 |
Center for Nonlinear Studies |
The fundamental thermodynamic limits of conventional computation are near enough to be an area of concern when contemplating future computing technologies. Several thermodynamic arguments imply lower limits on the energy required for computation, when conventionally construed. However, several of the known limits may be circumvented by using unconventional computing paradigms. Thermal noise limits on signal energies can potentially be circumvented in appropriately designed chaotic systems with sub-unity signal-to-noise ratios. And, limits on energy dissipation due to Landauer’s Principle can be circumvented using reversible computing. We review some recent work in these areas, including a new general theoretical framework for reversible computing, and a framework for asynchronous reversible computation.
Host: Francesco Caravelli