ChanceCoats123
|
|
March 21, 2012, 01:04:13 AM |
|
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.
|
|
|
|
m3ta
|
|
March 21, 2012, 02:24:22 AM |
|
this just popped up does it mean anything. 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') 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. Looks like hex obfuscation of malicious code.I could be wrong though, it could just be a warning. I LOL'd. A lot. That is all.
|
|
|
|
ChanceCoats123
|
|
March 21, 2012, 03:45:31 AM |
|
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.
|
|
|
|
ancow
|
|
March 21, 2012, 04:49:03 AM |
|
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).
|
BTC: 1GAHTMdBN4Yw3PU66sAmUBKSXy2qaq2SF4
|
|
|
Gabi
Legendary
Offline
Activity: 1148
Merit: 1008
If you want to walk on water, get out of the boat
|
|
March 21, 2012, 06:40:56 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.
|
|
|
|
Ente
Legendary
Offline
Activity: 2126
Merit: 1001
|
|
March 21, 2012, 08:11:33 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
|
|
|
|
sharky112065
|
|
March 21, 2012, 11:16:55 AM Last edit: March 21, 2012, 11:29:04 AM by sharky112065 |
|
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.
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.
|
Donations welcome: 12KaKtrK52iQjPdtsJq7fJ7smC32tXWbWr
|
|
|
ChanceCoats123
|
|
March 21, 2012, 12:20:43 PM |
|
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!
|
|
|
|
1onevvolf
Newbie
Offline
Activity: 43
Merit: 0
|
|
March 21, 2012, 12:51:23 PM |
|
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.
|
|
|
|
Gabi
Legendary
Offline
Activity: 1148
Merit: 1008
If you want to walk on water, get out of the boat
|
|
March 21, 2012, 01:14:00 PM |
|
Try driver 11.11
They don't have the sdk 2.6 and don't have the 100%cpu bug
|
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
March 21, 2012, 01:24:00 PM Last edit: March 21, 2012, 02:06:02 PM by DeathAndTaxes |
|
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.
|
|
|
|
sharky112065
|
|
March 21, 2012, 01:47:04 PM |
|
Try driver 11.11
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.
|
Donations welcome: 12KaKtrK52iQjPdtsJq7fJ7smC32tXWbWr
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
March 21, 2012, 02:08:39 PM |
|
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).
|
|
|
|
forrestv (OP)
|
|
March 21, 2012, 02:34:02 PM Last edit: March 21, 2012, 02:57:24 PM by forrestv |
|
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.
720 shares, which should normally be 720*10 seconds = 2 hours. This should be more self-documenting.. maybe return a JSON Object with an "averaging period" attribute. --- Thanks for the recomendations about handling the donation amount. I will implement the warning-on-p2pool-start idea. Also, the RRD graphs are deprecated. The new JavaScript-based ones have a lot more potential and flexibility. They don't have support for measuring individual miners yet, but that will come soon. 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.
BIP16 is already final. It has >50% of he Bitcoin network's hash rate behind it. Any attempt at opposing it can only invalidate your own work. Note that during the last few months, P2Pool has passed through your bitcoind's "coinbaseflags" into mined blocks, which let you vote during the period where it mattered.
|
1J1zegkNSbwX4smvTdoHSanUfwvXFeuV23
|
|
|
Ente
Legendary
Offline
Activity: 2126
Merit: 1001
|
|
March 21, 2012, 02:55:14 PM |
|
Also, the RRD graphs are deprecated. The new JavaScript-based ones have a lot more potential and flexibility. They don't have support for measuring individual miners yet, but that will come soon.
That's great news, thank you! Ente
|
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
March 21, 2012, 03:16:01 PM |
|
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.
The support is based on the client that each p2pool user is running, and whoever solves the block is the one that gets the "vote" That being said however, voting is pretty much over and P2SH is going ahead anyway, so if you don't support it you risk your blocks becoming orphan, and you not getting any payout at all. That would include all p2pool miners' payouts from a non supported bitcoind, since it is split.
|
|
|
|
sharky112065
|
|
March 21, 2012, 03:55:54 PM |
|
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).
I am fully aware that the DRIVER is what affects the CPU bug. I am also fully aware, that every combination of AMD-APP Runtime 2.5 variation I tried with newer drivers that fix the CPU bug, resulted in 10 Mh/s less per card on the three rigs I tried them on. This may not be the case on other OS's and card combinations. I tested it on Windows 7 X64 with 6970's and 5870's (cgminer 2.3.1-2 and deleting .bin files every time a change was made [20 minute run time]). Were you on 2.5.732.1 before you got rid of the CPU bug? If not you may not have noticed any drop in Mh/s.
|
Donations welcome: 12KaKtrK52iQjPdtsJq7fJ7smC32tXWbWr
|
|
|
Schleicher
|
|
March 21, 2012, 04:26:06 PM |
|
P2pool keeps growing. This is a new record I think: 2012-03-21 17:27:27.558000 Pool: 369GH/s Stale rate: 8.3% Expected time to block: 4.8 hours 2012-03-21 17:27:29.960000 New work for worker! Difficulty: 0.999985 Share difficulty: 900.936232 Total block value: 50.016434 BTC including 27 transactions
Oh, by the way, bitcoin-qt 0.6 rc4 doesn't crash under Windows. Hooray.
|
|
|
|
Ente
Legendary
Offline
Activity: 2126
Merit: 1001
|
|
March 21, 2012, 07:02:46 PM |
|
I have more orphans than I like.. Orphans point to bad connection? I forwarded port 9333, but I still only have 10 peers, 0 incoming.. Other hints? Can I change the '10' limit? The internet connection isnt that fast, I fear..
Ente
|
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
March 21, 2012, 07:05:13 PM |
|
Depends on how many? Some % of orphans is simply unavoidable but if your avg is similar to the network avg then your compensation is the same (i.e. you having 0% orphans and network having 0% orphans pays the same as you have 8% orphans and the network having 8% orphans).
If it is significantly more than the average then more connections may help. The 10 is only on outbound connections. If you portforward you can have more than 10. 10 outbound + unlimited inbound.
|
|
|
|
|