The Need for Speed: From Pipelines to Massively Parallel Machines

3 credits · Undergraduate

This course discusses why a modern CPU represents a massive parallel machine and what impact this has on how applications view the CPU. Topics include: evaluation of CPU performance; instruction level parallelism (ILP) and the various consequential complexities such as out-of- order execution, branch prediction, and speculative execution; design of memory hierarchies that support ITP; and complexities of multi-core design, including cache consistency and special purpose processors such as the Graphics Processing Unit. PREREQUISITES: CSCI 2121.03 or CSCI 2122.03, with a grade of C or better. EXCLUSIONS: ECED 3403.03