| Highlights of the ASCAC March 2006 Meeting |
| Dr. Jill Dahlburg |
| Dr. Ray Orbach, Department of Energy Office of Science (DOE SC) Director, termed the 2006 State of the Union address as "historic", because President Bush announced the American Competitiveness Initiative that is intended to boost critical basic research in the physical sciences by doubling funding to the DOE SC, the National Science
Foundation, and the National Institute of Standards and Technology core over the next decade. During his Advanced Scientific Computing Advisory Committee (ASCAC) meeting overview presentation in March 2006, he called this commitment a testimony to ASCAC leadership. Ultrascale scientific computing, the centerpiece initiative of the Office of Advanced Scientific Computing Research (ASCR), is the second priority in the DOE SC 20 year plan, "Facilities for the Future of Science: a Twenty-Year Outlook". Dr. Orbach noted that ASCR will receive a 35.7% funding increase this year for research towards this initiative;
this is the largest percentage rise within DOE SC. |
| In his overview of ASCR following Dr. Orbach's opening talk, Dr. Michael Strayer (DOE SC ASCR Associate Director) expressed that his aim for ASCR with SciDAC is that they will form the nexus of a global village for high end computing (HEC), which will enable whole new levels of scientific discovery to be realized. The goal is to sponsor the best-in-class HEC research and capabilities – including facilities, simulation, middleware and collaboratory tools, applied mathematics, and computational and computer science – to advance forefront DOE SC science and technology, with an emphasis on the petascale. |
| Both successes and challenges abound. Dr. Strayer described the recent discovery of a new instability important to core collapse supernovae (figure 1) in two and three dimensions. Briefing about progress toward developing Blue-Gene/P and /Q, Prof. Rick Stevens (Argonne National Laboratory; University of Chicago) noted that significant innovations are needed in order to achieve sustained petascale computing in the next decades (see table). |
Figure 1. The Supernova Shock Wave Instability (SASI) in three dimensions.
|
| Future Scaling Without Innovation |
| Scaling current peak performance numbers for various architectures and allowing system peak doubling every 18 months |
| |
Projected Year |
BlueGene/L |
Earth Simulator |
Mare Nostrum |
| 250 TF |
2005 |
1.0 MWatt |
100 MWatt* |
5 MWatt |
| 1 PF |
2008 |
2.5 MWatt |
200 MWatt* |
15 MWatt |
| 10PF |
2013 |
25 MWatt* |
2000 MWatt* |
150 MWatt* |
| 100 PF |
2020 |
250 MWatt* |
20,000 MWatt* |
1500 MWatt* |
|
| Networking researchers are preparing for the tremendous quantities of future HEC data: ESNet has stably connected SC assets to scientists worldwide since 1987, and ASCR-developed middleware recently enabled a 10-day flow of 500 TBytes from CERN to its associated sites. The ASCR plans to participate in the new DOE Global Nuclear Energy Partnership by providing 100-Tflop computing resources for simulation of reactor designs, reprocessing, fuel fabrication, and waste disposal. |
| Dr. Orbach issued two new charges to the ASCAC during the March meeting. ASCAC members Gordon Bell (Microsoft®) and James Hack (National Center for Atmospheric Research) are spearheading a committee response to the first charge, on the topic of ASCR facilities science-based performance metrics. They expect to provide an interim report at the ASCAC August 8-9, 2006 meeting, and a final report at the November 8-9, 2006 ASCAC meeting. ASCAC members Stephen Wolff (Cisco®) and Ellen Stechel (Sandia National Laboratories) are leading development of the response to the charge on the roles and efficacy of networking and networking research within the DOE SC. The interim report for this second charge is to be presented at the November 2006 ASCAC meeting, with the final report due in the autumn of 2007. |
|
Dr. Jill Dahlburg, ASCAC Chair; Naval Research Laboratory, Washington DC. On behalf of the ASCAC
|
Acknowledgements
These highlights were largely summarized from the March 2006 ASCAC Meeting Minutes (Fred O'Hara, recording secretary) |
Further Reading
All talks refer to those presented at ASCAC Meeting, March 15, 2006; http://www.er.doe.gov/ASCR/ASCAC/ASCAC.html |
