|
|||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||
|
Spaces, Software, and Services –
|
Back to top |
Emerging interdisciplinary computational requirements are driving faculty to look for high-performance and high-throughput computing resources for their teaching and research. In response, Academic Computing has started a pair of projects to shape HPC services for teaching. The first is a Website to support a community of practice for those interested in high-performance parallel computing (HPPC) . The HPPC Website serves as a focal point for information exchange and reference materials related to clustering hardware and software technology, as well as a forum to guide decisions regarding hardware and software procurements.
The second project leverages MIT's computing infrastructure in two pilot, undergraduate HPC teaching clusters. One cluster has Intel hardware running Red Hat Linux (the ROCKS Beowulf distribution), and the other cluster has Apple PowerPC hardware running Mac OS X (using the Sun-grid engine). The Athena student computing space in 4-035 was repurposed to house the Linux Beowulf cluster; the Mac OS X cluster is located in Building W91, and is accessible remotely. The implementation of these two HPC pilots in classes this year will provide information about what is needed to effectively use and support these systems.
Prof. Dave Darmofal, of Aeronautics and Astronautics, intends to use the Beowulf cluster for his 16.100 class this fall. In the spring, Profs. Gerd Ceder and Nicola Marzari, both in Materials Science and Engineering, will be trying it out with their students.
The Apple HPC cluster uses OKI authentication and authorization, and focuses on specific applications such as gridMathematica, a parallel implementation of Mathematica. Prof. Jim Elliot, from EAPS, and Prof. Raul Radovitsky, from Aeronautics and Astronautics, are among those considering use of this resource for their classes in the spring.
Faculty who are interested in learning more about the HPC clusters, or using them for their class work, should contact Phil Long, longpd@mit.edu.
MIT's collaboration with the University of Michigan, Indiana University, and Stanford University on the Sakai project provides direction for the on-going work and sustainability of Stellar.
Supported by the Mellon Foundation, Sakai will develop and share open-source software for learning tools based on service specifications from OKI.
While there will be limited pilots of Sakai tools in fall 2004, the production implementation of a Sakai-based course management system (CMS) is planned for fall 2005. The Sakai CMS will include new features, plus tools and applications developed in collaboration with other universities and commercial vendors. The framework will ultimately support new tools and services developed by MIT faculty and staff, such as an online gradebook that promises to make the grading process easier for MIT faculty.
To help faculty "shop" for technology which might support particular pedagogical goals, Academic Computing is creating an online catalogue of educational technology projects at MIT.
We are beginning to inventory such projects in close collaboration with MITCET and the Teaching and Learning Lab. The inventory seeks to identify and describe both large and small-scale uses of educational technology. The collected data will take advantage of the Carnegie Foundation's Snapshot tool to display details about each project's pedagogical goals, technical information, and assessment outcomes.
The new initiatives described above are by no means the only activities keeping IS&T Academic Computing busy in the coming months. We continue to maintain and update third-party software for the Athena computing environment, offer spatial data and GIS services in conjunction with the MIT Libraries, provide faculty consulting services through our newly named Educational Technology Consultants (formerly know as the Faculty Liaisons), and offer training to faculty in the use of the educational technologies available to them.
The period ahead will be characterized by an assessment of academic computing priorities as Academic Computing, collaborating with other groups engaged in supporting educational technology, shapes new services and responds to the sustainability needs of new initiatives. Faculty and student input is paramount to our efforts in identifying and building a rich and enduring ecosystem for educational technology. We welcome and invite your feedback. If you would like to find out more about any of the services or projects offered by Academic Computing, or to send us your thoughts, feel free to contact the Educational Technology Consultants at x3-0115,
et-consult@mit.edu.
Back to top | |
Send your comments |
home this issue archives editorial board contact us faculty website |