2012 Student Middleware Training - Pilot Program

EPSCoR RII Track I & II assistantship support was provided for 13 graduate and 5 undergraduate students in the Middleware Training Program for Spring/Summer 2012. Middleware knowledge and experience has become necessary as computational STEM research projects outgrow desktop capabilities. This is especially true as more dynamics and higher dimensional considerations are incorporated into engineering and scientific simulations, and as more complex algorithms are devised for extracting intricate information from data. Middleware components focused upon in this project involves the basics of programming and scripting languages, cluster use and emerging computer architectures, parallel and distributed programming techniques, multi-core programming including GPGPU, and the migration of desktop computing to the national computer grid.

Student participants in the program were nominated by selected faculty mentors drawn from a wide variety of disciplines, ranging from tissue fabrication, geonomics, fuel cells, computational biology, electrical engineering, analysis of sensor data, and digital humanities. Faculty researchers were identified based both upon the quality of their research and as being strong candidates for enhancement by middleware support. The program supports those faculty by off-loading middleware training responsibilities and allowing those faculty to instead concentrate on discipline-specific aspects of their Research Computing projects. (more)

The Program's goal is that students develop sufficient background in programming and middleware methods to enable them to design and implement an HPC research project with their faculty mentors at their home institutions. Initial funding for the pilot program was made available through an EPSCoR RII Track II grant to USC. Further Track I funding was made available by a competitive proposal to the State EPSCoR office as outreach to include students from Benedict College (HBCU) and Coker College (PUI).

The pilot training program has three primary objectives to directly address the South Carolina RII Track I goals:

  • provide student training in scientific middleware methods and tools with expected benefits to include:
    • major contribution toward professional development of computationally oriented STEM students in knowledge and new skills necessary for the scientific workforce of the future.
    • faculty mentors are freed to concentrate on computational activities more specific to their research disciplines rather than general HPC training.
  • attract talented faculty mentors to more fully utilize existing HPC infrastructure in South Carolina and at national supercomputer centers, by providing local middleware assistance performed by their students, as junior researchers.
  • create a working community of students selected from disparate disciplines who can share experiences and novel approaches to computationally related problems.

The participating students are individually assessed in terms of their computational background by the Program's Faculty Coordinator (Professor Colin Bennett in 2012) and our professional middleware staff (Drs. Ebalunode and Sergueev). Based upon the assessments, the coordinator recommends to each student any preparatory work necessary to complete their projects. This preparation includes on-line tutorials, seminars by our middleware staff and the coordinator, and tailored supervised reading. During the Summer I session, the students, in collaboration with their faculty mentors, should complete the design of their projects, make significant progress toward implementation, and deliver a 20 minute presentation of their work. (less)

An abstract for each 2012 student/mentor project can be found on our User Research Page.

General Training for High Performance Computing

The Computational Research and CyberInfrastructure Support Initative in USC's Office of Research offers a series of High Performance Computing Workshop tutorials. The topics covered include cluster computing, parallel message passing (MPI), GPU computing, parallel matlab, and interactive remote desktop analysis and visualization of running applications. The goal of the HPC tutorials is to get users familiarized with the HPC and visualization resources available at CR-CISI and provide the necessary operational principles and programming techniques for effective use of the resources.

CR-CISI HPC Workshop tutorials are free to the USC research community as well as individuals and groups that engage in collaborative activities with the USC researchers and CR-CISI personel.

The list of the HPC tutorials that the CR-CISI regularly offers can be found here. If the timing of the tutorials is not indicated they will be announced periodically depending on the current needs for such tutorials and the availability of the CR-CISI staff.

For all the inquiries regarding the training sessions please contact Dr. Jerry Ebalunode.