Design and performance analysis of a novel nanoscale associative memory

This paper proposes a nanoscale associative memory architecture. The type of associative memory considered is adept at handling the irregular patterning of nanoscale components but requires an encoding scheme to be useful. We show how a randomly configured interface between the micro and nano scales can accomplish this encoding. The resulting associative memory eliminates some of the supporting microscale components required by the original crossbar architecture at the cost of reduced storage capacity. If the projections made by Kuekes et al. about the potential size of their memory prove correct, this associative memory configuration may be a valuable alternative.