Response Time Analysis for Thermal-Aware Real-Time Systems under Fixed-Priority Scheduling

Abstract

This paper investigates schedulability analysis for thermal-aware realtime systems. Thermal constraints are becoming more and more critical in new generation miniaturized embedded systems, e.g. medical implants. As part of this work, we adapt the P F PASAP algorithm proposed in [1] for energy-harvesting systems to thermal-aware ones. We prove its optimality for non-concrete 1 fixed-priority task sets and propose a responsetime analysis based on worst-case response-time upper bounds. We evaluate the efficacy of the proposed bounds via extensive simulation over randomly-generated task systems.

Testbed for thermal and performance analysis in MPSoC systems

“…Many researchers analyze the thermal properties of computer systems. A common approach is to create a model and examine the thermal behavior using system simulation [2]- [4]. The disadvantage of such an approach is that the simulation precision highly depends on the input parameters.…”

“…To suppress the effect of ambient temperature changes, it might be best to have a model that counts the impact of the ambient temperature on the on-chip temperature readings and can produce delays of the heat propagation from the ambient environment to the chip. Such a model is referred to as a transfer function and can be estimated by system identification methods based on models such as OE, ARX, and ARMAX 4 . It is a well-known fact that for these methods to produce good results, it is necessary to have a high variation on the system inputs.…”