معرفی
Joseph Devietti is an Associate Professor at the University of Pennsylvania in the Department of Computer & Information Science. His research focuses on improving multiprocessor programmability through architecture and systems innovations, including work on cache optimization, deterministic execution, GPU programming, and race detection.
- PhD, University of Washington (2012)
- MS, University of Washington (2009)
- BSE, University of Pennsylvania (2006)
His research spans computer architecture and systems, with significant contributions to:
- Cache Optimization: Runtime prefetching (RPG2), code layout (OCOLOS), and branch target buffer design (Twig)
- Deterministic Execution: Hardware-software hybrid models (RADISH, RCDC), managed runtime systems (DetFlow), and scalable mechanisms (DMP)
- GPU Programming: Memory access analysis (GPUDrano), race detection (BARRACUDA), and deterministic architectures (GPUDet)
He has received multiple honors including the 2024 Penn Engineering Ford Motor Company Award for Faculty Advising and 2018 Radhia Cousot Young Researcher Best Paper Award. His work has been recognized in IEEE Micro Top Picks (2023, 2009, 2008). Devietti has graduated numerous PhD and Master's students, including Yuxuan Zhang, Omar Navarro Leija, and Akshitha Sriraman, who have taken roles at Google, Bolt Labs, and academic institutions.
