Authors: R.W. Schmieder
Affilation: NanoLogic, Inc., United States
Pages: 640 - 643
Keywords: computation, computer design, nanoelectronics, simulation
We describe here a new paradigm for computation and computer design that has the potential for performance many orders of magnitude higher than conventional digital computers, especially for large-scale simulation of complex systems. The paradigm combines three emerging technologies: (1) analog nanoelectronics; (2) adaptive analog arrays; and (3) topologic. A significant application of this paradigm is for simulating dynamical systems, including quasi-periodic and chaotic systems such as turbulent fluid dynamics. The introduction of nanoelectronic devices presents the opportunity to develop circuits with very high behavioral complexity while maintaining low device count. For instance, stacks of resonant tunneling diodes (RTD) can be used to generate nonlinear and discontinuous maps that represent mathematical catastrophes, which are used to describe many types of physical systems that exhibit sudden changes, such as eddies in turbulent fluids, deformation of structures, and rapidly changing social systems. The approach described here will enable developing relatively simple, fully-analog circuits that generate complex behavior, with appropriate precision and self-adaptivity. The overall efficiency of these circuits will depend on the extent to which nanoelectronics can be included. Estimates based on experience with analog arrays, nanoscale devices, and topologic indicate that enhancements greater than 10**6 over conventional digital approaches are feasible.