[CLOSED] Massively Merged Mining

[markm]

UPDATE: POOL CLOSED DUE TO LACK OF USE/DEMAND.

The DevCoin pool bounty has been made more general, now allowing a pool that has devcoins as one of its payout options instead of having to use devcoins as its only payout option.

Thus I can open my pool's payout option to actually putting pool shares on a market, where people can buy and sell pool shares; or I can simply offer to pay out in any type of coin or other asset traded on my Open Transactions server.

Currently my pool mines Bitcoins, Devcoins, Groupcoins, Namecoins, Ixcoins, I0coins and Coiledcoins. However the hope of various other altcoins such as Britcoins, Botcoins, bitNicKeLs and so on and so on and so on is that if we can get enough hashing power into Massively Merged Mining pools they too will be able to move back to being public blockchains instead of having to hide in the Open Transactions server from threats such as 51% attacks to which blockchains with low hashing power are vulnerable.

Note that this is alpha stage: I have p2pool running, it has been working for me for months, however few third parties has done much mining on it  it yet so I do not know how its statistics work, how vulnerable it is to miners who cheat and so on.

Since we mine more chains than any other pool, and do not intend to charge any fee/commission, we should be able to pay miners more than any other pool other, perhaps, than the one that resells hashing at a premium on demand for unknown purposes.10332/static/graphs.html

The pool should be reachable as port 10332 at p2pool.knotwork.com (That is, tell your miner program http://p2pool.knotwork.com:10332)

Use a bitcoin address as username; password is optional. Payouts go to the bitcoin address you used as username. Payouts are automatically created as part of the mined blocks.

It looks like the most general measure of contribution currently is on the graphs page at

http://p2pool.knotwork.com:10332/static/graphs.html

The so called "area" value seems to be the product of the hash rate and the amount of time spent hashing.

It does not mention difficulty though so I am still looking through the data directory trying to find something that will be more useful to compute payouts from. EDIT: I have now set commission to ZERO so miners will get ALL the bitcoins they mine, and that means the bitcoin blockchain will serve as a record of miners' contribution to the mining overall. I can thus periodically add up how many bitcoins the pool awarded them as shown on the blockchain and figure out from that how much "altchain bonus" to award them in their choice of altcoin.

I already have archives happening of historical values of various assets, see http://galaxies.mygamesonline.org/digitalisassets.html for some tables derived from those archives showing asset values expressed in a few selected assets. (I should expand the selection, there isn't really much reason other than not having to gotten around to it not to provide such tables expressed in terms of each of the assets.) So conversion rates between asset types are not a problem, the problem currently is how to figure out how many of each type of coin either was mined or should on average statistically have been mined so that those values can be converted into each user's preferred payout asset.

-MarkM-

[doublec]

Since we mine more chains than any other pool, and do not intend to charge any fee/commission

http://p2pool.knotwork.com:10332/fee returns 100. Is that a 100% fee?

[markm]

Since we mine more chains than any other pool, and do not intend to charge any fee/commission

http://p2pool.knotwork.com:10332/fee returns 100. Is that a 100% fee?

That is an artifact of using p2pool.

By default, p2pool pays out only in the primary chain (in this case that would be bitcoin), paying out in microtransactions it puts into the blocks it creates.

By setting the commission rate to 100%, I presumably prevent this automatic payout of bitcoins, which in any case does not take into account all the merged chains also being mined at the same time.

Unfortunately p2pool does not let a negative commision rate cause people to be payed out more than they mined themselves on that primary chain, otherwise I could have set a negative commission to let miners be sent extra bitcoins to account for the merged chains. (But would not be able to offer to pay in devcoins instead of bitcoins...)

Since one can qualify for the devcoin-pool bounty without paying out all the owed money in devcoins, I suppose one arrangement that miners might find attractive could be to set the commission to ZERO so they get full no-commission payout in bitcoins for the bitcoin mining, then worry separately about all the merged chains...

-MarkM-

EDIT: Since I am not seeing any responses, I am going with that ZERO COMMISSION model for now to see if I get responses to that. I can always open more ports and add more machines to end up running more different commission-models. Lets see if full payout of bitcoins automatically will attract some miners...

[doublec]

EDIT: Since I am not seeing any responses, I am going with that ZERO COMMISSION model for now to see if I get responses to that. I can always open more ports and add more machines to end up running more different commission-models. Lets see if full payout of bitcoins automatically will attract some miners...

You might want to include instructions on how users mine there for those not familiar with p2pool. For example, what username, password, how they get rewards, etc.

[markm]

Thanks. Added

Use a bitcoin address as username; password is optional. Payouts go to the bitcoin address you used as username. Payouts are automatically created as part of the mined blocks.

to the original post.

-MarkM-

[matthewh3]

Hi looks interesting and I was looking at using P2Pool or P2Pool based pool once I get my ASIC's.  The one thing I don't understand is how you get your other merged-mined coins apart from BTC.  Sorry if I've not read it proper.   

[markm]

Hi looks interesting and I was looking at using P2Pool or P2Pool based pool once I get my ASIC's.  The one thing I don't understand is how you get your other merged-mined coins apart from BTC.  Sorry if I've not read it proper.   

Yeah that is going to be the hard part.

Really the plan was never to try to give everyone all the types of coins. Originally, in line with the old version of the DeVCoin pool bounty, the plan was to pay out only in DeVCoins, converting all the other types of coin into DeVCoins for that purpose.

Now that the "must pay out only in devcoins" requirement for the bounty has gone, I have set it up to send out all the bitcoins as bitcoins automatically mainly because that ensures instant payouts for the bitcoin part of it, hopefully "buying time" for figuring out the technical details of how to figure out how many of all the other types of coins that same hashing either did mine or statistically on average would have mined.

Another factor here too is a recent thread in this subforum claiming that merged mining altcoins along with bitcoins is not actually paying pool miners as much as they get at a bitcoins-only pool. So I also had in mind that as long as I pay out at least as much as, and hopefully more than, the pool that chap was claiming to be so profitable, miners should be well pleased.

I would prefer to pay the altcoins reward in DeVCoins, partly because promoting DeVCoin is the original motivation for the pool and partly because DeVCoin addresses are exactly the same as bitcoin addresses, so that I can simply send people DeVCoins to the exact same address their bitcoins get sent to.

I am still trying to figure out the best way to actually compute or tally the amount the altcoin mining adds up to. I had an idea that ZI could look in the bitcoin blockchain to see how much bitcoin a miner earned and compute from that, but it occurs to me now that maybe they could p2pool mine at other pools using the same bitcoin address there and it would not be possible to see from the blockchain how much of their mining was done here thus also applicable to the altchains. So the method of figuring out the altcoin part is still not really finalised.

The fact that I am charging NO COMMISSION on the bitcoins hopefully by itself makes up for a lot; that alone might make just the bitcoin payouts alone come to more than what a miner would make at many other pools.

Another goal of the Massively Merged Mining project is to try to get so much hashing power running that it will be possible for all the chains that went into hiding mode when people started attacking merged mine altchains isntead of using merged mining as intended, to support merged altchains, to come out of hiding. The value reports on those, online at http://galaxies.mygamesonline.org/digitalisassets.html show they are doing well enough that adding them to the merged mining mix should turn out to be quite worthwhile, though it will also add to the complexity of calculating mining rewards.

I was actually tempted by the thread about merged mining pools not paying as much as a bitcoin-only pool to make it really simple and just pay per share more than any other pool, as that would eliminate the need to actually figure out how much the the merge actually brings in. But, pay per share is maybe not such a great way to pay, being too easy to cheat, and in any case p2pool doesn't pay that way.

If there is strong resistance/objection to receiving the altchains portion of miner awards as DeVCoins, I guess the other method would be for each miner to pick one altcoin they want their altchain awards to be paid in.

-MarkM-

[matthewh3]

Thanks.  If you add payouts in alt-coins on your P2Pool pool you can definitely count me in once I have my ASIC's.

[markm]

BiTCoins go out automatically, I don't know how many of those people got.

Our hashing rate is low so we expect massive random luck variation in the merged chains; for example we have not found any devcoin, namecoin, ixcoin or coiledcoin blocks yet.

Currently so far we have mined 4450.00000000 GRouPcoin, 456.00000000 I0Coin, according to the actual mature balances readouts of the daemons. I took the extra time to look at the actual transactions to see if we found blocks that have not matured yet; with I0Coin that will usually be unlikely at our current low hash rate, and sure enough there are none; with GRouPcoin we usually get all the blocks so yes of course there are many many immature blocks in the GRouPcoin pipeline.

The latest valuations in terms of devcoins, as shown at http://galaxies.mygamesonline.org/digitalisassets.html is 6.72493150 DVC per GRP and 6.68054794 DVC per I0C, at which rates our gross altchain income so far would be 32972.27503564 DVC (3046.32986064 from the I0C plus 29925.94517500 from the GRP).

Supposedly some code for divvying up mining proceeds from pools based on p2pool is to be open-sourced this month so I am somewhat inclined to wait to see that code before finalising how best to go about divvying up these proceeds.

Since already p2pool has everyone's bitcoin address, and devcoin addresses are identical to devcoin addresses, I still think paying it all in devcoins is best simply because that involves no extra overhead such as finding out and recording and looking up and using different addresses for people.

(Although, actually, it is I think technically possibly to convert addresses even between other chains, so in principle it maybe would be do-able to calculate groupcoin, namecoin, i0coin, ixcoin and coiledcoin addresses directly from people's bitcoin addresses, that they would use by importing their private key into the appropriate other coin's software...)

-MarkM-

[markm]

Thanks.  If you add payouts in alt-coins on your P2Pool pool you can definitely count me in once I have my ASIC's.

Once upon a time I read somewhere that Butterfly Labs stuff does not play well with p2pool, and that they had no plans to correct that when they do ASICs. Has p2pool maybe done something meanwhile to fix it at their end, or is it that you plan to use some other brand of ASIC?

-MarkM-

[matthewh3]

Thanks.  If you add payouts in alt-coins on your P2Pool pool you can definitely count me in once I have my ASIC's.

Once upon a time I read somewhere that Butterfly Labs stuff does not play well with p2pool, and that they had no plans to correct that when they do ASICs. Has p2pool maybe done something meanwhile to fix it at their end, or is it that you plan to use some other brand of ASIC?

-MarkM-

Oh I thought you was a normal PPS pool being a node off P2Pool 

[markm]

What miners onnect to is, simply, p2pool, I have not modified p2pool it is standard version fresh from github.

I do not know whether what I read once upon a time about Butterfly Labs products not being p2pool-friendly was true or not, nor whether, even if it was true once upon a time, it is still true now or will be true by the time ASICs are delivered to miners.

Also, p2pool is not straight PPS, I think? I think it is same first letters to start, but more letters than that in the acronym for what it does use? Some kind of pay for last X number of shares or something?

-MarkM-

[irritant]

from FPGA-README.txt from cgminer:

Bitforce

--bfl-range         Use nonce range on bitforce devices if supported

This option is only for bitforce devices. Earlier devices such as the single
did not have any way of doing small amounts of work which meant that a lot of
work could be lost across block changes. Some of the "minirigs" have support
for doing this, so less work is lost across a longpoll. However, it comes at
a cost of 1% in overall hashrate so this feature is disabled by default. It
is only recommended you enable this if you are mining with a minirig on
p2pool.



, long polling works but maybe not as good on p2pool... but i'm not sure if this is still .. up to date , with ASICS i don't know if they support --bfl-range,

i dont really know anything 

[gnpeer]

I'm testing the pool. How all of the merged coins would be transferred?

[markm]

I'm testing the pool. How all of the merged coins would be transferred?

Converted to DeVCoins and sent to the miner's bitcoin address, since devcoin address and bitcoin address are the same.

In principle other means are technically possible but this is the simplest most straightforward method.

If that method does not suit you, then it is probably a good idea to try to get an Open Transactions client set up and running so you can make use of my Digitalis Open Transactions server, which handles so far all of the currently merged chains except for coiledcoin (which might, arguably, be dead) and also all the chains that decided, in the face of the massive hostility being shown to merged chains, to retreat to Open Transactions format leaving GRouPcoin still out in the world as a blockchain so they can watch how GRouPcoin gets along and use it to guage at what point it will become reasonable to move back to blockchain format.

-MarkM-

[gnpeer]

A bit complicated thing.

And I have 10% rejects :/

[markm]

Hmm, p2pool's console output is saying it has 8.4% stale rate and the p2pool network as a whole has 7% stale rate, and our percent has slowly been creeping down toward to overall network's rate though when running so many merged chains it is to be expected that there will be some impact on the primary chain from all the extra processing the system is doing to handle all the additional chains.

-MarkM-

[doublec]

Hmm, p2pool's console output is saying it has 8.4% stale rate and the p2pool network as a whole has 7% stale rate, and our percent has slowly been creeping down toward to overall network's rate though when running so many merged chains it is to be expected that there will be some impact on the primary chain from all the extra processing the system is doing to handle all the additional chains.

p2pool has 10 second longpoll times. High stales are likely to be expected. Does it consider a result a stale if it's stale on the altchain but not the primary chain?

[markm]

Hmm, p2pool's console output is saying it has 8.4% stale rate and the p2pool network as a whole has 7% stale rate, and our percent has slowly been creeping down toward to overall network's rate though when running so many merged chains it is to be expected that there will be some impact on the primary chain from all the extra processing the system is doing to handle all the additional chains.

p2pool has 10 second longpoll times. High stales are likely to be expected. Does it consider a result a stale if it's stale on the altchain but not the primary chain?

Hmm probably people on p2pool's own thread would be more likely to know the answer to that. Interesting question, presumably the author had some particular reason for doing it one way or another way, its been worked on a long time with lots of fixes. I guess that doesn't mean its about as good as it can get yet though I suppose.

-MarkM-

[doublec]

Hmm probably people on p2pool's own thread would be more likely to know the answer to that. Interesting question, presumably the author had some particular reason for doing it one way or another way, its been worked on a long time with lots of fixes. I guess that doesn't mean its about as good as it can get yet though I suppose.

Yeah I assume they don't treat alt stales as the whole result being stale. I seem to recall that high stales on p2pool is normal but doesn't mean the same thing as high stales on standard pools so it's not a big issue.