Title :
Fine-grain stochastic modelling of dynamic power management policies and analysis of their power - latency tradeoffs
Author :
Chen, Yuanfeng ; Xia, Feng ; Shang, D. ; Yakovlev, Alex
Author_Institution :
Asynchronous Syst. Lab., Newcastle Univ., Newcastle upon Tyne, UK
fDate :
12/1/2009 12:00:00 AM
Abstract :
Dynamic power management (DPM) is one of the main system-level low-power techniques for portable devices. This study presents a fine-grain Markov modelling approach that enables accurate analysis of system power and latency characteristics with full consideration of mode switching overheads in both processor and power controller. The new approach also makes it possible to incorporate latency analysis in terms of deadline satisfaction.
Keywords :
Markov processes; low-power electronics; microprocessor chips; power control; dynamic power management analysis; dynamic power management policy; fine-grain Markov modelling; fine-grain stochastic modelling; mode switching overheads; portable devices; power-latency tradeoffs; system latency characteristics; system power characteristics; system-level low-power techniques;
Journal_Title :
Software, IET
DOI :
10.1049/iet-sen.2009.0003