Title :
Generic algorithm based voltage Islands-Aware Multiple Voltage Assignment
Author :
Du, Shimin ; Xia, Yinshui ; Chu, Zhufei ; Qi, Lixia
Author_Institution :
Sch. of Inf. Sci. & Eng., Ningbo Univ., Ningbo, China
Abstract :
Power consumption has become one of most serious challenges for SoC design. Among the various methods, Multiple Supply Voltage (MSV) is an effective method for power reduction. In this paper, Genetic Algorithm (GA) is used to optimize the Voltage Islands-Aware Multiple Voltage Assignment (VIAMVA). A cost function is developed to take the overhead of level shifters and the complexity of power network into account. The proposed algorithm is tested on six GSRC benchmarks. Experimental results show that the proposed method can save power consumption up to 47.60% with voltage island generation.
Keywords :
circuit optimisation; genetic algorithms; low-power electronics; system-on-chip; generic algorithm based voltage islands-aware multiple voltage assignment; level shifters; power consumption; power network; system-on-a-chip design; voltage island generation; Artificial neural networks; Benchmark testing; Multiple voltage assignment; genetic algorithm; power reduction; voltage island generation;
Conference_Titel :
Computer Application and System Modeling (ICCASM), 2010 International Conference on
Conference_Location :
Taiyuan
Print_ISBN :
978-1-4244-7235-2
Electronic_ISBN :
978-1-4244-7237-6
DOI :
10.1109/ICCASM.2010.5620181