Quasi-Static Assignment of Voltages and Optional Cycles for Maximizing Rewards in Real-Time Systems with Energy Constraints
42nd Design Automation Conference, Anaheim, CA, June 13-17, 2005, pp. 889-894
There exist real-time systems for which it is possible to trade off precision for timeliness. In these cases, a function assigns reward to the application depending on the amount of computation allotted to it. At the same time, many such applications run on battery-powered devices with stringent energy constraints. This paper addresses the problem of maximizing rewards subject to time and energy constraints. We propose a quasi-static approach where the problem is solved in two steps: first, at design-time, a number of solutions are computed and stored (off-line phase); second, one of the precomputed solutions is selected at run-time based on actual values of time and energy (on-line phase). Thus our approach is able to exploit, with low on-line overhead, the dynamic slack caused by tasks executing less number of cycles than in the worst case. We conduct numerous experiments in order to show the advantages of our approach.
[AEP05] Luis Alejandro Cortes, Petru Eles, Zebo Peng, "Quasi-Static Assignment of Voltages and Optional Cycles for Maximizing Rewards in Real-Time Systems with Energy Constraints", 42nd Design Automation Conference, Anaheim, CA, June 13-17, 2005, pp. 889-894