• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2007, Vol. 29 ›› Issue (6): 81-83.

• 论文 • Previous Articles     Next Articles

  

  • Online:2007-06-01 Published:2010-06-03

Abstract:

This paper provides a cache leakage power estimation model based on statistics. By using the statistics of cachelines' access intervals, this model c  an estimate the cache leakage power at different values of decay intervals. Compared with the Hotleakage leakage power simulator, the average variation   of the estimation results from the simulator designed by our statistics-based leakage power model is smaller than 3. 46%. This model can be used to esti mate the ratio of normalized leakage power at different values of decay intervals and select the best decay interval to reduce the cache leakage power a t most.

Key words: (statistics, Cache, leakage power, estimation model)