Building on the DOE Office of Science's past achievements in computational science, SciDAC will help create a new generation of scientific simulation codes that take full advantage of terascale computers and Grid technologies.

Berkeley Lab will help lead the development of a new generation of tools and technologies for scientific computing under a new $57 million program announced on August 14, 2001 by the DOE. Under the Scientific Discovery through Advanced Computing (SciDAC) program, scientists and mathematicians in Berkeley Lab's Computing Sciences organization are leading four of the projects and are collaborating in another seven projects.

SciDAC is an integrated program that will help create a new generation of scientific simulation codes. The codes will take full advantage of the extraordinary computing capabilities of terascale computers to address ever larger, more complex problems. The program also includes research on improved mathematical and computing systems software that will allow these codes to use modern parallel computers effectively and efficiently. Additionally, the program will develop collaboratory software to enable geographically separated scientists to effectively work together as a team, to control scientific instruments remotely, and to share data more readily. In all, 51 projects were funded nationwide, involving collaborations among 13 DOE national laboratories and more than 50 colleges and universities.

Here are the SciDAC programs that Computing Sciences staff will be leading:

• DOE Science Grid — This collaboratory will define, integrate, deploy, support, evaluate, refine, and develop the persistent Grid services needed for a scalable, robust, high performance DOE Science Grid. It will create the underpinnings of the software environment that the SciDAC applications need to enable innovative approaches to scientific computing through secure remote access to online facilities, distance collaboration, shared petabyte datasets, and large-scale distributed computation. (PI: Bill Johnston)
  
  
• Performance Evaluation Research Center (PERC) — As one of several SciDAC Integrated Software Infrastructure Centers (ISICs), this project will focus on how specific scientific applications can best be run on high-performance computers. The results of this research are expected to permit the generation of realistic performance levels, and to determine how applications can be written to perform at the highest levels and how this information can be applied to the design of future applications and computer systems. (PI: David Bailey)
  
  
• Scientific Data Management ISIC — Terascale computing and large scientific experiments produce enormous quantities of data that require effective and efficient management, a task that can distract scientists from focusing on their core research. The goal of this project is to provide a coordinated framework for the unification, development, deployment, and reuse of scientific data management software. (PI: Ari Shoshani)
  
  
• Applied Partial Differential Equations ISIC — The goal is to develop a high-performance algorithmic and software framework for solving partial differential equations arising from problems in magnetic fusion, accelerator design, and combustion-key mission areas for the DOE. (PI: Phil Colella)

Berkeley Lab computer scientists and mathematicians will also be collaborating with other national laboratories and universities on the following SciDAC projects:

• Advanced Computing for 21st Century Accelerator Science and Technology, — This project will work on simulating particle accelerators, some of the largest and most complex scientific instruments. A new generation of accelerator simulation codes will help us to use existing accelerators more efficiently and will strongly impact the design, technology, and cost of future accelerators. (Lead insitutions: Stanford Linear Accelerator Center and Berkeley Lab's Accelerator and Fusion Research Division)
  
  
• Collaborative Design and Development of the Community Climate System Model for Terascale Computers, — A multi-institutional team will develop, validate, document, and optimize the performance of this coupled climate model using the latest software engineering approaches, computational technology, and scientific knowledge. (Lead institution: Los Alamos National Laboratory)
  
  
• Earth Systems Grid II: Turning Climate Databases into Community Resources — This project will create a virtual collaborative environment that links distributed centers, users, models, and data, significantly increasing the scientific productivity of U.S. climate researchers by turning climate datasets into community resources. (Lead institution: Argonne National Laboratory)
  
  
• The National Fusion Grid — This collaboratory project will promote more effective integration of experiment, theory, and modeling by enabling networked real-time data analysis and instantaneous communication among geographically dispersed teams of experimentalists and theoreticians. (Lead institution: General Atomics)
  
  
• Particle Physics Data Grid Collaborative Pilot — This pilot project will develop, acquire, and deliver vitally needed Grid-enabled tools for the data-intensive requirements of particle and nuclear physics. (Lead institutions: University of Wisconsin, California Institute of Technology, and Stanford Linear Accelerator Center)
  
  
• Scalable Systems Software ISIC — This project will address the lack of software for effective management and utilization of terascale computing resources. This project will develop an integrated suite of machine-independent, scalable systems software needed for the SciDAC program. The goal is to provide open-source solutions that work on systems ranging in size from small to large. (Lead institution: Oak Ridge National Laboratory)
  
  
• Terascale Optimal PDE Solvers (TOPS) ISIC — Large-scale simulations of importance to the DOE often involve the solution of partial differential equations (PDEs). In such simulations, continuous (infinite-dimensional) mathematical models are approximated with finite-dimensional models. To obtain the required accuracy and resolve the multiple scales of the underlying physics, the finite-dimensional models must often be extremely large, thus requiring terascale computers. This project focuses on developing, implementing, and supporting optimal or near-optimal schemes for PDE simulations and closely related tasks. (Lead institution: Old Dominion University)

In addition to Berkeley Lab's participation in the projects mentioned above, NERSC is providing specialized consulting and algorithmic support for SciDAC projects. A consulting project facilitator has been assigned to each scientific discipline to help define project requirements, get resources, and tune and optimize codes, as well as coordinate special services (queues, throughput, increased limits, etc.). Algorithmic project facilitators have been assigned to develop and improve algorithms, enhance performance, and coordinate software development with the various ISICs.

NERSC's strategic plan for the next five years involves two new elements—Support for Scientific Challenge Teams and the Unified Science Environment—specifically designed to support the kinds of scientific collaborations envisioned by SciDAC.