BESAC Subcommittee on Theory and Computation

1
BESAC Subcommittee on Theory and Computation

• Co-Chairs
• Bruce Harmon Ames Lab and Iowa State University
• Kate Kirby ITAMP, Harvard Smithsonian Center
  for Astrophysics
• Bill McCurdy University of California, Davis,
  and Berkeley Lab

2
Charge to the Subcommittee

• The subcommittee is to identify current and
  emerging challenges and opportunities for
  theoretical research within the scientific
  mission of Basic Energy Sciences, with particular
  attention paid to how computing will be employed
  to enable that research. A primary purpose of
  the subcommittee is to identify those investments
  that are necessary to ensure that theoretical
  research will have maximum impact in the areas of
  importance to Basic Energy Sciences, and to
  guarantee that BES researchers will be able to
  exploit the entire spectrum of computational
  tools, including the leadership class facilities
  contemplated by the Office of Science.

3
Timeline for preparation of the full subcommittee
report

• February 22, 2004 First meeting of the
  subcommittee, prior to the February meeting of
  BESAC.
• April 17-18, 2004 Subcommittee meeting in
  Chicago to take testimony and discuss preliminary
  ideas and findings
• June 4, 2004 Letter report of the committee,
  delivered to John Hemminger and Pat Dehmer, for
  discussion at the August 5,6 meeting of BESAC.
• July 30, 2004 First draft extended outline
  delivered to entire subcommittee on Theory and
  Computing in the Basic Energy Sciences
• August 5 6, 2004 BESAC discussion of the
  preliminary report.
• Fall meeting of BESAC Proposed final draft of
  the full report to be delivered to BESAC for its
  evaluation.
• End of January, 2005 Final bound report to be
  delivered to the Office of Science and BES

4
Subcommittee Members

• Roberto Car, Princeton U.
• Peter Cummings, Vanderbilt U.
• Jim Davenport, BNL
• Thom Dunning, UT/ORNL
• Bruce Garrett, PNNL
• Chris Greene, U. of Colorado
• Bruce Harmon, Ames Lab
• Rajiv Kalia, USC
• Kate Kirby, Harvard-Smithsonian Center for
  Astrophysics
• Walter Kohn, UC-Santa Barbara

• Carl Lineberger, U. of Colorado
• Bill McCurdy, UC- Davis/LBNL
• Mike Norman, ANL
• Larry Rahn, Sandia/Livermore
• Tony Rollett, Carnegie Mellon
• Douglas Tobias, UC-Irvine
• Stan Williams, Hewlett-Packard
• Margaret Wright, Courant Institute, NY

5
Opportunities for Discovery Theory and
Computation in Basic Energy Sciences

• Subcommittee on Theory and Computation
• of the Basic Energy Sciences Advisory Committee
• U.S. Department of Energy

6
Executive Summary I. A Confluence of Scientific
Opportunities Why Invest Now in Theory and
Computation in the Basic Energy Sciences?
A. Dramatic Progress in Theory and Modeling in
Chemistry and Materials Sciences B. New
Scientific Frontiers C. New Experimental
Facilities D. New Computational
Capabilities II. BES Community Input and
Assessment A. Subcommittee Expertise B. Testimony
of the Theory Community C. Questions Solicited of
the BES Community D. A Consensus Observation The
Unity of Theory and Computation in the Basic
Energy Sciences
7

• III. Emerging Themes in BES Complexity and
  Control
• A. Opportunities and Challenges in Complex
  Systems
• B. Opportunities and Challenges in Quantum
  Control
• Opportunities and Challenges in Control of
  Complex Systems
• IV. Connecting Theory with Experiment at the DOE
  Facilities Accelerating Discoveries and
  Furthering Understanding
• A Major Theme Expressed by Experimentalists and
  Theorists in the Basic Energy Sciences
• V. The Resources Essential for Success in the BES
  Theory Enterprise
• A. The Full Spectrum of Computational Resources
• B. Supporting New Styles of Theory and
  Computation in the BES Portfolio Scientific
  Codes As Shared Instruments
• The Human Resources Training Future Generations
  of Theorists
• VI. Findings and Recommendations

8
I. A Confluence of Scientific Opportunities Why
Invest Now in Theory and Computation in the Basic
Energy Sciences?

• ? Striking recent scientific successes of theory
  and modeling
• ? The appearance of specific new scientific
  frontiers
• ? The development and construction of new
  experimental facilities
• ? The ongoing increase of computational
  capability, including the promise of new
  leadership-scale computational facilities.

9
Dramatic Progress in Theory and Computation

• Density functional theory (DFT) has transformed
  theoretical chemistry, surface and materials
  science
• Large-scale classical molecular dynamics has been
  able to treat motion of gt a million atoms
• Discrete grid and wave-packet methods for
  treating atoms/molecules, e.g. in intense fields
• A range of electronic structure methods have
  evolved coupled cluster, MBPT, QMC
• First-principles spin dynamics elucidated
  mechanism of giant magnetoresistance and
  spintronic devices
• Dynamical mean field theory (DMFT) successful in
  describing strongly correlated electronic states
• Ab Initio molecular dynamics (Car-Parinello)
  treats motion of atoms and changes in electronic
  structure

10
New Scientific Frontiers

• Nanoscience
• Ultrafast Chemistry and Physics
• Biomaterials and Biomimetic Systems
• Coherent Control
• Control of Quantum Coherence
• Spintronics

11
New Experimental Facilities

• Existing Light Sources APS (Argonne), ALS
  (LBL), and NSLS (Brookhaven), together with the
  new Linac Coherent Light Source under
  construction at SLAC, have created a growing wave
  of new experiments in chemistry, physics and
  materials science
• Construction of the Spallation Neutron Source at
  ORNL (sched. Completion 2006)
• Five Nanoscale Science Research Centers under
  design or construction
• Needed an overall strategy and increased
  support for theoretical research to guide and
  respond to the experiments at these facilities.

12
New Computational Capabilities

• Desktop workstations -- rapid growth in
  microprocessor speed (Moores Law)
• Cluster computing -- tens or hundreds of
  processors linked together, and run by a single
  research group or department have helped to
  ready many disciplines within BES for massively
  parallel computing
• Large-scale computing facilities -- operated by
  DOE and NSF (and others). Centers at NERSC
  (LBL), ORNL, and Argonne new facility at PNNL
  leadership-class facility at ORNL
• BES research -- a major user of these facilities
• BES community has demonstrated READINESS

13
II. BES Community Input Obtaining Testimony
from the Community

• ? Open meeting, April 17, 2004 in Chicago area
  16 invited talks, plus panel discussions
• ? Website established to collect input
  https//besac.nersc.gov
• ? E-mails inviting input to website, or to
  co-chairs directly, to DAMOP, DCP, DMP, DCMP of
  APS
• ? Announcement inviting input on ACS Division of
  Physical Chemistry home page

14
Questions asked of BES Community

• In your field, what are the major scientific
  challenges?
• In your area, do theory and computational science
  drive progress and/or partner with experiment?
• How might progress in your field impact other
  areas within BES?
• Are computing resources (hardware software) a
  limiting factor in your field?
• Would support for development of new algorithms
  for high-end computer architectures be important?
• Are there opportunities in your area to assemble
  interdisciplinary teams for attacking large
  problems?

15
Consensus Observation The Unity of Theory and
Computation in BES

• Theory and computation should be viewed as a
  unity, not as competing parts of the BES
  portfolio
• Theory enterprise in BES is heterogeneous, with
  respect to scientific problems, research group
  size and computational resources required
• Ensuring the highest quality scientific return
  requires the complete spectrum of theory activity
  (from the single-PI groups to the large,
  interdisciplinary teams), coupled with access to
  appropriate computational resources.