High Performance Computing

Catalog description

This class focuses on extracting performance from modern systems. The type of systems considered include uniprocessors (where cache performance is crucial), modern chips that have shared memory parallelism and SIMD features, distributed memory parallel systems, and accelerators like GPUs. Programming model appropriate for each of these is presented. Application of concepts to scientific kernels and practical development are emphasized. Lecture Hours: 3 Repeatability: May be repeated for credit one time. Grade Mode: Normal, Audit College Restrictions: School of Comp & Cyber Science Course Level Restrictions: Graduate Semester Schedule Type (Primary): Lecture Schedule Type (Additional): Asynchronous Instruction Click here for the Schedule of Classes.