Nav Button - Home

This issue...

  Brieflies

  View from the Inside

  Discovery of New Genes...

  Friendlier Gases...

  Working Science

  People

  About

  Subscribe Free





























This issue...

  Brieflies

  View from the Inside

  Discovery of New Genes...

  Friendlier Gases...

  Working Science

  People

  About

  Subscribe Free





























This issue...

  Brieflies

  View from the Inside

  Discovery of New Genes...

  Friendlier Gases...

  Working Science

  People

  About

  Subscribe Free





























This issue...

  Brieflies

  View from the Inside

  Discovery of New Genes...

  Friendlier Gases...

  Working Science

  People

  About

  Subscribe Free





























This issue...

  Brieflies

  View from the Inside

  Discovery of New Genes...

  Friendlier Gases...

  Working Science

  People

  About

  Subscribe Free

Advanced Scientific Computing—the Mover and Shaker of Scientific Research

by Nona Shepard

C. Edward OliverIn an interview with C. Edward Oliver, new Director of the Office of Advanced Scientific Computing Research, he points to the use of computer simulations throughout the Office of Science's Strategic Plan as evidence that advanced scientific computing has become a major enabler for scientific research.



ES News: The Office of Computational Sciences and Technical Research has been renamed the Office of Advanced Scientific Computing Research. What new message are you sending by changing the name of your Office? And, since scientists in every discipline already do their work by computer simulation, why the new emphasis on advanced computing now?

Oliver: We want to emphasize advanced computational research as a basic science.

It's always worth emphasizing that computing is very important to the Department of Energy, and it always has been; ever since the department was born, back in the AEC [Atomic Energy Commission] days. It's probably easy to verify that DOE has done more with computing than any other agency. When a manufacturer builds the fastest machine, DOE is the first in the cue to buy it. Then, we make the computer work to suit our specific scientific needs. That's just the flat fact-people who are familiar with DOE know that, but people outside DOE don't.

And we've made many innovations with the computer that people don't know about-if the computer needed software to make it work, we generated the software; if it needed a change in the algorithms to make it work, we funded the activity that made that happen. Then, the scientists used it to get their results.

It's exciting to the scientific community to see DOE work with these machines. When they see what we can do, universities want them, and other parts of the federal government are interested. So, we are the leaders not only in terms of actual procurement but also in the spirit of computing. Our "essence" is science, not computing—if we didn't want to do the science, we wouldn't be buying the machines.

ES News: What enticed you to leave Oak Ridge National Laboratory (ORNL), where you were already doing interesting things.

Oliver: Well, Martha Krebs [former Director of the Office of Science] rang me back in January or February—she and Ernie Moniz [Under Secretary of Energy] wanted to make simulation science a more important part of the department than it has been. If you look through the Office of Science's Strategic Plan, what do you see? Look at all the computer simulations—they are everywhere. So how important is computing to the essence of the Office of Science? When you get that level of commitment—an undersecretary and an assistant secretary—to support something new, it's pretty hard to say no.

ES News: What did they have you doing when you first arrived?

Oliver: I came to help with strategic planning. We worked long and hard, having many meetings, bringing hundreds—probably a thousand-people together to look at what the department could do with simulation. We looked to see where the sweet spots were, to find areas where we could make major gains if we had an infusion of new money; what we could do if we had better facilities than we presently have—and we have some terrific facilities in computing.

ES News: During that time, didn't PITAC (the Presidential Information Technology Advisory Committee) issue a report that became a Presidential initiative, affecting all the research agencies, including the Energy Department?

Oliver: Yes, after that we went back to the drawing board to see how to join with other agencies to make it an important interagency initiative. In particular, we joined with NSF [National Science Foundation] because we had things in common and already did some things together.

ES News: You've had enthusiastic support from all of the Associate Directors within the Office of Science, haven't you?

Oliver: Yes, because simulation is an important tool to the ways they all do business. Advanced Scientific Computing has got threads throughout the entire Office of Science; and throughout the Department as a whole. For example, we work with Defense Programs and we're getting more involved with the people in Energy Efficiency and Fossil Energy to solve computational problems—it's really a Department-wide initiative.

So, it makes it a fun job to come to because you aren't fighting any kind of internal battles inside your Department. It only becomes a matter of convincing OMB [Office of Management and Budget] and then trying to win over our appropriations committee in Congress. That's where we didn't succeed last year. But, we are optimistic that we can make simulation important in the 2001 budget.

ES News: What are some of the workshops you've held?

Oliver: We've had workshops on predictability of nonlinear problems, on optical interconnect technology, on combustion simulation, and many other topics.

For example, the workshop on Optical interconnection brought together researchers and managers from government, industry, and academe, including the Defense Advanced Research Projects Agency, National Aeronautics and Space Administration, National Security Agency, Air Force, Lawrence Livermore National Laboratory, Sandia National Laboratory, Oak Ridge National Laboratory, Allied Signal, Sematech, Lucent, Semiconductor Research Corp., and Irvine Sensors, MIT, Rensselaer, Georgia Tech., and the Universities of Arizona, Alabama, and New York State. It was a very interesting workshop, as you can imagine.

ES News: What are some of the things you're planning to do next?

Oliver: We're generating a strategic plan for computing for the whole Office of Science, not just for this office that I run. Writing a strategic plan is new for us! And gathering the information that goes into the strategic plan is a lot of fun. The beauty is that we are involving all the computational people in the Office of Science in writing this plan and we want to make it readable and understandable so that…well, it's going to be more than a guidebook for us. It's going to help explain to people outside of DOE and certainly outside the Office of Science, where we are going and how important we think computing and simulation is to our business of science.

Where do we intend to go in five years? That's what we are going to try to put into a document that will be as easy to read as the Strategic Plan for the Office of Science. We will put it on the web. We're going to keep it alive; we're going to have the program managers in biology, chemistry, physics, materials—in all the disciplines—working together all the time, looking at new technology. We'll go to the web and modify it. We can have a true, up-to-date plan that anyone can read. That is an important thing on my agenda for the moment.

Also, all of the offices have advisory committees, and we are about to put in place an advisory committee for Advanced Scientific Computing.

ES News: At ORNL, you were also involved in education. What are your thoughts about computational education?

Oliver: Well, DOE is a leader in computational science—of that, there is no doubt. And I think we will have a tremendous influence on some choices that universities may make. However, universities have a lot of inertia; it's difficult to change things. But, it wasn't that long ago that there were no computer science departments in any university in the country because computing just wasn't that important. But look at it today—computer science has blossomed in various and sundry colleges, sometimes it's in [the school of] Arts and Sciences, sometimes it's in Engineering. Wherever that flower was allowed to bloom, it popped up. Hopefully, computational science will do the same kind of thing. Maybe it will grow in physics in one university; it might grow in some part of engi-neering in another university; or in chemistry; or in biology. Biology has been the sleeping giant for the use of computers.

We would like to see degrees in computational science somewhat different than students have gotten in the past. That is to say, we'd like to see physicists and chemists and biologists take more courses in computer science and mathematics; and we'd like to see an actual degree in computational chemistry, for example.

We actually have a computational science graduate fellowship program that this Office funds that does in fact require the students at the graduate level to take a specified number of computer science courses and math courses as it leads to their degree in science and engineering. It's been a very successful program, and we'd like to see that program grow. And we'd like to see universities take up the opportunity to offer degrees spelled out in meaningful ways in their diploma.

ES News: People are getting very excited over computers and technology and ever faster computations.

Oliver: You ain't seen nothing yet.

ES News: You've been here such a short time, but have already said you are having fun.

Oliver: That really is an operative word, you know—fun. Some of the people on the advisory committee asked, "Am I going to have fun if I join this committee?" You do have to have fun, and that means you've got to let people have interesting problems.

The number one fun thing is that DOE funds the very best people in the country in computer science, applied mathematics, and many areas of the computational science arena. It is at the forefront of the infrastructure of the business, buying the very best machines—NERSC just procured a 3.6 teraflop machine that will come online about 12 months from now—and having very good networks to allow the scientists to get on them. That goes back 50 years and the Department still has a tremendous reputation among all the agencies. So it is fun and exciting getting together with all these top people in the country and tracking what they're doing.

Having the support from the bottom to the top makes it exciting and fun—because few things are worse than having a contentious situation with your own people since you need energy to work the outside part of the funding game. That amount and level of support is truly unusual: Peña [past Secretary of Energy] supported the simulation game; Betsy Moler [former Deputy Secretary] was very supportive, and Secretary Richardson is supportive; that must mean we're onto something.

ES News: We'll look forward to hearing more about that "something" in the near future. Thanks Ed.


Related News about Advanced Scientific Computing:

New! PrePRINT Network

PubSCIENCE—New Desktop Access to Journals

Supercomputers Help Solve a 50-Year-Old Quantum Physics Problem

Contents
Search
Comments
previous    next



Oliver officially became the new Associate Director of the Office of Science's Office of Advanced Scientific Computing Research on August 15, 1999. He joins SC from the Oak Ridge National Laboratory where he was the Associate Laboratory Director for Computing, Robotics, and Education. Oliver brings knowledge and expertise of the computing environment from the Federal and laboratory perspectives.

   previous     next


• Energy Science News • Energy Science News • Energy Science News • Energy Science News •
www.pnl.gov/energyscience/