[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 664.

Smoovious

hero member

Activity: 504

Merit: 500

Scattering my bits around the net since 1980

Quote from: kwukduck on July 24, 2012, 08:09:42 PM

Yes, i'm running 24/7.

p2pool tells me my hash rate is much lower then my actual mining speed (about 60%) and whenever my miner finds a new share it's not added to the p2pool count.
I recently started it and my miner tells me i found 50 shares (all accepted), p2pool says 15?
very confused...

What miner is being used... got failovers set up?

-- Smoov

kwukduck

legendary

Activity: 1937

Merit: 1001

Quote from: mdude77 on July 24, 2012, 08:23:25 PM

Quote from: kwukduck on July 24, 2012, 08:09:42 PM

Yes, i'm running 24/7.

p2pool tells me my hash rate is much lower then my actual mining speed (about 60%) and whenever my miner finds a new share it's not added to the p2pool count.
I recently started it and my miner tells me i found 50 shares (all accepted), p2pool says 15?
very confused...

How much did you get from the last block?

I know my hash rate swings all over the place, between 4500m/h and 5100m/h, depending on when I look at it. Are you checking the local graphs to give an average, or just at the console?

I've read a few times that there are actually two block chains, one for p2pool, and one for bitcoin, or something like that. I don't follow the logic at all, but what you are describing makes sense with what others have said. Maybe someone will chime in who can explain this better.

M

Last block was 0.068, i've seen worse

mdude77

legendary

Activity: 1540

Merit: 1001

Quote from: kwukduck on July 24, 2012, 08:09:42 PM

Yes, i'm running 24/7.

p2pool tells me my hash rate is much lower then my actual mining speed (about 60%) and whenever my miner finds a new share it's not added to the p2pool count.
I recently started it and my miner tells me i found 50 shares (all accepted), p2pool says 15?
very confused...

How much did you get from the last block?

I know my hash rate swings all over the place, between 4500m/h and 5100m/h, depending on when I look at it. Are you checking the local graphs to give an average, or just at the console?

I've read a few times that there are actually two block chains, one for p2pool, and one for bitcoin, or something like that. I don't follow the logic at all, but what you are describing makes sense with what others have said. Maybe someone will chime in who can explain this better.

M

kwukduck

legendary

Activity: 1937

Merit: 1001

Yes, i'm running 24/7.

p2pool tells me my hash rate is much lower then my actual mining speed (about 60%) and whenever my miner finds a new share it's not added to the p2pool count.
I recently started it and my miner tells me i found 50 shares (all accepted), p2pool says 15?
very confused...

mdude77

legendary

Activity: 1540

Merit: 1001

Quote from: kwukduck on July 24, 2012, 07:43:54 PM

Quote from: mdude77 on July 24, 2012, 07:26:15 PM

Quote from: kwukduck on July 24, 2012, 07:16:10 PM

i've been mining with p2pool for a while now and today i thought of counting what i had mined, the average of the last month is about 0.17 btc/day, i compared this to what i should get at my speed (625MHash) which is about 0.34 btc/day. I haven't used a regular pool for a while now so i can't really say if that would be any better.
I knew p2pool had slightly lower rewards then regular pools but i didn't expect to lose 50%...
Is there something wrong with my config or is this normal?

This is not normal. What's your stale rate?

M

p2pool window says 9%
I also noticed a big variance in payout 0.02~0.09 btc per payout.

I'm far from an expert in p2pool, so I'll throw some simple math out here that I think is right.

Assuming a perfect world, at the current pool hash rate and difficulty, p2pool gets a block ~10 hours. That's 2.4 blocks a day.

Current pool hash rate is 235g/h, or 235,000 m/h. Your piece of that hash rate is 625/235000. Multiply that times block payout, say 50btc even, and you get ~0.133 btc/block. At 2.4 blocks a day, that 0.32 btc/day. So I agree with your calculation.

So, we just got a block. What was your payout? I'm currently at 4870m/h, pool is as 243,000m/h. Using the above math, my piece of the pie should be around 1btc, I got 0.89. I think my shares are low recently - 24 hours ago my payouts were real close to 1btc.

Are you running 24/7? I think the payout is based on your share of the shares over the last 24 hours. If you aren't running 24/7, you are cutting yourself short.

Come to think of it, since the pool seems to fluctuate between 200g/h and 300g/h, there's either something funny going on or a lot of people are shorting themselves.

M

kwukduck

legendary

Activity: 1937

Merit: 1001

Quote from: mdude77 on July 24, 2012, 07:26:15 PM

Quote from: kwukduck on July 24, 2012, 07:16:10 PM

i've been mining with p2pool for a while now and today i thought of counting what i had mined, the average of the last month is about 0.17 btc/day, i compared this to what i should get at my speed (625MHash) which is about 0.34 btc/day. I haven't used a regular pool for a while now so i can't really say if that would be any better.
I knew p2pool had slightly lower rewards then regular pools but i didn't expect to lose 50%...
Is there something wrong with my config or is this normal?

This is not normal. What's your stale rate?

M

p2pool window says 9%
I also noticed a big variance in payout 0.02~0.09 btc per payout.

mdude77

legendary

Activity: 1540

Merit: 1001

Quote from: kwukduck on July 24, 2012, 07:16:10 PM

i've been mining with p2pool for a while now and today i thought of counting what i had mined, the average of the last month is about 0.17 btc/day, i compared this to what i should get at my speed (625MHash) which is about 0.34 btc/day. I haven't used a regular pool for a while now so i can't really say if that would be any better.
I knew p2pool had slightly lower rewards then regular pools but i didn't expect to lose 50%...
Is there something wrong with my config or is this normal?

This is not normal. What's your stale rate?

M

kwukduck

legendary

Activity: 1937

Merit: 1001

i've been mining with p2pool for a while now and today i thought of counting what i had mined, the average of the last month is about 0.17 btc/day, i compared this to what i should get at my speed (625MHash) which is about 0.34 btc/day. I haven't used a regular pool for a while now so i can't really say if that would be any better.
I knew p2pool had slightly lower rewards then regular pools but i didn't expect to lose 50%...
Is there something wrong with my config or is this normal?

mdude77

legendary

Activity: 1540

Merit: 1001

Quote from: nibor on July 24, 2012, 02:04:46 PM

Quote from: Smoovious on July 24, 2012, 10:31:03 AM

Quote from: nibor on July 24, 2012, 02:16:15 AM

Shame I can't use my BFL single (or has anyone got a solution for the 5sec issue?)!

5sec issue?

-- Smoov

BFL FPGA singles search all 4b nounces before returning control to the miner with a list of 1 difficulty shares it has found. This takes about 5 seconds. In this 5 seconds when mining on a "normal" pool the chance of a block being created is very small so you get very few stales.

On p2pool there is a new share every 10 seconds. So with BFL FPGA single you get about 50% stales. BFL need to fix this issue.

Increasing difficulty doesn't fix it?

M

racerguy

sr. member

Activity: 270

Merit: 250

From what I can gather the bfl's can only be given new work every 10 secs which for p2pool means they produce 50% stales.

nibor

sr. member

Activity: 438

Merit: 291

Quote from: Smoovious on July 24, 2012, 10:31:03 AM

Quote from: nibor on July 24, 2012, 02:16:15 AM

Shame I can't use my BFL single (or has anyone got a solution for the 5sec issue?)!

5sec issue?

-- Smoov

BFL FPGA singles search all 4b nounces before returning control to the miner with a list of 1 difficulty shares it has found. This takes about 5 seconds. In this 5 seconds when mining on a "normal" pool the chance of a block being created is very small so you get very few stales.

On p2pool there is a new share every 10 seconds. So with BFL FPGA single you get about 50% stales. BFL need to fix this issue.

Smoovious

hero member

Activity: 504

Merit: 500

Scattering my bits around the net since 1980

Quote from: nibor on July 24, 2012, 02:16:15 AM

Shame I can't use my BFL single (or has anyone got a solution for the 5sec issue?)!

5sec issue?

-- Smoov

nibor

sr. member

Activity: 438

Merit: 291

p2pool is better than Deepbit over last 90 days (assuming deepbit had 100% luck) according to p2pool.info!

90 Luck has climbed above above 97% so if you add in the 0.2% of transaction fees you get from p2pool and subtract the 3% fee deepbit would have charged you, you would have been better off on p2pool over the last 3 months.

Shame I can't use my BFL single (or has anyone got a solution for the 5sec issue?)!

twmz

hero member

Activity: 737

Merit: 500

Quote from: kano on July 23, 2012, 06:01:14 PM

Quote from: twmz on July 22, 2012, 05:04:47 PM

Quote from: mdude77 on July 22, 2012, 02:08:18 PM

I was wondering how this page calculates the performance of finding blocks (7 days, 30 days, 90 days). Is it based upon when the block is found, or the average for the block?

Pool hashrate is sampled every 5 minutes, and the luck estimate/calculation uses the average of those samples over the duration of the round.

The fact that pool hashrate is sampled every 5 minutes is actually the reason the page loads so slowly (there are a zillion 5 minute data points between now and last february when I started collecting 5 minute samples). The 5 minute sample are useful for the luck calculation, but the graph doesn't really need to be that precise so I will be changing JSON API to return fewer data points to the graph which will cut down the time for the page to load (block luck calculations happen on the server and will still use the 5 minute samples).

Yes but hash rate is an estimate ...
Really it should keep the history of shares and use that - which is accurate.
Shares is the only accurate measurement of a pool.

I agree completely that in a perfect world that would be better. I just don't care about the difference enough to invest my free time to make the substantial changes required to make that work. If someone else wants to do it and provide me a JSON feed of p2pool blocks along with the actual number of diff 1 shares that went into that block, I'm happy to replace the current hashrate based estimates.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Quote from: twmz on July 22, 2012, 05:04:47 PM

Quote from: mdude77 on July 22, 2012, 02:08:18 PM

I was wondering how this page calculates the performance of finding blocks (7 days, 30 days, 90 days). Is it based upon when the block is found, or the average for the block?

Pool hashrate is sampled every 5 minutes, and the luck estimate/calculation uses the average of those samples over the duration of the round.

The fact that pool hashrate is sampled every 5 minutes is actually the reason the page loads so slowly (there are a zillion 5 minute data points between now and last february when I started collecting 5 minute samples). The 5 minute sample are useful for the luck calculation, but the graph doesn't really need to be that precise so I will be changing JSON API to return fewer data points to the graph which will cut down the time for the page to load (block luck calculations happen on the server and will still use the 5 minute samples).

Yes but hash rate is an estimate ...
Really it should keep the history of shares and use that - which is accurate.
Shares is the only accurate measurement of a pool.

rav3n_pl

legendary

Activity: 1361

Merit: 1003

Don`t panic! Organize!

Look at week stats. One day is not enough to make any close estimations.
On my crappy 150MH there was days i not get even one share, then 3 in hour. Just luck

zvs

legendary

Activity: 1680

Merit: 1000

https://web.archive.org/web/*/nogleg.com

Quote from: louisBSAS on July 23, 2012, 11:23:38 AM

Great day yesterday. I have been mining for a month now, and P2Pool seems to beat other pools. Cudos!

I have been running just shy of 2Gh for over a week- non-stop. "current payout" listed at around .15at first, then .41 (after 4 days), now .28 to .33.

I have been trying to figure out how the payout amount is calculated and such. All in all, it seems fair according to pool speed.

Question:

Over 8 hours while I slept, I gained 22 shares in 8 hours.. I had 418 when I woke up 5 hours ago - now I have 419 (in 5 hours time).
Orphans and 'Dead' have not changed and 'expected time to share:' continues to be between 18 and 22 minutes.

Any ideas?

--
Running 2 rigs piped into P2Pool/Bitcoin running on a non-mining machine - all Linux
No changes in configuration at all - so, that can't be the cause.

Thanks ahead of time.

Just crap luck, I'd guess. I run between 4-6GH/s on p2pool and there have been times where I've gone over an hr between shares.

louisBSAS

member

Activity: 623

Merit: 11

Proof-of-Stake Blockchain Network

Great day yesterday. I have been mining for a month now, and P2Pool seems to beat other pools. Cudos!

I have been running just shy of 2Gh for over a week- non-stop. "current payout" listed at around .15at first, then .41 (after 4 days), now .28 to .33.

I have been trying to figure out how the payout amount is calculated and such. All in all, it seems fair according to pool speed.

Question:

Over 8 hours while I slept, I gained 22 shares in 8 hours.. I had 418 when I woke up 5 hours ago - now I have 419 (in 5 hours time).
Orphans and 'Dead' have not changed and 'expected time to share:' continues to be between 18 and 22 minutes.

Any ideas?

--
Running 2 rigs piped into P2Pool/Bitcoin running on a non-mining machine - all Linux
No changes in configuration at all - so, that can't be the cause.

Thanks ahead of time.

twmz

hero member

Activity: 737

Merit: 500

Quote from: mdude77 on July 22, 2012, 02:08:18 PM

I was wondering how this page calculates the performance of finding blocks (7 days, 30 days, 90 days). Is it based upon when the block is found, or the average for the block?

Pool hashrate is sampled every 5 minutes, and the luck estimate/calculation uses the average of those samples over the duration of the round.

The fact that pool hashrate is sampled every 5 minutes is actually the reason the page loads so slowly (there are a zillion 5 minute data points between now and last february when I started collecting 5 minute samples). The 5 minute sample are useful for the luck calculation, but the graph doesn't really need to be that precise so I will be changing JSON API to return fewer data points to the graph which will cut down the time for the page to load (block luck calculations happen on the server and will still use the 5 minute samples).

mdude77

legendary

Activity: 1540

Merit: 1001

Quote from: twmz on July 22, 2012, 09:15:04 AM

Quote from: joemde on July 22, 2012, 07:43:42 AM

On a different note, the stats webpage will not load for me today from any comp I have. Does anyone know what's up?

Temporary fix is in place. A better fix in the works that should also improve performance.

I was wondering how this page calculates the performance of finding blocks (7 days, 30 days, 90 days). Is it based upon when the block is found, or the average for the block?

For example, let's say the hash rate for the pool was 300g/h. It was a long block, and a bunch of people gave up before it was found. When it was found, we were at 250g/h. If the performance is based upon when it was found, it would look like we did better than reality, because it would take longer at 250 than it would at 300, when in reality, we were at 300 for most of the time of the block.

Conversely, if were at 250g/h and then suddenly jumped to 300 and found a block, it would look like we did worse than reality, because it would be shorter at 300 than it was at 250.

Since the hash rates seems to swing wildly from 200 to 300, I thought I'd ask. If it's based on the time the block was found, it needs to be taken with a grain of salt.

Right?

M

Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 664. (Read 2591920 times)