Author

Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 553. (Read 2591920 times)

sr. member
Activity: 344
Merit: 250
Flixxo - Watch, Share, Earn!
WOW! How?

1) actual bitcoin from git ( moves  variables ...Fees.. from main.h to main.cpp)
 2)  set in src/main.cpp the parameters to
int64 CTransaction::nMinTxFee = 1000000000;    # Override with -mintxfee
int64 CTransaction::nMinRelayTxFee = 1000000000;





3) compile bitcoin

EDIT:
Or.. not tested:  set  -mintxfee  -minrelaytxfee to 1000000000  without editing main.cpp
 

4) in bitcoin.conf:
blockmaxsize=5000
blockprioritysize=0
blockminsize=0

Greets
Basically your saying that pool mining is 100 times better for bitcoin than your p2pool settings coz some pools can commit up to 100x the transactions that you do.
Transactions are the other part of what is necessary to keep bitcoin alive.
No transactions means no bitcoin.

I wouldn start a discussion over the pro and cons of TX handling with P2pool.

i  would only show the impact of TX integration on the Performance in an actual setup.
legendary
Activity: 1540
Merit: 1001
I'm not sure there'd be much benefit to C compared to some other languages which are more easily cross-platform such as Perl or Java. More power to the elbow of anyone who wants to though.
There is a benefit in using a compiled language that does not have garbage collection and gives you access to low-level OS primitives - you can write really tight high performance code. I think that C++ would be a better choice over C, though.

Don't get me wrong, I'm a long time C/C++ advocate and user. But is something like this really in that much need of that level tight, high-performance code? (Maybe it is, I just don't know that much about it).

It seems to me there's a scaling issue, which is why it isn't working properly with an asic as small as a jalapeno.  I'm concerned about cross platform compatibility too.

M
legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
WOW! How?

1) actual bitcoin from git ( moves  variables ...Fees.. from main.h to main.cpp)
 2)  set in src/main.cpp the parameters to
int64 CTransaction::nMinTxFee = 1000000000;    # Override with -mintxfee
int64 CTransaction::nMinRelayTxFee = 1000000000;





3) compile bitcoin

EDIT:
Or.. not tested:  set  -mintxfee  -minrelaytxfee to 1000000000  without editing main.cpp
 

4) in bitcoin.conf:
blockmaxsize=5000
blockprioritysize=0
blockminsize=0

Greets
Basically your saying that pool mining is 100 times better for bitcoin than your p2pool settings coz some pools can commit up to 100x the transactions that you do.
Transactions are the other part of what is necessary to keep bitcoin alive.
No transactions means no bitcoin.
full member
Activity: 147
Merit: 100
Do you like fire? I'm full of it.
100% agreed. A few pages back I posted I was having trouble with p2pool's latency going up over time, now it seems more and more people are reporting it. Reverting to 11.3.

Something that might help you debug it: I found that p2pool returns to normal latency not when you restart it, but when you restart the bitcoin daemon. It might not be purely a latency issue with p2pool itself but rather the way it communicates with bitcoind.

Old P2Pool: http://u.forre.st/u/drxlzowj/p2pool_win32_11.3.zip

PS: Forrestv it really doesn't help that your binaries repository donesn't have a standardized folder structure and the builds are inside random-named folders whose root is unbrowsable. I had to use google's cache. Just sayin. Cheesy
legendary
Activity: 3248
Merit: 1070
can anyone give me the link for 11.3 version?

thank you
sr. member
Activity: 344
Merit: 250
Flixxo - Watch, Share, Earn!
WOW! How?

1) actual bitcoin from git ( moves  variables ...Fees.. from main.h to main.cpp)
 2)  set in src/main.cpp the parameters to
int64 CTransaction::nMinTxFee = 1000000000;    # Override with -mintxfee
int64 CTransaction::nMinRelayTxFee = 1000000000;





3) compile bitcoin

EDIT:
Or.. not tested:  set  -mintxfee  -minrelaytxfee to 1000000000  without editing main.cpp
 

4) in bitcoin.conf:
blockmaxsize=5000
blockprioritysize=0
blockminsize=0

Greets
legendary
Activity: 1361
Merit: 1003
Don`t panic! Organize!
sr. member
Activity: 344
Merit: 250
Flixxo - Watch, Share, Earn!
newbie
Activity: 52
Merit: 0
Quote
Could you post the line from the p2pool program feed that looks like:
New work for worker! Difficulty: 0.000200 Share difficulty: 0.994552 Total block value: 50.210500 LTC including 10 transactions

It's still the same as mentioned above:  "New work for worker! Difficulty: 0.232827 Share difficulty: 5.000000 Total block value: 53.123000 LTC including 45 transactions"  This is using: Usrnme/5+5

P2pool version is: 11.3 (I tried 11.4, that seemed to decrease my share rate further)  Cgminer version: 3.1.0.

I'm receiving payouts.  The odd behavior is receiving consistent share/payout per block of .45 LTC for 12+ hours, then it drops to .10 for another 12 + hours.  This happens when the p2pool node has been operating fine, along with all the miners connected.  Smooth & even hashrates.  The overall pool hashrate has a mean score of 580MH/s.  I would anticipate a much lower payout if the pool hashrate jumped to 1,200MH/s, but it stays within 100-150+- MH/s variance of the mean score.  My efficiency is normally between 110%-120%.

Thank you for the reply. 

Thanks - I think this shows what's wrong:

1) your +5 difficulty is way to high so p2pool has already reduced it down to .232 (otherwise it would get no statistics)
2) the /5 difficulty is causing 200% higher variance than is typical. The payouts change so much because you are only submitting a rare high difficulty share instead of frequent low difficulty shares. p2pool is not a PPS payout system, but a PPLNS. When you find a share, you will get paid out every time p2pool finds a block for some period of time. If you find shares infrequently, there will be much variance. With a /5 difficulty, you'll find shares infrequently.

Both of those settings won't affect your expected or average payout, just change the variance and frequency of payouts.

i'd try one of these two:
1) don't use /difficulty+difficulty at all and let p2pool tune it for you (especially good if you're a small miner)
2) If you're a bigger miner and want to tune your difficulty, perhaps trying /2+.0002 as a starting point and tune as you see fit. If you're payout variance is too high, get rid of the /2. If you're statistics reporting to p2pool is too much (crashing the server) then increase +.0002 to something higher (or lower if you want more accurate statistics)

Thank you, I'm starting to get a better understanding of how P2pool LTC difficulty works now.  There isn't much info regarding the subject (at least that I've found).  It's still a bit murky on how p2pool assigns shares & what the requirement/per difficulty setting.  Ideally, it would be nice to know that I'm only submitting hashes that meet the difficulty requirement to receive a share from p2pool.  For example:

usrnme/2 = would require 200k share (cgminer etc..) to receive a share from P2pool.
usrnme/3 = would require 300k share
usrnme/4 = etc....

I've used the usrnme/2 difficulty and submitted (accepted) a 250K share from cgminer, and looked in P2pool to see if it counted, and it was not counted.  Me being a pain in the arse, I like to understand how things work, so that I can apply them to my needs as best I can. 

I really appreciate you sharing your knowledge, it's been a wealth of information!
newbie
Activity: 37
Merit: 0
A mean average of 0.1 went up to a mean average of over 0.3.....

My 11.3 was at ~0.2-0.3 and my 11.4 seems to be at ~0.5-0.6.

I'm out of town this week, but will downgrade to see if it helps when I get back.
legendary
Activity: 2576
Merit: 2267
1RichyTrEwPYjZSeAYxeiFBNnKC9UjC5k
I'm not sure there'd be much benefit to C compared to some other languages which are more easily cross-platform such as Perl or Java. More power to the elbow of anyone who wants to though.
There is a benefit in using a compiled language that does not have garbage collection and gives you access to low-level OS primitives - you can write really tight high performance code. I think that C++ would be a better choice over C, though.

Don't get me wrong, I'm a long time C/C++ advocate and user. But is something like this really in that much need of that level tight, high-performance code? (Maybe it is, I just don't know that much about it).

donator
Activity: 1617
Merit: 1012
I'm not sure there'd be much benefit to C compared to some other languages which are more easily cross-platform such as Perl or Java. More power to the elbow of anyone who wants to though.
There is a benefit in using a compiled language that does not have garbage collection and gives you access to low-level OS primitives - you can write really tight high performance code. I think that C++ would be a better choice over C, though.
legendary
Activity: 2576
Merit: 2267
1RichyTrEwPYjZSeAYxeiFBNnKC9UjC5k
I'm not sure there'd be much benefit to C compared to some other languages which are more easily cross-platform such as Perl or Java. More power to the elbow of anyone who wants to though.
hero member
Activity: 924
Merit: 1000
Watch out for the "Neg-Rep-Dogie-Police".....
A mean average of 0.1 went up to a mean average of over 0.3.....
sr. member
Activity: 263
Merit: 250
Also, memory usage graph still does not work on windoze.
Memory graph only works if there is the WMI thing installed:
http://timgolden.me.uk/python/wmi/index.html
Then it looks like this:
https://dl.dropboxusercontent.com/u/34410241/p2pool-leak.jpg

Yeah, saw that.

I prefer having normal latency times than being able to see how much memory I'm wasting....... Cheesy Cheesy

What measure of latency do you claim becomes worse in 11.4?  It seems fine here.
hero member
Activity: 924
Merit: 1000
Watch out for the "Neg-Rep-Dogie-Police".....
Sooooo, anyone fancy re-writing p2pool in C yet? Please??

I'm contemplating trying it.

M

Do it my man. It looks like it's the only way anything will be fixed - I'd do it myself if I had any notion how to...... Wink
full member
Activity: 172
Merit: 100
Upgraded to p2pool 11.4 and my latency trebled! Also, memory usage graph still does not work on windoze.

Downgraded back to 11.3 and all is good again. Suggest everyone running 11.4 check their latency values......

Sooooo, anyone fancy re-writing p2pool in C yet? Please??

Thanks for the heads up! Note to self...stay on 11.3!
legendary
Activity: 1540
Merit: 1001
Sooooo, anyone fancy re-writing p2pool in C yet? Please??

I'm contemplating trying it.

M
hero member
Activity: 924
Merit: 1000
Watch out for the "Neg-Rep-Dogie-Police".....
Also, memory usage graph still does not work on windoze.
Memory graph only works if there is the WMI thing installed:
http://timgolden.me.uk/python/wmi/index.html
Then it looks like this:
https://dl.dropboxusercontent.com/u/34410241/p2pool-leak.jpg

Yeah, saw that.

I prefer having normal latency times than being able to see how much memory I'm wasting....... Cheesy Cheesy
hero member
Activity: 675
Merit: 514
Also, memory usage graph still does not work on windoze.
Memory graph only works if there is the WMI thing installed:
http://timgolden.me.uk/python/wmi/index.html
Then it looks like this:
https://dl.dropboxusercontent.com/u/34410241/p2pool-leak.jpg
Jump to: