INTERCONNECTS

REvolution Computing's Director of Community, Dr. David Henderson, will chair a session entitled "High Performance III" at this year's useR! Conference in Germany August 12-14. The discussion will focus on the challenges associated with using R across multiprocessor or multicore computers. "Base R does not natively support parallel processing, forcing users to wait while computationally intensive work is performed on a single processor or core," said David Henderson. "At a time when R users have increasingly greater access to multiprocessors it is especially important to now understand how R users can enhance analyses using parallel processing," Henderson added. In addition, Dr. Henderson will present an overview of REvolution Computing's technology for parallelizing R scripts, High Performance Computing with NetWorkSpaces for R. This technology has been freely available for the R community on CRAN and is now available for the first time as part of REvolution's commercial ParallelR product. NetWorkSpaces features shared workspaces along with a web interface that displays the workspaces and their contents, a helpful feature when debugging or developing a program or monitoring the progress of an application. NetWorkSpaces is easy to learn, accessible from many development environments, and deployable on ad hoc collections of CPUs. Also newly added to ParallelR, REvolution Computing has contributed an optimized transport option based on NetWorkSpaces to the SNOW package on CRAN and has worked with Max Kuhn of Pfizer Research to develop and contribute caretNWS, a parallel classification and regression training package. To learn more about REvolution Computing, please visit its Web site.