In case you missed it, the National Science Foundation issued the request for proposals (RFP) for the next ‘Towards a Leadership-Class Computing Facility – Phase 1″ last week. It’s for $60 million and, among other things, specifies a system to be at least a two- to three-fold time-to-solution performance improvement over the Blue Waters machine at the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign’s (UIUC).
The RFP is for the acquisition and deployment of a HPC system (Phase 1) with the option of a possible future upgrade to a leadership-class computing facility. As described by NSF, the phase 1 system would be a robust, well-balanced, and forward-looking computational asset for a broad range of research topics and “also serve as an evaluation platform for testing and demonstrating the feasibility of an upgrade to a leadership-class facility five years following deployment.”
Obviously the high stakes are likely to draw many responses. Bill Gropp, interim director of the NCSA, told a local Urbana news outlet “we plan to win.” No doubt others do too. The fact that there’s a possible Phase 2 called out in the RFP suggests NSF is taking a slightly longer term planning approach than is typical to the competition.
- A detailed acquisition plan for deploying a reliable and well-balanced HPC system with at least two- to three-fold time-to-solution performance improvement over the current state of the art, the University of Illinois at Urbana-Champaign’s (UIUC)Blue Waters system, for a broad range of existing and emerging computational and data intensive applications;
- A thorough operations plan for the Phase 1 system to ensure that it will serve as an effective computational tool for the broad scientific and engineering community, and for the Nation at large;
- A detailed three- to five-year project plan for scientific and technical evaluation of the Phase 1 system that will lead to an upgrade design of a leadership-class system, called the Phase 2 system, as well as the physical facility that will host it: the Phase 2 system is expected to have a ten-fold or more time-to-solution performance improvement over the Phase 1 system;
- Clear and compelling science and engineering use cases, as well as detailed strategic project goals for a leadership-class computing facility; and
- A persuasive articulation of educational and industry outreach, and the achievement of other broader societal impact goals, in the long-term strategic plan for the leadership-class computing facility.
The $60,000,000 awarded in FY 2018 will be used “to fund one award” and at least “95% of the proposal amount” should be for the system acquisition cost. Up to $2,000,000 in additional funds “are anticipated to be available in FY 2019” planning activities associated with the conceptual design phase for Phase 2 of the award.
Note that this solicitation requests proposals for the acquisition and operation of a Phase 1 system as well as a project plan for the design of a potential upgrade or replacement to a leadership-class computing facility at the end of the five-year deployment period, subject to the availability of funds. Support for subsequent preliminary design and final design phases for Phase 2 will be provided in separate funding actions.
Link to RFP: https://nsf.gov/pubs/2017/nsf17558/nsf17558.htm