Sunday: Impact of Evolving Hardware on Applications
Session Chair: Robert George
- Danny Cohen, Myricom, Clustering techniques for multicomputers
- Andrew Chien, University of Illinois, Urbana-Champaign, A perspective on software communication overhead
Monday: Real-Time
Morning
Session Chair: Arkady Kanevsky
- Ashok K. Agrawala, University of Maryland, Role of Time in Systems
- Arkady Kanevsky, MITRE, The issues for software, hardware, and networks to enable technology to support real-time system of systems
Photographs at 9:45 AM
Evening
Session Chair: Ashok Agrawala
- Jean-Dominique Decotignie, Event driven vs. time driven real-time systems
- Kirstie Bellman, DARPA, real-time, virtual time, and behavioral time
- Chris Landauer, The Aerospace Corporation, What if we take time seriously?
Tuesday: Petaflops/Futureflops
Morning
Session Chair: Anthony Skjellum
- David Greenberg, Sandia National Laboratories, Letting MPPs out of the box
- Don Heller, Ames Laboratory, Statistical Variations in Performance Measurement
- Anthony Skjellum, Many Level Multicomputers, Active Networks, and Bumps in Wires
Evening
Session Chair: Peter Kogge
- Peter Kogge, University of Notre Dame, Processing memory: chips to petaflops (and deep space too)
- INFORMAL PANEL - Petaflops/Futureflops
Wednesday: Applications Day
Business Meeting after Evening Panel
Morning
Session Chair: Edward Luke
- Olaf M. Lubeck, Los Alamos National Laboratory, Scalability issues in large scale technical computing from the ASCI program
- Thomas H. Cormen, Dartmouth College, The spin on parallel disks: They may be a RAID to you, but they're just a bunch of disks to me
- Paul Plassman, Argonne National Laboratory, The impact of evolving architectures on unstructured mesh calculations
Evening
- Paul Woodward, University of Minnesota, Programming techniques to exploit advantages of SMP clusters and DSM
- Evening Panel: The impact of evolving architectures on applications
Panel Chair: Pat Worley
- Thomas H. Cormen
- Olaf M. Lubeck
- Paul Plassman
- Pat Worley
- Edward Luke
- Paul Woodward
Thursday: Abstract vs. Performance in Software
Morning
Session Chair: Anthony Skjellum
- Andrew Lumsdaine, University of Notre Dame, Must abstraction of barriers be barriers to performance?
- Lenore Mullins, SUNY - Albany, The psi calculus and seeing the world in arrays: unifying the specification and efficient implementation of array-based computation or can F90 ever perform as well as F77 and what HPF?
- William Gropp, Argonne National Laboratory, Do we need memory abstractions?
Evening
Session Chair: Andrew Lumsdaine
- Bob Numrich, Cray Research - SGI, The F-- Extension to Fortran 90: An alternative model for parallel Programming
- Jim Browne, University of Texas, Intelligent agent implementation of dynamic computations on metacomputers?
|