A Workload-Aware, Eco-Friendly Daemon for Cluster Computing
Shi‐Yu Huang, Wu-chun Feng · Computer Science Technical Reports (Virginia Tech) · 2008
Abstract—This paper presents an eco-friendly daemon that reduces power consumption while better maintaining high per-formance via a novel behavioral quantification of workload. Specifically, our behavioral quantification achieves a more ac-curate workload characterization than previous approaches by inferring “processor stall cycles due to off-chip activities. ” This quantification, in turn, provides a foundation upon which we con-struct an interval-based, power-aware, run-time algorithm that is implemented within a system-wide daemon. We then evaluate our power-aware daemon in a cluster-computing environment with the NAS Parallel Benchmarks. The results indicate that our novel behavioral quantification of workload allows our power-aware daemon to more tightly control performance while delivering substantial energy savings. I.