GETTING UP TO SPEED THE FUTURE OF SUPERCOMPUTING

Susan L. Graham, Marc Snir, and Cynthia A. Patterson, Editors

Committee on the Future of Supercomputing

Computer Science and Telecommunications Board

Division on Engineering and Physical Sciences

NATIONAL RESEARCH COUNCIL OF THE NATIONAL ACADEMIES

THE NATIONAL ACADEMIES PRESS
Washington, D.C.
www.nap.edu



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page R1
Getting up to Speed the Future of Supercomputing GETTING UP TO SPEED THE FUTURE OF SUPERCOMPUTING Susan L. Graham, Marc Snir, and Cynthia A. Patterson, Editors Committee on the Future of Supercomputing Computer Science and Telecommunications Board Division on Engineering and Physical Sciences NATIONAL RESEARCH COUNCIL OF THE NATIONAL ACADEMIES THE NATIONAL ACADEMIES PRESS Washington, D.C. www.nap.edu

OCR for page R1
Getting up to Speed the Future of Supercomputing THE NATIONAL ACADEMIES PRESS 500 Fifth Street, N.W. Washington, DC 20001 NOTICE: The project that is the subject of this report was approved by the Governing Board of the National Research Council, whose members are drawn from the councils of the National Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine. The members of the committee responsible for the report were chosen for their special competences and with regard for appropriate balance. Support for this project was provided by the Department of Energy under Sponsor Award No. DE-AT01-03NA00106. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the organizations that provided support for the project. International Standard Book Number 0-309-09502-6 (Book) International Standard Book Number 0-309-54679-6 (PDF) Library of Congress Catalog Card Number 2004118086 Cover designed by Jennifer Bishop. Cover images (clockwise from top right, front to back) 1. Exploding star. Scientific Discovery through Advanced Computing (SciDAC) Center for Supernova Research, U.S. Department of Energy, Office of Science. 2. Hurricane Frances, September 5, 2004, taken by GOES-12 satellite, 1 km visible imagery. U.S. National Oceanographic and Atmospheric Administration. 3. Large-eddy simulation of a Rayleigh-Taylor instability run on the Lawrence Livermore National Laboratory MCR Linux cluster in July 2003. The relative abundance of the heavier elements in our universe is largely determined by fluid instabilities and turbulent mixing inside violently exploding stars. 4. Three-dimensional model of the structure of a ras protein. Human Genome Program, U.S. Department of Energy, Office of Science. 5. Test launch of Minuteman intercontinental ballistic missile. Vandenberg Air Force Base. 6. A sample of liquid deuterium subjected to a supersonic impact, showing the formation of a shock front on the atomic scale. The simulation involved 1,320 atoms and ran for several days on 2,640 processors of Lawrence Livermore National Laboratory’s ASCI White. It provided an extremely detailed picture of the formation and propagation of a shock front on the atomic scale. Accelerated Strategic Computing Initiative (ASCI), Department of Energy, Lawrence Livermore National Laboratory. 7. Isodensity surfaces of a National Ignition Facility ignition capsule bounding the shell, shown at 200 picosec (left), 100 picosec (center), and near ignition time (right). An example of ASCI White three-dimensional computer simulations based on predictive physical models. ASCI, Lawrence Livermore National Laboratory. Copies of this report are available from the National Academies Press, 500 Fifth Street, N.W., Lockbox 285, Washington, DC 20055; (800) 624-6242 or (202) 334-3313 in the Washington metropolitan area; Internet, http://www.nap.edu Copyright 2005 by the National Academy of Sciences. All rights reserved. Printed in the United States of America

OCR for page R1
Getting up to Speed the Future of Supercomputing THE NATIONAL ACADEMIES Advisers to the Nation on Science, Engineering, and Medicine The National Academy of Sciences is a private, nonprofit, self-perpetuating society of distinguished scholars engaged in scientific and engineering research, dedicated to the furtherance of science and technology and to their use for the general welfare. Upon the authority of the charter granted to it by the Congress in 1863, the Academy has a mandate that requires it to advise the federal government on scientific and technical matters. Dr. Bruce M. Alberts is president of the National Academy of Sciences. The National Academy of Engineering was established in 1964, under the charter of the National Academy of Sciences, as a parallel organization of outstanding engineers. It is autonomous in its administration and in the selection of its members, sharing with the National Academy of Sciences the responsibility for advising the federal government. The National Academy of Engineering also sponsors engineering programs aimed at meeting national needs, encourages education and research, and recognizes the superior achievements of engineers. Dr. Wm. A. Wulf is president of the National Academy of Engineering. The Institute of Medicine was established in 1970 by the National Academy of Sciences to secure the services of eminent members of appropriate professions in the examination of policy matters pertaining to the health of the public. The Institute acts under the responsibility given to the National Academy of Sciences by its congressional charter to be an adviser to the federal government and, upon its own initiative, to identify issues of medical care, research, and education. Dr. Harvey V. Fineberg is president of the Institute of Medicine. The National Research Council was organized by the National Academy of Sciences in 1916 to associate the broad community of science and technology with the Academy’s purposes of furthering knowledge and advising the federal government. Functioning in accordance with general policies determined by the Academy, the Council has become the principal operating agency of both the National Academy of Sciences and the National Academy of Engineering in providing services to the government, the public, and the scientific and engineering communities. The Council is administered jointly by both Academies and the Institute of Medicine. Dr. Bruce M. Alberts and Dr. Wm. A. Wulf are chair and vice chair, respectively, of the National Research Council. www.national-academies.org

OCR for page R1
Getting up to Speed the Future of Supercomputing COMMITTEE ON THE FUTURE OF SUPERCOMPUTING SUSAN L. GRAHAM, University of California, Berkeley, Co-chair MARC SNIR, University of Illinois at Urbana-Champaign, Co-chair WILLIAM J. DALLY, Stanford University JAMES W. DEMMEL, University of California, Berkeley JACK J. DONGARRA, University of Tennessee, Knoxville, and Oak Ridge National Laboratory KENNETH S. FLAMM, University of Texas, Austin MARY JANE IRWIN, Pennsylvania State University CHARLES KOELBEL, Rice University BUTLER W. LAMPSON, Microsoft Corporation ROBERT F. LUCAS, University of Southern California PAUL C. MESSINA, Distinguished senior computer scientist, Consultant JEFFREY M. PERLOFF, University of California, Berkeley WILLIAM H. PRESS, Los Alamos National Laboratory ALBERT J. SEMTNER, Naval Postgraduate School SCOTT STERN, Northwestern University SHANKAR SUBRAMANIAM, University of California, San Diego LAWRENCE C. TARBELL, JR., Technology Futures Office, Eagle Alliance STEVEN J. WALLACH, Chiaro Networks Staff CYNTHIA A. PATTERSON, Study Director PHIL HILLIARD, Research Associate (through May 2004) MARGARET MARSH HUYNH, Senior Program Assistant HERBERT S. LIN, Senior Scientist

OCR for page R1
Getting up to Speed the Future of Supercomputing COMPUTER SCIENCE AND TELECOMMUNICATIONS BOARD DAVID LIDDLE, U.S. Venture Partners, Co-chair JEANNETTE M. WING, Carnegie Mellon University, Co-chair ERIC BENHAMOU, Benhamou Global Ventures, LLC DAVID D. CLARK, Massachusetts Institute of Technology, CSTB Chair Emeritus WILLIAM DALLY, Stanford University MARK E. DEAN, IBM Almaden Research Center DEBORAH ESTRIN, University of California, Los Angeles JOAN FEIGENBAUM, Yale University HECTOR GARCIA-MOLINA, Stanford University KEVIN KAHN, Intel Corporation JAMES KAJIYA, Microsoft Corporation MICHAEL KATZ, University of California, Berkeley RANDY H. KATZ, University of California, Berkeley WENDY A. KELLOGG, IBM T.J. Watson Research Center SARA KIESLER, Carnegie Mellon University BUTLER W. LAMPSON, Microsoft Corporation, CSTB Member Emeritus TERESA H. MENG, Stanford University TOM M. MITCHELL, Carnegie Mellon University DANIEL PIKE, GCI Cable and Entertainment ERIC SCHMIDT, Google Inc. FRED B. SCHNEIDER, Cornell University WILLIAM STEAD, Vanderbilt University ANDREW J. VITERBI, Viterbi Group, LLC CHARLES N. BROWNSTEIN, Director KRISTEN BATCH, Research Associate JENNIFER M. BISHOP, Program Associate JANET BRISCOE, Manager, Program Operations JON EISENBERG, Senior Program Officer RENEE HAWKINS, Financial Associate MARGARET MARSH HUYNH, Senior Program Assistant HERBERT S. LIN, Senior Scientist LYNETTE I. MILLETT, Senior Program Officer JANICE SABUDA, Senior Program Assistant GLORIA WESTBROOK, Senior Program Assistant BRANDYE WILLIAMS, Staff Assistant For more information on CSTB, see its Web site at <http://www.cstb.org>, write to CSTB, National Research Council, 500 Fifth Street, N.W., Washington, DC 20001, call (202) 334-2605, or e-mail the CSTB at cstb@nas.edu.

OCR for page R1
Getting up to Speed the Future of Supercomputing This page intentionally left blank.

OCR for page R1
Getting up to Speed the Future of Supercomputing Preface High-performance computing is important insolving complex problems in areas from climate and biology to national security. Several factors have led to the recent reexamination of the rationale for federal investment in research and development in support of high-performance computing, including (1) continuing changes in the various component technologies and their markets, (2) the evolution of the computing market, particularly the high-end supercomputing segment, (3) experience with several systems using the clustered processor architecture, and (4) the evolution of the problems, many of them mission-driven, for which supercomputers are used. The Department of Energy’s (DOE’s) Office of Science expressed an interest in sponsoring a study by the Computer Science and Telecommunications Board (CSTB) of the National Research Council (NRC) that would assess the state of U.S. supercomputing capabilities and relevant research and development. Spurred by the development of the Japanese vector-based Earth Simulator supercomputer, the Senate’s Energy and Water Development Appropriations Committee directed the Advanced Simulation and Computing (ASC) program of the National Nuclear Security Administration (NNSA) at DOE to commission, in collaboration with DOE’s Office of Science, a study by the NRC. Congress also commissioned a study by the JASONs1 to identify the distinct requirements of the stockpile stewardship program and its relation to the ASC acquisition strategy. 1   Formed in 1959, the JASONs are a select group of scientific advisors who consult with the federal government, chiefly on classified research issues.

OCR for page R1
Getting up to Speed the Future of Supercomputing CSTB convened the Committee on the Future of Supercomputing to assess prospects for supercomputing technology research and development in support of U.S. needs, to examine key elements of context—the history of supercomputing, the erosion of research investment, the changing nature of the problems demanding supercomputing, and the needs of government agencies for supercomputing capabilities—and to assess opportunities for progress. The 18 distinguished members of the study committee (see Appendix A for their biographies) were drawn from academia, industry, and government research organizations in the United States. Several committee members have had previous government and/or industry service. Their collective expertise includes software, computer architecture, performance assessment, applications using supercomputing, economics, and policy matters. The committee did its work through its own expert deliberations and by soliciting input from key officials in its sponsoring agency (DOE) and numerous experts in both the United States and Japan, including government officials, academic researchers, supercomputer manufacturers, software vendors, supercomputer center managers, and application users of supercomputing systems (see Appendix B). In addition to meeting six times, the committee hosted a workshop attended by more than 20 scientists from a broad range of disciplines to explore the supercomputing needs and opportunities of key scientific domains in the coming decade and to discuss the supercomputing technologies that will facilitate supercomputer use in these domains. Many of the workshop participants provided white papers (see Appendix C for a list) expressing their views on computational challenges in supercomputing, which informed both the workshop and this report. The committee also visited five DOE supercomputer centers and the National Security Agency’s (NSA’s) Supercomputer Center (see Appendix B). A subset of the committee received classified briefings from the Department of Energy on stockpile stewardship and from the NSA on signals intelligence that helped illuminate how these mission requirements drive supercomputing needs now and in the future. Given that a significant fraction of government funding of supercomputing is for classified national security programs, the committee believed such briefings were needed to ensure that its report would be useful for the entire supercomputing community. Having received the briefings, the committee believes that the needs of the classified supercomputing applications reinforce, but do not change, the committee’s findings and recommendations for the future of supercomputing. This unclassified report does not have a classified annex, nor is there a classified version. To facilitate communication within the broader community, the committee hosted a town hall meeting at the annual 2003 Supercomputing

OCR for page R1
Getting up to Speed the Future of Supercomputing Conference in Phoenix, Arizona. In addition, a subset of the committee spent one week in Japan meeting with senior colleagues from the Japanese government, industry, and academia to discuss scientific, technical, and policy issues of mutual interest and to better understand both the similarities and the differences in how the two countries approach supercomputing. They visited several sites in Japan, including the Earth Simulator; the government ministry responsible for funding the Earth Simulator; a university supercomputer center; Japan’s Aerospace Exploration Agency; and an auto manufacturer. On the committee’s behalf, the National Academy of Engineering co-sponsored with the Engineering Academy of Japan a 1-day forum in Tokyo on the future of supercomputing. Twenty-five Japanese supercomputing experts participated in the forum. The sharing of ideas in those meetings provided important perspectives that contributed to the completeness and accuracy of this report. It is the hope of the committee that activities such as the Tokyo forum will lead to future collaboration between Japan and the United States in areas that will advance supercomputing in both countries. In July 2003, the committee released an interim report2 that provided a high-level description of the state of U.S. supercomputing, the needs of the future, and the factors that contribute to meeting those needs. That report generated a number of comments that helped to guide the committee in its work for this final report. Additional inputs helpful to committee members and staff came from professional conferences, the technical literature, and government reports. The committee is grateful to the many people who contributed to this complex study and its comprehensive report. First and foremost, the committee thanks the sponsors, DOE’s Office of Science (Fred Johnson and Dan Hitchcock) and DOE’s NNSA (Dimitri Kusnezov, Edgar Lewis, and José Muñoz), not only for their financial support but also for their help in facilitating meetings with people with whom its members wished to speak. The committee appreciates the thoughtful testimony received from many individuals at its plenary sessions (see Appendix B for a complete list of briefers). The NSA and DOE site visits provided critical input to the committee deliberations. These site visits would not have been possible without the assistance of people at each locale. The committee and staff thank the following people for their help: Gary D. Hughes (NSA), Lynn Kissel (Lawrence Livermore National Laboratory), James S. Peery (Los Alamos National Laboratory),Horst D. Simon (Lawrence Berkeley Na- 2   National Research Council (NRC). 2003. The Future of Supercomputing: An Interim Report. Washington, D.C.: The National Academies Press.

OCR for page R1
Getting up to Speed the Future of Supercomputing tional Laboratory), Robert Thomas (Sandia National Laboratories), Rick Stevens (Argonne National Laboratory), and Thomas Zacharia (Oak Ridge National Laboratory). The committee thanks the workshop participants for the insights they contributed through their white papers (see Appendix C for a list of papers), discussions, breakout sessions, and subsequent interactions. The committee is particularly grateful to Warren Washington (National Center for Atmospheric Research), Charles McMillan (Lawrence Livermore National Laboratory), Jeffrey Saltzman (Merck Research Laboratory), and Phillip Colella (Lawrence Berkeley National Laboratory) for their thoughtful plenary presentations. Many people were instrumental in making the trip to Japan a success. The committee is extremely grateful to Kenichi Miura (Fujitsu fellow) and Tadashi Watanabe (NEC) for their assistance before and during the trip. The 1-day Japan–U.S. Forum on the Future of Supercomputing would not have been possible without the support of the Engineering Academy of Japan and the National Academy of Engineering. The committee learned a lot from insightful presentations and discussions from all the Japanese forum participants. The committee and staff also thank the individuals at each site who took time to meet with the committee. In particular, they thank Tetsuya Sato at the Earth Simulator Center and Harumasa Miura at the Ministry of Education, Culture, Sports, Science, and Technology. Maki Haraga provided excellent translation services and logistical help for the committee’s entire trip. The committee was fortunate to receive many thoughtful and perceptive comments from the reviewers as well as from the Monitor and the Coordinator of this report. These comments were instrumental in helping the committee to sharpen and improve its report. Finally, the committee thanks the various members of the NRC staff who helped to move this report from vision to reality. Cynthia Patterson provided continuing wisdom, guidance, encouragement, and friendship, in concert with her hard work on the report. Margaret Huynh’s skills in organizing the committee’s meetings and supporting its efforts and Phil Hilliard’s research support were key contributions to the work of the committee. Liz Fikre edited the final manuscript for publication. Kevin Hale and Machelle Reynolds successfully facilitated the security clearances and security review necessary to complete this study in a timely manner. Janice Mehler and Liz Panos were very helpful in facilitating and expediting the review process. Susan L. Graham and Marc Snir, Co-chairs Committee on the Future of Supercomputing

OCR for page R1
Getting up to Speed the Future of Supercomputing Acknowledgment of Reviewers This report has been reviewed in draft form by individuals chosen for their diverse perspectives and technical expertise, in accordance with procedures approved by the National Research Council’s (NRC’s) Report Review Committee. The purpose of this independent review is to provide candid and critical comments that will assist the institution in making the published report as sound as possible and to ensure that the report meets institutional standards for objectivity, evidence, and responsiveness to the study charge. The review comments and draft manuscript remain confidential to protect the integrity of the deliberative process. We wish to thank the following individuals for their review of this report: Mark E. Dean, IBM, Steven Gottlieb, Indiana University, Shane Mitchell Greenstein, Northwestern University, Sidney Karin, University of California, San Diego, Ken Kennedy, Rice University, Richard Loft, National Center for Atmospheric Research, J. Andrew McCammon, University of California, San Diego, Kenichi Miura, National Institute of Informatics, Michael Norman, University of Illinois, Urbana-Champaign, Richard Proto, National Security Agency (retired), Daniel A. Reed, University of North Carolina, Chapel Hill, Ahmed Sameh, Purdue University, Gary Smaby, Smaby Group, Inc.,

OCR for page R1
Getting up to Speed the Future of Supercomputing Burton J. Smith, Cray Inc., Allan Edward Snavely, San Diego Supercomputing Center, William W. Stead, Vanderbilt University, and Paul Tackley, University of California, Los Angeles. Although the reviewers listed above have provided many constructive comments and suggestions, they were not asked to endorse the conclusions or recommendations, nor did they see the final draft of the report before its release. The review of this report was overseen by Elsa Garmire of Dartmouth University and Samuel H. Fuller of Analog Devices, Inc. Appointed by the NRC, they were responsible for making certain that an independent examination of this report was carried out in accordance with institutional procedures and that all review comments were carefully considered. Responsibility for the final content of this report rests entirely with the authoring committee and the institution.

OCR for page R1
Getting up to Speed the Future of Supercomputing Contents     EXECUTIVE SUMMARY   1 1   INTRODUCTION AND CONTEXT   11      Study Context,   12      Computenik,   15      About the Interim Report,   18      Organization of the Report,   19 2   EXPLANATION OF SUPERCOMPUTING   20 3   BRIEF HISTORY OF SUPERCOMPUTING   28      The Prehistory of U.S. Supercomputing,   28      Supercomputers Emerge as a Market,   31      Control Data and Cray,   33      Enter Japan,   36      Innovation in Supercomputing,   38      Recent Developments in Supercomputing,   43      The U.S. High-Performance Computing Industry Today,   44      An Industrial Revolution,   53      Impacts,   61      Direct Contributions,   62      Spillover Effects,   65

OCR for page R1
Getting up to Speed the Future of Supercomputing 4   THE DEMAND FOR SUPERCOMPUTING   67      Compelling Applications for Supercomputing,   70      Common Themes and Synergies Across Applications Areas,   72      Selected Application Areas,   74      Stockpile Stewardship,   74      Signals Intelligence,   76      Defense,   78      Climate Modeling,   79      Plasma Physics,   84      Transportation,   85      Bioinformatics and Computational Biology,   89      Societal Health and Safety,   92      Earthquakes,   93      Geophysical Exploration and Geoscience,   94      Astrophysics,   96      Materials Science and Computational Nanotechnology,   97      Human/Organizational Systems Studies,   100      Projected Computing Needs for Applications,   101 5   TODAY’S SUPERCOMPUTING TECHNOLOGY   104      Supercomputer Architecture,   105      Scaling of Technology,   105      Types of Supercomputers,   111      Performance Issues,   113      Trade-offs,   118      Trends in Supercomputer Architecture,   121      Supercomputing Algorithms,   125      Solving Partial and Ordinary Differential Equations,   126      Mesh Generation,   128      Dense Linear Algebra,   129      Sparse Linear Algebra,   129      Discrete Algorithms,   130      Fast Transforms,   131      New Algorithmic Demands Arising from Supercomputing,   131      Disciplinary Needs,   131      Interdisciplinary Needs,   132      Synthesis, Sensitivity Analysis, and Optimization Replacing Analysis,   133      Huge Data Sets,   133      Changing Machine Models,   133      Supercomputing Software,   134      Operating Systems and Management Software,   135      Programming Models, Programming Languages, and Tools,   138

OCR for page R1
Getting up to Speed the Future of Supercomputing      Libraries,   142      Applications Software,   142      Reliability and Fault Tolerance,   145      Performance Estimation,   146      Performance Benchmarks,   146      Performance Monitoring,   148      Performance Modeling and Simulation,   148      Performance Estimation and the Procurement Process,   150      The Imperative to Innovate and Barriers to Innovation,   151      Systems Issues,   151      Issues for Algorithms,   153      An Example from Computational Fluid Dynamics,   154      Software Issues,   155 6   SUPERCOMPUTING INFRASTRUCTURES AND INSTITUTIONS   157      Supercomputing Ecosystem Creation and Maintenance,   161      How Ecosystems Get Established,   162      Potential Barriers for New Ecosystems,   166      Ecosystem Workforce,   170      Consumer Institutions,   174      Supercomputing Centers,   174      Industrial Supercomputing,   176 7   SUPERCOMPUTING ABROAD   180      Japan,   182      Similarities,   182      Differences,   183      The Earth Simulator,   183      Other Japanese Centers,   187      China,   188      Europe,   189      United Kingdom,   189      Germany,   190      France,   190      Spain,   191      Application Software,   191 8   A POLICY FRAMEWORK   192      The Government as the Leading User and Purchaser of Supercomputer Technology,   193      Supercomputer Technology Investments as Public Goods,   194      Potential Costs of Government Intervention,   195

OCR for page R1
Getting up to Speed the Future of Supercomputing      Alternative Modes for Government Intervention,   196      Government Incentives,   197      Government Research,   199      Competing Government Objectives,   201      Coordination Versus Diversification,   201      Commitment Versus Flexibility,   201      Secrecy Versus Spillovers,   202 9   STEWARDSHIP AND FUNDING OF SUPERCOMPUTING   206      Satisfying Current Supercomputing Needs,   208      Ensuring Future Supercomputing Leadership,   209      The Need for Hardware and Software Producers,   209      The Need for Stability,   210      The Need for a Continuum from Research to Production,   211      The Need for Money,   216      The Need for People,   218      The Need for Planning and Coordination,   219      A Supercomputing Roadmap,   220      Responsibility and Oversight,   222 10   THE FUTURE OF SUPERCOMPUTING—CONCLUSIONS AND RECOMMENDATIONS   225      Conclusions,   225      Recommendations,   230     APPENDIXES         A   Committee Member and Staff Biographies   249     B   Speakers and Participants at Meetings and Site Visits   263     C   List of White Papers Prepared for the Applications Workshop   276     D   Glossary and Acronym List   278