Cluster Expansion for Applied and Computational Mathematics

点击翻译按钮获取中文摘要

This effort will acquire and install compute cluster components to support several research activities at Caltech's Applied and Computational Mathematics (ACM) department. The components will be used to expand the Shared Heterogeneous Cluster (SHC) at Caltech's Center for Advanced Computing Research (CACR). SHC is a teraflops class machine that is optimized for high precision parallel codes. The additional components will increase its peak capacity by about 400 gigaflops and 160 gigabytes. The expansion of SHC will provide computational support to the research of three faculty members in ACM: Drs. Oscar Bruno, Emmanuel Candes, and Tom Hou. These faculty members and their ACM colleagues work on some of the most intellectually challenging problems in mathematical algorithms. These problems are characterized by multiple scales, inhomogeneity, and the sampling of very complex spaces. The applications are wide ranging, including electromagnetic modeling, optimal sampling for MRI biological imagery, and computing underground material flows for geophysics problems. In each case, progress in algorithmic development depends on access to large-scale computer resources.This effort will expand an existing shared compute cluster (named SHC) to support several research and education activities at Caltech's Applied and Computational Mathematics (ACM) department. The expansion of SHC will provide computational support to three faculty members working on some of the most intellectually challenging problems in mathematical algorithms. The applications are wide ranging, including electromagnetic modeling, optimal sampling for MRI biological imagery, and computing underground material flows for geophysics problems. The expansion will have broader impacts in education and, potentially, public outreach. ACM has a vital educational role at both the graduate and undergraduate level at Caltech. As computational science emerges as the third pillar of scientific discovery, this role will increase. SHC will also be used to produce compelling images of numerical calculations to help better explain the science. The SHC expansion leverages significant existing resources and will support several promising research areas, provide students with advanced computational capabilities, and help show the public the value of computational mathematics.