 |
Salishan 2012
Theme: 2015 Petascale Systems: Next Steps
Keynote Address: DOE's Plan for Exascale Computing, Bill Harrod, DOE/Office of Science; ThucHoang, DOE/NNSA-ASC
Session 1: Computing at Scale: how are we doing at mixing Big Data with traditional Parallel Scientific Computing?
- Where is coherency needed? Can we mix High Throughput and Data-intensive computing with traditional Parallel Scientific Computing on large HPC systems?, Ilene Carpenter, National Renewable Energy Laboratory
- Big iron for big data and other unnatural acts,
Steve Plimpton, Sandia National Laboratories
- The Interdependence of Data, Modeling, Simulation and Analytics in the Exascale Era, James Sexton, IBM Thomas J. Watson Research Center
- Data Analysis and Data Intensive Computing at LLNL, Daniel Laney, Lawrence Livermore National Laboratory
Session 2: Error Avoidance, Detection and Recovery
Working Dinner/Speaker
Session 3: Application/Machine Hierarchy
Session 4: Performance and Portability on Low Power Processors
- Processing Nearer to Memory, J. Thomas Pawlowski, Chief Technologist & Micron Fellow, Micron
- Improving Permeability in System Architecture, Doug Carmean, Intel Fellow and Director of the Efficient Computing Lab, Intel Corporation
- Towards High Performance Computing Application Energy Efficiency,
James H. Laros, Sandia National Laboratories
- Nanophotonic Interconnection Networks for System Wide Performance-Energy Optimized Computing, Keren Bergman, Columbia University
Session 5: Experiences and Lessons learned with New Architectures
- Tracking the Effects of Technology and Architecture on Energy through the Top500, Green500, and Graph500, Peter Kogge, University of Notre Dame
- Effective performance of K computer in Life Science Applications, Ryutaro Himeno, Next Generation Computational Science Program, RIKEN
- Peta-scale GPU applications on TSUBAME 2.0, Takayuki Aoki, Global Scientific Information and Computing Center, Tokyo Institute of Technology
- Aggressively Applying Lessons from the Cell Processor to Emerging Exascale Computing Technologies, Paul R. Woodward, Laboratory for Computational Science & Engineering, University of Minnesota
*will not be posted
|