Jangwoo Kim

Friday 3 May 2019:
Jangwoo Kim “Next-Generation Intelligent, Cost-Effective Computer Systems”
16:00 CAB G51

Abstract:  The next-generation computer systems must satisfy many critical design goals (e.g., performance, energy efficiency, scalability, reliability), while running emerging applications (e.g., artificial intelligence, big data, multimedia). However, it is extremely difficult for a system architect to come up with the best server architecture and system software to achieve all the critical design goals in the most cost-effective way. In this talk, I will start my talk by describing key limitations of conventional systems for such next-generation applications. I will then introduce my recent work to address these goals (e.g., performance modeling, high-performance system, intelligent system, low-power system). Next,  I will describe some of those work in more detail which include ssd-based storage system brain-inspired neuromorphic system, and AI accelerator exploiting heterogeneous units (if time permits).

Bio:  Jangwoo Kim is an associate professor in the Department of Electrical and Computer Engineering at Seoul National University. Before joining SNU, he served as assistant/associate professor for the Department of Computer Science and Engineering at Pohang University of Science and Technology (POSTECH). He earned his PhD degree in Computer Engineering from Carnegie Mellon University, and his BS degree in Electrical Engineering and MEng degree in Computer Science all from Cornell University. Prior to his academic career, he worked as a CPU/system architect at Sun Microsystems (later acquired by Oracle Corporation) in Santa Clara, USA, to develop UltraSPARC T4 CPUs and servers. His research interests include computer architecture, system software, and full-system development.