Cellular Automaton-based nanoelectronic hardware

Computing-in-Memory has recently attracted increasing interest because of the expected limitations to be faced by the traditional von Neumann architecture under further extensions of Moore's law. Cellular architectures are especially well positioned as candidates in this context. The most well-known among these are Cellular Automata (CA), which are computing devices with a regular structure of cells that are locally interconnected to each other. Their regularity facilitates manufacturing methods, like Interference Lithography (IL), that can produce high-density patterns of extreme regularity. When this technology is combined with bottom-up methods based on self-assembly, architectures become possible with extremely high numbers of identical cells that can be configured to conduct a wide variety of functions. This paper gives a focused introduction to cellular designs over the last 50 years and discusses their suitability for nanoelectronic implementations.