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

[DeathAndTaxes]

My GPU makes 300MHash/s in solo (2 threads) and 285MHash/s in p2pool (1 thread). Thats 5% slower. Does that mean I lose 5% revenue compared to solo mining?

Yes but I don't show a lower hashrate (well maybe <2% lower) but that might just be noise.  I consider the 1%-2% worth it for the reduced variance (and the support it provides other miners who wouldn't choose solo over conventional pool).

[DiabloD3]

My GPU makes 300MHash/s in solo (2 threads) and 285MHash/s in p2pool (1 thread). Thats 5% slower. Does that mean I lose 5% revenue compared to solo mining?

This implies you're using the pool to measure mhash: DO NOT DO THIS. This is measured by shares submitted, which is a random event. The ONLY true measurement of mhash is the miner itself.

[tynt]

Look what I mean. Using the latest cgminer:
--intensity 6 --gpu-threads 1


--intensity 6 --gpu-threads 2


Thats like 10% slower.

[DeathAndTaxes]

Why is your intensity a 6?  Try an 8.  multiple threads hides the latency in such a tiny batch size  (intensity sets how many nonces are check in one "batch").

Also 90C?  Wow.  Not good for GPU longevity.

[forrestv]

Using "--gpu-threads 1" isn't at all essential for P2Pool. I get fine reject rates without it.

There's an optimum point that balances between a lower rejection rate and a lower hash rate, you just have to find it. However, you might be below that point.. maybe try dropping --gpu-threads 1 or a slightly higher intensity. (Bitcoin has a balance in the same manner, it's just at a very different point since work needs to be updated every 10 minutes instead of 10 seconds.)

[tynt]

You are right. 2 threads is better. I see at least 5% improvement. Raising intensity didn't work for me, whole thing started to throttle down.

1 thread
Total rate: 3.22G, stale rate: 35M

2 threads
Total rate: 3.42G, stale rate: 55M

[DiabloD3]

You are right. 2 threads is better. I see at least 5% improvement. Raising intensity didn't work for me, whole thing started to throttle down.

1 thread
Total rate: 3.22G, stale rate: 35M

2 threads
Total rate: 3.42G, stale rate: 55M

DiabloMiner uses 3, it isn't a settable option. This is for a reason.

[kano]

You are right. 2 threads is better. I see at least 5% improvement. Raising intensity didn't work for me, whole thing started to throttle down.

1 thread
Total rate: 3.22G, stale rate: 35M

2 threads
Total rate: 3.42G, stale rate: 55M

DiabloMiner uses 3, it isn't a settable option. This is for a reason.

What is that reason?

[DiabloD3]

You are right. 2 threads is better. I see at least 5% improvement. Raising intensity didn't work for me, whole thing started to throttle down.

1 thread
Total rate: 3.22G, stale rate: 35M

2 threads
Total rate: 3.42G, stale rate: 55M

DiabloMiner uses 3, it isn't a settable option. This is for a reason.

What is that reason?

Because drivers suck dick. I thought it was self evident.

[broken]

Look what I mean. Using the latest cgminer:
--intensity 6 --gpu-threads 1
http://img696.imageshack.us/img696/5831/1thh.png

--intensity 6 --gpu-threads 2
http://img526.imageshack.us/img526/3117/2th.png

Thats like 10% slower.

you should speed up your fans to above 4000RPM so your cards cool of to <80°C. Your cards will live a lot longer with lower temperatures.
try:

Code:

--auto-gpu --auto-fan

note that when you compare hashrates, with "--auto-gpu" cgminer takes a few minutes after restart untill gpus run at full load. It slowly increases clocks in the beginning.

[jme621]

so i followed the guide for windows and got my local machine to mine and run  p2pool, but how would i point my other machine running bamt to it? i tried using the username:password i setup in the rpc conf and pointed it to 127.0.0.1:9332, yes the ports open, but it did nothing. thanks in advance

[twmz]

so i followed the guide for windows and got my local machine to mine and run  p2pool, but how would i point my other machine running bamt to it? i tried using the username:password i setup in the rpc conf and pointed it to 127.0.0.1:9332, yes the ports open, but it did nothing. thanks in advance

127.0.0.1 is the wrong IP address if you are trying to point a miner on a different machine to your windows p2pool node.  127.0.0.1 is "localhost" and always refers to the local computer you use it on.  You need to use the IP address or hostname of your windows machine:  mywindowsmachine:9332  or x.x.x.x:9332

[DeathAndTaxes]

so i followed the guide for windows and got my local machine to mine and run  p2pool, but how would i point my other machine running bamt to it? i tried using the username:password i setup in the rpc conf and pointed it to 127.0.0.1:9332, yes the ports open, but it did nothing. thanks in advance

127.0.0.1 is a special address for localhost

127.0.0.1 on your other machine is look back at itself not looking at the machine running p2pool.

you just need to replace 127.0.0.1 with the local IP address of the machine running p2pool.  likely it is 192.168.x.x

[Krak]

so i followed the guide for windows and got my local machine to mine and run  p2pool, but how would i point my other machine running bamt to it? i tried using the username:password i setup in the rpc conf and pointed it to 127.0.0.1:9332, yes the ports open, but it did nothing. thanks in advance

127.0.0.1 is a special address for localhost

127.0.0.1 on your other machine is look back at itself not looking at the machine running p2pool.

you just need to replace 127.0.0.1 with the local IP address of the machine running p2pool.  likely it is 192.168.x.x

Also make sure you set a static IP for the PC running p2pool.

[spiccioli]

My first block ever!

Code:

2012-03-30 08:31:56.657079 GOT BLOCK FROM MINER! Passing to bitcoind! http://blockexplorer.com/block/0000000000000613569f5ce021350e9fc584f4c148d74f5209077824fa3a1626

go p2pool, go!



spiccioli

[Gabi]

P2pool is at 390 GH/s  now, nice job!

[Ente]

P2pool is at 390 GH/s  now, nice job!

Nice! So its finally speeding up again!

OT:



Anyone likes to explain that "alternate universe" graph?

Also, the green "found blocks" and red "hashes hashed" look similar, but are spreading. Is that from the ~10% stale/orphans etc the whole p2pool has in average? Then it will continue just like this, and wont converge again. Its hard to believe that the diverging is just the variance?

Ente

[forrestv]

The "Cumulative hashes reflected in blocks in an alternate universe" line was the result of a simulation done using P2Pool's hash rate data that assumes that P2Pool is working perfectly. It was useful to get an idea of the variance to be expected in the green line. It changed every hour.

I just changed it to display 30 of the same thing, so now it gives a better idea of the expected variance in the blue line.

EDIT: I wouldn't say that they're definitely spreading.. and no, assuming nothing is wrong (no bugs or people maliciously dropping block solutions), it's due to bad luck. They should eventually converge.

[forrestv]

URGENT: According to gathered statistics, 40% of P2Pool's hashing power still has not upgraded to P2Pool 0.10.3, which is required by April 1st UTC (28 hours remain!) to ensure that everyone is using a BIP16-compatible Bitcoin client.



Bitcoin 0.6.0 was released today, so there is no longer any reason to wait to upgrade. Please help by telling anyone that you know that hasn't upgraded to do so. Thanks!

Bitcoin 0.6.0: http://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.6.0/
P2Pool 0.10.3 for Windows: http://u.forre.st/u/jtjrvlxm/p2pool_win32_0.10.3.zip

[kano]

...
Bitcoin 0.6.0 was released today, so there is no longer any reason to wait to upgrade. Please help by telling anyone that you know that hasn't upgraded to do so. Thanks!
...

Is that a Bitcoin requirement or a P2Pool requirement?

(and are there any issues with following your recommendation to upgrade to 0.6.0?)