Bitcoin Forum

This is a bit of a random thought that has come up while I've been researching ASIC and CPU mining... I'm hoping that someone a lot smarter than me can quickly figure this out so I don't have to waste nine months on calc. Or perhaps this has already been discussed and my attempted searches just missed it.

Let's assume for the purpose of this thought experiment that Coinlab keeps the vast majority of its GPU power concentrated on mining BTC for at least the next couple years... and that all the ASICs get delivered as promised and perform as estimated.

What would happen if coinlab signed a bunch of large MMOGs, or say one or two of the top ones? They have billions of playing hours a month... I don't know what the number of playing hours vs the number of hours that are allocated to coinlab mining looks like, but under any estimates it looks like they do have a lot of potential there, with those millions of beafy GPU cards and no hardware costs.

I suppose the easiest way to frame this question is... how many total GPU mining hours would coinlab have to control to compete with the TOTAL amount of ASIC power that will be live within the next six months? How much of the MMOG market would coinlab have to corner to get those kinds of hours?

Come to think of it, what kind of hash rate could coinlab see if they blow up to a degree no one predicted, and sign WoW and league of legends or other top MMOG like that? I apologize if this has been calculated already, but believe me I have searched.

Gamers will need to have not only GPU's but BPU (Bitcoin Processing Unit) too. Please realize that upcoming ASIC tech is by 2 !!! orders of magnitude more power efficient than GPU's. This means that GPU's will not make enough bitcoins to cover electricity costs. It is not so much about cost of hardware as about costs of electricity.

A quick illustration for you. A typical household electric circuitry cannot handle more than 20-30 Ghps worth of GPU mining rigs. While this is something a relatively tiny and low powered ASIC miner can handle. GPU is about to get outclassed by 100 times more efficient hardware.

There is no future for GPU mining, whether it is dressed in gaming closes or not.

Thank for the response. My question relates specifically to coinlab.com... their model is not GPU mining itself, as far as I understand, it is essentially to pay MMOGs for the spare GPU time of their (non-paying) customers. Electrical and hardware costs don't really come in to this equation, I suppose the real variable we are missing right now is the market value for that commodity; spare GPU time from MMOG players. I have very little information on that subject... if they already have a revenue stream for that GPU power, it may not be worth it for them to partner with coinlab. However, if it's true that the MMOG only profit from the 10% of players who are "whales" and spend money in game, and they aren't currently monetizing the majority of their players... it stands to reason that coinlab may be able to negotiate deals with these companies.

That is one of the variables, how much more efficient asic will be. If it's only an 100 to 1 advantage for ASIC miners, seems to me like that would be great for coinlab, since they have no hardware or electrical costs and could end up simply paying the MMOGs a percentage.

Right now the 22.4 Thash/s of capacity is over 30,000 AMD 5970s (or 6990s, or 7970s, which are all roughly about the same). 

Or take a mix of 5970s, 6970s, 5830s, etc., and the current capacity is roughly around 50,000 GPUs (or more).

Most gamers bought NVidias, and if all hashing capacity came from decent NVidia cards, you would need well over 100,000 of them.

That's existing capacity, not even considering the addition of the hashing capacity from GPUs, which will likely double to quadruble total capacity in a matter of weeks, just from the pre-paid ASIC purchases.




CoinLab has two initiatives.  The original was a client that mined bitcoins using the gamer's GPU hardware.  The proceeds go towards game subscription and in-game purchases   
The second, initiative (and now, main initiative, it appears) is to have GPU owners run a client that will do computing tasks on-demand, including scientific or other types of computations, or mining if that yields better revenue.  The presumption is that the revenue from mining using any GPUs will not be sufficient to pay for the cost of electricity, thus most will be retiring their GPUs from Bitcoin mining duty.

Thanks for the info. So if I am understanding your analysis correctly, if coinlab gets the equivalent of an average of 100,000 NVIDIA cards online multiplied by four ("double or quadruple total capacity")... so if you could average 400k GPU cards, you could match all pre-paid ASIC orders? Or am I misunderstanding this information?

Yes this is what I thought, but if doing other types of calculations do not start paying off for coinlab in the near-term, they will continue mining with the bulk of their capacity, right?

Unless I am missing something, coinlab doesn't have to consider cost of electricity or GPU hardware, simply the market price at which they can purchase spare GPU power from the MMOGs.



With all due respect I don't think that is true, I can't find evidence that it is, and I can't imagine why it would be. The in game incentives offered to players for running the coinlab software while idle neednot be costly.

Most gamers have mid range cards like 5770/6770 doing at most 200MHs (nVidia half that)

Each individual ASIC chip is capable of doing 7500MHs, low end Jalapeno 4500MHs (average 6000MHs)

Results in 30x average GPUs to 1x average ASIC.

But the game may not run well without at least half the GPU (30fps/60fps) so make that 60x GPU's v 1x ASIC if it's done while they are paying.

Then double it again, because who really will game 24x7.

60x GPUs should be able to generate about $1.18 a day or $430.70 this year, even after difficulty rises, and reward drops, while at 1BTC=$12 assuming if they played 24hrs a day.

120x GPUs should be able to generate about $1.18 a day or $430.70 this year, even after difficulty rises, and reward drops, while at 1BTC=$12 assuming if they played 12hrs a day.

I hope Coinlabs have more aces in their sleeves. Otherwise this is a doomed project, and it has been doomed from the start. There is simply no way Coinlab would make any money at all, let alone millions they would need to recoup "investments" on any of the mentioned above initiatives. Not in environment where specialized hardware that is by two orders of magnitude more efficient and is widely available.

Initiative 1. Mining for gamers. No sane gamer will do it, at least not after his parent's see the electricity bill and then there will be lawsuits for failure to notify about their "shift paying for your games to your electricity bill and pay four times more" idea. (think not about now but about 2013)

Initiative 2. Back to usual pre-bitcoin age folding and SETI and such. Nothing new here. Same "do charity via paying to your utility company" scheme.

These "initiatives" are ok for a typical 17yo and not very bright Bitcoin enthusiast who frequents these forums. Not ok, by any stretch of imagination for a VC funded startup. But maybe, hopefully even, I do not know something about Coinlabs and they actually do something smart with that VC funding.


Gatorhex: Please add to your calculations cost of electricity consumed and calculate gross profit from GPU mining (hint: it is going to be negative). This would make it abundantly clear why this simply cannot be profitable and why "ohh but the hardware is free" argument is a not such a good one.

+1

Before ASICs, Coinlab idea was awesome!

After ASICs, I don't see how Coinlab can still operate and make money unless they plan to use that GPU power for other purposes.

As per your point, I thought that was their goal, to basically have our GPU's do HPC work for projects they contract with other companies that need HPC work. Think of us GPU owners being an amazon compute instance. They would pay us GPU owners in BTC for the work attributed to their HPC contracts.

If that is what their plan was, I don't see why they still wouldn't have a nice business idea. There could be a great number of other projects for which GPU's would be useful for other then mining Bitcoin. ASICS will mine only bitcoin, they won't Crack passwords or run simulations. It sounds a for project Seti@home type of situation.

Interesting.

Why would they need "gamers" to do that? Why not just create a software that "mines" the HPC work that needs to be done so that you don't need to be in a game to do it. And not just be in the game, but be idle in the game.

Actually if there are enough requests for high performance computing in the world, this might be a good business idea: start a company that develops mining software for specific HPC tasks and acts as a mining pool for the GPU/CPU miners that join the pool.

You just described CoinLab's second initiative.

You are only talking about one or two GPUs per gamer, which will not be noticed on the electricity bill by 99% of parents, especially if it is mining only when the game is on.  The bigger question is when difficulty increases, will they be able to provide the game companies with enough incentive to even bother with it.

I think the biggest problem for coinlab is the move to tablet and mobile gaming.  MMOGs were the hot thing "last decade", which is why you see so many of them going free to play nowadays.

what are you guys all "I hope they have more ace in their sleeve"?

botting gamers boxes because they clicked some bullshit EULA nobody reads is W R O N G

coinlab can die in a fire

Where did you get that idea about us?

When we partner with a game company, we create a stand-alone mining client, branded with that game's logo, etc.  Mining is opt-in only and the mining player earns in-game virtual currency for their GPU's efforts.  We convert the Bitcoins they earn into that game's virtual currency, so players earn proportionally to their hashrate.


So far we are focusing on F2P (free to play) games because they have in-game purchases which their players can earn by running the client.  We won't partner with a game company that just wants to mine BTC and wont give their players anything in return.



Yep, this is exactly what we are working on. Earn BTC for "mining" paid HPC work.

Actually, I was hoping it would be visible in game, like click a button in the corner and you see the premium currency start incrementing while all you are doing is chatting with others, browsing the market, or grinding xp.  Then when you are about to do some important things like a raid or a boss fight, you click it to turn it off, dedicating all performance to the game.

I would never want them to bot gamers' boxes like how Skype hogs up your bandwidth because of some fine print in the EULA about the peer to peer nature of the technology.


I figured that you would be focusing on F2P because that's where you can find in-game rewards to give out in return for mining.  What I didn't realize is that you create a separate client, so it can be run on a more powerful machine while the game itself might be run on portable devices that are not effective miners.

To definitively answer OP's question:

If we assume 100,000 of the average gamer's GPUs to be equal to the current network hashrate (23TH/s), and we estimate ASICs will increase that difficulty to 250TH/s (a good average of the various estimates being thrown around), then we can conclude it would take 1,087,000 gamer's GPU's to equate to the network hashrate after the ASIC bump.

But, that would assume 24/7 usage.  Most gamers don't run their games 24/7 - maybe 2 hours a day, on average.  Now you need 13M GPU's.  But the hashrate while gaming drops by half, typically.  Now you need 26M GPU's.

Hmmm... well good luck with that coinlab!

Honestly, I think it is an awesome project, and I hope it is somehow successful.  ;)

So you aren't that guy from the Seattle meetup proposal?

perhaps I spoke too soon

Since this is the Speculation subforum...  ;) :D

I don't see the point of CoinLab's second half. To me, they started as a bitcoin pool, but wanted to be more than that. Now they have to be more than that to survive, but I can't see how they actually plan on doing that.

Actually coinlab was never going to be a bitcoin pool as far as I can tell. I'm guessing the pool was found to be a good stepping stone for people with GPUs to get out of the bitcoin mining scene when it becomes unprofitable.

Disclaimer: I work for coinlab but only as a software contractor and have very little knowledge of the goings on in the company.

Their model seems to be to persuade the gamer to run their machine and a stand-alone client 24/7, instead of it running concurrently while gaming.

Quoted for hilarity. You core bitcoiners are even more delusional than I assumed.
Sorry if that meant as a joke, I reference poes law as excuse in that case.