Grand Challenge Proposal Guidance

Submit your proposals via email to: computing-grand-challenge [at] llnl.gov (computing-grand-challenge[at]llnl[dot]gov)

Watch committee chair Erik Draeger give a brief overview of Computing Grand Challenge program objectives before describing the proposal writing and review process (internal only): https://llnlfed.webex.com/webappng/sites/llnlfed/recording/8dc6c0fb1f77103c9dfe00505681f366/playback

Please read the content and length guidelines closely, proposals that omit requested information will be scored lower or rejected outright.

Proposal Content and Length Guidelines 

This call is for High Performance Computing (HPC) allocations on Lassen and/or Dane/Ruby. Two types of awards are available: 

  • Tier 1: Highest visibility and impact, allocations of ~100k-200k node-hours 
  • Tier 2: High visibility and impact, allocations of ~25k-50k node-hours 

All proposals should be written assuming Tier 2 allocations. A detailed justification for how additional resources would enhance the impact and outcome will be used to select a small number of proposals for Tier 1 awards. Proposals should be put through IM before submission.

Projects will be evaluated using three primary criteria: (1) the scientific importance and visibility of the results, including the publications that are likely to result; (2) the significance and impact of the computational approach used, including any novel algorithms, methods or implementations; and (3) the quality of the HPC research plan. The quality and expertise of the team and any external collaborations as well as the alignment with the laboratory S&T strategic vision will also be evaluated. 

Proposals should be written at an appropriate level of detail for a technically diverse review team (e.g., as for Scientific American). The proposals are limited in length to a cover page and six pages of text (eight for continuing proposals) plus references, using the following format: 

Cover Page [1 page] 

This section should include project title, the names and contact information for the principal investigator and the members of the research team (internal and external) who will require access to the computing resource, and the machine(s) on which the proposed runs will take place (Dane, Lassen, and/or Ruby). 

Proposers must also describe, for each member of their team, the level of effort and the non-programmatic source of funds that will be devoted to this project. Please state whether LDRD support will be sought for the effort of some team members. Foreign nationals requiring access to any LLNL computing resource must have an approved computer security plan in place prior to being granted access. Please contact the LC Hotline (925-422-4531) for details on current policy regarding foreign national access. 

Significance and Impact of Proposed Work [2-3 pages] 

This section, of utmost importance during the selection process, should describe clearly the significance and singular impact of the anticipated scientific result and the computational approach taken. Examples include impact on the following: a scientific, computational or engineering discipline or field of research; demonstration and acceptance of high-fidelity numerical simulations as a scientific or decision-making tool; missions of Laboratory, NNSA or Department of Energy; demonstration of the value of supercomputing on a new field; and impact on the future development of high-performance computing methods or technology. Included in this section should be a description of what publications are expected to result. 

High Performance Computing Research Plan [1-2 pages] 

This section should outline a detailed research plan for this Grand Challenge campaign at the Tier 2 award level, including the computational approach; codes to be used (including evidence that they can run efficiently at the proposed scale); estimated job mix (number of jobs and number of nodes per job); and an approximate timeline of usage between January 1, 2025 and December 31, 2025. Please list any specific architecture, memory, disk, and archival storage (HPSS) requirements. Projects requesting time on Lassen must show evidence that they can run efficiently on the GPU. 

Justification for a Tier 1 Award [1 page] 

This optional section should describe how Tier 1 award level resources would increase the scientific importance and visibility of the results and/or the significance and impact of the computational approach. It should be clear to the reader what specific additional benefits would be gained from the larger allocation, e.g. additional and/or higher impact publications, demonstrating new HPC capabilities or algorithms at scale, etc. 

Summary of Results & Publications from Previous Computing Grand Challenge Work/HPC Experience [2 pages maximum] 

This section is for previous Computing Grand Challenge awardees only. Please summarize the results and provide references for all publications stemming from previous CGC awards. In cases where the proposed work was not fully completed and/or planned publications were not submitted, please explain why.