NSF looks to future-proof high-performance computing
Connecting state and local government leaders
The National Science Foundation said it is looking for a high-performance system that can grow into a leadership-class computing facility.
The National Science Foundation is looking for home for the next generation of high-performance computing.
In a request for proposals to colleges and universities, NSF said it is looking for a high-performance computing system that can grow into a leadership-class computing facility.
Given scientists’ increasing demand for computational resources and the rapidly evolving technology environment, NSF wants to ensure large-scale computational support will be available for researchers whose long-term projects extend beyond the lifetime of a single HPC system.
“The deployment of a leadership-class facility can provide this visibility into longer-range plans for large-scale computing, enabling science teams to confidently adapt, evolve, and create new approaches in their computational techniques to push the frontiers of science and engineering,” NSF said in its RFP.
As a first step toward such a facility, NSF is asking for a Phase 1 system that supports a broad range of computational and data-intensive workflows requiring the highest capabilities in terms of scale, throughput and data analytics. The system must offer at least a two- to three-fold time-to-solution performance improvement over the Blue Waters supercomputer installed at the University of Illinois at Urbana-Champaign.
The petascale-level, Cray supercomputer-based Blue Waters is deployed for “grand challenge” problems in areas as diverse as astrophysics, space weather modeling, neuroscience, advanced materials and drug discovery.
Besides delivery of a robust HPC system and facility, NSF also wants information on how such a system can lead to the design of a Phase 2 leadership-class system. Proposers must describe how the Phase 1 system will support as future frontier applications and connect with external resources and national networks.
Subject to availability of funds, a total of $60 million is available for this five-year project.
More information is available here.
NEXT STORY: IoT fuels next-gen law enforcement