetics. In contrast, computing advances all of science and engineering, because all disciplines benefit from high- resolution model predictions, theoretical validations and experimental data analysis.

HPCwire: Please outline your speech this year at SC04 entitled "Computing - An Intellectual Lever for Multidisciplinary Discovery." Where do you see art and technology merging most prominently in 2005?

DR: I believe the challenges of the 21st century are interdisciplinary. I recently asked a group to describe what happened in New York City on 9/11. They gave a technical answer about steel deformation under thermal stress and concomitant building collapse. Although narrowly true, the answer didn't capture the larger issues of the impact on financial markets, transportation systems, communication networks, and social structures. Building and validating such computational models, based on complex, multidisciplinary data, is one of our great 21st century challenges.

Similarly, our burgeoning knowledge of the genetic basis for disease raises profound ethical, social and public policy issues. Do you want to know if you are likely to develop Alzheimer's disease? Do you want your insurance company to know? How do we plan health care policies in such a world? As the rate of knowledge discovery accelerates, the social challenge is creating mechanisms that allow us to collectively debate and determine the appropriate and ethical use of the new technologies.

I also believe the arts, humanities, and sciences are all manifestations of the same creative impulses; only the media of expression differ. Understanding this continuum is our challenge and opportunity. My SC04 talk will describe emerging opportunities in the arts, humanities, science and engineering where interdisciplinary approaches can have profound impact on discovery and creative expression. I will also touch on some of our social challenges in capitalizing on these opportunities.

HPCwire: How will these advances change the HPC community?

DR: Community building and creating interdisciplinary reward structures will reshape and broaden participation. We need the insights from a broad and diverse community, as our new challenges touch the entire fabric of modern society.

HPCwire: What do you think will be the most impressive thing at SC04? Why?

DR: We're in the midst of a sea change, from data poor to data rich computational science. The explosive growth of new sensors and instrumentation, in domains as diverse as biology, environmental science and astronomy, means that petabytes must become as much a part of our lexicon as petaflops. Indexing, mining and federation will be central to our success, as storage capacity is rising far more rapidly than storage bandwidth.

I believe we are at the beginning of a phase transition in computing capability. We are near the point where we have the potential to model (accurately) a wide area of complex phenomena at extraordinary resolution. Whether these be in silico models of cell activity or the simulations of the "universe in a box," the future will be amazing. I would love to see us embrace one of these big visions, as a driver for the future of computational science. Working together, the future can be bright – grab your shades.

HPCwire: Exciting stuff, Dan! Thanks again for making some time to speak with us.

Dr. Reed is listed in HPCwire's "Top People to Watch" (http://www.tgc.com/hpcwire/features/topwatch04.html), and it is obvious that he has not slowed down at all over the past year! Catch Dr. Reed's presentation, ""Computing - An Intellectual Lever for Multidisciplinary Discovery" at SC2004 on November 11. The presentation will be held in Ballroom B-C from 9:15 to 10:00am.