SCREMS: Building a Statistical Computing Environment to Support Scientific Research

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

The Scientific Computing Research Environments for the Mathematical Sciences grant provides crucial resources to the computing infrastructure within the Department of Statistics at the University of California, Berkeley. The grant will fund the acquisition of a four-processor dual-core AMD Opteron compute-server, an Apple X-serve, a 1.5 terabyte Raid Tray with Network Attached Storage Appliance and tape drive. The compute server will be able to handle the heavy I/O and memory requirements associated with processing and analyzing large data sets. The heterogeneous nature of the Department's research computing environment places complex demands on network file transfer, and the network attached storage device will provide greater reliability and functionality by supporting native file system integration across all platforms. The Apple Xserve running an Xgrid controller will enable researchers to incorporate the Department's Intel dual-core iMacs into the existing distributed computer infrastructure. The computing equipment acquired under this grant will support several interdisciplinary research projects in astronomy, bioinformatics, genomics, and neuroscience. These projects bring with them heavy computational demands due to processing large data sets, and they push the frontiers of statistical methodology and theory in support of the advances in the application-specific problems. It is critical that researchers have the ability to move, store, and rapidly process large data sets to support these collaborative research projects. These collaborations include: the Taiwanese American Occultation Survey to determine the distribution of comet-sized bodies in the Kuiper Belt; the Gamma-ray Large Area Telescope (GLAST), a $NASA$ mission to detect gamma-ray pulsars; the development of fast algorithms for data compression for addressing the increasing prevalence of high-volume data sets in modern scientific contexts; applications in neuroscience pertaining to in vivo experimentation of cortical neurons; and several bioinformatics and genomics projects. The computing equipment will support the research programs of approximately twenty graduate students and seven post doctoral fellows. In addition, it will support the research of undergraduates involved in the Department's research apprenticeship program.