The memory system will be a growing bottleneck for many workloads running on high-end servers. Performance improvements from technology scaling are also expected to decline in the coming decade. Therefore, new capabilities will be required in memory devices and memory controllers to achieve the next big leaps in performance and energy efficiency. Some of these capabilities will be inspired by emerging workloads (e.g., in-memory big-data, approximate computing, co-scheduled VMs), some will be inspired by new memory technologies (e.g., 3D stacking). The talk will discuss multiple early-stage projects in the Utah Arch lab that focus on DRAM parameter variation, near-data processing, and memory security.
Rajeev Balasubramonian is an Associate Professor at the School of Computing, University of Utah. He received his B.Tech in Computer Science and Engineering from the Indian Institute of Technology, Bombay in 1998. He received his MS (2000) and Ph.D. (2003) degrees from the University of Rochester. His primary research areas include memory hierarchies and interconnects. Prof. Balasubramonian is a recipient of an NSF CAREER award, an IBM Faculty Partnership award, an HP IRP award, and various teaching awards at the University of Utah. He has co-authored papers that have been selected as IEEE Micro Top Picks (2007 and 2010) and that have received best paper awards (HiPC'09 and PACT'10).