|
Yeah, I mine on USEast and have been having a ton of problems this morning.
|
|
|
|
Ok, So I kinda solved my first issue. Still cant get decent speeds with my working 1x to 16x adaptor for the life of me. Bu Also, I cannot get a full 300 out of my saphires. I am running diablo with -v 2 and -w 256 as that seems to work best for me, but for two cards, I get around 500 instead of 600. Any ideas?
I have to assume you are talking about 5830's. If you are, just OC them to around 950/300 and you should pull just about 300. 975/300 is even better. 1000/300 crashes the card for me. |
|
|
|
Question is: Would the miner inform you if you found a block or wouldn't it tell you anything, anyway? That way you could at least get annoyed because you come to the conclusion that solo mining maybe actually would have been more profitable for you.
As I understand it: Every time you see an accepted share, your client (phoenix, poclbm, etc) thinks it has solved a block. The pool requests an artificially lower difficulty setting, this allows you to "solve" blocks more easily. Every time you solve one of these easier blocks, you are credited with a share, a proof of work. Since the method to solve both difficulties is identical, the same work would be done to solve both. So the only real difference is your client reporting back more often, the pool interprets these "solved" blocks as shares, and knows the actual difficulty, so when a true solution is found the pool can submit the solved block, dividing the proceeds to everyone based on the amount of shares they have into the round. tl;dr Your client couldn't know when you solved the actual block, only the pool. I think this is right. I know (or rather, I believe) that BTC Guild recently raised the difficulty from 1 to 2 to halve the strain on the servers. |
|
|
|
|
If you allow difficulty decreases more quickly, then you allow real DDoS attacks against large pools (or maybe even individual nodes in the BTC Network) to be more effective.
Hit all of the major pools, and then you get a bigger piece of the pie when the difficulty drops.
I might be missing something, but I think this is the counterargument.
Also, if someone wanted to be devious, and had that much hashing power, they could just fork the block chain at will, basically stopping mining in it's tracks, forever. So this scenario isn't really one you need to guard against (since worse attacks exist with that much hashing).
|
|
|
|
Finally got one of my custom shelves built last night!
Holding 4x5830
I have a couple more machines, but they are more boring.
Very nice, but are you not worried about the cards getting very hot in direct contact with the wood? Seriously, even if the cards got up to 100C, wood doesn't self ignite until 300C or above http://www.engineeringtoolbox.com/fuels-ignition-temperatures-d_171.htmlPretty much this... but also it doesn't touch the really hot bits either. Even if the card was at 100C, where it touches would be way less than that. |
|
|
|
It looks like "someone" (you?) ran poclbm as root. poclbm creates an .elf file on every startup (I think).
You can try -
sudo rm -f a1e83b97617e24d438cfc4107fb4f147.elf
Then try running poclbm again.
Yeah, I have been. I don't think I did that the first time. I have had permissions issues with LinuxCoin since the first time I ran it, for whatever reason. They are easy to get around, but involve running things as root that I probably don't have to. |
|
|
|
is there any GUI miner for linux?
i am trying to get linuxcoin to work and i keep getting this error when i try to start a miner from the command line.
Exception in thread Thread-2:
Traceback (most recent call last):
File "/usr/lib/python2.6/threading.py", line 532, in __bootstrap_inner
self.run()
File "/usr/lib/python2.6/threading.py", line 484, in run
self.__target(*self.__args, **self.__kwargs)
File "/opt/miners/poclbm/BitcoinMiner.py", line 272, in miningThread
self.loadKernel()
File "/opt/miners/poclbm/BitcoinMiner.py", line 373, in loadKernel
binaryW = open(cacheName, 'wb')
IOError: [Errno 13] Permission denied: 'a1e83b97617e24d438cfc4107fb4f147.elf'
the rest of it seems to be working great, i just cant get the miners to start.
I have the same problem. I run poclbm directly with sudo.
I do:
1. cd opt/miners/poclbm
2. sudo chmod +x ./poclbm.py
3. sudo ./poclbm.py -d blah blah blah
Maybe you can just sudo the batch that he made though... I would try that.
|
|
|
|
I get a lot of idle messages, but my GPUs keep running at 99% and the Hashrate doesnt go down at all. Does it mean I can just ignore the warnings?
Pretty much yes. If your hashrate is unaffected its unlikely that your rig is being affected by tiny amounts of idles. Ignore the "fatigue" guy, sounds like he has some weird problem with his individual setup. The only problem I'm having with BTCGuild is a slightly overhigh stale share rate. Still floating around 2%, whereas before I used to be at 1/2%. Perhaps its just an inevitable effect of higher difficulty I dunno. I don't think difficulty and stales would be related... |
|
|
|
Finally got one of my custom shelves built last night!
Holding 4x5830
I have a couple more machines, but they are more boring.
What is that, 3x PCI-e 8x extender, plus one PCI-e 16 x extender? Where did you get those cables, and how much did they cost? 2x PCIe x16 extender 2x PCIe x1 extender https://cablesaurus.com/ |
|
|
|
Finally got one of my custom shelves built last night! Holding 4x5830 I have a couple more machines, but they are more boring.     |
|
|
|
*push*
I know people have 4 on 7x64. Not sure about 5. With windows you will need dummy plugs or reg editing to get the cards to show up though. |
|
|
|
eventually the diff level will increase to the point that you will be mining for weeks before you get a block.. that will crush mining all in all.
No... there will always be pools... so it will make no difference... |
|
|
|
yeah 5770s are getting fairly cheap, usually under $100.
It's going to be more and more expensive to build mining rigs until the 7-series is actually released, then you'll be able to grab a bunch of 5-series for cheap hopefully
Plus you can get them single slot, which could be nice for some people. |
|
|
|
As we all know, all of 5850 and 5830 are sold out.
What do you think might be the next best GPU?
Excluding all cards that you can't readily buy from major e-tailers.
If you are power conscious... the 5770 is pretty good. 200 mhash for 100W. If not... probably 6xxx |
|
|
|
what about 4 GPUs, that's safe right?
I have a system with 4 GPU's... two on x16 and two on x1. It's working perfect |
|
|
|
i think you need the flags
+99 -v -w 128 Try it and let me know |
|
|
|
Video cards are designed to go into the long slots (PCIe x4,x8, and x16). But you can buy cables that allow you to use the short slots (PCIe x1) for mining. Or you can take a file to it.
That scares my inner child. Well, the card won't physically fit unless you remove the end of the slot  |
|
|
|
Ok so I haven't built a machine myself since AGP was the architecture of choice for GPU interfaces. All this PCIe stuff has me a bit confused. I went looking at mobos yesterday at Fry's and saw a bunch that had full sized PCIe and really short (maybe 1") PCIe slots. Now supposing I wanted to build a 3-4 video card rig... would I have to find a mobo that has 3-4 of the full-size PCIe slots? Or would I be able to use like an HD5870 in one of the short slots?
Sorry for the dumb question. It's been years since I've even looked at the inside of a computer, literally.
Video cards are designed to go into the long slots (PCIe x4,x8, and x16). But you can buy cables that allow you to use the short slots (PCIe x1) for mining. Or you can take a file to it. |
|
|
|
The server is on the west coast! It's 04:45 there, so I'd imagine SC doesn't even know yet!
I think he said he was from TN once. |
|
|
|
So if i understand correctly, a simple situation would be this:
If Person 2 joined the pool at when the weighting was 5.
And the number means how many shares have been submitted at that weighting (would decrease within a given time period).
Person 1 Person 2 Share Weighting (Decreases with time until block is solved)
1 10
1 9
1 8
1 6
1 3 5
1 3 4
1 3 3
1 3 2
1 3 1
Person 1 submitted 10 shares. Has a total weighting of 55.
Person 2 submitted 15 shares. Has a total weighting of 45.
This means that Person 1 minimises their losses (or even equals or gets more than) Person 2 as he started mining earlier in the pool, and so the weighting starts canceling out the fact that Person 2 has a higher share of the reward.
No, that's what the scoring is meant to prevent. Share # Person a (Pool Hopper) Person b (loyal member) Share weight 1 1mh 1mh 1 100 1mh 1mh 1.1 10000 1mh 1mh 2 100000 0mh 1mh 4 1000000 0mh 1mh 10 Please note these numbers are purely made up, but they might help get the point across. Since you never know when a block will end and pool hopping exploits the start of a new block, it causes anyone trying to cheat to not get a disproportionate reward on short rounds. The share weights are not 1-1, and reset with each block, creating a constant sliding scale. Your rewards should be the same or better than a proportional share if you are not trying to exploit pool hopping. Doesn't this just change the incentive for pool hoppers from exploiting the first half of the round, to exploiting the second one? *joins for first 1/3 of round, hops pools, re-joins for last 1/3 of current difficulty*  Well... you never know when you are going to find a block though. So you can't know when to rejoin. It's just good to be there right before it is found. |
|
|
|
|
Show Posts
