Third Workshop on Massively Parallel Processing (WMPP)

Saturday 26 April 2003 at the

International Parallel and Distributed Processing Symposium

IPDPS'03

Nice, FRANCE

Message from the General and Program Chairs

This year's workshop builds on the success of the First and Second Workshops on Massively Parallel Processing, held as part of IPDPS'01 and IPDPS'02. The first and second workshops featured invited keynote talks by Peter Kogge of Notre Dame, and David Bader of the University of New Mexico, respectively. Each workshop had around 10 papers, and following the usual practice at IPDPS, all papers were published with the paper abstracts in a hardcopy volume and the complete paper in an accompanying CD-ROM.

As you can see from the schedule below, the workshop has grown to 15 papers this year, and features an invited keynote talk by Thomas Sterling (Cal Tech and JPL) that reflects recent new initiatives by Caltech, Cray, and other institutions under the DARPA High Productivity Computing Systems Program.

We hope that you enjoy the workshop!

Schedule

Session 1 (Opening Session) -- 8:00-9:10
(Session Chair -- Johnnie Baker, Kent State University)

• Opening Remarks
  Johnnie Baker, Kent State University
• KEYNOTE ADDRESS: Next Generation MPP Architecture in the Petaflops Decade
  Thomas Sterling, California Institute of Techology & NASA Jet Propulsion Laboratory
  Abstract: The MPP has replaced all other previous forms of parallel processing architecture to dominate high end computing for more than a decade. But low efficiencies for some important classes of applications experienced with the largest ASCI MPPs and the rapid growth of commodity cluster computing with their high performance-to-cost challenge the future of the MPP even as it leads the world in sustained performance with the Japanese Earth Simulator. Commodity clusters (including Constellations) now comprise approximately half of the Top-500 list of computers measured by the Linpack benchmark and benefit from much of the same technology employed by MPPs without the development costs or lead time. However, the source of the strength of clusters, their sole use of COTS subsystems, is also their principal weakness and is not shared by MPPs. The future generation of MPPs as represented by the DARPA High Productivity Computing Program is transforming high end computer architecture with dramatic new strategies addressing the drivers of performance degradation resulting from low efficiency and exploiting new opportunities in VLSI computing structures as yet untapped by conventional microprocessor architecture to deliver unprecedented performance-to-cost. This talk will present a set of innovations in MPP architecture being explored by the Cray Cascade project that may catalyze a renaissance in high end computer architecture and determine the next generation of MPP architecture as it enters the trans-Petaflops performance regime of the next decade.

Break -- 9:10-9:25

Session 2 (Architecture) -- 9:25-10:05
(Session Chair -- David Andrews, University of Kansas)

• A Fine-Grained Parallel Pipelined Karhunen-Loeve Transform
  Martin Fleury, Bob Self, Andy Downton, University of Essex
• Trident: Technology-Scalable Architecture for Data Parallel Applications
  Stanislav Sedukhin and Mostofa Soliman, University of Aizu

Break -- 10:05-10:30

Session 3 (Architecture) -- 10:30-12:05
(Session Chair -- Robert Walker, Kent State University)

• Parallel Cellular Programming for Devloping Massively Parallel Emergent Systems
  Domenico Talia, Universita della Calabria
• Architectural Frameworks for MPP System on a Chip
  David Andrews and Douglas Niehaus, University of Kansas
• Importance of SIMD Computation Reconsidered
  Will Meilander, Johnnie Baker, and Mingxian Jin, Kent State University
• Multiple Instruction Stream Control for an Associative Model of Parallel Computation
  M. Scherger, J. Baker, and J. Potter, Kent State University
• Implementing a Scalable ASC Processor (*)
  Hong Wang and Robert A Walker, Kent State University

Break (lunch on your own) -- 12:05-1:30

Session 4 (System Management in MPP) -- 1:30-3:10
(Session Chair -- Jie Wu, Florida Atlantic University )

• System Management in the BlueGene/L Supercomputer
  The IBM BlueGene/L Team, IBM
• An Executable Analytical Performance Evaluation Approach for Early Performance Prediction
  A. Jacquet, V. Janot, R. Govindarajan, C. Leung, G. Gao, and T. Sterling, University of Delaware
• Automatic Resource Management using an Adaptive Parallel Environment
  David Wangerin and Isaac Scherson, University of California-Irvine
• Partitioning with Space-Filling Curves on the Cubed-Spher
  John Dennis, National Center for Atmospheric Research
• A Decentralized Hierarchical Scheduler for a Grid-based Clearinghouse
  Xavier Percival, Cai Wentong, and Francis Lee Bu Sung, Nanyang Technological University

Break -- 3:10-3:40

Session 5 (Algorithms and Models for MPP) -- 3:40-4:30
(Session Chair -- Johnnie Baker, Kent State University)

• Parallel Algorithms to Find the Voronoi Diagram and the Order-k Voronoi Diagram (*)
  Christian Trefftz and Joseph Szakas, Grand Valley State University
• GCA: A Massively Parallel Model (*)
  Wolfgang Heenes, Rolf Hoffmann, and Klaus-Peter Volkmann, Darmstadt University of Technology
• On Self-similarity and Hamiltonicity of Dual-Cubes
  Changfu Wu and Jie Wu, Florida Atlantic University

Post-Workshop Wrap-up & Planning for Next Year -- 4:30-5:00

Workshop Organizers

• Organizing Committee:
  • General Co-Chairs:
    Johnnie Baker, Kent State University ( jbaker@cs.kent.edu )
    Robert Walker, Kent State University ( walker@cs.kent.edu )
  • Program Co-Chairs:
    Philip A. Wilsey, University of Cincinnati ( paw@ececs.uc.edu )
    Raymond Hoare, University of Pittsburgh ( hoare@ee.pitt.edu )
• Program Committee:
  • Nael Abu-Ghazaleh, SUNY Binghamton
  • Johnnie Baker, Kent State University
  • Thomas Braunl, The University of Western Australia
  • Ray Hoare, University of Pittsburgh
  • Mahmut Kandemir, Penn State University
  • Peter Kogge, Notre Dame University
  • H. J. Siegel, Colorado State University
  • Theo Ungerer, University of Karlsruhe
  • Robert Walker, Kent State University
  • Philip A. Wilsey, University of Cincinnati
• Publicity Committee:
  • Nael Abu-Ghazaleh, SUNY Binghamton
  • Ray Hoare, University of Pittsburgh