2006 · 66 citations · 12 references
ComputingComputer ScienceHigh-performance ComputingComputer ArchitectureParallel ComputingPerformance MeasurementBenchmarkingCost-performance TradeoffsParallel Computing SystemsPower ManagementHeatHigh-performance Computing PerformanceHigh-performance Computing SystemsMetricsParallel Computing PerformancePerformance Benchmarking
For decades, supercomputing performance has been measured in FLOPS, but this focus has led to power‑hungry machines requiring extensive cooling, neglecting other metrics like reliability, availability, and usability, thereby driving up total cost of ownership. To raise awareness of other performance metrics such as energy efficiency, the authors propose a Green500 List. The paper discusses potential metrics that would rank supercomputing systems on such a list. Despite the importance of the TOP500 List, the authors argue it hampers focus on metrics beyond speed.
For decades now, the notion of has been synonymous with (as measured in FLOPS, short for floating-point operations per second). Unfortunately, this particular focus has led to the emergence of supercomputers that consume egregious amounts of electrical power and produce so much heat that extravagant cooling facilities must be constructed to ensure proper operation. In addition, the emphasis on speed as the performance metric has caused other performance metrics to be largely ignored, e.g., reliability, availability, and usability. As a consequence, all of the above has led to an extraordinary increase in the total cost of ownership (TCO) of a supercomputer. Despite the importance of the TOP500 List, we argue that the list makes it much more difficult for the high-performance computing (UPC) community to focus on performance metrics other than speed. Therefore, to raise awareness to other performance metrics of interest, e.g., energy efficiency for improved reliability, we propose a Green500 List and discuss the potential metrics that would be used to rank supercomputing systems on such a list.
12