Proceedings of the Second Workshop on Cluster-Based Computing

Saturday 6 May 2000
Santa Fe, New Mexico


We take great pleasure in introducing these papers presented at the Second International Workshop on Cluster-Based Computing, held on 6 May 2000 in Santa Fe, New Mexico in conjunction with the 14th ACM-SIGARCH International Conference on Supercomputing.

Clusters are now the platform of choice in scalable, high-performance computing, and the current industrial efforts to further reduce communication latencies by bringing interconnects closer to processors promises will further enhance their importance as an architecture. Clusters thus play a critical role in defining current and future supercomputing. With this workshop, we hope to bring together leaders in cluster-based research in a productive forum both for exchanging knowledge and for fostering collaborations.

We received many good submissions on a wide variety of topics. We accepted nine of these submissions on a range of topics from application studies and cluster management to runtime systems and performance tuning. We hope that you find them as enjoyable and as useful as have we.

We wish to thank the people that helped to make this workshop possible. First, we thank Constantine Polychronopoulos, who encouraged us to organize the first workshop and who gave freely of his help and advice throughout the process. We next thank the authors for contributing their work and for their cooperation and patience with the process. Finally, Alex Veidenbaum, program chair for ICS'00, was helpful in bringing this second conference into being.

Nicholas P. Carter
Assistant Professor
Steven S. Lumetta
Assistant Professor
Department of Electrical and Computer Engineering
Coordinated Science Laboratory
University of Illinois at Urbana-Champaign

Table of Contents

Session I:Application Studies and Management Tools
  "Parallel FEM Simulation of Crack Propagation on the AC³ Velocity Cluster"
George Coulouris, Gerd Heber, David Lifka, Keshav Pingali, David Schneider, Paul Stodghill, Paul Wawrzynek, and John Zollweg
Cornell University
  "Scalability and Performance of CTH on the Computational Plant"
Ron Brightwell, H. Eliot Fang, and Lee Ward Sandia National Laboratories
  "The Scalable System Administrator: via C3 & M3C Tools"
Ray Flanery, Al Geist, Brian Luethke, Jens Schwidder, and Stephen Scott
Oak Ridge National Laboratory
Session II: Software Infrastructure
  "Overlapping Data Transfer with Application Execution on Clusters"
Karen L. Reid and Michael Stumm
University of Toronto
  "Evaluation of Two BSP Libraries through Parallel Sorting on Clusters"
Christophe Cérin, Université de Picardie Jules Verne
Jean-Luc Gaudiot, University of Southern California
  "Latency Reduction using a Polling Scheduler"
Jørgen S. Hansen and Eric Jul
University of Copenhagen
Session III: Performance Analysis and Tuning
  "An Agent-Based Architecture for Tuning Parallel and Distributed Applications Performance"
Sherif A. Elfayoumy and James H. Graham
University of Louisville
  "Performance Analysis of PC-CLUMP Based on SMP-Bus Utilization"
Taisuke Boku, Ken'ichi Itakura, and Shigehiro Yoshikawa, University of Tsukuba
Masaaki Kondo, University of Tokyo
Mitsuhisa Sato, Real World Computing Partnership
  "Parallel Program Traces for Accurate Prediction of Proposed Cluster Performance"
William E. Cohen, University of Alabama at Huntsville
W. D. Garrett, Mitre Corporation
R. K. Gaede, University of Alabama at Huntsville