In the future, all computers will be reversible. Reversible computation will be many times more energy efficient and computationally powerful than any possible irreversible computation. Reversible computation is the future.
Unfortunately, reversible computing has been mostly ignored by the media and academia. Hardware manufacturers are not proactively researching reversible computation, and few people are working on reversible programming languages and software. Reversible computation is not very popular because unlike 'artificial intelligence', 'GPT', 'quantum computation', 'nuclear fusion', 'Elon Musk', 'Theranos', 'information superhighway' , '2020' , the phrase 'reversible computation' is not a buzzword. Humans are attracted to buzzwords similar to how flies are drawn to sources of light that zap them. Yes. Humans are unwilling to educate themselves on what is truly important, but they instead follow all the other sheep to the end of the cliff like they did in Turkey
https://www.foxnews.com/story/450-turkish-sheep-leap-to-their-deaths. But the commenters here will probably ignore everything I say and just go on some non-sensical rant about how I should not cite Fox News because Fox News is very bad. Please focus. Fox News is completely irrelevant to the discussion.
Landauer's principle states that it takes k*T*ln(2) energy in order to delete a bit of information. Here, k=1.38*10^(-23) Joules/Kelvin, T is the temperature, and recall that ln(2)=0.69314... For example, if you have a gigabyte of random bits and you want to replace all of those bits with zeros, then you will need k*T*ln(2)*8*10^9 energy to replace all of those bits with zeros. In practice, it will cost much more energy than simply k*T*ln(2) per bit deleted in order to overcome thermal noise. If you use irreversible computation (regardless of whether it is CMOS or any other kind of hardware and regardless of technological improvements), then you should expect to spend more than 100*k*T energy per bit deleted. k*T*ln(2) is not a lot of energy at room temperature, but we are about to realize the limit of conventional irreversible computation since it becomes difficult to improve one's computational hardware well before reaching Landauer's limit. Up until about now, it has been easy to ignore reversible computation since
The only way to overcome Landauer's limit and compute without spending >>k*T*ln(2) energy per bit deleted is to try to compute while deleting as little information as possible. Energy efficient reversible computing is the type of computing where one computes while trying to delete as little information as possible to save the >>k*T*ln(2) energy that one would otherwise dissipate. Right now, we do not have profitable reversible computing hardware, but we have reversible programming languages, so we can develop more reversible software and reversible algorithms in anticipation for the reversible computing hardware.
Now, energy efficient reversible computing will be difficult to get off the ground. Part of the reason for this is that reversible computing has some computational overhead. But the computational overhead is quite surmountable. By applying strategies such as a good solution to Bennett's pebble game to reversible computation, one can turn any conventional irreversible algorithm into a reversible algorithm which means that the process of computing will require a little bit more space and time but as a tradeoff, it will require a lot less deletion of information. This means that reversible computation will eventually outperform conventional irreversible computation, but this also means that it will be difficult to make the first profitable reversible computers.
There is no theoretical limit to the energy efficiency of reversible computation. Reversible computation is good for general purpose computing (one should compare this with the hyped quantum computing which is only good for a few specific computations). Why don't more people know about reversible computing? Why are academia, the media, and technology corporations not yet interested in reversible computation?
It is now the time for people to become interested in reversible computation since we are getting to the point where our computation is energy efficient enough and close enough to Landauer's limit that we will need to start thinking about reversible computing.
I have brought up the topic of reversible computation before, but I have decided to bring up this topic again because reversible computation is too important to only mention once.
P.S. Do you have any good suggestions for buzzwords that describe reversible computing? Could we use something like 'ultragreen computing' , 'green computing' , 'adiabatic computing' , 'pre-quantum computing' , 'isentropic computing' , 'perpetual motion computing' (ok that was a joke), 'injective computing', or 'vladimir computin' (I apologize for this joke)?
-Joseph Van Name Ph.D.