DNA self-assembled parallel computer architectures

New varieties of computer architectures, capable of solving highly demanding computational problems, are enabled by the large manufacturing scale expected from self-assembling circuit fabrication (1012–1019 devices). However, these fabrication processes are in their infancy and even at maturity are expected to incur heavy yield penalties compared to conventional silicon technologies. To retain the advantages of this manufacturing scale, new architectures must efficiently use large collections of very simple circuits. This paper describes two such architectures that are enabled by self-assembly and examines their performance. (Some figures in this article are in colour only in the electronic version)