Josephson junction binary oscillator computing

The first modern computers were built at Manchester University in the late 1940s and amongst the principal components used were vacuum tube oscillators. Following the development of the transistor, such oscillator-based computers quickly became obsolete. Here, a novel application of superconducting Josephson Junction (JJ) technology based on neural dynamics is proposed. Neuron-type oscillators can be constructed from JJs and they can be connected together to form computer circuits in a manner entirely consistent with modern architectures. SIMetrix models of both computation and memory are presented and metrics are discussed. It has been estimated that JJs switch 100 times faster than their transistor counterparts and use 300 times less power (including the super-cooling). Utilizing the circuits described in this paper, it is also estimated that computers could be constructed using at least 10 times fewer components. Thus, theoretically these JJ-based computers could be up to 300,000 times more efficient than their CMOS counterparts.