Performance Analysis and Distributed Computing

Dagstuhl Seminar No. 02341
August 19-23, 2002
Schloss Dagstuhl
Wadern, Germany

Agenda


Please see the official Dagstuhl web site for more information.

Note that this schedule is preliminary and is likely to change. Please send email to the organizers if you have a correction.


Monday, August 19, 2002
Session 0: Introduction
08:30 - 09:00 Welcome and Opening Remarks
Hans-Michael Gerndt, Adolfy Hoisie, Al Malony, Vladimir Getov, Barton Miller
Session 1: Grid Programming (Bart Miller)
09:00 - 10:30 Domain Decomposition in Grid Appilcations
Cliff Addison - FECIT
Speculative and Optimistic Grid Computations
Craig Lee, Aerospace Corp. - El Segundo, California
Pipelining for Performance Optimizations in Scientific Computing
Gudula Rünger, TU Chemnitz
Break
Session 2: Grid Applications (Bart Miller)
11:00 - 12:00 Performance Analysis of Interactive Grid Applications
Marian Bubak, AGH - Krakow
Performance Analysis for Grid Systems and Applications
Hans Christian Hoppe, Pallas GmbH - Brühl
Lunch
Session 3: Grid Projects (Bart Miller)
14:30 - 16:00 The UNICORE Grid and its Options for Performance Analysis
Mathilde Romberg, Forschungszentrum Jülich GmbH
Performance Analysis on the Grid - Activities in the CROSSGRID Project
Roland Wismüller, TU München
Performance Analysis in the Realitemgrid Project
Graham Riley, Manchester University
Break
Session 4: Grid Performance (Bart Miller)
17:00 - 18:00 Benchmarking the Grid
Robert T. Hood, NASA - Moffett Field
Web Performance, Caching
Arun Iyengar, IBM Almaden Research Center


Tuesday, August 20, 2002
Session 5: Scalable Tools (Al Malony)
09:00 - 10:30 1000 Nodes: Serious Scalability for Parallel Tools
Barton Miller, University of Wisconsin
Scalable Parallel Program Analysis
Dieter Kranzlmüller, University of Linz
Distributed Performance Analysis
Hans-Michael Gerndt, TU München
Break
Session 6: Performance Data and Management (Al Malony)
11:00 - 12:00 Uniform Resource Visualization - Software and Services
Diane Rover, Iowa State University
Comparative Profiling with PPerfDB
Karen Karavanic, Portland State University
Lunch
Session 7: Monitoring and Instrumentation (Al Malony)
14:30 - 16:00 Dynamic Instrumentation and Optimization of Distributed Java Applications
Paul H.J. Kelly, Imperial College - London
Java Virtual Machine Profiler Interface (JVMPI)
Michael Philippsen, Universität Karlsruhe
Monitoring of Grid Applications
Norbert Podhorszki, Hungarian Academy of Sciences
Break
Session 8: Mobile Agents (Al Malony)
16:30 - 17:30 Performance Evaluation of Mobile-Agent Middleware
Marios Dikaiakos, University of Cyprus
Mobile Agents for Cluster and Grid Computation
Beniamino DiMartino, Seconda Universitá di Napoli


Wednesday, August 21, 2002
Session 9: Architecture (Vladimir Getov)
09:00 - 10:30 Performance Characteristics of Cellular Architectures
Jose Castanos, IBM T.J. Watson Research Center
The Simulator-guided Memory Analyzer-SIGMA
Luiz DeRose, IBM T.J. Watson Research Center
Collective Communication Pattern on the Quadrics Interconnection Network
Salvador Coll Arnau, Univ. Politèc. de Valencia
Break
Session 10: Architecture (Vladimir Getov)
11:00 - 12:00 Techniques for Comparing Performance of Parallel and Distributed Systems
John Gurd, Manchester University
Pinpointing and Quantifying Scalability Bottlenecks in DSM Multiprocessors
Josep Torrellas, Univ. of Illinois - Urbana
Lunch
Excursion and Open Discussions


Thursday, August 22, 2002
Session 11: Modeling (Michael Gerndt)
09:00 - 10:30 Performance Modelling of Parallel Applications and Architectures
Adolfy Hoisie, Los Alamos National Laboratory
Performance Prediction on Large Systems
Darren Kerbyson, Los Alamos National Laboratory
Performance Modelling of Task-Parallel Programs
Thomas Rauber, Martin-Luther-Universität Halle-Wittenberg
Break
Session 12: Scheduling and Resource Management (Michael Gerndt)
11:00 - 12:00 Economic Scheduling
Carsten Ernemann, Universität Dortmund
Performance and Fungibility of Resources
Scott Baden, Univ. California - San Diego
Lunch
Session 13: Performance Prediction (Michael Gerndt)
14:30 - 16:00 Performance Prediction and Scheduling on Heterogeneous Grids
Larry Carter, Univ. California - San Diego
Automatic, Symbolic Performance Prediction of Distributed-Memory Programs
Arjan van Gemund, Delft University of Technology
Break
Session 14: Scheduling and Resource Management (Michael Gerndt)
16:30 - 17:30 Agent-supported Performance Management in Grid and Peer-2-Peer Systems
Omer F. Rana, University of Wales - Cardiff
Programming the Grid with Distributed Objects
Thierry Priol, IRISA - Rennes


Friday, August 23, 2002
Session 15: Performance Analysis Tools (Adolfie Hoisie)
09:00 - 10:30 AKSUM: A Tool for Multi-Experiment Automated Searching of Bottlenecks in Parallel and Distributed Programs
Thomas Fahringer, Universität Wien
Advances in the TAU Performance System
Allen Malony, University of Oregon
Automatic Analysis of Hybrid Programs with EXPERT
Bernd Mohr, Forschungszentrum Jülich GmbH
Break
Session 16: Peformance Analysis (Adolfie Hoisie)
11:00 - 12:00 Automatic Performance Analysis of Distributed Applications with Kappa-Pi
Tomàs Margalef, Autonomous University of Barcelona
Performance Analysis of Hybrid Programming Models
Vladimir Getov, Univ. of Westminster - London
Lunch

Last modified: Sun Aug 18 13:05:58 CDT 2002 by bart