SagTree: Towards Efficient Mutation in Evolutionary Circuit Approximation

Approximate circuits that trade the chip area for the quality of results play a key role in the development of energy-aware systems. Designing complex approximate circuits is, however, a very difficult and computationally demanding process. Evolutionary approximation - in particular, the method of Cartesian Genetic Programming (CGP) - currently represents one of the most successful approaches for automated circuit approximation. In this paper, we thoroughly investigate mutation operators for CGP with respect to the performance of circuit approximation. We design a novel dedicated operator that combines the classical single active gene mutation with a node deactivation operation (eliminating a part of the circuit forming a tree from an active gate). We show that our new operator significantly outperforms other operators on a wide class of approximation problems (such as 16 bit multipliers and dividers)
and thus improves the performance of the state-of-the-art
approximation techniques. Our results are grounded on a rigorous
statistical evaluation including 39 approximation scenarios and 14,000 runs.