[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool

[zvs]

is it normal to always only have 6 out connections? i mean, every once in a while i will see 7 or 8, but the majority of the time its only 6.

Im trying to find ways to really improve performance. I had an older machine running my node with just ubuntu server, but my efficiency was terrible, less than 80% after 48 hours. So i switched up to my windows laptop and currently running over 100% at just under 48 hours. my latency has also decreased. but i have only produced 1 share with 13gh/s. When i was running on my slower server i had produced nearly 5 shares 1 orphan in 48 hours.

I just think something is wrong here but i cant pinpoint it.

71.91.202.165:9332 is my node running on a 30Mbps cable connection.

yeah, 6 is the default.  restart p2pool and use  --p2pool-node 5.9.24.81 and --p2pool-node 198.12.127.2   ... that'll give you 8 outgoing connections, with guaranteed connection to two good nodes

the shares sound perfectly normal for 13gh/s... just a little bit unlucky

(oh, and you got lucky before that.. isn't it like 12hrs per share on avg?)

[xyzzy099]

I'm about to install the bitcoin-qt client on the computer i'm running my miner through, I guess it takes a few days now... will it impair the efficiency of my asic machine?
I want to join p2pool but I have one computer I work on, and one that does the mining (which I had to replace recently - thus - no client  ).  

If your machine is reasonably fast, it can run P2Pool, your miners, the lot.  Anything with an i3 or better is fine.

The main restriction with running your own node is bandwidth.  The average DSL connection can't provide what p2pool and bitcoin need to keep your stale rate low.

M

Is bandwidth the isssue, or is it really latency?  I'm not using p2pool atm, but I don't recall it using much BW when I did...

[shogunmark]

yeah, 6 is the default.  restart p2pool and use  --p2pool-node 5.9.24.81 and --p2pool-node 198.12.127.2   ... that'll give you 8 outgoing connections, with guaranteed connection to two good nodes

the shares sound perfectly normal for 13gh/s... just a little bit unlucky

(oh, and you got lucky before that.. isn't it like 12hrs per share on avg?)

There is no doubt i had great luck over the weekend.. i had racked up 10 shares with 2 orphans in 3 days. I was using p2pool a few weeks ago and had similar luck, but bitcoind became corrupted and i had to start all over.

Yes, with 13Gh/s, its showing 13hrs per share on avg

[gyverlb]

I'm about to install the bitcoin-qt client on the computer i'm running my miner through, I guess it takes a few days now... will it impair the efficiency of my asic machine?
I want to join p2pool but I have one computer I work on, and one that does the mining (which I had to replace recently - thus - no client   ).  

If your machine is reasonably fast, it can run P2Pool, your miners, the lot.  Anything with an i3 or better is fine.

The main restriction with running your own node is bandwidth.  The average DSL connection can't provide what p2pool and bitcoin need to keep your stale rate low.

M

You should define what the average DSL connection is (this changes wildly between countries). I have what I consider one (in my country >10/1 Mbps ADSL connections are common) and I don't have any problem with my stale rate (obviously I changed the default settings of both bitcoind and p2pool).

or so you say...  You should define what you consider a problematic stale rate.

Obviously it's anything that brings your average efficiency lower than what you can expect to earn on a competing pool.

If you are above 100% you know your stale rate isn't a problem. The grey area depends on what your other options are.

In my case I wouldn't switch to another pool unless my average efficiency were below 97%. So for me a problematic stale rate is one that brings my efficiency below 97%.

[HellDiverUK]

I'm about to install the bitcoin-qt client on the computer i'm running my miner through, I guess it takes a few days now... will it impair the efficiency of my asic machine?
I want to join p2pool but I have one computer I work on, and one that does the mining (which I had to replace recently - thus - no client  ).  

If your machine is reasonably fast, it can run P2Pool, your miners, the lot.  Anything with an i3 or better is fine.

The main restriction with running your own node is bandwidth.  The average DSL connection can't provide what p2pool and bitcoin need to keep your stale rate low.

M

No, it isn't.  p2pool only needs a few tens of kilobytes per second.  You could do that on 512k DSL no problems.  It's getblocktemplate latency that's the issue, which is purely bitcoind hogging CPU. 

But, hey, don't let facts get in the way of a good argument.

[matthewh3]

is it normal to always only have 6 out connections? i mean, every once in a while i will see 7 or 8, but the majority of the time its only 6.

Im trying to find ways to really improve performance. I had an older machine running my node with just ubuntu server, but my efficiency was terrible, less than 80% after 48 hours. So i switched up to my windows laptop and currently running over 100% at just under 48 hours. my latency has also decreased. but i have only produced 1 share with 13gh/s. When i was running on my slower server i had produced nearly 5 shares 1 orphan in 48 hours.

I just think something is wrong here but i cant pinpoint it.

71.91.202.165:9332 is my node running on a 30Mbps cable connection.

yeah, 6 is the default.  restart p2pool and use  --p2pool-node 5.9.24.81 and --p2pool-node 198.12.127.2   ... that'll give you 8 outgoing connections, with guaranteed connection to two good nodes

the shares sound perfectly normal for 13gh/s... just a little bit unlucky

(oh, and you got lucky before that.. isn't it like 12hrs per share on avg?)

Won't forwarding the ports for bitcoin-qt give you a lot more nodes?

[shogunmark]

Won't forwarding the ports for bitcoin-qt give you a lot more nodes?

I have 8333 forwarded

[gyverlb]

I'm about to install the bitcoin-qt client on the computer i'm running my miner through, I guess it takes a few days now... will it impair the efficiency of my asic machine?
I want to join p2pool but I have one computer I work on, and one that does the mining (which I had to replace recently - thus - no client  ).  

If your machine is reasonably fast, it can run P2Pool, your miners, the lot.  Anything with an i3 or better is fine.

The main restriction with running your own node is bandwidth.  The average DSL connection can't provide what p2pool and bitcoin need to keep your stale rate low.

M

No, it isn't.  p2pool only needs a few tens of kilobytes per second.  You could do that on 512k DSL no problems.  It's getblocktemplate latency that's the issue, which is purely bitcoind hogging CPU. 

But, hey, don't let facts get in the way of a good argument.

The situation is usually more complicated (which is why people often don't agree here on what is enough/good for p2pool: they have widely different situations to compare).

getblocktemplate latency isn't much of an issue if you have a multi-core CPU and your system workload leaves at least one core free for p2pool even when bitcoind is hogging one core.
It's clearly a problem when you don't have this free core (p2pool slows down and your efficiency dives).

A 512k DSL is clearly enough, but this assumes you don't use it for anything else that makes your link latency rise.
If you can setup QoS on your DSL and eliminate the influence of other traffic then clearly you can use p2pool on a DSL link. Unfortunately it's often not simple to do: some routers actively shape the traffic the way they want without letting users have much to tune or making the rest of the traffic too unreliable.
Even with a decent knowledge and control of the QoS settings you can have problems. For example I shot myself in the foot once, my QoS settings were perfect for bitcoind and p2pool but they caused timeouts repeatedly in my irc client.

[mdude77]

The situation is usually more complicated (which is why people often don't agree here on what is enough/good for p2pool: they have widely different situations to compare).

getblocktemplate latency isn't much of an issue if you have a multi-core CPU and your system workload leaves at least one core free for p2pool even when bitcoind is hogging one core.
It's clearly a problem when you don't have this free core (p2pool slows down and your efficiency dives).

A 512k DSL is clearly enough, but this assumes you don't use it for anything else that makes your link latency rise.
If you can setup QoS on your DSL and eliminate the influence of other traffic then clearly you can use p2pool on a DSL link. Unfortunately it's often not simple to do: some routers actively shape the traffic the way they want without letting users have much to tune or making the rest of the traffic too unreliable.
Even with a decent knowledge and control of the QoS settings you can have problems. For example I shot myself in the foot once, my QoS settings were perfect for bitcoind and p2pool but they caused timeouts repeatedly in my irc client.

I made my statement based on my observations.  I have roughly 3mb down/768kb up connection.  I found from observation of using my own p2pool node over a period of a few weeks that my lowest stales occurred when I had no port forwarding turned on for bitcoin or for p2pool.   Even then it wasn't acceptable to me, as this isn't a dedicated link.  Download something and watch your stales/dead rate go through the roof.  I tried turning on qos, it made everything else unacceptable.

Now with the newer version of p2pool and ridiculous difficulty, for someone like me with ~12gh/s hash power, having that one share I get in an 18 hour period go stale is awful.  Probably wouldn't be so bad if you had 10x my hashrate, so I probably should have clarified my statement.  I thought I was responding to someone using about the same hashpower as me.

M

[zvs]

The situation is usually more complicated (which is why people often don't agree here on what is enough/good for p2pool: they have widely different situations to compare).

getblocktemplate latency isn't much of an issue if you have a multi-core CPU and your system workload leaves at least one core free for p2pool even when bitcoind is hogging one core.
It's clearly a problem when you don't have this free core (p2pool slows down and your efficiency dives).

A 512k DSL is clearly enough, but this assumes you don't use it for anything else that makes your link latency rise.
If you can setup QoS on your DSL and eliminate the influence of other traffic then clearly you can use p2pool on a DSL link. Unfortunately it's often not simple to do: some routers actively shape the traffic the way they want without letting users have much to tune or making the rest of the traffic too unreliable.
Even with a decent knowledge and control of the QoS settings you can have problems. For example I shot myself in the foot once, my QoS settings were perfect for bitcoind and p2pool but they caused timeouts repeatedly in my irc client.

I made my statement based on my observations.  I have roughly 3mb down/768kb up connection.  I found from observation of using my own p2pool node over a period of a few weeks that my lowest stales occurred when I had no port forwarding turned on for bitcoin or for p2pool.   Even then it wasn't acceptable to me, as this isn't a dedicated link.  Download something and watch your stales/dead rate go through the roof.  I tried turning on qos, it made everything else unacceptable.

Now with the newer version of p2pool and ridiculous difficulty, for someone like me with ~12gh/s hash power, having that one share I get in an 18 hour period go stale is awful.  Probably wouldn't be so bad if you had 10x my hashrate, so I probably should have clarified my statement.  I thought I was responding to someone using about the same hashpower as me.

M

I have 25/2 bonded DSL and even uploading at 50KB/s on this piece of shit slows my connection down.  I have a server up right now that I'm not mining on, so I don't really care about it (and it gets shut off when I'm doing anything important).

QoS is PoS

[IYFTech]

So nice it has been implemented already. But I will agree with comments made earlier on, it should be applied to miners with 500 GH/s (1.7% of total p2pool hashrate). Please reduce it to at least 2% from 5%. It's a simple fix, and it will help a lot to small miners.

From experience, having "only" 2% of the p2pool hashrate doesn't mean there's too much variance (the rewards oscillate +/-~10% around).

With 5%, ~20 large miners can take most of the pie, leaving crumbs to others. With 2%, this is raised to 50 large miners.

I'd be OK with a 1% limit, but 2% should limit complains of high variance from large miners (the reward may even move as much from the hashrate fluctuations of the whole pool as from the share frequency variance, people with 5% or more may be able to confirm this if they have variance around 10% in their rewards too).

BTW: I've not looked into the code, what would happen if there were only 10 miners on p2pool? Is the current algorithm able to converge on sane values or would the current 5% target raise the individual share difficulty without bonds.
From the one-liner extract above it seems there's a missing parameter to achieve this protection (the number of distinct addresses with valid shares in the recent sharechain).

I will decrease the percentage then. Maybe 1.67%? That's a share every half hour.

If there are very few miners, nothing insane happens. Each miner's share difficulty multiplier becomes the maximum, 30, and then the 30-second share period target decreases the minimum difficulty until there's a share every 30 seconds again.

Has this been implemented yet? I'm considering turning my node back on to see if it improves anything.....

Any news yet?

[zvs]

So nice it has been implemented already. But I will agree with comments made earlier on, it should be applied to miners with 500 GH/s (1.7% of total p2pool hashrate). Please reduce it to at least 2% from 5%. It's a simple fix, and it will help a lot to small miners.

From experience, having "only" 2% of the p2pool hashrate doesn't mean there's too much variance (the rewards oscillate +/-~10% around).

With 5%, ~20 large miners can take most of the pie, leaving crumbs to others. With 2%, this is raised to 50 large miners.

I'd be OK with a 1% limit, but 2% should limit complains of high variance from large miners (the reward may even move as much from the hashrate fluctuations of the whole pool as from the share frequency variance, people with 5% or more may be able to confirm this if they have variance around 10% in their rewards too).

BTW: I've not looked into the code, what would happen if there were only 10 miners on p2pool? Is the current algorithm able to converge on sane values or would the current 5% target raise the individual share difficulty without bonds.
From the one-liner extract above it seems there's a missing parameter to achieve this protection (the number of distinct addresses with valid shares in the recent sharechain).

I will decrease the percentage then. Maybe 1.67%? That's a share every half hour.

If there are very few miners, nothing insane happens. Each miner's share difficulty multiplier becomes the maximum, 30, and then the 30-second share period target decreases the minimum difficulty until there's a share every 30 seconds again.

Has this been implemented yet? I'm considering turning my node back on to see if it improves anything.....

Any news yet?

yeah, it's in a git commit from 3 or 4 days ago

[lenny_]

Indeed situation has been improved since commit
At the day of patch release, p2pool was on 36 TH/s with 161k share diff. Right now it's on 50.6 TH/s with 171k share diff. That's 40% hashrate increase with only 6% sharediff increase. Very nice surprise for small miners But we need more people to upgrade, so variance for small miners will go down even further. Please upgrade!

[Sanweb40]

Hello,

can everyone tell me, how i can find the startup commands to start p2pool with PPLNS?

Thank you very much for any suggestions.

King regards

[IYFTech]

Indeed situation has been improved since commit
At the day of patch release, p2pool was on 36 TH/s with 161k share diff. Right now it's on 50.6 TH/s with 171k share diff. That's 40% hashrate increase with only 6% sharediff increase. Very nice surprise for small miners But we need more people to upgrade, so variance for small miners will go down even further. Please upgrade!

Is the hash rate increase due to multipool.us hooking up to p2pool? Or does it not show on the p2pool graph?

[IYFTech]

Hello,

can everyone tell me, how i can find the startup commands to start p2pool with PPLNS?

Thank you very much for any suggestions.

King regards

https://en.bitcoin.it/wiki/P2Pool

[lenny_]

Indeed situation has been improved since commit
At the day of patch release, p2pool was on 36 TH/s with 161k share diff. Right now it's on 50.6 TH/s with 171k share diff. That's 40% hashrate increase with only 6% sharediff increase. Very nice surprise for small miners But we need more people to upgrade, so variance for small miners will go down even further. Please upgrade!

Is the hash rate increase due to multipool.us hooking up to p2pool? Or does it not show on the p2pool graph?

Spikes up to 60TH can be easily seen on the graph, just have a look.

[RoadTrain]

Indeed situation has been improved since commit
At the day of patch release, p2pool was on 36 TH/s with 161k share diff. Right now it's on 50.6 TH/s with 171k share diff. That's 40% hashrate increase with only 6% sharediff increase. Very nice surprise for small miners But we need more people to upgrade, so variance for small miners will go down even further. Please upgrade!

Is the hash rate increase due to multipool.us hooking up to p2pool? Or does it not show on the p2pool graph?

It likely is.
Their address is 146197ntzrBT41DDQRrB2STBb19dw7ct2F, the highest hashrate of all miners.

[Sanweb40]

https://en.bitcoin.it/wiki/P2Pool

Thank you, but on this side was no information about the needed parameters, that p2pool runs with PPLNS ...

Any Ideas, where i can found this information?

[forrestv]

Thank you, but on this side was no information about the needed parameters, that p2pool runs with PPLNS ...

Any Ideas, where i can found this information?

P2Pool always uses PPLNS. There is no way to disable it.