Thread about GPU-mining and Litecoin

[coblee]

The original purpose of Litecoin is to be a CPU coin where anybody with their computer can mine litecoins. What has happened with Bitcoin is that GPU mining on bitcoin was a lot more efficient, so a lot of people starting mining bitcoins with GPUs. This pumped up the difficulty and made CPU mining unprofitable and therefore pointless. I don't want this to happen to Litecoin and I think most people agree with me on this.

Recently, there has been some rumors that mtrlt has modified his GPU miner to work with Litecoin. And he claims to have been able to create a GPU miner that outperforms CPU miner by a lot. Of course, all this could be FUD thrown at Litecoin by Solidcoin supporters. But I have talked to mtrlt about this and he seems genuine. So I'd like to get to the bottom of this.

Here's what I'd like to accomplish:
1) Figure out if GPU mining litecoins is indeed more efficient. And if so how much better is it.
2) Do we want to switch to a new hashing algorithm that is more GPU-hostile.
3) If we do want to switch, there are a ton of other questions. Can we modify scrypt params or do we need something totally different. How far away do we do the algorithm switch? How do we get miners/pools/clients ready for the switch so that there's no downtime?

Everyone, please refrain from SolidCoin bashing in this thread. And SolidCoin supporters, please refrain from posting unless you have something constructive to say. Thanks.

[1715019608]

1715019608

[1715019608]

1715019608

[1715019608]

1715019608

[Ahimoth]

If you want to believe me, then I can vouch for mtrlt's gpu miner being significantly more efficient than any current cpu miner for scrypt.

From what I know of the gpu miner, option 3 of modifying the scrypt parameter will have minimal impact. The pad size did not seem to matter much, and can be compressed for lack of a better word, with on the fly value reconstruction. So any increase in pad size will have a relatively equal impact on cpu miners until you exceed their cache size, at which point, gpus may become even more efficient.

I think you will be stuck with option 2, finding a completely different hashing algorithm.

[Ahimoth]

Not in an attempt to troll the thread, but if you look at solidcoin's hash code, you will see it has random reads and writes that are of varying size, spread out over a large memory range, and are randomly aligned. These are key techniques in creating havoc with a gpu's memory access methods. I would suggest looking for code that has similar traits if you really want to defeat gpu's or at least keep them on a level playing field with cpus.

[DeLorean]

I can't really contribute to ideas that would make Litecoin more GPU-unfriendly... but I do agree that it is essential to Litecoin to be CPU-friendly and CPU-unfriendly. If this turns out to be true, something needs to be done to remedy it or Litecoin has no valid purpose.

[ssvb]

1) Figure out if GPU mining litecoins is indeed more efficient. And if so how much better is it.

I guess the only option is to just implement GPU miner and check how fast it runs. It is really surprising that so few people have actually tried this so far.

3) If we do want to switch, there are a ton of other questions. Can we modify scrypt params or do we need something totally different.

Maybe scrypt author can be contacted and asked about his opinion?

How far away do we do the algorithm switch? How do we get miners/pools/clients ready for the switch so that there's no downtime?

This is actually interesting. If I understand it correctly, bitcoin itself does not rule out a possible change of hashing algorithm in the future (if the need arises). Attempting this for litecoin now could be treated as some kind of rehearsal and provide a valuable experience.

[someguy123]

Wouldn't changing the algorithm force a new blockchain for litecoin? That'd screw up every pool, exchange, client, etc. which is probably going to annoy a lot of the network.
And if that did cause some sort-of "litecoin 2", that would send the value of litecoin downhill, meaning anyone who currently has a lot of their money in litecoins, ends up with nothing...

Another thing is that mtrlt still refuses to show any proof... You can't really say much when he could just be spreading FUD. At most the only proof I see is his 3MH/s on squidnet, but that doesn't show his stale rate to tell whether it's all just bullshit "fake" work, and plus I know of a few people who have managed to get similar hashrates by mining on their businesses clusters, or home clusters, which surprisingly a lot of people do.

[Ahimoth]

Since I was asked to clarify, "significantly more efficient", I guess I will post some hash per watt numbers.

According to litecoin wiki mining hardware comparison, an AMD Phenom X4 955 at 3.6ghz gets 24kh @ 125 watts. This translates to 0.192kh per watt.
A gpu rig consisting of 69xx series gpus can produce 998kh @ 920 watts at the wall. This translates to 1.08kh per watt.

So does at least a 5.6 factor increase in *efficiency* qualify as "significantly more"?

Consider the litecoin wiki entry for the Intel Core i7 860 which produces 25kh at 153 watts (a believable wattage consumption for the entire system). It gives a system kh/watt score of only 0.163. The gpu example is now a factor of 6.6 times more efficient.

PS, Mtrlt has gotten better kh/watt scores by playing with the clocks and voltages, but I figured I would give you an initial test result.

[CoinHunter]

Wouldn't changing the algorithm force a new blockchain for litecoin? That'd screw up every pool, exchange, client, etc. which is probably going to annoy a lot of the network.
And if that did cause some sort-of "litecoin 2", that would send the value of litecoin downhill, meaning anyone who currently has a lot of their money in litecoins, ends up with nothing...

Coblee has multiple options.

1) Drop support for litecoin1 and start litecoin2 or whatever, and give people all the existing coins in Litecoin1 into Litecoin2 (I did this with SolidCoin v2). Of course this also means the guys GPU mining still have their new coins in the new network but it potentially means you can start fresh with a new algorithm. It's a lot of work however.

2) Do a forking change in litecoin itself. The problem with this is the "old network" will continue in parallel with the "new litecoin". Which leads to a lot of support issues "Which litecoin are you on?".

3) Start an entirely new coin. Like he did with Litecoin over fairbrix. Probably the second easiest option to pull off since you don't have to support multiple old things and allows him to do some things from scratch a bit better. However the downside is the original Litecoins probably decrease in value due to no more developer (like tenebrix and fairbrix).

4) Do nothing. The easiest option.

The hardest thing is, unless you are well versed in understanding CPU and GPU architecture making a CPU hard coin is very difficult. Given Coblees failure to know if Scrypt was GPU hard are we going to believe he can now make one that is? This will require a lot of work and a lot of testing to verify. And you're going to need talented C++ and OpenCL coders to help you out I think.

[shawnp0wers]

If the efficiency is 5-6 times that of a CPU, it's still a far cry from Bitcoin's CPU-worthlessness. It would be interesting to see the math concerning how worthwhile it would be to mine BTC vs LTC on a theoretical GPU scrypt miner.

Perhaps the ability to mine LTC with a GPU will still be moot. If it's more economical to mine for BTC and buy LTC with the profits, it would be foolish to mine LTC with a GPU even if possible.

Anyone like crunching numbers? What would the price of LTC need to be in order to make GPU mining worthwhile? I don't think we're there, but like I said, I haven't run the numbers at all.

[someguy123]

If the efficiency is 5-6 times that of a CPU, it's still a far cry from Bitcoin's CPU-worthlessness. It would be interesting to see the math concerning how worthwhile it would be to mine BTC vs LTC on a theoretical GPU scrypt miner.

Perhaps the ability to mine LTC with a GPU will still be moot. If it's more economical to mine for BTC and buy LTC with the profits, it would be foolish to mine LTC with a GPU even if possible.

Anyone like crunching numbers? What would the price of LTC need to be in order to make GPU mining worthwhile? I don't think we're there, but like I said, I haven't run the numbers at all.

Well,

[04:12:10] <@PooL-X> The expected generation output, at 200 KHps, given current difficulty of 1.5926655, is 126.31 LTC per day, 5.26 LTC per hour, Estimated time to find a block is 9 hours 30 minutes 3 seconds

That's the amount of kh/s mtrlt was saying, or at least an average: that's $1.20 a day.
Now, with my 6870 I get 300mh/s on bitcoin...which earns me an avg of 0.2BTC a day, which is around $1/day
So really they're about equal, if bitcoin price rises any more, it's more worthwhile to mine bitcoins on a GPU.

A gpu rig consisting of 69xx series gpus can produce 998kh @ 920 watts at the wall

Okay that's

[04:15:36] <@PooL-X> The expected generation output, at 998 KHps, given current difficulty of 1.5926655, is 630.27 LTC per day, 26.26 LTC per hour, Estimated time to find a block is 1 hour 54 minutes 15 seconds

630 LTC a day, thats around $6.30 a day, no idea how much they'd get with bitcoin, probably a LOT more $ value in GPU mining on bitcoin with them...

[gmaxwell]

assisting in a reduction of simultaneous block finds.

Having two different algorithms doesn't reduce simultaneous block finds, only making blocks further apart does.  

Anyone like crunching numbers? What would the price of LTC need to be in order to make GPU mining worthwhile? I don't think we're there, but like I said, I haven't run the numbers at all.

Why does it matter?  For may (most? almost all who aren't stealing power/computing?) people it's long been more efficient relative to power costs to mine bitcoin with gpus and sell for litecoin if you really want litecoin.   I can't see how this would make it any _worse_ than that.  Perhaps efficient GPU mining will take litecoin back from the thieves and make it profitable for honest people to mine again?

You can't really say much when he could just be spreading FUD.

So, how much have you placed against this at http://betsofbitco.in/item?id=292  ?   I'd think anyone sure enough to call someone a liar would be willing to put up way more than 1.50 BTC.

[shawnp0wers]

Interesting. It's a shame it so close, it would be nice to have a *little* wiggle room.  That said, price parity based on GPU speeds would be a legitimate way to stabilize LTC prices. I just wish it wouldn't stablize the price so low, lol!

[shawnp0wers]

Why does it matter?

The advantage isn't for folks willing to invest hardware, the advantage with CPU mining is that people can mine without any initial investment. I totally understand what you're saying, and maybe the "accessible to the masses" isn't even a practical idea -- but that's the idea. Anyone can mine with what they have.

[Ahimoth]

Anyone like crunching numbers? What would the price of LTC need to be in order to make GPU mining worthwhile? I don't think we're there, but like I said, I haven't run the numbers at all.

easy.. the same rig on bitcoin would make $5.17 daily, whereas on litecoin it would make $6.11 daily. This is according to allchains.info.

[DeathAndTaxes]

assisting in a reduction of simultaneous block finds.

Having two different algorithms doesn't reduce simultaneous block finds, only making blocks further apart does. 

I think you misunderstand what I was getting at ... it reduces the chance that a single source can find 2 blocks in a row ... makes the 51% issue maybe a 75% or more issue or whatever the stats would be.

You misunderstand 51% attack (like many other things).  With 51% (technically 50% + 1 hash per second) you will eventually have the longest chain., that is a mathematical certainty. 

Instead of attacker buying x CPUs or y GPUs he will now simply buy 1/2x CPUs and 1/2y GPUs.

[DeathAndTaxes]

You just proved what I was stating as the benefit, the attacker has to compete with BOTH sources .. either by "51%" both sources or having immensely more power on 1 source.. enough more hash power to compete with the more efficient miners.  And i called it a defense, not a prevention of that attack vector, meaning it just makes it harder.

Fail again.  

The network won't be twice as strong.  It will have half the hashing power in CPU and half in GPU.  The cost to attacker hasn't risen.

Buying X CPU.
vs
Buying Y GPU.
vs
Buying 1/2 X CPU AND 1/2Y GPU.

The strength of the network is split between both components.   At any given economic value the network will reach equilibirum between price and difficulty. By having different block types the combined difficulty for both blocks (2 blocks total) will be no higher than 2x a single block now.

Lots of extra code for no benefit.  It will not be any harder for an attacker to get 2, 6, 20, 100,000 blocks ahead of the honest chain using 2 types of blocks or 50.

TL/DR:
You:  If difficulty doubles then the network is stronger.  See you proved my point.
Me:
You:  See 1 + 1 is obviously harder than 2.
Me:

Maybe you suggest an idea he/LTC community could consider doing?

Nothing.  There is no proof this magical miner exists.  Changing hashing algorithm is a massive change and certainly non-trivial.  If/when independently the efficiency of GPU mining is verified the solution will depend on how high that efficiency is AND how various methods degrade that.

Doing something based on no proof is exactly what you want to happen.


If you would like to continue trolling my simple idea that coblee could choose to run with or not, fine but please stick to the OP and be planning what should be done for the benefit of LTC... is this not a more productive plan?  Maybe you suggest an idea he/LTC community could consider doing?
[/quote]

[DeathAndTaxes]

Is there some reason you want just a single difficulty?  That would not make a lot of sense, you would need a difficulty per block type, otherwise having too much hashing from one source makes it impossible to mine efficiently on the other.

LOLZ.  Are you that dense.

I just used 1 + 1 = 2x as an example.  The split doesn't matter and obviously each block would have indepedent difficulty.

Let me make it simple.

THE ENTIRE LITECOIN NETWORK GENERATES X USD PER UNIT OF TIME.

X buys so much computee things (bleep, bleep boop boop -> hashs come out).   

Difficulty is a manifestation of X.  If X rises so will difficulty.  If X falls so will difficulty.  There may be fluctuations but equilibrium will be reached between difficulty and price.

Thus the COMBINED difficulty will always be a reflection on X.  Changing the number and types of blocks won't change anything.

Combined difficulty will always be a reflection on X.  Thus the cost to build enough hashing power will be a reflection on difficulty which is a reflection on X.

Only way the network gets stronger is if price goes up.  Period.  Moore's law doesn't help (it also helps attackers).  New tech breakthroughs (like GPU mining at higher efficiency) doesn't help because attacker can use it too.   Lower energy costs globally (say cold fusion) doesn't help because ... yup it helps attacker too.

The only thing that makes the network stronger is the revenue paid to miners in the form of fees/subsidies.  Since miners costs are in fiat what ultimately matters is the amount of fiat generated per day.

[coblee]

Viper's solution is not without merit. It would make it harder to 51% a coin in the earpy stages because you'd need both a bitcoin farm and a botnet.

[DeathAndTaxes]

Viper's solution is not without merit. It would make it harder to 51% a coin in the earpy stages because you'd need both a bitcoin farm and a botnet.

There are easier methods to protect a coin in the early stages than saddling it with a dual block structure long after it is no longer needed. 

[ThiagoCMC]

Don't give up of Litecoin!!
Everybody wants to see for sure that GPU-mining for Litecoins is a reality and efficient...