|Special Guest Lectures|
|Experiences Teaching Computational Science Remotely|
|Karen Tomko, Ohio Supercomputing Center|
|Johnston Hall 338
November 19, 2010 - 02:00 pm
Computational Science courses often lack the critical mass necessary to offer them regularly at a single institution and few universities provide comprehensive curricula in the computational science and engineering (CSE) disciplines. Inter-institutional offerings are one solution for providing students access to a wider selection of CSE courses. However, developing and offering such courses to a geographically distributed set of students requires overcoming technical challenges, bridging differing educational expectations and exploring new environments for interacting with students. In this talk Iâ��ll discuss two different experiences in the remote teaching of parallel computing skills to computational science students. The first course was co-taught with Juana Moreno of LSU Physics/CCT to a small group of students from LSU and from Germany (Uni. WÃ¼rtzberg, Uni. GÃ¶ttingen) as part of an NSF PIRE project. The second was offered to a large group of students at universities across the US as part of the Virtual School of Computational Science and Engineering summer school. Course format, technologies used and some reflections on this teaching paradigm will be presented.
Karen Tomko's research interests are in high performance computing, scientific application performance, accelerator technologies and domain decomposition (graph partitioning).