Reversible Computing
The term computer reversibility or reversible computing refers to an architecture for computers that is (at least approximately) reversible, so that the initial state could also be restored from the end result during the calculations.
The most important motivation for research on reversible computer architectures is the desire to circumvent the theoretical limits of energy efficiency in today's computer architecture. This is of particular interest with regard to the construction of quantum computers , since the heat generation of irreversible logic circuits would lead to malfunctions in these.
Neumann-Landauer border
Today's computer architecture is based on irreversible logic. This means that information is lost when performing logical operations. For example, a simple AND has two input signals but only one output signal. So one bit is lost and the number of possible logical states is reduced from 4 to 2.
If logical states are treated like physical ones, the rules of thermodynamics apply to them . So a reduction in the number of states from 4 to 2 leads to a reduction in entropy (with k being the Boltzmann constant ). The entropy reduction in turn requires that an amount of heat of at least at least is given off. This lowest limit for the energy that is used per logical operation in modern computers is called the Neumann-Landauer limit (after Rolf Landauer and John von Neumann ).
Reversible logical processes
Logical operations do not have to be irreversible. One can define reversible logical operations and as Landauer has shown, logically reversible processes are always physically reversible (Landauer's principle).
In particular, it can be shown that all logical operations can be represented as a combination of reversible operations. In theory, a computer can also be built with reversible logic. In addition, there are many publications by some well-known authors that deal with concepts and mechanisms on this topic. Nevertheless, reversible computing is only used for research purposes to this day.
disadvantage
Since the information must be retained with reversible logic, a large number of additional bits arise during a calculation and must be stored.