Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 691. (Read 2591920 times)

Some good luck today. 

And p2pool is sitting under 90% 90 day luck and seems to have been since January (when the pool actually started to grow).  The question still remains; Is this 90% really just "bad luck", or is it caused by some flaw in p2pool's code?

2 explanations for those not sure about things:

1) The E: value in cgminer really doesn't mean you are getting more BTC (higher E) or less BTC (lower E)
The easiest way to explain it is as an example.
This example isn't using exact numbers - in reality the E will be lower - but simplified for explanation.
... and by 'shares' I of course mean 1-difficulty pseudo shares as cgminer displays

A) If you have a computer with 2 devices that can each hash a full nonce range in 10 seconds (2^32 hashes) then on average (p2pool) it will complete each work item it gets and thus on average find 1 share per work item and have an E of 100%
B) If you have a computer with 4 devices that can each hash a full nonce range in 20 seconds (2^32 hashes) then on average (p2pool) it will complete only half of each work item and thus on average find 1 share per 2 work items and thus have an E of 50%

A) is of course 2 shares per 10 seconds and 2 work items per 10 seconds
B) is also 2 shares per 10 seconds but 4 work items per 10 seconds

Both the above computers have the same MH/s
Both of the 2 above will produce roughly the same amount of BTC
The second one just happens to get twice as much work to get the same result
There are other factors that come into the E value (e.g. roll-n-time) but the point is that E means how much work needs to be transferred to get your shares (not how many shares you expect to get over any given time frame)

2) Pool comparisons.
There is no point comparing pools over a day - that simply shows you don't understand how bitcoin mining works
Also saying "I got paid X BTC" in a short time frame (a day, a week or even a month) doesn't mean you will get that same amount the next day, week or even month.

Quite seriously, the only measurements of the performance of a pool that are valid, are:

A) the average of: "number of 1-difficulty shares" / "the block difficulty for each block found" ... over a LONG period of time (months)
 this is how you correctly calculate "LUCK" and should normally be close to 1.0 (100%) over a long period of time

B) the % of orphaned blocks

P2pool throws away share history ... so not everyone can check that.
Of course on P2Pool each share found actually counts for multiple 1-difficulty shares.
So if the current difficulty was say 650, then each share found while the p2pool difficulty was 650 would count as 650 "1-difficulty shares"

I only need People/worker who's use my node   

Well Done! Would love to see more p2pool Nodes in Europe

Hi together,

I'am new  to the P2Pool Project, but I will support the Idea.

So i have set up one Public Node and its ready to be used.

Who wants, ca Try it under

http://p2pool-cologne.dyndns.org:9332

The Stats can be seen on http://p2pool-cologne.dyndns.org:9332/static ,
The current Fee can be seen at  http://p2pool-cologne.dyndns.org:9332/fee


It's  an VM running on an VMware Cluster for HA at a location near Cologne / Germany.

The Server is also listed at http://nodes.p2pmine.com/


Have fun to test/use it

Greets

Stefan 

After about 3 days of mining my cgminer shows:
It is desktop PC, my kids (and me) are using it abut 12h/day ;]
I think 500-600MH/s card can give 100% (or more?) efficiency on my settings.

Because of fast longpooling efficiency shown by cgminer is much lower than on other pools (miner have less time to scan for solution).
Part of my cgminer.conf that I think can help:

What i realized is that a fast CPU is very essential for good share results, if the used machine isnt working with the right capacity usualy there are more stales/DOA (especialy when you run p2pool and cgminer on the same machine for example). Yes to move the p2pool/bitcoind to an other machine could bring the solution in my oppinion. My efficiency is usualy above 100% MPBM seems to reduce my CGMINER hashingspeed aswell...

Using cgminer I was getting 60-70% efficiency and 20-30 U/m. Configuration matters on p2pool.
-u bitcoinaddress+1  This helped to reduce D.O.A.
-g 1
-I 7

@Frizz23 Wow, both of your payouts are below the expected payout from Alloscomp's Bitcoin Calculator. Perhaps you have something misconfigured to get such bad results on both pools.

So?

In the next 24 hours it might be the other way round.

I tested p2pool for 1/2 year!! - not just 24 hours. Or a week. I used it for full 6 month.

And I only got about 80-90 percent of what I would have gotten using BTC Guilds PPS method.

- With BTC Guild PPS I get ~0.48 BTC / day for my 840 MH/s.
- Using p2pool I got ~0.39 BTC / day (average for the past 6 month) for my 840 MH/s.

And no, I did not have a faulty configuration. Or old version (btw. this constant need for updating p2pool - and bitcoind - was a pain in the ass too. No need to worry about that using deepbit, slush, etc.).

If someone tells me he's getting the same or more using p2pool compared to a "traditional" pool he is simply a liar. Or fool. Or both.

Given 100% luck, what would the BTC per gigahash be?

What about stats for the last 7, or 30, or 90 days?

Awesome sumarization! I absolutly agree!

P2pool rules! 

Thanks check_status!

Maybe all our "troubles" are from badly coded luck chart?

Your perception is flawed, just as is most peoples when it comes to know where they stand in terms of coin generation. The only way your pool can make more than p2pool is if your pool solves more blocks for the same amount of hashing power. So the only factor that matters is how many BTC per gigahash you are making per day and actually receive. Which scoring system is used doesn't matter when you are trying to compare coin generation quantity per gigahash per day. To determine where you stand, you need a relative position, PPS, like with http://www.alloscomp.com/bitcoin/calculator.php is currently at ⊅.63 per day per gigahash. What percentage are you in comparison to this value with each pool?

Here are my stats from May 19, 12
I have only used p2pool:
Daily average = ⊅0.66846 per Gigahash
The current expected ⊅BTC for the difficulty 1733208 is ⊅0.58016 per Gigahsh
My daily average is 15.24% better than the current expected payout.

The total network is producing ⊅7450 per 24hrs for 11.91 TH/s. How many ⊅ per GH/s is that? It's ⊅0.62552 per GH/s.
That means the total network is paying out 8.4% better than expected.

Where you at?

Edit: This is where you are at in the last 24 hours:

In the past 24 hours Deepbit has found 18 blocks at 3177 GH/s, giving them a value of ⊅0.28328612 per gigahash.
In the past 24 hours P2Pool has found 5 blocks at 229 GH/s, giving them a value of ⊅1.09170305 per gigahash.

How much worse is p2pool now?

bitcoind uses 8332 for JSON-RPC and 8333 for peer connections, and p2pool uses 9332 for RPC and 9333 for peer connections. You should open 8333, 9332, and 9333 for a public setup, but not 8332. For a private setup, just open 8333 and 9333.

If you restart p2pool shares count is starting over form 0
No worry, all you have shared is counted: you have payout > 0
Incomming connections will come over time, it is mostly 2-3 hrs after node start.
It should be port 9332 also if you want to connect miners from outside and allow other to see your stats

P2pool is taking all shares close to 1 (0.99xxxx) but accetping to payout only shares higher than pool diff. Pool diff is calculated to make 1 pool share every 10 second. It was over 600 when 350GH pool, now about 450 on 200GH pool.
More miners -> more power -> higher payout share diff.
That near to 1 shares are accepted to keep you informed that miner is working - see yours stats page.
Also this log indicating that you just have started pool and it is still checking last pool shares
Try open ports to get incomming connections.

My node stats:

CGminer solves difficulty 1 shares (that places like BTCguild and Deepbit accept). In an effort to keep a 10 second share distribution spread, P2pool uses much higher difficulty shares (last I checked, it was like ~600).

No, actually, it's running just fine.  The shares shown in cgminer are not the same shares as in p2pool.  To really get any sort of solid numbers from the p2pool client, you need to run it for about 24 hours.