Case Study in KSR Programming: Finding Outliers by the Minimum Volume Ellipsoid Method
Bergmark, Donna; Crovella, Mark
This case study describes the enablement of a parallel application on Cornell Theory Center's KSR1, a highly parallel machine from Kendall Square Research. The application is from statistics, and exposes some interesting facets of the KSR as well as some parallel programming tools and techniques. Statistics are a novel application area for supercomputing, at least at the Cornell Theory Center. First we describe the application itself, then the approaches to parallelizing it, and finally present some results.
Previously Published As