Program Details for Summer 2010 - June 21 to August 13, 2010

Overview and Mailing List

The REU Site: Interdisciplinary Program in High Performance Computing is located in the Department of Mathematics and Statistics at UMBC, in the heart of the Baltimore-Washington corridor of industry and government. The program is comprised of learning scientific, parallel, and statistical computing and of project work on interdisciplinary applications. All activities of the program are conducted by teams of students, closely supported by graduate teaching and reseach assistants as well as the faculty of the program.
We maintain a mailing list for news and updates on the REU Site. To add yourself to the mailing list as well as for general information about the program, please send e-mail to hpcreu@umbc.edu.

Key Elements:

• eight weeks for eight students
• UMBC is located in the Baltimore-Washington corridor of industry and government agencies
• team work with dedicated graduate student mentors and faculty instructors
• professional development workshops
• GRE preparation course
• poster presentation
• technical report and publication opportunities
• no cost to the participants

Unique Features:

• fully transferable three-credit course on parallel computing with MPI using the parallel cluster with high performance interconnect in the UMBC High Performance Computing Facility (HPCF)
• includes introduction to tools of scientific and statistical computing: C, Linux, MATLAB, R, LaTeX
• interdisciplinary projects with clients from industry and government agencies through the Center for Interdisciplinary Research and Consulting (CIRC)
• field trips to local industry and government labs

Formal Details:

• Summer 2010: Monday, June 21 to Friday, August 13, 2010; travel to UMBC and move-in to campus housing on Sunday, June 20, 2010. See detailed schedule and below for more information on the program and the activities planned.
• progam benefits: program stipend $3,600; UMBC campus housing and food allowance provided; travel allowance provided; free GRE preparation workshop and course book; paid tuition and course book for the fully transferable three-credit course Math 447 Introduction to Parallel Computing
• eligibility Requirements: Earned a minimum of 30 college credits with a GPA of at least 3.0 on a 4.0 scale; provide documentation of U.S. Citizenship or Permanent Residency; reside in campus housing at UMBC during the program; participate in the summer research experience full-time
• pre-requisites: courses in Calculus (differential and integral), Linear Algebra, Basic Probability and Statistics, Introduction to Programming completed successfully by June 2010.
• application deadline for full consideration: Feb 19, 2010 by online application

More Information on the Program and its Detailed Schedule

The detailed schedule gives an overview of the topics involved with the program and of the activities planned during the 8 weeks of this this REU Site program. The 8 weeks are comprised of a first phase that focuses on learning scientific, parallel, and statistical computing and a second phase focused on project work on interdisciplinary applications. The first phase occupies Weeks 1, 2, and 3 full-time and then starts the second phase overlapping with the first during Weeks 4 and 5, before Weeks 6 and 7 are fully dedicated to project work. Week 8 can be thought of a third phase that is practically fully dedicated to preparing a technical report and a poster presentation of the work.
All activities of the program will be conducted in teams of (about) 4 students, closely supported in all phases by graduate teaching and reseach assistants as well as by the instructors of the classes. The structured learning of scientific, parallel, and statistical computing will consist of classes, each of which is immediately followed by team work in a computer lab, during which the teams will work on assignments with the support of a graduate TA. The program includes credit for a senior-level three-credit course on parallel computing that corresponds to the 29 numbered classes in the schedule.
During Weeks 1 through 3, teams will learn about several possible interdisciplinary projects with clients from application areas outside of the mathematical sciences. Examples of clients include faculty from other departments such as mechanical engineering or biology, researchers from industrial companies such as Northrop-Grumman, researchers from government agencies or research labs such as the Environmental Protection Agency, the National Institutes of Health (NIH), or the National Institute of Standards and Technology (NIST). By the end of Week 3, each team will choose a project and propose a plan of attack. The project work will start during Weeks 4 and 5, while the class on high performance computing concludes, and be the sole focus of Weeks 6 and 7. The team work on the interdisciplinary projects will be supported closely by a graduate RA assigned to each team. Additionally, there will be many formal and informal opportunities to bounce ideas off of all graduate TAs and RAs as well as the clients and all faculty involved with the program.
The last week of the program, Week 8, is dedicated to assembling the results of the project work. A technical report that will be posted on the preprint server of the UMBC High Performance Computing Facility will serve as the citable result of the work as well as the basis for the product delivered to the client. Additionally, the work will be presented in poster form at the Summer Undergraduate Research Fest (SURF) hosted by the College of Natural and Mathematical Sciences (CNMS) on Wednesday of Week 8.
Throughout the program, there will be plenty of opportunities to interact with all members of the program both in formal project updates, informal brainstorming sessions, and more. We also plan on several field trips to interesting sites, for instance, a client's organization.
The program also includes additional material presented to help support the project activities, such as an overview of LaTeX and other tools for presentations and report writing, a discussion of academic integrity in scientific work, and a GRE preparation course. All material of this program is designed to include an introduction that assumes very little background in it but also very advanced material that will make it useful for experienced users. These broader aspects of the program are designed to make the project work more effective as well as to provide an excellent preparation for and impression of graduate studies in mathematics or statistics.
In total, the program of this REU Site, as summarized in the detailed schedule provides a combination of formal introduction to high performance computing in the mathematical sciences covering aspects of scientific, statistical, and parallel computing with team-based work on an interdisciplinary application project. This combination of aspects will give participants a powerful and exciting experience how to combine learning with applying material to project work, all in an atmosphere of mutual support by all members of the project from undergraduate students, graduate students, faculty, to clients.