Information Technology Frontiers for a New Millenium DOE's ASCI Program

	Introduction ASCI's role in the Nation's Stockpile Stewardship Program Computing and simulation with ASCI Academic Strategic Alliance Program ASCI computing platforms PathForward
Introduction	DOE's Accelerated Strategic Computing Initiative (ASCI) began in FY 1996, but this is the first year that ASCI has been explicitly described in the President's HPCC Budget. This is in part a response to the PITAC's suggestion, but it also reflects ASCI's world leadership position in scientific and engineering computing.   ASCI's mission is closely tied to national security. On August 11, 1995, President Clinton announced the United States' intention to pursue a "zero yield" Comprehensive Test Ban Treaty and thus reduce nuclear danger. This decision ushered in a new era in the way the U. S. ensures confidence in the safety, performance, and reliability of its nuclear stockpile.
ASCI's role in the Nation's Stockpile Stewardship Program	ASCI is an essential element of the Department of Energy's Stockpile Stewardship Program (SSP), which was established to build on existing capabilities while developing new means of assessing the performance of nuclear weapon systems, predicting their safety and reliability, and certifying their functionality. The SSP not only must fulfill its responsibilities without nuclear testing, but also must deal with constraints on non-nuclear testing, the downsizing of production capability, and the cessation of new weapon designs to replace existing weapons. Further complicating matters, weapon components will exceed their design lifetimes, and manufacturing issues and environmental concerns will force changes in fabrication processes and materials of weapon components. The DOE selected computer simulation and modeling as the way to fulfill its responsibilities and established ASCI to provide high fidelity computer simulations of weapon systems that will enable scientists to make the necessary judgements to maintain the credibility of the nuclear deterrent.   The new capabilities must be developed within a fast-approaching deadline. Both the scientists who form the basis of our experience with nuclear weapon design and testing and the weapons themselves are aging. This leads to the year 2004 timeframe as the target date for having working ASCI computing systems and codes available so a smooth transition from "test based" certification and assessment can be made.
Computing and simulation with ASCI	ASCI is a mission-driven large-scale computing and simulation effort to deliver a set of advanced capabilities within a constrained time period. Because of this the ASCI effort focuses on creating and developing: Advanced Applications Software. ASCI is developing the high performance applications software needed to implement simulation capabilities for weapon systems. The Highest End of Computing. Computers that are more powerful are needed for simulation applications. Through partnerships between the national labs and the U.S. computer industry, ASCI is stimulating the more rapid development of high performance computers with speeds and memory capacities much greater than are anticipated from current development trends. Collaborations with industry, academia, and other Government agencies is spurring development of high performance access, communications, and presentation "supercorridors," making an information-rich connection between weapon scientists and simulations. Problem-Solving Environments. ASCI is developing a computational infrastructure to allow applications to be developed and execute efficiently on the ASCI computing systems and allow accessibility through the high performance, information-rich supercorridors.   The year 2004 is the timeframe for having working ASCI computing systems and codes.
Academic Strategic Alliance Program	The ASCI Academic Strategic Alliance centers take the form of centers of Excellence at five universities, focusing on the high performance, simulation-based, scientific and computational and computer science areas that strongly support the ASCI and Science Based Stockpile Stewardship objectives. The centers were established with the intent of at least a five-year funding commitment, subject to a contract continuation review in the third year. At the end of five years, the program plan allows for renewal or re-bids for another five years.   Each center focuses on one or more national-scale, multi-disciplinary applications for which the coupling and integration of computer-based simulations from multiple disciplines offers unprecedented opportunities for major advances and discoveries in basic and applied science areas that are important to ASCI, to the DOE SSP, and to the application areas. These applications are unclassified and relevant to nationally significant scientific, economic, and/or social priorities. A primary goal of the centers is to enable advances in mathematical modeling, numerical mathematics, computer systems, and computer and information science to help meet the high performance computing and problem solving environment needs of ASCI. The centers are: center for Integrated Turbulence Simulations (CITS), Stanford University Computational Facility for Simulating the Dynamic Response of Materials, California Institute of Technology center for Astrophysical Thermonuclear Flashes, University of Chicago center for Simulation of Accidental Fires and Explosions, University of Utah/Salt Lake City center for Simulation of Advanced Rockets, University of Illinois at Urbana/Champaign Brief descriptions of the Academic Strategic Alliance centers and other information such as accomplishments, milestones, and personnel can be found in the ASCI Alliances FY 1999 Implementation Plan, which can be accessed on the Web at http://www.llnl.gov/asci-alliances/.
ASCI computing platforms	The "ASCI Red" machine, built by the Intel Corporation and Sandia National Laboratories, links more than 9,000 Pentium Pro desktop microprocessors. In December 1996, this machine captured international headlines with its world record-breaking speed of one trillion operations per second -- teraops -- the first break into terascale computing. The "Blue Pacific" system, built by IBM and delivered to Livermore National Laboratory in the fall of 1998, incorporates 5,856 processors to achieve a peak speed of 3.9 teraops with 2.6 terabytes of memory. Los Alamos National Laboratory installed the "Blue Mountain" system in October, 1998. This SGI/Cray computer has 6,144 processors and over 1.5 terabytes of memory.   Los Alamos National Laboratory installed the "Blue Mountain" system in the fall of 1998.
PathForward	The PathForward program is enabling U.S. computer companies to develop the technologies needed to produce the next generation ultra-scale computing systems for ASCI. PathForward draws on the capabilities, availability, expertise, and products currently being produced by leading computer companies, focusing on interconnect technologies, data storage technologies, systems software, and tools for large scale computing systems. These technologies, while critical to ASCI's platform needs, are areas in which private sector development would not otherwise take place, at least not in the time frame required by the SSP. At the same time, they are investments in which industry sees value for future products and markets -- essential scaling and integration technologies that enable ultra-scale computing systems to be engineered and developed out of commodity computing building blocks.