PROGRAM SPONSORS:
Oak Ridge Associated Universities (ORAU), Oak Ridge National Laboratory (ORNL)
BACKGROUND:
In recent years, ORNL’s commitment to providing the world’s most powerful open resource for scientific computing has resulted in an impressive computational capability in Oak Ridge, Tennessee (see section on ORNL’s HPC Resources below). ORAU is partnering with the laboratory to provide opportunities for faculty and staff to become more intimately engaged in high-performance computing through a grant program. This program would not only provide a monetary award, but more importantly, would provide valuable access to the computing resources and staff at ORNL with the goal of expanding the existing research initiatives of university faculty or creating new research areas
By the end of 2008, ORNL will deploy the Cray XT5 supercomputer, which is a 1-petaflops leadership-class system for science. By 2018, the laboratory expects to have an exaflops system in place, which would be a thousand times faster than the petaflops Cray system. The opportunities introduced with these ultrascale computing resources are revolutionizing the lab’s approaches to energy, environmental sustainability and security global research challenges.
According to Associate Laboratory Director for Computing and Computational Sciences Thomas Zacharia, the U.S. has held a dominant position in supercomputing and continuing that is “absolutely critical” not only for our scientific enterprise but critical to the U.S. sustaining a competitive edge and superiority in science and technology.
ORAU is committed to making investments in faculty-student programs for our member institutions that align with the science agenda at ORNL, specifically in areas that require the leadership-class computational resources available in the Leadership Computing Facility and the National Institute for Computational Sciences.
ORNL’S HPC RESOURCES
ORNL’s Computing and Computational Sciences Directorate (CSSD) features the following computing resources:
- Cray XT (Jaguar)— the 250+ teraflops system (http://www.nccs.org)
- ORNL Institutional Cluster (OIC)
- Kracken – National Institute for Computational Sciences (NICS)
- High Performance Storage System for software, servers and storage hardware in the NCCS
- Visualization capabilities including a large PowerWall display, a medium-sized visualization cluster and a visualization laboratory
Access to these computing resources to be used to enhance educational and research experiences for faculty and students, making scientific discovery possible in such areas as superconductivity, computational biology, climate, combustion, and astrophysics is the ORNL commitment to this program.
ORNL’S CROSS-CUTTING SCIENCE AGENDA
Strengthening ORNL’s leadership position in computing and the computational sciences is at the core of the CCSD mission ( http://computing.ornl.gov and www.jics.utk.edu/index.html). Measures of success are breakthrough contributions in science and engineering with broad impacts to U.S. scientific leadership and national security. Dramatic advances in the power and performance of computing resources, networks, and data-storage facilities are opening new pathways to the modeling and simulation of physical, chemical, and biological systems and providing new insights into a host of complex science and engineering problems in climate change impacts, biomass supply chain, energy assurance, and knowledge discovery.
PROGRAM OVERVIEW/GUIDELINES
ORAU will sponsor a series of grants to provide for meaningful faculty engagement with ORNL’s leadership computing capabilities. Proposals for cross-cutting research projects from faculty are welcomed. Opportunities exist for both students and faculty.
- Purpose:
- Provide the opportunity for faculty to create or expand collaborative research with ORNL in scientific areas of discovery requiring HPC capabilities.
- Provide ORAU member institutions the opportunity to create a strategic partnership with ORNL through alignment of faculty research with ORNL’s science agenda.
- Enhance the development of the future workforce in scientific discovery through computing by providing real-world experiences for students.
- Roles/Responsibilities:
- ORAU: Provide yearly funding for grants up to three years.
- ORNL: Provide access to HPC resources and staff collaboration on research projects.
- ORAU Member Institutions: Demonstrate institutional commitment as appropriate in support of faculty research proposed.
- Faculty-Student Teams: Propose research to be performed using ORNL’s HPC resources that
- has significant scientific merit.
- develops or expands the faculty member’s research program.
- aligns with ORNL’s cross-cutting science agenda.
- Grant Details:
- Anticipated Funding Amount: $75,000 grant—$25,000 per year—for up to three years.
- Type and Number of Grants:
- Competitive grant renewable for up to three years.
- Two grants awarded per year.
- Significant progress, as demonstrated in a report to ORAU/ORNL on research outcomes at the end of each grant year, will determine what warrants second- or third-year funding.
- Ultimately, in steady state, six awards would be in place in an individual year. Desired mix would be three awards to principal investigators experienced in the application of high-performance computing and three awards to principal investigators with a desire to become involved in high-performance computing, with one of these three grants being awarded to a historically black college or university.
- Team Eligibility Information:
- Program is open to faculty-student teams proposing research in any scientific area. The proposed research does not have to be in computer science, but rather could be research in any scientific discipline that would benefit in some way from the application of high-performance computing resources.
- Team makeup and number would be at the discretion of the faculty member, however preference would be to support a team consisting of a faculty member with a graduate or undergraduate student(s).
- Team must be from an ORAU member institution.
- Timeline:
- June 2008—Call for proposals
- September 2008—Proposals due
- October 2008—Peer review of proposals
- November 2008—Grants awarded for FY2009
Thank you for reviewing this program content. We welcome any and all feedback. If you would like to post a response that all other readers can view, please click either the “Comments” link or begin typing your comment in the “Leave a Reply” box (depending on your view). Comments will appear for others to read after they have been reviewed by ORAU staff.
April 30, 2008 at 9:57 pm
John, as an ORAU Board member – is this offer limited to “Membership Institutions” only, or does it also include “Associate Membership Institutions”?
May 1, 2008 at 10:08 pm
Why are you limiting to only two awards? Is the reason capacity or funding? What would the funds be used for, if your objective is to provide access to HPC do you envision technical support as in programming?
May 2, 2008 at 3:44 pm
Like Sarita Soni I do not understand the need for the limitation to 2 grants per year. If the purpose is to bring ORNL computational resources to ORAU member faculty, then it seems to me that a larger number of grants of smaller monetary value would be preferable. I am not quite sure what the money part of the grants is meant to cover — salaries (students/postdocs/PI summer salaries), equipment at PI’s university?
May 3, 2008 at 2:01 pm
I’m excited to know of this opportunity. Like the two previous respondents, I think that two awards may actually discourage people from proposing when we weigh the potential odds despite the good intentions. Perhaps there may be another mechanism where potential users get certain allocations as is done in various centers. Like Rikvold, perhaps a smaler monetary value for attending workshops/roundtable discussions for users to share experiences may be good. Perhaps some restriction on how the money is spent, as Rikvold mentioned.
Personally, I just like to get access to a big computer.
May 4, 2008 at 9:48 pm
We are certainly interested in seeing expanding computational research at a number of levels — university, statewide, and nationwide. So this is a welcome initiative. Given the modest funding levels (support for approx 1 graduate student), and the administrative process required by those submitting the proposals, you may want to consider shortening the turnaround of awards to get the process running sooner with all 6 awardees in-place.
May 6, 2008 at 8:25 am
I also support expansion of computational research and initiatives to partner with Oak Ridge National Laboratory. I too was surprised by the small number of grants/year. Also, a $25K/yr grant is not sufficient to support a graduate student–it amounts to about 1/2 grad student/yr (based on the current minimum stipend in the OSU-MSE Dept, tuition and fees, fringe benefits, and indirect costs.)
May 6, 2008 at 12:25 pm
I believe it would be useful to give possible/appropriate budget items. Could the 75K be used for GRA salaries, travel to ORNL, etc or does the 75K represent “time on one of the ORNL HPC systems” and/or ORNL research computing support services.
As others have mentioned, two awards each year could lead researchers to believe that the odds of getting an award are not good.
This is a great first step to open a discussion that should lead to more interactions between ORNL and the core universities.
May 8, 2008 at 7:57 am
I agree with all of the comments posted to date. This program is great for allowing facuty-student teams from ORAU universities to use the great facilities at Oak Ridge. Better clarification of what resources will be available if the grant is awarded and more grants at lower levels would be perfect. The more people that can use the computer facilities at ORAU the better. In my view, the funding should be used for travel to use the facilities, while faculty salaries and student stipends should be supported by already funded grants or university resources. Therefore grants of $10K per year would allow 5 teams to participate the first year and up to 15 in two years.
May 28, 2008 at 7:45 am
We appreciate the comments and questions so I would like to provide a response to clarify several issues raised.
Our rationale for the number of grants—two per year—was based primarily on two factors. First, we wanted to provide a grant of sufficient size so that a faculty member could plan on at least a three year relationship with ORNL. This drove us to a larger dollar amount for fewer grants than a smaller amount for a greater number of grants. Second, there is limited discretionary computing time at ORNL and limited staff time to commit to working with faculty under these grants. Virtually all high performance computing time at ORNL is allocated by a competitive, peer reviewed process. The Associate Lab Director for computing has limited discretionary computer time. The Associate Lab Director has agreed to allocate some of his discretionary computer time to these grants. In addition, the lab has made a commitment to allow staff to work with faculty and students under these grants to provide scientific collaboration as well as technical/programming support. So these factors combined led us to commit to two grants per year, each grant three years in duration.
We based the ORAU funding on the concept that a faculty member and student would spend the summer at ORNL. For most of the programs that ORAU manages in collaboration with the lab, funding for a graduate student to spend the summer in Oak Ridge is about $8,000. This includes travel, lodging, and a stipend. For a faculty member to spend the summer at ORNL, the funding is typically about $17,000. This also includes travel, lodging, and a stipend.
Funding in the amount of $25,000 per year for each of three years was based on the notional concept of a faculty member and a graduate student spending time at ORNL during the summer. While this is our model, we will not require this. A proposal may call for the $25,000 in annual funding to be spent however the faculty member desires. The funding could support three graduate students at ORNL, or the funding could go toward supporting a graduate student. A faculty member may have other needs where the funding could be applied. While we are not restricting how the funding will be spent, I need to emphasize that preference will be given to proposals that include faculty-student teams spending time at ORNL.
The primary purpose of this grant initiative is to provide faculty and institutions the opportunity to develop or grow a scientific research program that relies on high performance computing and leverages a relationship with Oak Ridge National Laboratory. We recognize and share a desire for a larger number of grants. We want to go through this first cycle to assess interest and see how the program will work operationally before exploring the possibility of a greater number of grants.
Associate members are eligible to submit a proposal under this grant program.
July 18, 2008 at 2:17 pm
The official call for proposals has been announced, and details about the grant program as well as proposal submission guidelines can be found at:
http://www.orau.org/consortium/programs/hpc/index.htm
April 15, 2009 at 7:26 am
Anyone interested in the HPC Grant progam can read about this year’s first class of grant winners on the new ORAU web site: http://www.orau.org/university-partnerships/faculty-student-programs/hpc/default.aspx