Office of Science
FAQ
User Access

Guidelines and Criteria for Submitting a CIR

Contents

Overview | Facility | Computationally Intensive Research | Guidelines | Submission and Selection Dates | Proposal Submission | Proposal Evaluation and Selection | Awards | Publication of Awards

Additional Information

Overview

EMSL is soliciting one year proposals for allocations of computer time for Computationally Intensive Research (CIR) projects renewable up to three years duration. Research areas sought are in environmental molecular science, basic and applied research areas that address the environmental problems and research needs facing the U.S. Department of Energy (DOE) and the Nation. The Computing Research Projects page lists current projects using EMSL's computing resources. Some potential environmental molecular science proposal topics are listed below.

EMSL Science Themes represent EMSL's focus areas, and provide insight into the kinds of projects being solicited. The MSC "Greenbook" [.pdf, 2.37Mb], describes the science drivers used in procurement of the new high performance computer, Chinook.

Molecular Science Computing at EMSL

EMSL is located at the Pacific Northwest National Laboratory (PNNL) and is funded by the U.S. Department of Energy's (DOE) Office of Biological and Environmental Research (OBER). Production computing resources for Fiscal Year 2010 comprise:

Computing staff are available to help users access EMSL and make efficient use of its resources. A next-generation computational resource is expected to become available within the three-year scope of projects awarded time as part of this call. New computing hardware and software will be made available to the CIR once it is production ready.

The Chinook cluster was specifically designed for large scale computational chemistry problems that require a balanced mix of memory, processor speed, large amounts of local disk space, and a high speed interconnect between nodes. Calculations can be done faster on Chinook than on other large high performance computers because of Chinook’s large amounts of memory and disk space.

Computationally Intensive Research

To obtain optimal impact of EMSL’s computational resources in addressing the DOE's critical environmental problems, only a few large proposals will be awarded computer time allocations. This approach is based on the "Grand Challenge Applications" concept used successfully by EMSL and the Mathematical, Information, and Computational Sciences (MICS) Division within the Office of Advanced Scientific Computing Research (ASCR) of DOE. Grand-Challenges are complex, large-scale problems in science and engineering with broad scientific and environmental or economic impacts whose solution can be advanced by applying high-performance computing techniques and resources. The use of teams of scientists with different backgrounds and capabilities provides the largest scientific impact for solving the complex environmental and scientific problems facing the DOE and the Nation. Such teams also enable the training of students and postdoctoral fellows in the use of high-performance computing to tackle complex problems. The idea of teams of researchers undertaking such large computational problems is also consistent with the basic tenets of EMSL, which is a collaborative national user facility.

Guidelines

CIRs are expected to utilize EMSL's computing resources predominantly for large-scale parallel calculations that scale efficiently with both the number of processors and size of problem. On the current system, CIR calculations typically utilize from 16 to 512 nodes (128 to 4096 processor-cores). It is recognized and allowable that projects may periodically need to run calculations with less than 16 nodes as part of a proof-of-concept or software development activity. This type activity will be allowed but it is expected that they will be a minor component of each proposal. Applications are encouraged that require more than 512 nodes.

Project proposals are expected to comprise multiple investigators working collaboratively as a team. A collaborative team approach provides an excellent mechanism to engage the expertise of researchers from a variety of institutions to be focused on the basic and applied research projects needed to solve the DOE's and the nation's environmental problems. Teams should consist of computational scientists from universities; national laboratories; and, where appropriate, industry. The teams should be balanced in such a manner to make the most effective use of EMSL's computing resources in accomplishing their proposed CIR objectives. Proposals that include experimental work also done at EMSL are highly encouraged.

Projects will have a Team Leader who will have the responsibility to make certain the allocated computing resources are utilized in an optimal way and the efforts of the team remain focused on the proposed research project. The Team Leader will also be responsible for reports and reviews of the team's research progress and accomplishments and for managing the allocation of computer resources to team members. Team Leaders will work with EMSL’s computing staff management to optimize overall effectiveness and use of resources. Changes in the Team Leader will need the approval of EMSL computing staff for continuing time allocations, and may require peer review if the change is viewed to be a substantive change in direction for the CIR.

Coupling of experiments with theory, modeling, or simulation is strongly encouraged and such proposals will be preferred over those that do not have an experimental component.

Submission and Selection Dates

The following dates should be noted. Note that Letters of Intent are no longer required. Instead the proposals are to be entered directly into the EMSL User Portal.

Proposal Submission

The PI is required to submit the CIR proposal for their team. All proposals are required to be submitted electronically. No paper proposals will be accepted. All parts of the proposal form must be completed and submitted together at the same time for a proposal to be considered complete. Incomplete proposals will not be accepted or reviewed. All proposals must follow the instructions and format specified in the CIR Proposal Guide (PDF or Microsoft Word). Proposals that do not follow the instructions, omit required information, or deviate from the proposal format will be considered as incomplete proposals, and will not be accepted or reviewed. This is especially important because the format has changed from previous years.

The complete proposal must be submitted via the EMSL User Portal (the user interface to the EMSL Usage System [EUS]). Part of the EUS requires details about team members and affiliations; these details are listed in the Proposal Guide. Large text sections can be attached as Microsoft Word files or Adobe PDF files. Once a proposal is submitted, the PI can no longer make changes, however, a read-only copy is available via the EMSL User Portal.

Questions about submission of the CIR proposal or changes that need to be made should be directed to the EMSL Computing Contact (see Additional Information section below).

An acknowledgment for receipt of a complete proposal will be sent by e-mail to the PI.

Proposal Evaluation and Selection

We seek proposals in environmental molecular science research that require large numbers of processors and a significant portion of our computational resources. Proposals will be reviewed and selected based upon the following criteria:

All complete proposals will be reviewed by external scientific peer reviewers. Scientific peer reviewers will be chosen with regard to both their scientific expertise and the absence of conflict-of-interest issues. The project proposals selected for allocation awards will be reviewed internally before final approval.

Awards

It is expected that only 10 to 12 CIR proposals will receive allocation awards in this round of proposals. Computer resource allocations on EMSL’s production computers will be made to selected projects for a one-year period, contingent upon the availability of resources (expected to be 9 million node-hours for FY10). Allocation award sizes are expected to be on the order of 350,000-2,000,000 node-hours for a one-year period, each node having 8 processor-cores available. Allocations for the second and third year of the project are contingent upon the progress-to-date, accomplishments, and projected needs for continued research. A project Summary with 2-3 pages of accomplishments and 2-3 pages of the following year’s research activity will be requested in August, 2010. Only allocations of time on EMSL’s production computer resources will be awarded. No awards of funding for salaries, travel, equipment, or supplies will be made. EMSL reserves the right to award allocations in whole or in part to any or none of the proposals received in response to this notice.

All awarded proposals are subject to both Terms and Conditions for research done at EMSL and the User Policies for EMSL's Computing resources

Publication of Awards

CIR proposals that are selected and awarded allocations will have the following information published electronically on the Web and possibly in press materials or reports to sponsoring agencies.

The PI will be responsible for periodically providing highlights, publications, recognitions, and awards that are a result of the research activities.

Manager, Computational Grand-Challenge and Pilot Projects Erich R. Vorpagel, Ph.D. | , 509-376-0751