Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 707. (Read 2591919 times)

Test version seems to be much more stable

Dead again
http://bitpoppool.geekgalaxy.com:9332/static/graphs.html
http://screencast.com/t/mIYH7UNBd

Testing now

Pretty sure p2pool_win32_1f87c32 worked good

Wait its scrolling again and the graph works, seems to freeze up but maybe comes back. I just am too inpatient to wait.

Here's a build that doesn't use IOCP: http://u.forre.st/u/rhkxgsda/p2pool_win32_0.11.0-dirty.zip Can you test that?

Do you remember specifically which version last worked fine?

PS it always does in with the last line showing the hours like 6.2 hrs

Stop answering to my site graphs and to the miners.
Been having this issue for several versions, I usually restart like 15 times until it works. I'm using 2008 r2 x64

What do you mean by "stop answering"? Versions since 0.10.4 on Windows use the IOCP API, which may be causing problems. What version of Windows are you using, and does 0.10.3 run without problems?

Is it normal to get stuck at this screen and stop answering? I've already restarted a few times. All the versions do this. Do i wait or keep restarting?
http://screencast.com/t/uVYh9xNKgB5

Beautiful http://bitpoppool.geekgalaxy.com:9332/static/graphs.html

P2Pool release 0.11.1 tag: 0.11.1 UPGRADE REQUIRED SOON

Windows binary: http://u.forre.st/u/xyewbxrx/p2pool_win32_0.11.1.zip
Source ZIP: https://github.com/forrestv/p2pool/zipball/0.11.1
Source TAR: https://github.com/forrestv/p2pool/tarball/0.11.1

This release's sole purpose is to test possible fixes to our luck problems and provide useful data that may also help. It is a mandatory upgrade because it breaks compatibility with previous releases.

Major changes:
* Broadcast all block solutions, bitcoin/p2pool-stale or not, until they're 5 blocks old
* Submit blocks over P2P connection to bitcoind, in addition to RPC
* Automatically report caught exceptions as bug reports

This switchover is different from previous ones - Instead of a scheduled time, all nodes will automatically switch 22 hours after 95% of the P2Pool hash rate has upgraded. Therefore, please upgrade as soon as possible so you're not part of the last 5%!

EDIT: The current proportion of the hash rate upgraded is visible as the "2" line on the last chart at http://forre.st:9332/static/graphs.html
EDIT: Released 0.11.1 with IOCP disabled by default.

Another thing I wanted to point out, perhaps the most obvious of all which is why I forgot to post it before, is I wanted to run a solid month with p2pool to see how my payouts were compared with another pool. In theory, due to variance, over that period of time my payout compared to another pool should be the same.

My payouts with p2pool have been less, around 80% of what I would have if I had mined with deepbit. To clarify, in the past I used the 'Reward estimation' feature on the deepbit site while I was mining with them and it very closely matched my payouts. I have been entering my hashrate (phoenix hashrate not p2pool hashrate) and the estimated payout from deepbit has been between 2.5 and 2.7 btc/day for the whole month. When I work out 2.6 btc * the number of days I've been mining solidly with p2pool, it's about 80% of my actual p2pool payout.

I guess the best way to really confirm this is to set half my miners to p2pool and half to deepbit and compare balances over that period, but I am quite certain that over the past month I have earned less with p2pool mining than if I had been with another pool. I know there are factors like difficulty changing etc so really the approach I have made to this experiment is not a very scientific one, mostly because I planned to mine with p2pool no matter the 'cost', but mining is an economic activity and it doesn't seem like my payout with p2pool is anywhere near the same as with other pools, when in theory it should be.

The point of me posting this is not to detract other users away from p2pool, I just want to try to understand a) whether something is incorrect with the pool and b) if there is, why that is so. I want to see p2pool be as good or better than the other pools so I can join again.

you can use a pps p2pool. luck is statistical, you will always make more over a long run, eg. a month.

p2pool for the past one month solid (on and off with p2pool for a couple of months before that).

started mining about a year ago, using ozcoin and deepbit pools depending on my mood, until the switch to p2pool a few months ago. Sadly I have left p2pool for ozcoin again, I just don't see how a month straight can have that sort of luck. I would rather support p2pool but the luck is a killer. I can take some amount of worse luck to support p2pool, but ozcoin is a great pool so I don't feel bitcoin is loosing anything by my supporting them rather than p2pool.

There is where it's different for me and you.

My actual is 1250mhash. P2pool.info will give me a range (depending upon when I check) between 1100 and 1500. So that makes sense for me. How long have you been mining?

But to be lucky all the time makes it seem like it's an incorrect value and not due to luck. I never saw my p2pool hashrate reflect my phoenix hashrate.

Is the hashrate used in the calculation of luck or is it just 'an indicator that things are happening' for the user?

I had mentioned this before, and I believe someonw responded saying that the reported hashrate from p2pool.info is based upon the shares you have per the time, so you might just been lucky in your share finding which raised the "hashrate" that p2pool saw.

Actually ... the luck calculation should be based on shares per block ...
Which I presume it is anyway.

I have a theory that hashrate is being reported incorrectly by p2pool and is skewing the luck calculation. I realise that luck is a comparison of hashes by shares vs hashes in blocks and may not be related to the reported GH/s value, but I thought the numbers below were too compelling not to mention.

My phoenix thread says my hashrate per gpu core is always 380-390 MH/s. It's very consistent. I have 12 cores so that makes for a hashrate range between 4.56 GH/s and 4.68 GH/s

In my p2pool thread I see my total local hashrate usually around 4.8 to 5 GH/s and often up to 5.1 GH/s

Taking the average phoenix hashrate vs the average p2pool hashrate and the result is about 4.62 / 4.9 = 0.94

The most compelling part of all is that I recently switched to ozcoin for a while and they also report my hashrate at around 4.6 GH/s - the reported value by p2pool is never that low.

Also, when comparing actual payout to expected, make sure not to forget the author donation.  0.5% isn't much, but it still can matter.

Another reduction will come from the fact that some % of miners won't have 'submit stale' enabled in whatever they mine with.
No idea how big or small that is, but, of course, that relates directly to what I was mentioning before.