I showed no -10 MH/s per card on a farm of 24 5970s. You sure you weren't just seeing variance. Still even if it is -10 MH/s if it saves you 20 watts you likely come out ahead. Using affinity still keep a core burning away at max clock rate and current.
You are aware the DRIVER is what affects the cpu bug. The SDK is what affects OpenCL performance. You can use any driver (including one with no CPU bug) with any SDK runtime (including the one you believe is +10MH/s better).
Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 740. (Read 2591916 times)
March 21, 2012, 09:08:39 AM
March 21, 2012, 08:47:04 AM
Try driver 11.11
They don't have the sdk 2.6 and don't have the 100%cpu bug
They don't have the sdk 2.6 and don't have the 100%cpu bug
And has a nice reduction of 10 Mh/s per card as apposed to 11.9 (Windows 7 X64 6970's and 5870's) Better to use Affinity IMO and not loose the 10 Mh/s per card.
March 21, 2012, 08:24:00 AM
p2pool shows a global hashrate stat shown at "/rate" or on the p2pool command line. Anyone know over what time period that is calculated? Last hour? Last day? Last x shares? I got an idea for a promo / contest involving p2pool.
March 21, 2012, 08:14:00 AM
Try driver 11.11
They don't have the sdk 2.6 and don't have the 100%cpu bug
They don't have the sdk 2.6 and don't have the 100%cpu bug
March 21, 2012, 07:51:23 AM
Does anyone know if there are plans for a p2pool WITHOUT bip16 support come april 1st? Would p2pool be robust enough to work in a mixed client scenario?
It seems to me like there could be quite a few miners who like the idea and lower variance of p2pool but would also like some say in which changes get accepted into the bitcoin protocol.
It seems to me like there could be quite a few miners who like the idea and lower variance of p2pool but would also like some say in which changes get accepted into the bitcoin protocol.
March 21, 2012, 07:20:43 AM
I agree with Sharky. I'll just be manually setting affinity because I'm on 2.5 and 11.6 which for my 5870 and 3x 5830's is what I've found to work best. 
Thanks for the advice, fellas!
Thanks for the advice, fellas! March 21, 2012, 06:16:55 AM
Yeah, looks like the 100% cpu bug
It's not a p2pool bug, it's about driver and using OpenCL (miner is in OpenCL of course) How to fix: update your driver.
It's not a p2pool bug, it's about driver and using OpenCL (miner is in OpenCL of course) How to fix: update your driver.
Updating your AMD-APP Runtime and Driver is not always the answer. I spent hours trying different AMD-APP Runtimes and Drivers only to discover that on the combinations where I was able to get rid of the 100% CPU bug, I lost 10 Mh/s per card. I ended up going back to 11.9 (AMD-APP Runtime 2.5.732.1, Driver 8.892.0.0000) and using Affinity to only use one core.
March 21, 2012, 03:11:33 AM
Small info:
I experimented with merged mining, and did so with GeistGeld. This alt-chain has a blockrate of only a few seconds (15, or 7 if it halved already). It did work, I solved some GeistGeld blocks. The reported hashrate by miners and p2pool was right as well. But after solving 10 shareblocks, I had 5 orphans and one dead. Must be because p2pool creates new work for the miners every few seconds, when GeistGeld reports a new block.
I switched Geistgeld merged mining off again, now I have around 10% orphans.
Of course this is an extrema, since GeistGeld is the fastest chain by far.
Ente
I experimented with merged mining, and did so with GeistGeld. This alt-chain has a blockrate of only a few seconds (15, or 7 if it halved already). It did work, I solved some GeistGeld blocks. The reported hashrate by miners and p2pool was right as well. But after solving 10 shareblocks, I had 5 orphans and one dead. Must be because p2pool creates new work for the miners every few seconds, when GeistGeld reports a new block.
I switched Geistgeld merged mining off again, now I have around 10% orphans.
Of course this is an extrema, since GeistGeld is the fastest chain by far.
Ente
March 21, 2012, 01:40:56 AM
Yeah, looks like the 100% cpu bug
It's not a p2pool bug, it's about driver and using OpenCL (miner is in OpenCL of course) How to fix: update your driver.
It's not a p2pool bug, it's about driver and using OpenCL (miner is in OpenCL of course) How to fix: update your driver.
March 20, 2012, 11:49:03 PM
when I opened taskmanager, I was confused to find cpu usage at 100%... My miner was consuming almost all of the cpu resouces
That would be the "famous" AMD 100% CPU usage bug. If you can, try using the 2.5 SDK instead of 2.6. More info in the cgminer thread (and a whole bunch of other places, I assume).
March 20, 2012, 10:45:31 PM
Just wanted to mention that I looked into the bitcoind responding issue and when I opened taskmanager, I was confused to find cpu usage at 100%... My miner was consuming almost all of the cpu resouces and left little to none for p2pool.exe and bitcoind.exe. I dropped the affinity of cgminer from 2 to 1 core (no penalty on hashing speed as far as I can see) and haven't seen an issue with the local speeds reported by p2pool or the failure to communicate issue. So my advice for anyone seeing bitcoind communication errors or a lower than expected local hash rate reported by p2pool, check your affinities in taskmanager and allocate a few more resources to bitcoind and p2pool.
March 20, 2012, 09:24:22 PM
this just popped up does it mean anything.
Quote
2012-03-20 02:10:18.525000 ALERT: ('\x01\x00\x00\x00\xf6\xfacO\x00\x00\x00\x00\x
f2*EQ\x00\x00\x00\x00\xf3\x03\x00\x00\xf2\x03\x00\x00\x00P\xc3\x00\x00|\xc4\x00\
x00\x00\x88\x13\x00\x00\x00JURGENT: security fix for Bitcoin-Qt on Windows: http
://bitcoin.org/critfix\x00', '0E\x02!\x00\xf4\xbd>d\xa1\x07w\x99\xc8\xd7Z\xa5\xb
fGUd\x93\xceXF\x96P\x98\xea"\xa68\xbfW\xde\x8c\xa9\x02 -\xda\xe4\xa4\xc9-N\xe4\x
fa&G\xdb\x94dQ\x819\xdd\xd4\x8fW\xe0\xa0Q\x04Ew\tr\xa2Y\x0b')
f2*EQ\x00\x00\x00\x00\xf3\x03\x00\x00\xf2\x03\x00\x00\x00P\xc3\x00\x00|\xc4\x00\
x00\x00\x88\x13\x00\x00\x00JURGENT: security fix for Bitcoin-Qt on Windows: http
://bitcoin.org/critfix\x00', '0E\x02!\x00\xf4\xbd>d\xa1\x07w\x99\xc8\xd7Z\xa5\xb
fGUd\x93\xceXF\x96P\x98\xea"\xa68\xbfW\xde\x8c\xa9\x02 -\xda\xe4\xa4\xc9-N\xe4\x
fa&G\xdb\x94dQ\x819\xdd\xd4\x8fW\xe0\xa0Q\x04Ew\tr\xa2Y\x0b')
This isn't an attempt at an exploit, its more like a public service announcement letting you know about a vulnerability so you can upgrade.
I believe the stuff around the text is a PGP signature, but I could be wrong.
I could be wrong though, it could just be a warning.
I LOL'd. A lot.
That is all.
March 20, 2012, 08:04:13 PM
Will do. I had a power outage last night which presented the perfect time to get everything updated. P2pool is showing a much better local speed now. I was on CGmine 2.2.1 and got bitcoind 5.3.1 running today so it appears to like that a little better.
March 20, 2012, 06:56:48 PM
Has anyone had issues with bitcoin timeouts? I have tried 6.0 RC2 and 5.3.1 and have had some pretty common timeout issues where the P2pool.exe windows says that it lost connection bitcoin for x.x minutes. I am running the bitcoind version and not -qt so that is all setup and I've not had issues like this in the past. I've also been using Cgminer for my miner. According to that, my local hash rate is 1300+. Though on p2pool.info and in the p2pool window, it reports my local as 1150 or so. Is there anyway that I could actually get my true hashrate to reflect in my mining efforts? I'd rather not lose more than 10% of my hashing power.
Try Bitcoin 0.6.0 RC4. March 20, 2012, 06:50:28 PM
Has anyone had issues with bitcoin timeouts? I have tried 6.0 RC2 and 5.3.1 and have had some pretty common timeout issues where the P2pool.exe windows says that it lost connection bitcoin for x.x minutes. I am running the bitcoind version and not -qt so that is all setup and I've not had issues like this in the past. I've also been using Cgminer for my miner. According to that, my local hash rate is 1300+. Though on p2pool.info and in the p2pool window, it reports my local as 1150 or so. Is there anyway that I could actually get my true hashrate to reflect in my mining efforts? I'd rather not lose more than 10% of my hashing power.
March 20, 2012, 05:11:17 PM
alright thanks.
about the payment, is it send directly to my wallet? i mean, i didnt put an address for it anyway, like on deepbit or other pools. Thats what im unsure about.
Yuo, p2pool get the address from the bitcoin client. If you notice when you start p2pool it tell you an addressabout the payment, is it send directly to my wallet? i mean, i didnt put an address for it anyway, like on deepbit or other pools. Thats what im unsure about.
As for shares, 1 p2pool share is like 600 difficulty 1 shares, so, be patient.
March 20, 2012, 02:18:59 PM
Ah, I would love individual graphs for each miner.. Thats the last missing piece for my setup, to easily check if all miners behave well..
Any thoughts on this?
Ente
Any thoughts on this?
Ente
The RRDs are there to do this. And, I have it on my list to dig into the code to implement as I too want individual graphs. But, time is what I lack. But, with the data there already, this is possible with not much work.
It looks like the graphs will be done in html/javascript going forward:
http://p2pool.hopto.org:9332/static/graphs.html
might be easier to improve. I've been meaning to take a look at them too.
March 20, 2012, 10:38:44 AM
Ah, I would love individual graphs for each miner.. Thats the last missing piece for my setup, to easily check if all miners behave well..
Any thoughts on this?
Ente
Any thoughts on this?
Ente
The RRDs are there to do this. And, I have it on my list to dig into the code to implement as I too want individual graphs. But, time is what I lack. But, with the data there already, this is possible with not much work.
March 20, 2012, 09:10:25 AM
When you mine, the VIP password = your miner's password.
That's when your address shows up on the local pool data.
That's when your address shows up on the local pool data.
Aaaah! I would never have guessed that! Thank you!
Ente
March 20, 2012, 08:51:54 AM
There are no individual graphs. You can set each miner to a unique username, and use the VIP password to get something like this:
by going to http://:9332/graphs
Now, if anyone knows how to edit the order, or remove old names from these graphs, that would be great.
by going to http://
Now, if anyone knows how to edit the order, or remove old names from these graphs, that would be great.
Seems like your image isnt embedded here, the direct url does work though.
Thanks for that! I see, all miners are in one graph..
I have the global and local hashrate graphs, but with no individual miners combined in the local hashrate. And I didnt see an obvious place to put my VIP password in anyway.. Will have a look again at home.
Ah, I would love individual graphs for each miner.. Thats the last missing piece for my setup, to easily check if all miners behave well..
Any thoughts on this?
Ente
That's when your address shows up on the local pool data.
Jump to: