Software and Hardware Co-Design for Distributed Systems

Professor: Philip Brisk

Description: Deploying distributed systems at the data center scale creates significant performance challenges for networks, networking cards, CPUs, GPUs, operating systems, and other accelerators. This project aims to characterize performance bottlenecks that impede distributed system performance and explore software and/or hardware mechanisms to remedy these performance bottlenecks. Participating students can focus on different topics of interest including, but not limited to, system performance modeling and characterization, microarchitectural simulation, hardware accelerator design, and/or distributed algorithms and applications.

Preferred Qualifications: The only official requirement is programming skills comparable to a 3rd year undergraduate student. Most of our work is done in Python or C/C++. If you are unable to program without ChatGPT doing all of the thinking for you, then please do not apply.