11:00 AM-1:00 PM
Nicollet D2
Many of the grand-challenge applications in sciences and engineering require the solution of large sparse eigenvalue problems. While these applications could benefit substantially from parallel processing, it can be noted that parallelism is making a slow impact relative to comparable technologies in sparse linear systems. Recent advances in parallel computing, such as massively parallel computers, cluster computing, and communication standards, as well as advances in eigenvalue methods, such as preconditioners, restarting, and improved robustness, are setting new computational challenges. Among these are the need for parallel algorithms to compute a large number of interior eigenvalues and the need for efficient parallel preconditioners. This minisymposium will focus on how new eigenvalue methods and parallel computing technologies are combined to solve the new and traditional types of matrix eigenvalue problems which arise in a few applications. It will also attempt to address scalability issues of current techniques, and discuss the incorporation of these new technologies into efficient parallel software.
Organizers: Yousef Saad and Andreas Stathopoulos
University of Minnesota, Minneapolis
PP97 Homepage | Program Updates|
Registration | Hotel Information | Transportation | Program Overview | Program-at-a-Glance