An Efficient DVS Algorithm for Fixed-Priority Real-Time Applications

Due to the effectiveness of slack time analysis methods for dynamic voltage scaling (DVS), we propose a low power fluid slack analysis (lpFSA) for enhancing energy efficiency of the existing methods. The on-line method is for hard real-time systems that execute periodic, fixed-priority tasks. lpFSA can be applied by most slack estimation methods without increasing their time complexities. The key technique of the heuristic method lengthens the analysis scope and estimates the amount of movable slack. Experimental results show that the proposed method based on lpWDA reduces the energy consumption by 21%~45% over the lpWDA-based methods and other existing rate-monotonic DVS algorithms.