The new Cryptocurrency Circcash has been recently launched and is available for you to compile on Ubuntu (I have created but I have not yet released the Windows executable since it gets flagged and quarantined by anti-virus software and has not been signed yet).

Github:

https://github.com/jvanname/circcashWhitepaper:

https://github.com/jvanname/Zammazazzer/blob/master/CirclefishICO.pdfAlternate Circcash discussion:

https://bitcointalk.org/index.php?topic=5290467.msg55617409#msg55617409Website:

https://circcashcore.com (under development)

Relevant discussion from 2017 (the general principle behind using reversible computation for cryptocurrency mining was there, but the details have been modified quite a bit since then, so that discussion is outdated):

https://bitcointalk.org/index.php?topic=2041169.0;all**What is Circcash?**Circcash is the new cryptocurrency with a mining algorithm that is designed to make it easier for hardware manufacturers to develop reversible computing hardware and therefore solve one of the most important outstanding technological problems. Since it takes a degree of intelligence, altruism, and other virtuous qualities to appreciate Circcash, Circcash will attract much higher quality users than most other cryptocurrencies. Furthermore, since Circcash will do a public good, governments will be much less likely to ban or restrict Circcash.

Right now, Circcash is one of the few cryptocurrencies with a mining algorithm that even attempts to repurpose the mining towards solving an important computational problem, and right now there are not very many serious proposals for such useful mining algorithms.

**What is reversible computation?**Purely reversible computation is the type of computation that is time reversible in the sense that the state of the machine at time s can be recovered from the state of the machine in time t for t>s simply by running the computation in reverse. In particular, in the process of the computation, no information has been deleted and replaced with zeroes. Partially reversible computation is the type of computation where most but not all of the computational process is reversible (in practice, reversible computation refers to partially reversible computation instead of fully reversible computation).

For example, the AND gate is not reversible since one cannot recover the input from the output since if the output of the AND gate is 0, then one does not know if the input was (0,0),(0,1), or (1,0) since 0=0 AND 0=0 AND 1=1 AND 0. For a similar reason, the OR, XOR, NAND gates are not reversible either. However, the NOT gate is a logically reversible gate since one could always recover the output of a NOT gate from its input. Since reversible logic gates cannot delete any information, every reversible logic gate must have the same number of inputs as outputs.

Consider the following gates:

CNOT gate-(x,y)->(x,x XOR y)

Toffoli gate-(x,y,z)->(x,y,(x AND y) XOR z)

Fredkin gate-(0,y,z)->(0,y,z), (1,y,z)->(1,z,y).

Then the CNOT, Toffoli, and Fredkin gates are all examples of reversible gates. Now, every 2 bit input and 2 bit output reversible gate must be an affine transformation, so the 2 bit input and 2 bit output gates cannot be used for universal reversible computation, but the Toffoli gate is universal for reversible computation, and the Fredkin gate is universal for reversible computation.

The (hopefully near) future use case of reversible computation is that there is a limit to the energy efficiency of irreversible computation but this limit does not apply to the energy efficiency of reversible computation. By CPT-symmetry (charge, parity, time), the laws of physics are time reversible. Therefore whenever information is deleted in a computation, by the laws of physics, that information is not truly deleted but it must go somewhere, and this information is converted into entropy (i.e. waste heat). Furthermore, Landauer's limit gives a precise quantity for the amount of waste heat that must be produced whenever one deletes a bit of information. Landauer's principle states that every time a bit of information is deleted, k*T*ln(2) energy is consumed and transformed into waste heat. Here, k is Boltzmann's constant (k=1.38*10^(-23) Joules/Kelvin) and T is the temperature. To put this into perspective, the quantity (3/2)*k*T is the average kinetic energy of an atom in a monatomic (this means there no molecules consisting of more than one atom) ideal gas, and k*T is quite close to the kinetic energy of an atom moving at the speed of sound (the ideal gas law also gives an intuitive idea about the size of the quantity k*T). Now, this k*T*ln(2) limit is only applicable to computation when one deletes energy. The k*T*ln(2) limit does not give any bound on the efficiency of reversible computation where one limits the amount of information that is deleted.

Now, Boltzmann's constant and Landauer's limit are both extremely small quantities, but right now some forms of computation are so energy efficient that one must take Landauer's limit into consideration when considering how computers may perform such calculations in the not-too-distant future.

In general, reversible computation will require new algorithms, and reversible computation will generally require more space, time, and/or parallelism. Fortunately, the space/time overheads arising from reversible computation are surprisingly reasonable [BENNETT, 89] even in a worst case scenario, and in a best case scenario (for example with an encryption algorithm designed for reversibility), there is no significant computational complexity overhead incurred from reversibility. Therefore, one should expect for general purpose computation to be reversible in the far future. However, in the near future, it will be difficult to make reversible computation energy efficient enough to pay for the space/time overheads and be profitable.

**How does Hashspin accelerate reversible computation development?**Hashspin is a cryptocurrency mining algorithm that is designed to be as reversible as possible so that one does not incur a computational complexity overhead simply by mining Circcash on a physically nearly reversible hardware. Hashspin has also been designed to be as simple as possible, so that it is as easy as possible to make physically reversible hardware for Hashspin and/or hardware for Hashspin whose energy efficiency is as close to the Landauer limit as one can get.

Hashspin is relies on a cryptographic technique called the Solution Lottery Technique (SLT), and the SLT allows Hashspin to both have a high level of cryptographic security and contain the qualities that will allow the hardware manufacturers to most easily make energy efficient physically reversible mining machines for mining Hashspin. In particular, the SLT allows over 99.9% of the computation for Hashspin to be spent running two interconnected linear feedback shift registers (LFSRs) of lengths 15 and 17 [LN] (LFSRs have been used in quite a few stream ciphers). If you know anything about LFSRs, then you would know that LFSRs require very little hardware and very little energy. This will reduce the burden on ASIC manufacturers so that it is easier for them to make reversible ASICs, and this will get the energy efficiency of the LFSRs much closer to Landauer's limit where reversible computation becomes a necessity in order to mine competitively.

**Is Hashspin safe?**Hashspin is an entirely new cryptographic function that has not been tested very much by the cryptographic community. Nevertheless, there are reasons to believe that Hashspin should remain secure even though it has not been through all the rigorous tests that one should expect from a cryptographic function. Hashspin relies on the SLT which means that the LFSRs for Hashspin only require 16 bits of security. it is not feasible for a miner to mine Hashspin more efficiently than simply performing a brute force search since any cryptographic break needs to be faster than simply running the LFSR based function (a function that will be very fast and efficient in dedicated hardware) 65536 times.

The miners will not benefit personally from the advances resulting from Circcash mining. Furthermore, the use of Hashspin over SHA-256 mining will not harm the cryptocurrency since the miners, users, and everyone else in the cryptocurrency ecosystem would behave nearly the same way with Hashspin as the mining algorithm as they would if SHA-256d were their mining algorithm. If there is still any doubt as to whether mining algorithms should only serve to secure the network or whether they should solve a scientific problem, then the best way to answer these doubts is to see how Circcash performs in the long-term to determine whether it is a good idea to use a cryptocurrency mining algorithm that is designed to advance science.

**What is Circcash better than previous attempts at a useful mining algorithm?**Cryptocurrency mining algorithms must satisfy some stringent cryptographic conditions. This means that most computational problems cannot be made into suitable cryptocurrency mining algorithms. Furthermore, it is no good for a cryptocurrency mining algorithm to solve a scientific problem and provide 1 million dollars of scientific value when it costs 1 billion dollars in mining to solve that scientific problem. A good cryptocurrency mining algorithm needs to be able to provide billions of dollars of scientific benefit as long as billions of dollars are spent on mining. Since reversible computation is the only way to make computation more energy efficient in the long-term, the problem of developing reversible computing hardware is an extremely important problem, and it is well worth spending potentially billions of dollars on mining hardware to help solve this problem. This cannot be said of other so-called useful mining algorithms. Right now there are only a few serious proposed and implemented useful cryptocurrency mining algorithms, and it is unlikely that a new mining algorithm can be as useful as Hashspin without compromising security in any way.

The difference between Hashspin and other proposed useful mining algorithms is that Hashspin behaves much more like a cryptographic hash function (that only needs to jumble bits up as efficiently as possible) and less like a typical algorithm for solving a useful problem. From a miner's perspective, Hashspin will behave just like SHA-256; the miners will not benefit personally in any way from mining other than the block reward. In any case, if there is any question about whether having a useful mining algorithm is inheritently detrimental to the currency, one should support experimentation so that one can back up this hypothesis with evidence, so Circcash has experimental value since it is a cryptocurrency with a useful mining algorithm.

**Information about Circcash**Circcash is a fork of Litecoin with 2 minute block intervals. There is no halving period for Circcash. Circcash will issue the same amount of coins for all time. Under this protocol, the inflation rate for Circcash will approach zero as the time approaches infinity, and the inflation rate will be inversely proportional to the amount of time since the Genesis block was created. 12.5% of the mining reward will go towards the development of Circcash, but this small tax is only temporary (the Circcash code will automatically drop this reward under the right conditions and the Circcash development will voluntarily drop the 12.5% development fund when it becomes unnecessary). Circcash is denominated by CIRCs which stand for Certificate of Innovation in Reversible Computation.

The genesis block for Circcash was mined about 4 months ago, and the first few blocks were mined on July 27, 2020 (Genesis time: Wednesday July 01, 2020: I waited a few weeks between the Genesis block and the time when the next few blocks were mined). The Circcash source code was made publicly available a couple of weeks after the genesis block was mined.

This is the first official Bitcointalk post about Circcash since its launch. However, I have many times in the past for the past three years since May of 2017 both online and in person have been promoting the use of cryptocurrency mining to help accelerate the development of the reversible computer, so people should have been anticipating the release of Circcash.

**References**[BENNETT, 89] TIME/SPACE TRADE-OFFS FOR REVERSIBLE COMPUTATION. CHARLES H. BENNETT. SIAM J. COMP Vol. 18, No. 4, pp. 766-776, August 1989

https://www.math.ucsd.edu/~sbuss/CourseWeb/Math268_2013W/Bennett_Tradeoffs.pdf[LN] Finite Fields (Encyclopedia of Mathematics and its Applications) 2nd Edition by Rudolf Lidl (Author), Harald Niederreiter (Author)