Predict and act

Raid Zuhair Ayoub, Tajana Simunic Rosing
2009 Proceedings of the 14th ACM/IEEE international symposium on Low power electronics and design - ISLPED '09  
In this paper, we propose a proactive dynamic thermal management scheme for chip multiprocessors that run multi-threaded workloads. We introduce a new predictor that utilizes the band-limited property of the temperature frequency spectrum. A big advantage of our predictor is that it does not require the costly training phase like ARMA [7]. Our thermal management scheme incorporates temperature prediction information and runtime workload characterization to perform efficient thermally aware
more » ... hermally aware scheduling. Our results show that applying our algorithm considerably improves the average system temperature, hottest core temperature, product MTTF and performance by 6 o C, 8 o C, 41% and 72% respectively.
doi:10.1145/1594233.1594256 dblp:conf/islped/AyoubR09 fatcat:spqdafsufnfvlcbiyzzlneau3e