Pitfalls and Opportunities for Main Memory Compression, Mattan Erez (UT Austin)

Friday October 11, 2019
Location: CIC 4th floor Panther Hollow Conference Room
Time: 12:00PM-1:00PM

Main memory compression has been researched for decades and even commercially implemented in an IBM machine. However, we identify little-discussed overheads associated with additional data movement required for compression that significantly curtails any of its potential advantages. I will explain why these issues exist, why they may have been somewhat-ignored in the past, how to overcome the movement challenges, and why a new evaluation methodology is required to holistically demonstrates the impact of main memory compression on both the microarchitecture and memory management. I will also highlight new directions in how memory compression may aid in managing heterogeneous memory systems with slower and faster memories and offer general thoughts about memory-systems research directions.

Mattan Erez is a Professor at the Department of Electrical and Computer Engineering at the University of Texas at Austin. His research focuses on improving the performance, efficiency, and scalability of computing systems through advances in memory systems, hardware architecture, software systems, and programming models. His current focus areas are architectures for machine learning, large-scale and high-performance computing, and memory systems. His work aims to improve cooperation across system layers and develop flexible and adaptive mechanisms for proportional resource usage. Mattan received a BSc in Electrical Engineering and a BA in Physics from the Technion, Israel Institute of Technology and his MS and PhD. in Electrical Engineering from Stanford University. He was awarded a Presidential Early Career Award for Scientists and Engineers from President Obama and received an Early Career Research Award from the Department of Energy and an NSF CAREER Award.