High Performance Computing Systems and Networking is a study loosely-coupled systems from exotic supercomputers to commodity Off-The-Shelf clusters. Design rationales including computational, high availability, and load balancing are investigated. Networking capabilities and limitations are explored along with appropriate security considerations. Operating system support from kernel-level to file systems is analyzed. Programming considerations including message passing libraries are put into practice. Implementation specifics such as scheduling, job management, monitoring, and administration tools are reviewed in theory and practice. (3 cr. hrs.) Prerequisites: CSIT 1320 and CSCS 1730.
HPC Systems & Networking