9:15 AM-10:00 AM
Chair: Rosemary E. Chang, Silicon Graphics Computer Systems
Recent directions in high performance computers for scientific computing have been driven by two key factors: the advantages of employing commodity technology as building blocks and the understanding of how to build scalable multiprocessors that efficiently support shared-memory programming. The first capability can lead to high performance computers that are cost competitive with workstations and servers, while the second capability allows high performance machines to be software compatible with small and mid-range computers that uniformly use shared-memory multiprocessing. While these trends may increase the usability and generality of high performance computing, other technology trends will require users to be more aware of issues such as locality of access and data distribution. The challenge in the future will be in creating hardware and software that helps the user to achieve high performance without undue programming effort.
John L. Hennessy
School of Engineering, Stanford University
AN97 Homepage | Program Updates|
Registration | Hotel and Dormitory Information | Transportation | Program-at-a-Glance | Program Overview