WEDNESDAY, APRIL 14
9:00 AM - 10:00 AM
KEYNOTE ADDRESS
The Evolution of Multithreaded Architecture
Burton Smith, Tera Computer Company
10:30 AM - 12:30 PM
SESSION 10
Routing and Broadcasting II
Chair: José D.P. Rolim
University of Geneva
Oblivious Deadlock-Free Routing in a Faulty Hypercube
JinSuk Kim, Eric Lehman, and Tom Leighton, Massachusetts Institute of Technology
Sparse Hypercube -- A Minimal k-Line Broadcast Graph
Satoshi Fujita, Hiroshima University, Japan, Arthur M. Farley, University of Oregon
All-to-All Broadcast on Switch-Based Clusters of Workstations
Matt Jacunski, P. Sadayappan, and D.K. Panda, Ohio State University
VBMAR: Virtual Network Load Balanced Minimal Adaptive Routing
Xicheng Liu, Timothy J. Li, and Wen Gao, Motorola-NCIC Joint R&D Laboratory, Chinese
Academy of Sciences, P.R. China
10:30 AM - 12:30 PM
SESSION 11
Scientific Engineering Systems
Chair: David Martinez
MIT Lincoln Laboratory
Portable Parallel Programming for the Dynamic Load Balancing of Unstructured Grid
Applications
Rupak Biswas, MRJ Technology Solutions, Inc., Sajal K. Das and Daniel J. Harvey,
University of North Texas at Denton, Leonid Oliker, RIACS
Parallel Algorithms for Singular Value Decomposition as Applied to Failure Tolerant
Manipulators
Tracy D. Braun, Anthony A. Maciejewski, and Howard Jay Siegel, Purdue University
A Parallel Adaptive Version of the Block-Based Gauss-Jordan Algorithm
N. Melab, E-G. Talbi, and S. Petiton, Universite' des Sciences et Technologies de Lille,
France
Sparse Matrix Block-Cycle Redistribution
Gerardo Bandera and Emilio L. Zapata, University of Malaga, Spain
A New Approach to Parallel Dynamic Partitioning for Adaptive Unstructured Meshes
Gerd Heber, University of Delaware, Rupak Biswas, NASA Ames Research Center, Guang R. Gao,
University of Delaware
An Object-Oriented Environment for Sparse Parallel Computation on Adaptive Grids
S. Filippone, IBM Semea, Italy, M. Colajanni and D. Pascucci, University of Rome "Tor
Vergata," Italy
10:30 AM - 12:30 PM
SESSION 12
Performance
Chair: Jan Cuny
University of Oregon
A Network Status Predictor to Support Dynamic Scheduling in Network-Based Computing
Systems
JunSeong Kim and David J. Lilja, University of Minnesota
Performance Evaluation of an Infrastructure for Worldwide Parallel Computing
Thomas T. Kwan and Daniel A. Reed, University of Illinois
BRISK: A Portable and Flexible Distributed Instrumentation System
Aleksandar Bakic, Matt W. Mutka, and Diane T. Rover, Michigan State University
An Efficient Logging Algorithm for Incremental Replay of Message-Passing Applications
Franco Zambonelli, Universita di Modena e Reggio Emilia, Italy, Robert. H.B. Netzer, Brown
University
Lazy Logging and Prefetch-Based Crash Recovery in Software Distributed Shared Memory
Systems
Angkul Kongmunvattana and Nian-Feng Tzeng, University of Southwestern Louisiana
Visualization and Performance Prediction of Multithreaded Solaris Programs by Tracing
Kernel Threads
Magnus Broberg, Lars Lundberg, and Hakan Grahn, University of Karlskrona/Ronneby, Sweden
2:00 PM - 4:00 PM
PANEL DISCUSSION
Prospects for High Performance Computing with Java and Middleware
MODERATOR
Gul Agha, University of Illinois at Urbana-Champaign
PANELISTS
Dennis Gannon, Indiana University
Satoshi Matsuoka, Tokyo Institute of Technology
Traian Muntean, University of Marseille, France
Nalini Venkatasubramanian, University of California at Irvine
By supporting portability, heterogeneity, and security, Java provides a new level of
abstraction for Internet computing. Middleware architectures allow a more modular,
component-based approach to building distributed systems. But these technologies show
relatively poor performance, with the Java approach faring much worse than conservative
middleware approaches. To what extent is high performance a conflicting goal with
portability and interoperability? Is high performance computing feasible without special
purpose high-speed networks? Can virtual machine code run as fast as compiled code? Is
Java inherently slow? Are CORBA-like architectures a preferable route to network based
high performance computing? Or will Java RMI, JavaBeans and Jini replace CORBA and DCOM?
Or will application specific solutions render general approaches irrelevant for important
commercial markets such as multimedia, enterprise integration, and federated databases.
4:40 PM - 5:00 PM
Industrial Track Presentation
Mercury Computer Systems, Inc.
Web site: http://www.mc.com
Delivering On Standards: Balancing Portability and
Performance
John Robinson, Media Relations Manager
5:00 PM - 7:00 PM
SESSION 13
Mesh Architecture
Chair: Afonso Ferreira
CNRS-I3S-INRIA, France
Better Deterministic Routing on Meshes
Jop F. Sibeyn, Max-Planck-Institut fuer Informatik, Germany
Efficient Parallel Algorithms for Selection and Multiselection on Mesh-Connected Computers
Hong Shen, Griffith University, Australia
Constant-Time Algorithm for Medial Axis Transform on the Reconfigurable Mesh
Amitava Datta, University of Western Australia, Australia
2.5n-Step Sorting on n*n Meshes in the Presence of o(n^{1/2}) Worst-Case Faults
Chi-Hsiang Yeh, Behrooz Parhami, Hua Lee, and Emmanouel A. Varvarigos, University of
California at Santa Barbara
The Recursive Grid Layout Scheme for VLSI Layout of Hierarchical Networks
Chi-Hsiang Yeh, Behrooz Parhami, and Emmanouel A. Varvarigos, University of California at
Santa Barbara
5:00 PM - 7:00 PM
SESSION 14
Signal Processing
Chair: George Cybenko
Dartmouth College
Multi-Threaded Design and Implementation of
Parallel Pipelined STAP on Parallel Computers with SMP Nodes
Wei-keng Liao, Syracuse University, Alok Choudhary, Northwestern University, Donald Weiner
and Pramod Varshney, Syracuse University
A Parallel Phoneme Recognition Algorithm Based on Continuous Hidden Markov Model
Sang-Hwa Chung, Min-Uk Park, and Hyung-Soon Kim, Pusan National University, Korea
Load Adaptive Algorithms and Implementations for the 2D Discrete Wavelet Transform on
Fine-Grain Multithreaded Architectures
Ashfaq A. Khokhar, Gerd Heber, Parimala Thulasiraman, and Guang R. Gao, University of
Delaware
Application of Parallel Processors to Real-Time Sensor Array Processing
David R. Martinez, MIT Lincoln Laboratory
Mapping Media Streams Onto a Network of Servers
Reinhard Lueling, University of Paderborn, Germany
A Systolic Algorithm to Process Compressed Binary Images
Fikret Ercal, Mark Allen, and Hao Feng, University of Missouri at Rolla
5:00 PM - 7:00 PM
SESSION 15
Program Optimization, Resource Allocation, Scheduling
Chair: Fran Berman
University of California at San Diego
Optimizations for Language-Directed Computational Steering
Jeffrey Vetter, University of Illinois, Karsten Schwan, Georgia Institute of Technology
Optimization Rules for Programming with Collective Operations
Sergei Gorlatch, Christoph Wedler, and Christian Lengauer, Universitaet Passau, Germany
A Flexible Clustering and Scheduling Scheme for Efficient Parallel Computation
S. Chingchit and M.J. Kumar, Curtin University of Technology, Australia, L.N. Bhuyan,
Texas A&M University
Mechanisms for Just-in-Time Allocation of Resources to Adaptive Parallel Programs
Arash Baratloo, New York University, Ayal Itzkovitz, New York University and Technion,
Israel, Zvi M. Kedem and Yuanyuan Zhao, New York University
Exploiting Application Tunability for Efficient, Predictable Parallel Resource Management
Fangzhe Chang, Vijay Karamcheti, and Zvi Kedem, New York University
Supporting Priorities and Improving Utilization of the IBM SP Scheduler Using Slack-Based
Backfilling
David Talby and Dror G. Feitelson, The Hebrew University of Jerusalem, Israel
COMMERCIAL EXHIBITS