MemcachedGPU: Scaling-up Scale-out Key-value Stores
Friday, October 2, 2015
11:30 am - 1:00 pm
Social Sciences 119
Tor Aamodt, University of British Columbia
Abstract: This work tackles the challenges of obtaining more efficient data center computing while maintaining low latency, low cost, programmability, and the potential for workload consolidation. We introduce GNoM, a software framework enabling energy-efficient, latency bandwidth optimized UDP network and application processing on GPUs. We use GNoM to develop MemcachedGPU, an accelerated key-value store, and evaluate the full system on contemporary hardware. MemcachedGPU achieves 10 GbE line-rate processing of 13 million requests per second (MRPS) while delivering an efficiency of 62 thousand RPS per Watt (KRPS/W) on a high-performance GPU and 84.8 KRPS/W on a lowpower GPU. At 8 MRPS, MemcachedGPU achieves a 95-percentile RTT latency under 300¿s. An offline limit study on the low-power GPU suggests that MemcachedGPU may continue scaling throughput and energy efficiency up to 28.5 MRPS and 127 KRPS/W respectively. Bio: Tor Aamodt is an Associate Professor in the Department of Electrical and Computer Engineering at the University of British Columbia. His current research focuses on the architecture of general purpose GPUs and energy efficient computing. Several of his papers have been selected as "Top Picks" by IEEE Micro Magazine and one as a "Research Highlight" in Communications of the ACM magazine. He was a visiting faculty in the CS Department at Stanford during his 2012-2013 sabbatical. Tor received his BASc, MASc and PhD at the University of Toronto.