What has a lower long term cost per hash: LTC, BTC or other alt coin?

4

1

I'm trying to understand the cost per hash operation with Litecoin vs Bitcoin over time. The theory is that the network that has the best prospects for long-term security (via hashing) will set a standard that may influence its own price.

Specifically I'd like to compare

  • Capital outlays (reasonable mining hardware ASIC vs GPU)
  • OPEX (Electrical, etc)

Suppose I were to compare BTC/Litecoin mining to ore mining. Once the ore is extracted and purified, no more cost is needed other than to secure the facility where the gold is stored.

Its clear the mining process is technology dependent (ASIC vs GPU). What isn't clear to me is once the technology matures, what approach has a greater chance for long term efficiency for maintaining and securing the network.

I suspect that the scrypt -vs- sha256 implementations will have different power requirements, and I'd like to understand which one is "greener" for the environment in the long haul.

Any information relating to

  • Cost of current mining rigs to power efficiency

  • Cost of future mining rigs and projected efficiency

Would be very interesting.

goodguys_activate

Posted 2013-03-08T18:14:04.110

Reputation: 11 898

"Cost of current mining rigs to power efficiency" This depends directly on the exchange rate, so this is less of a question about which will use less energy and more about human irrationality. – placeybordeaux – 2013-03-08T21:19:36.403

@PeterMichealLacey-Bordeaux I'm not talking about profit, I'm referring to capital and operational expense, which should be fixed. Revenue is a different topic I didn't want this confused with. – goodguys_activate – 2013-03-08T21:32:15.567

I am a bit confused, you are talking about captial and operational expense, but for what? You can mine both of them on your phone if you really wanted to, but the return would be really really small. – placeybordeaux – 2013-03-08T21:44:42.657

2@PeterMichealLacey-Bordeaux - You're thinking ROI which takes price into consideration. I'm thinking only of return on hash power. BTC was once a lose-only proposition. It has appreciated to profit for various reasons. – goodguys_activate – 2013-03-08T23:06:09.073

What has a lower long term cost per hash It's not particularly relevant what the cost per hash is - if the hash is 100 times easier to compute, then it will be computed a 100 times more. – Nick ODell – 2013-03-09T03:44:28.730

@NickODell can you define easier? A technology that makes it easy? I'm trying to measure cost of ease. In a long term view, that may come down to electricity consumption. I worry that an ever growing need for energy would be needlessly devastating for the planet and result in centralization of power/control. e.g. Power Plants are in control of Bitcoin – goodguys_activate – 2013-03-09T03:50:01.263

Suppose I make my own cryptocurrency, NickCoins. Instead of the SHA256 hash that bitcoin uses for proof-of-work, I use SHA256 a hundred times. It now costs a hundred times as much to compute a single hash, but that doesn't mean it takes a hundred times as much computing power to secure my network. – Nick ODell – 2013-03-09T03:56:19.647

@makerofthings7: Bitcoin/Litecoin mining is artificially difficult. It's difficult on purpose. There's no point in trying to find an 'easier" hash function since it would be easier for a potential attacker too. The Litecoin hash function is much harder than Bitcoin's (by a factor of a few thousands) but that's irrelevant for what you want to know. If the hash function is easier it just means more hash calculations will be needed per block. For the purpose of being "green", what matters is the long-term ratio between the cost per device and its power consumption. – Meni Rosenfeld – 2013-09-01T21:03:25.430

@MeniRosenfeld - I don't care about the cost per device per-se, since that can be artificially inflated by the manufacturer. So what math problem is computationally hard, without expending tons of electricity scrypt, sha256, or something else? – goodguys_activate – 2013-09-02T12:26:44.647

@makerofthings7: The cost can be inflated by a manufacturer. But there is more than one manufacturer and the equilibrium price will be determined by fundamentals, not by any single manufacturer's whims. The (almost) only costs of computation are the device and the power, so to make a problem hard without requiring more power you need the manufacture of the device to be expensive. So I repeat - you need a computation that, with state of the art technology and optimally leveraging existing infrastructure, has the highest ratio of (manufacture cost per device / energy consumption per device). – Meni Rosenfeld – 2013-09-02T13:52:11.953

Answers

3

One of scrypt design goals was to make CPU and GPU performance balanced, which meant that it was far more dependent on memory speed.

The GPU had a solid work around and as far as I know CPUs don't have any advantage.

But that did cause it basically impossible to do scrypt mining with an ASIC unless it has a ton of memory and memory channels added in, which is a huge design problem. Unlikely unless GPU mining for scrypt is very profitable.

This also makes Scrypt far more energy intensive for the same amount of useful hashing, but the network adapts as it requires far less hashes to solve problems (Difficulty number)

Which is greener? By far ASIC mining, it's power usage is tiny compared to performance, 4500MH for 4.5w vs 275MH for 100w.

But again, the network balances.

Liquid5n0w

Posted 2013-03-08T18:14:04.110

Reputation: 95

Can you elaborate on why or how the network adapting requires less hashes to solve problems? – Rob Kielty – 2013-09-08T11:55:40.913

1

@RobKielty see https://en.bitcoin.it/wiki/Difficulty

– Tim S. – 2014-01-17T18:08:36.380

1

The network adapts by changing the difficulty level, which means adjusting the chance that any single hash computation is accepted as finding the next block. If the hash rate for the network goes down, the difficulty decreases which means any single hash calculation has a greater chance of finding a block. If the hash rate of the network goes up, the difficulty goes up and the lower chance there is of a single hash calculation finding a block. The system is designed this way so that, no matter how many hashers there are in the network, a block is going to be found every X minutes on average. (The value of X depends on the fixed design of the currency but is hard-coded so doesn't fluctuate once the design is fixed.)

Said another way, a hash is accepted as finding a block if the result of the hash is less than some value Y. What happens by adjusting the difficulty is that this Y value goes up or down. So more hashes on the network increases the difficulty and lowers Y. Fewer hashes decreases the difficulty and raises Y.

davmp

Posted 2013-03-08T18:14:04.110

Reputation: 111

Asked: 2013-03-08T18:14:04.110

Viewed: 1 326 times

Active: 2016-02-21T10:34:06.470