Minimal energy dissipation in logic

TLDR: In this article, it was shown that any required reliability can be obtained with this device, without increased energy expenditure, but at the expense of an increasing time per computational step, and that the system has dissipations larger than but of the same order of magnitude as, the original minimal quantities.

Abstract: Minimal energy dissipations for the logic process based on thermodynamics and general phase space considerations are known. The actual availability of these minimal dissipations has not, however, been demonstrated. These minimal dissipation sources in a computing system also act as noise sources and thereby lead to questions about the ultimate available reliability of the computing process. A new and hypothetical device is presented in this paper and used to construct a physically analyzable computing system. It is demonstrated that this system has dissipations larger than, but of the same order of magnitude as, the original minimal quantities. It is also shown that any required reliability can be obtained with this device, without increased energy expenditure, but at the expense of an increasing time per computational step.