An examination of Design and Performance trade-offs in modern parallel computer architectures. Includes an introduction to the basic concepts of parallel computing including message passing and shared memory programming models. An examination of a number of parallel architectures, including pipeline architectures, shared memory multiprocessor systems including both bus based and CC NUMA distributed memory systems, message passing systems, focusing on interconnect issues will also be examined. Also covers cache architectures and strategies to ensure cache coherency in shared memory systems.
This course may not be repeated for credit.

Sections