HPC RESOURCES
- High Performance Computing (HPC) is a collective term to describe computational facilities that have the capacity to outperform the current high-end standalone machines. HPC facilities can be vary from massively parallel processing supercomputers to clusters of commodity personal computers. Rutgers has a rich history of providing facilities and expertise to encourage the adoption of HPC technologies in support of its research and education goals. OIRT serves as a coordination point at the university for this important technology.
-
Each Spring, Mechanical and Aerospace Engineering conducts a
course in Linux. This course covers general aspects of operating a Linux environment
and includes a section on Linux clustering. The course is limited to sixteen students due to the availability of workstations
in the Engineering Unisys Lab.
The Linux course is offered on a first-come, first-served basis to qualified students
(i.e., students with the requisite background in Unix at a "user" level).
OIRT conducted a seminar in Fall 2005 on how to gain access to the national supercomputing centers. The National Science Foundation's Partnership for Advanced Computational Infrastructure (PACI) serves as a coordination point for these facilities. This seminar can be brought to your particular workgroup. Contact oirt@rutgers.edu to arrange.
- OIRT, as part of the UHR's IT Certificate Program, conducts a course in High Performance Computing in Research Applications. The course covers the historical background of HPC at universities, current and emerging technologies, and how to find HPC resources and expertise internal/external to Rutgers. Visit http://uhr.rutgers.edu/profdev/calendar.asp and search on "High Performance Computing."
-
Research groups interested in exploring HPC technology are encouraged to contact oirt@rutgers.edu.
OIRT will help researchers identify the type of HPC technology best suited for their research.
For those wishing to explore clustering for the first time, OIRT has a small demonstration cluster that can be configured to match a researchers needs, serving as a mechanism for a proof of concept. Groups with existing clusters will be able to make use of the test cluster to explore the behavior of their research programming with different operating systems, compilers, etc. Please contact OIRT at oirt@rutgers.edu.
- HPC Administrators and faculty have their own mailing list- hpc_admin. Announcements, events and discussion surrounding HPC will be done on this list. To subscribe, send email to hpc_admin-subscribe@email.rutgers.edu.
- OIRT conducted a technology meeting on HPC cluster computing performed on campus. Several groups aligned with the original HPCP project presented their current clustering environments. Discussions included planning for the future. A summary on what was presented and discussed is located at http://oirt.rutgers.edu/archives/techmtg-mar-summary.html.
-
Cyberinfrastructure Technology Watch is an online resource where current
HPC and networking technology, events and initiatives are discussed.
CTwatch is a joint effort of the National Center for Supercomputing Applications and the San Diego Supercomputing Center.
The Globus Alliance is an organization of researchers involved with developing the software tools to make grid computing effective.
Partnership for Advanced Computational Infrastructure (PACI) serve as a coordination point for accessing the national supercomputing centers.