Someone once told me that with reversible computers there is no energy cost for the computation itself, there is only a cost for running the hardware. But is this true in practice, or is it only an idealization?

  • 6
    $\begingroup$ "In theory, theory and practice are the same. In practice, they are different." $\endgroup$
    – John Doty
    Mar 19 at 20:52
  • $\begingroup$ Maybe a better fit for Computer Science. $\endgroup$ Mar 19 at 20:54
  • $\begingroup$ You are asking about the Landauer principle. Imho this is an ineradicable myth coming from a confusion of Shannon entropy with thermodynamical entropy. $\endgroup$
    – Kurt G.
    Mar 19 at 21:06
  • $\begingroup$ Nobody has built a reversible computer yet. In principle, we know there is a theoretical minimum energy cost for non-reversible computations, and we know reversible computers could be a loophole to bypass it. But nobody's actually built one that does. Our best computers today still use zillions of times more energy than the theoretical limit $\endgroup$
    – user253751
    Mar 20 at 1:24