5

I'm interested in mining altcoins, mainly to turn some servers into (literal) space heaters, since it's winter. One thing that really interests me is if an altcoin has been designed to only be mineable by traditional CPUs. The big thing I can think of to cause this is large memory requirements. A large memory requirement for hash processing would cause current GPUs to be slower than CPUs (due to the individual "cores" of the GPU not having enough direct memory). It would also cause ASICs and FPGAs to be very difficult to develop to be faster than CPUs, and expensive. This is because you effectively reduce the mining complexity so that it bottlenecks at the memory bus (which is much harder to significantly speed up)

I'm sure I'm not the only one to have thought of such a scheme. Is there an altcoin out there that is designed with this kind of ideal behind mining?

  • Perhaps designing an altcoin that has limits on how much work can be done per miner and per ip address per day?? Making it rather useless to spend big dollars on mining rigs and forcing the distribution of the coins more widely across the planet. – user11386 Dec 25 '13 at 10:59
  • but then that just sounds like it would benefit botnets – Earlz Dec 26 '13 at 7:22
6

Litecoin is the 2nd most popular currency in the same family as Bitcoin, and:

It is based on the Bitcoin protocol but differs from Bitcoin in that it can be efficiently mined with consumer-grade hardware. Litecoin provides faster transaction confirmations (2.5 minutes on average) and uses a memory-hard, scrypt-based mining proof-of-work algorithm to target the regular computers with GPUs most people already have.

https://litecoin.org/

The intention was to allow CPU mining to be viable, but according to Wikipedia it hasn't quite met this goal:

Litecoin uses scrypt in its proof-of-work algorithm: a sequential memory-hard function first conceived by Colin Percival.[14] The original intended benefit of using scrypt was to avoid giving advantage to GPU, FPGA and ASIC miners over CPU miners, which occurs in the Bitcoin protocol. However, this turned out to be an incorrect assumption: GPU mining in Litecoin's implementation of scrypt is currently ten times more efficient than CPU mining

http://en.wikipedia.org/wiki/Litecoin

  • Litecoin is almost the antithesis of a reply to this question. Even ASICs exist for it. – user36303 Jul 1 '16 at 7:21
4

Yep there are three good options for CPU mining that I know of:

XPM: PrimeCoin

QRK: Quark

PTS: Protoshare - very very new and exactly what you described.

Google search those and you'll find plenty of info.

Seems like there was a fourth but I can't remember it off the top of my head.

  • 1
    There is also my Cuckoo Cycle proof-of-work which hasn't been adopted in any coin yet. Whitepaper and implementation at github.com/tromp/cuckoo – John Tromp Mar 7 '14 at 15:41
1

Either you had a very good idea, some years behind, or you read about Litecoin and wanted to be answered with its name :)

In principle it's impossible to design an algorithm that can only be run on a generalistic CPU installed in a normal computer - you can always design purpose-specific hardware for whatever application though normally this is limited by cost-benefit reasons.

Litecoin's designers' decision to use scrypt rather than SHA-256 was precisely because they wanted mining not to be over-taken by GPUs. In the end this was a partial defeat, since GPUs do can be used to mine scrypt-based crypto-currencies, albeit the efficiencies vs. CPUs are not as enormous as with SHA-256: a GPU is between some dozens to a hundred times more efficient than a CPU, rather than several thousands more efficient.

The reason is scrypt forces you to either have large amounts of memory or be able to re-compute previous values of the hash function rapidly, thus making GPUs way less efficient for this and complicating the design of an ASIC.

Primecoin's algorithm (looking for Cullingham chains of prime-numbers) can only be done right now by a CPU. Should this über-neue crypto-currency rise as sharply in value as Bitcoin, you could see a new generation of purpose-specific chips designed solely to look for prime numbers (or to test the primality of one), which would mean the end of CPU-mining for it as well as have a deep impact both on scientific and security computing (for the good and bad, respectively).

  • 1
    I've actually started to think that possibly having a non-zero barrier to mining is a good thing. Otherwise anyone with the money to buy a 10,000 computer botnet would be a millionaire over night. ugh understanding economies is hard – Earlz Dec 6 '13 at 20:28
  • Mining Litecoin with a GPU is more efficient, though. – Nick ODell Mar 2 '15 at 18:55
1

There are three problems with this idea:

  1. It focuses on how to distribute wealth rather than how to create it, when the latter part is the hard part. If you don't create any wealth to distribute, it doesn't matter whether you have a good way to distribute it.

  2. This would produce a very insecure crypto-currency because anyone who wanted to attack it could rent or steal general purpose computers, which are widely available. You want a mining algorithm that gives legitimate miners (who invest in the coin) an advantage over attackers, and CPUs give the advantage to attackers because botnets have lots of CPUs and general-purpose computing power can easily be rented.

  3. It's basically economically infeasible to create a coin with profitable mining. If it's especially profitable, more people will do it until it becomes just barely profitable. Eventually, only those with some special advantage (such as the cheapest power) will do it. You can't use mining as a way to spread wealth.

  • What you mentioned in point #2 is what happened to primecoin, right at its beginning: it was very profitable to rent virtual machines at shared-time servers and punt them to mine the coin all day long, to the point that difficulty skyrocketed after the first few days and some companies dedicated to that business saw their servers collapse (or had to kick miners). Fortunately all those people were honest miners rather than attackers. – Joe Pineda Dec 26 '13 at 20:19
0

Try Burstcoin.Can be mined by simply using your extra space in your HDD.If you want to gain more then buy TB HDDs to plot and surely you will earn much. You can use your computer while it is mining since it uses less processor time. The forum members are also very supportive and will guide to the whole process. Free burstcoins are given out every day from faucet.

Go to the official site and go to forum for help: https://forums.burst-team.us/

0

It is not possible for a coin to be mineabe only by CPU, since GPUs are also general purpose processors. What can be done, however, is to try to decrease the advantage that GPUs, FPGAs, and ASICs get when mining compared to CPUs. There are several attempts to do this. Two of then are Cryptonight (used in most Cryptonote coins), and Cuckoo Cycle (I don't think it's used in a coin yet).

The more low level ways to decrease the gaps between those include:

  • making memory hard algorithms (ASICs can get expensive quickly per hash when you add memory, since it's a lot more silicon than the compute part)
  • using circuitry already present on many CPUs, such as AES-NI instructions
  • changing the PoW algorithm from time to time

The last one is used by Myriadcoin. Cryptonight uses the first two. Litecoin uses the first one (and failed at it, since there is a memory/speed tradeoff inheren in its scrypt PoW).

Cryptonote coins are probably the best bet now if you want something where CPUs are competitive. Monero, Digitalnote, AEON. Possibly Boolberry, I'm not sure how its PoW (Wild Keccak) does on the CPU/GPU gap.

-1

There is a Russian coin called Sifcoin that was the basis for Quarkcoin. Because it is inflationary not many coins have been mined thus far (the reward is not decreasing expoentially) and so can be mined with a CPU / standard laptop.

Not the answer you're looking for? Browse other questions tagged or ask your own question.