TY - GEN
T1 - Three scalable approaches to improving many-core throughput for a given peak power budget
AU - Sartori, John
AU - Kumar, Rakesh
PY - 2009/12/1
Y1 - 2009/12/1
N2 - Recently proposed techniques for peak power management [18] involve centralized decisionmaking and assume quick evaluation of the various power management states. These techniques suffer from two limitations. First, they do not prevent instantaneous power from exceeding the peak power budget, but instead trigger corrective action when the budget has been exceeded. Second, while these techniques may work for multi-core architectures (processors with small number of cores), they are not suitable for many-core architectures (processors with tens or possibly hundreds of cores on the same die) due to an exponential explosion in the number of global power management states. In this paper, we look at three scalable techniques for peak power management for many-core architectures. The proposed techniques (mapping the power management problem to a knapsack problem, mapping it to a genetic search problem, and mapping it to a simple learning problem with confidence counters) prevent power from exceeding the peak power budget and enable the placement of several more cores on a die than what the power budget would normally allow. We show up to 47% (33% on average) improvements in throughput for a given power budget. Our techniques outperform the static oracle by 22%.
AB - Recently proposed techniques for peak power management [18] involve centralized decisionmaking and assume quick evaluation of the various power management states. These techniques suffer from two limitations. First, they do not prevent instantaneous power from exceeding the peak power budget, but instead trigger corrective action when the budget has been exceeded. Second, while these techniques may work for multi-core architectures (processors with small number of cores), they are not suitable for many-core architectures (processors with tens or possibly hundreds of cores on the same die) due to an exponential explosion in the number of global power management states. In this paper, we look at three scalable techniques for peak power management for many-core architectures. The proposed techniques (mapping the power management problem to a knapsack problem, mapping it to a genetic search problem, and mapping it to a simple learning problem with confidence counters) prevent power from exceeding the peak power budget and enable the placement of several more cores on a die than what the power budget would normally allow. We show up to 47% (33% on average) improvements in throughput for a given power budget. Our techniques outperform the static oracle by 22%.
UR - http://www.scopus.com/inward/record.url?scp=77952222285&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77952222285&partnerID=8YFLogxK
U2 - 10.1109/HIPC.2009.5433221
DO - 10.1109/HIPC.2009.5433221
M3 - Conference contribution
AN - SCOPUS:77952222285
SN - 9781424449224
T3 - 16th International Conference on High Performance Computing, HiPC 2009 - Proceedings
SP - 89
EP - 98
BT - 16th International Conference on High Performance Computing, HiPC 2009 - Proceedings
T2 - 16th International Conference on High Performance Computing, HiPC 2009
Y2 - 16 December 2009 through 19 December 2009
ER -