The 2015 submission deadline will be announced soon.
NCAR computational scientists are encouraged to submit requests for the next round of NCAR Strategic Capability (NSC) projects to run on the Yellowstone system at NWSC. The deadline for submissions is February 18, 2014, for projects to begin a 12-month computing effort on May 1, 2014. An estimated 100 million core-hours will be available for NSC projects.
To submit a request, fill out the NCAR NSC Allocation Request Form
and attach the required Request Summary document in PDF format.
Because of the competitive nature of these allocations, labs may have chosen to coordinate submissions from each lab. Potential applicants are encouraged to contact their lab's or division's allocation representative prior to submitting a request.
Please send any questions to firstname.lastname@example.org.
NCAR Strategic Capability (NSC) project allocations target large-scale projects lasting one year to a few years (but not indefinitely long) that align with NCAR’s scientific priorities and strategic plans. The NCAR Executive Committee (EC) reviews and approves a set of NSC projects for the coming 12-month period each year. NSC projects requiring more than one year’s allocation must submit continuation requests each year to report progress toward the project’s objectives.
To be considered for an NSC allocation, a proposed project:
- should directly relate to one or more specific priorities in the new NCAR strategic plan. Requests that address the grand challenges in the Path Forward, including the development of global models with limited-area higher-resolution grids and the use of data assimilation techniques and statistics with model development and verification, are especially encouraged.
- may be linked to an agency funding award or awards separate from the NCAR Base funding.
- must require significant computational resources, above and beyond the minimum level defined by the NCAR Executive Committee.
- should have a well-defined scope and completion timeline.
- must be technically ready to begin production runs from the start of the allocation period.
An estimated 100 million core-hours—17% of the Yellowstone HPC resource—are allocated among NSC projects each year, plus a similar fraction of analysis and visualization resource use and of GLADE project space.
NSC projects can request allocations of HPSS storage for data related to their activities, within the same budget-constrained growth limits of the CISL archive. However, NSC use of HPSS will follow the data management policies for university and CSL allocations. That is, NSC projects can store data in HPSS for the duration of the project and may ask to keep the data for an additional three years. After that point, the data need to be transferred to a lab allocation or deleted.
The progress of NSC projects will be monitored quarterly to ensure that projects are making sufficient progress and that all projects can complete their work in the remaining allocation period. After each quarter, projects may be subject to losing a portion of their unspent, prorated allocation. Exceptions can be made for projects that identify in their submissions that they have compressed or alternative timetables due to external factors, such as the need to align computations with a planned field campaign or calendar season.
A permanent NCAR staff member from any lab is eligible to be project lead on an NSC request. Labs may choose to implement policies to coordinate the submissions from the lab in each request period. Joint work with university collaborators is eligible, although the project lead must hold a permanent NCAR appointment. Projects that span labs are encouraged, though a single project lead should be identified.
NSC allocations have a minimum request size of 5 million core-hours; the minimum size may be revised by the NCAR EC, and the EC may choose to permit exceptions for cause. There is no maximum size limit, though in practice the review process will attempt to accommodate approximately a dozen large-scale projects each year. Consistent with the NSC objectives, NSC requests should not aggregate many smaller projects out of the same lab to meet the minimum request limit.
NSC requests must prepare Proposal Documents, which should follow the guidance and structure for large University allocation requests. Notably, NSC requests must include a five-page summary along with relevant supporting documentation.
Submit NSC requests by filling out this NCAR NSC Allocation Request Form and attaching the required Request Summary document in PDF format.
NSC requests are subject to two-phase review, with final approval by the NCAR EC.
The first phase is a technical readiness review conducted by CISL’s Consulting Services Group (CSG). A technical readiness review assesses the project’s readiness to conduct productions run, the efficiency of the codes and/or models to be used, and any special needs that require variances from the standard environment or policies on CISL resources. A readiness review also evaluates the accuracy of the computing estimates and confirms that the request meets the NSC minimum levels. CSG may recommend improvements in the technical approach to be used. The CSG readiness assessment is passed along to the next review phase.
The second phase is a scientific and strategic evaluation of the project. The panel conducting this review is composed of representatives of each NCAR lab and program who are appointed by the lab and program directors. This panel evaluates proposals according to several criteria:
- The three criteria used by the CHAP in its evaluation of large university requests – the effectiveness of the methodology, the appropriateness of the research plan, and the efficiency of resource use. The technical readiness assessment provides input into this aspect of the evaluation.
- The scientific appropriateness of the project and its relationship to NCAR strategic priorities.
- Whether the large-scale computational resources needed are commensurate with the project’s strategic priority.
The panel recommends allocation levels and identifies priority ranking for submitted requests. A written review summary is made available to requestors following the final decision on awards. All panel representatives from labs and programs not involved in the proposed project review each request.
In the final phase, the EC approves or agrees to modify the recommendations of the review panel. CISL then establishes the final allocation awards. Should any awarded project encounter issues that require it to stop work or be unable to complete its proposed work, additional allocation awards may be made to unawarded requests in order of priority rank and as resource availability permits.
NSC requests are reviewed once per year. Projects arising too late for NSC consideration will either need to wait, identify bridging allocations from NCAR labs, or apply for startup allocations via the NCAR Director’s Reserve. Such NSC pre-awards must satisfy the criteria for a Director’s Reserve award.
Continuation and reporting
NSC allocations come with commensurate reporting requirements. For those projects requiring more than one year’s allocation, a continuation request will need to be submitted as part of the next year’s NSC request and review process. The continuation request should include a short write-up (approximately one page) for CISL’s portion of the NCAR Annual Report.
At the completion of each project, the project lead documents the work conducted, resulting outcomes, and contributions toward the strategic priority by preparing both a short write-up (approximately one page) for CISL’s portion of the NCAR Annual Report and a brief, 15-minute presentation to the Executive Committee.