Improving Energy Efficiency on Partially Reversible Pipelined QCA Circuits
Marco A. Ribeiro, Iago Augusto de Carvalho, Jeferson F. Chaves, Omar P. Vilela Neto · 2018
Recent studies show that Field-Coupled Nanocomputing devices, such as Quantum-dot Cellular Automata (QCA), can reach ultra-low power consumption, notably when operating with reversible computing techniques. Partially Reversible Pipelined QCA circuits are one of such examples where improving energy efficiency can be balanced with throughput degrading. In this work, we propose an algorithm to divide those circuits into stages, reducing energy without meaningful changes in its throughput when compared to a näive division. Furthermore, we manipulate the fanouts' positions in the circuit's netlist to improve even more the energy efficiency. Computational experiments with a state-of-the-art benchmark suite show that our algorithm's best solutions dominate about 56% of the best solutions from the näive method, while our fanout's manipulation achieves an average energy reduction of 47%.