Molecular dynamics simulation of protein folding. See page 48 for more details.

Allocations of computer time and archival storage at NERSC are awarded to research groups,regardless of their source of funding, based on an annual review of proposals. As proposals are submitted, they are subjected to peer review to evaluate the quality of science, how well the proposed research is aligned with the mission of DOE SC, and the readiness of the specific application and applicant to fully utilize the computing resources being requested.

Figure 1. NERSC FY00 massively parallel processing (MPP) allocations by site.

	Figure 2. Percentage of NERSC MPP users and usage (computing time) by institution type.

The NERSC Program Advisory Committee (see Appendix B) is responsible for the scientific peer review process used to allocate 40 percent of NERSC’s computing resources. The peer review and resource allocation process for the remaining 60 percent is managed directly by the DOE SC programs, reflecting their mission priorities.

Two other groups provide general oversight: the NERSC Policy Board (Appendix A) advises the Berkeley Lab Director on the policies that determine the impact and performance of the NERSC Center, and the NERSC Users Group (Appendix C) advises the NERSC management and provides feedback from the user community. This section of the Annual Report gives an overview of the research supported by NERSC and points out some of the year’s achievements, which are described further in the Science Highlights section.

Advanced Scientific Computing Research

DOE’s Office of Advanced Scientific Computing Research (OASCR) supports a number of projects in computer science and applied mathematics. NERSC staff take the lead or participate in several of these projects, including research on turbulent reacting flow by our Center for Computational Sciences and Engineering; development of software packages that enable ordinary C or Fortran computer programs to perform arithmetic with 32 or 64 decimal digit accuracy (this software is being used to explore an unresolved question regarding the regularity of vortices); developing computational tools for linear algebra problems that are ubiquitous in computational science and engineering; and researching subspace-based techniques for information retrieval, such as latent semantic indexing.

Figure 3. NERSC MPP users and usage by scientific discipline.

Other OASCR-funded research includes testing the scalability of parallel discrete event simulations, which are used in a wide variety of fields, from switching of cellular communications networks to studies of material failure. Another group is studying instabilities in turbulent mixing, an important issue for fiuid dynamics that impacts such questions as the rate of heat transfer by the Gulf Stream, resistance of pipes to fiuid flow, combustion rates in automotive engines, and the late time evolution of a supernova. Research jointly funded by OASCR and other DOE offices includes developing global optimization approaches to protein fold refinement, and developing a new generation of electron-atom and electron-molecule scattering codes that are capable of treating all details of electron impact ionization.

Basic Energy Sciences

NERSC provides computational support for a large number of materials sciences, chemical sciences, geosciences, and engineering research projects sponsored by DOE’s Office of Basic Energy Sciences.

Research in the chemical sciences will have important impacts on energy efficiency, pollution prevention, and environmental restoration. This year researchers showed that molecular-based simulations of complex fiuids can be used to predict properties such as the viscosity index and pressure-viscosity coefficient of lubricants. The ability to predict these properties via simulation will lead to the molecular design of improved lubricants, which will result in better energy efficiency.

A new area of research at NERSC this year was numerical simulation of combustion in homogeneous charge compression ignition (HCCI) engines. HCCI engines are an attractive alternative to diesel engines, offering the potential for diesel-like efficiencies, while producing extremely low emissions without expensive aftertreatment. Computational simulations will provide crucial direction to HCCI design efforts.

Electronic structure theory has emerged as a valuable counterpart to direct experiments for the study of reactive species that may not be characterized easily (if at all) in the laboratory — for example, large polycyclic aromatic hydrocarbon (PAH) cations, which arise in combustion processes as intermediates to the formation of soot particles, and which are also believed to play a significant role in interstellar carbon chemistry. Electronic structure research on the chemical and conformational transformations of biomolecules is beginning to yield a novel microscopic picture of biochemical dynamics. Understanding biological chemical processes at the atomic level with this and other methods will have a major impact on the drug and biotechnology industries.

In the materials sciences, Georgia Tech physicist and NERSC user Uzi Landman won this year’s Feynman Prize in Nanotechnology (Theoretical) for his pioneering work in computational materials science for nanostructures (see page 39). Such computer modeling provides deep insights into the nature and properties of matter at the nanoscale, and is essential in predicting what could be built at the molecular level, reducing time spent on expensive laboratory experiments.

New modeling tools developed by Landman and other researchers are also making it possible to begin bridging the gaps between scales, so that the effects of atomic and microscopic phenomena can be seen at the macroscopic scale. As these modeling tools mature, they will make important contributions in both technological and environmental areas, including carbon sequestration, the development of high-temperature superconductors, miniaturization of electronic and mechanical devices, development of lasers and sensors, design of novel logic gates and information storage strategies, control of friction under extreme conditions, and the design of electric motors with reduced weight and improved performance.

Biological and Environmental Research

DOE’s Office of Biological and Environmental Research is a major supporter of global climate studies as well as computational medical and biological research using NERSC resources. In addition to providing computational support, NERSC is also the repository for the archive of the Program for Climate Model Diagnosis and Intercomparison (PCMDI). Complete data sets from PCMDI are publicly available to researchers through NERSC.

Ongoing improvement in the precision of climate modeling codes is resulting in a growing understanding of the factors contributing to global climate change as well as better modeling of regional changes. DOE, NASA, and the National Center for Atmospheric Research are jointly developing a next-generation Community Climate System Model (CCSM), which will incorporate a higher degree of physical consistency than current models and enable longer and more detailed simulations. Development and testing of CCSM components are under way. The Parallel Climate Model (PCM) has also been improved with better sea ice and ocean components, a river transport component, and a higher resolution atmosphere component that has better definition of the continent-ocean boundaries and an improved treatment of mountains.

Steady progress is being made in determining the level of natural climate variability and distinguishing it from anthropogenic changes. Contributing to this effort is a large ensemble set of PCM simulations, showing the global climate changes due to increased greenhouse gases and changes in sulfate aerosols, for the years 1870–2100.  And basic research is under way to understand the potential effectiveness and the environmental impacts of carbon sequestration in the oceans.

One of the most significant findings of the past year challenges widespread speculation that particulate pollution may offset the effects of greenhouse gases by increasing the cloud cover and reflecting solar energy back into space. Field observations of the dark haze that covered the Indian Ocean in February and March 1999, and subsequent computational analysis of the data, showed that the haze absorbed solar heat, and an atmospheric temperature increase of only 1º per day at noon was enough to significantly reduce the cloud cover (see page 52).

In the medical and life sciences, much of the research done using NERSC computers is focused on developing the computational methods needed to interpret and make use of genomic data — for example, identifying and classifying protein folds in complete genomes, predicting protein structures on the basis of genomic data, and elucidating the mechanisms of protein folding through simulations. Another major project explores the details of the chemical mechanisms employed by enzymes to serve as catalysts of biochemical reactions — events that are too fast to be measured by experiment.

Fusion Energy Sciences

The Office of Fusion Energy Sciences has historically been in the forefront of promoting computational science. Recent progress in fusion research has been accelerated by a strong coupling between theory, computation, and experiments. Three-dimensional modeling contributes to the developing understanding of plasma physics, improves the analysis of experimental results, and suggests new ways to improve magnetic and heavy ion fusion reactor designs.

The Numerical Tokamak Turbulence Project this year completed the lengthy process of benchmarking turbulence simulations from several different codes. They presented the first toroidal electromagnetic simulations of tokamak microturbulence, and also discovered that electron temperature gradient turbulence (ETG modes) can, under some conditions, cause transport comparable to that resulting from ion temperature gradient (ITG) modes. Simulations that quantified the transition from electrostatic to electromagnetic turbulence with increasing   called into question the validity of the electrostatic approximation commonly employed in turbulent transport studies. The new simulations found that microturbulence takes on an electromagnetic character even at low values of  , and that significant electromagnetic effects on turbulent transport occur.

In the work of other research groups, the importance of nonlinearly generated zonal flow for the reduction of ion thermal transport was demonstrated, as well as the role played by ion-ion collisions in the bursting behavior observed in tokamak experiments. Studies of two-stream instabilities in space-charge-dominated beams in accelerators helped to explain the beam loss observed in various machines. A new working model was developed of edge localized modes (ELMs), which have been observed but poorly understood for two decades; the model successfully described ELM behavior in the DIII-D tokamak.

Gyrokinetic growth rate calculations analyzing the drift-wave stability of a variety of tokamak plasmas found that discharges with neon injection had improved energy confinement due to the suppression of ITG-mode turbulence.

High Energy and Nuclear Physics

The DOE Office of High Energy and Nuclear Physics sponsors important theoretical studies, computational simulation and analysis of experimental data, and simulations that are helping design the next generation of experimental facilities.

NERSC’s leadership in astrophysics data analysis yielded headline-making results for the second time. Two years ago, analysis of supernova data led to the conclusion that the Universe will continue expanding forever. This year, analysis of the BOOMERANG cosmic microwave background data supported the earlier finding as well as the new conclusion that the geometry of the Universe is flat (see pages 6 and 67).

On the theoretical side, NERSC provides computational resources for several large research efforts in lattice quantum chromodynamics (QCD). Lattice QCD provides the most promising approach to understanding the behavior of quarks and gluons, the building blocks of strongly interacting particles. QCD studies are relevant to the physics of the early Universe and are crucial to interpreting the results of experimental attempts to produce a quark-gluon plasma. QCD calculations test the Standard Model and may provide clues to a new physics.

Nuclear physics researchers are working to explain the properties and reactions of nuclei in terms of interacting nucleons (protons and neutrons). A team from Argonne National Laboratory has achieved calculations of six- through ten-nucleon systems that use realistic interactions and that are accurate to 1% for the binding energies. The resulting wave functions can be used to compute properties measured at electron and hadron scattering facilities and to compute astrophysical reaction rates, many of which cannot be measured in the laboratory.

The Advanced Computing for 21st Century Accelerator Science and Technology project (the former Computational Accelerator Physics Grand Challenge) is developing a comprehensive terascale accelerator simulation environment for the U.S. particle accelerator community. The design and construction of the next generation of accelerators will involve greater complexity than ever before, and will require unprecedented precision in accelerator design and beam control. For all these accelerator systems, terascale simulation will play a key role by facilitating important design decisions, increasing safety and reliability, optimizing performance, and helping to ensure project completion within budget and on schedule. The project will add new codes and modules to an existing set that is being used for simulations at accelerator facilities around the world.