Council on Academic Computing
March 28, 2007
Ralph Roskies, Co-Scientific Director, of the Pittsburgh Supercomputing Center, discussed the future of
high-performance computing. He focused
upon petascale computing. Increases in computing power will be achieved
not through novel computing architecture, as once believed, but through
additional processors. The petascale machines will present several operational
challenges, such as additional power requirements and load balancing issues.
The NSF will award $200 million to a consortium to build a
high-performance computer with sustained, petascale
capability. The machine, which will be
operational by 2011, will enable scientists to model phenomena that they have
heretofore been unable to model adequately.
Other U.S. agencies, such as the Department
of Energy, and other countries, such as Japan and the member states of the
European Union, are sponsoring similar initiatives.
Members of the Council discussed several other issues:
- the academic computing
- multidisciplinary simulation
and modeling project
- multidisciplinary energy
- Center for Molecular and