Review and Evaluation

Review Process

All allocation requests must be submitted through InfoReady. Applications will be reviewed by the Research Computing Advisory Board (RCAB) four times a year.

Evaluation Criteria for Research Allocations (>300K SUs)

The RCAB will evaluate proposals based on the following criteria, focusing on the necessity, technical viability, and efficient utilization of the requested large-scale resources. The committee reviews for computational merit and resource justification, not for the general scientific merit of the underlying research.

Necessity and Justification (Why this Allocation is Needed)

1. Necessity of Scale and Exhaustion of Default Allocation:

   1. Is the request for more than 300K SUs clearly justified by the scale of the problem (e.g., large data volume, high-resolution runs, extensive parameter space)?
   2. Has the PI demonstrated that the work cannot be feasibly completed within the standard 300K SU annual allocation?

2. Alignment and Institutional Return:

   1. Does the project align with NJIT's strategic research goals?
   2. Does the project provide a high return on investment, particularly by supporting or generating new external grant funding?

Technical Viability and High-Efficiency Scaling

1. Appropriateness of Methodology:

   1. Is the proposed methodology a suitable and efficient use of the parallel HPC cluster architecture (i.e., does the problem require concurrent processing)?

2. Code Readiness and Demonstrated Performance:

   1. Has the PI provided quantitative evidence (e.g., benchmark data, scaling plots) that the code is ready to run and will perform efficiently when utilizing a large number of cores/nodes?
   2. Is the requested amount of compute time logically justified based on problem size and demonstrated scaling efficiency?

3. Optimal Resource Utilization:

   1. Are the requested ancillary resources (e.g., high-memory nodes, GPU nodes, specific storage) precisely matched to the technical needs of the application, avoiding waste or over-provisioning?

Project Management, Outcomes, and Stewardship (Crucial for Renewals)

1. Coherence of Computational Plan:

   1. Does the proposal present a clear, executable, and well-phased plan for utilizing the substantial allocation within the one-year timeframe?
   2. Does the PI or team possess the necessary technical expertise to manage large-scale, parallel computing jobs?

2. History of Use and Demonstrated Stewardship (For Renewals Only):

   1. Was the previous allocation time utilized efficiently and responsibly?
   2. Did the PI meet all reporting requirements, and is the progress reported in the mandatory Progress Report (Section 7) consistent with the resources consumed and the project goals?

3. Deliverables and Data Management:

   1. Are the specific deliverables (e.g., manuscripts, thesis chapters) that will result from this allocation clearly defined?
   2. Is there a feasible and detailed plan for managing the large data I/O, transfer, and archiving/cleanup upon project completion?