NERSCPowering Scientific Discovery Since 1974

Call for Proposals: High-Impact Science at Scale on Cori

March 18, 2019 by Brandon Cook

NERSC is seeking proposals to conduct high-impact science campaigns using NERSC’s Cori supercomputer at scale. A successful proposal would require the use of at least 2,000 KNL nodes to solve a problem, with a preference for proposals that exploit Cori’s unique capabilities (e.g., using all or most of the 9,688 KNL nodes or making heavy use of the burst buffer). Previous awards were in the range of 10-100M NERSC hours. Data-intensive projects requiring large scale analytics (statistics, machine learning, deep learning) are encouraged to apply.

The number of awarded projects will be determined based on the number, quality and diversity of responses, as well as by the aggregate amount of time requested. This call will remain open on an ongoing basis and submissions will be evaluated every two weeks. Time awarded through this program will be available for use through September 30, 2019. (Note that a new repository will be created for this campaign; awards are not intended to simply augment an existing project.)

 To fill out the online response form you will need to enter:

  • A descriptive title for the proposed project

  • PI name, institution, and email address

  • Short description of the scientific result you hope to achieve and its relevance to the DOE Office of Science mission, which includes advancing the state of the art in large-scale scientific computation and data analytics

  • Description of your code and evidence that it is capable of running efficiently at scale on KNL

  • Requirements from NERSC (e.g., software libraries, tools, workflow support)

  • NERSC Hours requested

  • Data requirements beyond those provided by default

  • Proposed usage timeline (can be an approximate estimate, but please include any known deadlines)

To apply, enter the above information at the following form:

 Allocation awards made through this program are valid through September 1, 2019.