Circuit Approximation Using Single- and Multi-objective Cartesian GP
In this paper, the approximate circuit design problem is formulated as a multi-objective optimization problem in which the circuit error and power consumption are conflicting design objectives. We compare multi-objective and single-objective Cartesian genetic programming in the task of parallel adder and multiplier approximation. It is analyzed how the setting of the methods, formulating the problem as multi-objective or single-objective, and constraining the execution time can influence the quality of results. One of the conclusions is that the multi-objective approach is useful if the number of allowed evaluations is low. When more time is available, the single-objective approach becomes more efficient.
KeywordsGenetic programming Cartesian genetic programming Evolutionary design Approximate computing Approximate circuits Multi-objective approach
This work was supported by the Czech science foundation project Advanced Methods for Evolutionary Design of Complex Digital Circuits 14-04197S. The authors would like to thank Jiri Petrlik for useful discussions on multi-objective evolutionary optimization.
- 1.Chakradhar, S.T., Raghunathan, A.: Best-effort computing: Re-thinking parallel software and hardware. In: Proceedings of the 47th Design Automation Conference - DAC, pp. 865–870. ACM (2010)Google Scholar
- 5.Han, J., Orshansky, M.: Approximate computing: An emerging paradigm for energy-efficient design. In: Proceedings of the 18th IEEE European Test Symposium, pp. 1–6. IEEE (2013)Google Scholar
- 6.Hilder, J., Walker, J., Tyrrell, A.: Use of a multi-objective fitness function to improve Cartesian genetic programming circuits. In: NASA/ESA Conference on Adaptive Hardware and Systems, pp. 179–185. IEEE (2010)Google Scholar
- 9.Nepal, K., Li, Y., Bahar, R.I., Reda, S.: Abacus: A technique for automated behavioral synthesis of approximate computing circuits. In: Proceedings of the Design, Automation and Test in Europe, pp. 1–6. DATE 2014, EDA Consortium (2014)Google Scholar
- 10.Sekanina, L., Vasicek, Z.: Approximate circuits by means of evolvable hardware. In: IEEE Int. Conf. on Evolvable Systems, SSCI-ICES. pp. 21–28. IEEE CIS (2013)Google Scholar
- 11.Vasicek, Z., Sekanina, L.: Evolutionary approach to approximate digital circuits design. IEEE Tran. on Evolutionary Computation, pp. 1–13 (2015 to appear)Google Scholar
- 12.Venkataramani, S., Roy, K., Raghunathan, A.: Substitute-and-simplify: a unified design paradigm for approximate and quality configurable circuits. Design. Automation and Test in Europe, DATE 2013, pp. 1367–1372. EDA Consortium San Jose, CA, USA (2013)Google Scholar
- 13.Venkataramani, S., Sabne, A., Kozhikkottu, V.J., Roy, K., Raghunathan, A.: Salsa: systematic logic synthesis of approximate circuits. In: The 49th Annual Design Automation Conference 2012, DAC 2012, pp. 796–801. ACM (2012)Google Scholar