Let me know if there is anything other information I can give you to help with.
I just pushed out beta2. This fixes the crashes on Windows, you can go crazy disconnecting and reconnecting devices if you want.  Could you try it on Mac and see if works now? (Same URL/beta button above) And yes, thanks for the crash log, it was very helpful. |
|
|
|
Am I supposed to install drivers for these? I'm kind of at a loss...
On Windows it should install the necessary drivers automatically when you plug it in. Not sure about Mac and Linux yet. It's just a standard virtual-serial-port-over-USB kind of thing. I'm looking into the issue with crashing under Mac OS X. If you have access to a Windows machine you could try that so they don't sit around idle. |
|
|
|
Trying it out on OSX Lion (10.7.3) Server w/ 3x BitForce singles. When I try to scan for FPGAs, it just crashes. After it gets through with trying the Bluetooth modem (which fails), it dies. Mooing to try rebooting and giving it another shot.
The whole app crashes, and possibly Java leaves a crash log on the desktop? Probably the RXTX library (for Java access to serial ports) that's doing it. I'm starting to hate that buggy thing. Could you try adding them one by one using the menu to probe a specific port? |
|
|
|
|
The BitMinter miner also works fine on Macs. As long as you have Java installed, you just click on the webpage and off it goes.
|
|
|
|
In theory it should be faster than cgminer and the other miners, because of faster work dispatch. In practice the difference is negligible (less than 1 Mhps per BFL single). This may prove more useful with the rig boxes. Instead of checking whether the device is done with 10 ms pauses in between checks, it will first leave the device in peace until it gets close to done (cgminer does the same), then start 10 ms polling, and when it gets very close to done it starts 1 ms polling. This allows quick reaction pushing out new work once the device is idle. Right after it goes "ok, I'm all done, I think I'll just read the newspaper now" you have to detect that situation quickly and put it back to work. You can click the wrench and change the 10 ms and 1 ms values. You could even change the 1 ms to 0 ms. Running with 0 ms might be ok. It won't hog the CPU as much as the AMD drivers do/did. I just thought it might be a bit brutal for a default setting. |
|
|
|
v1.2.0 beta1 just released. New in this release: support for BFL BitFORCE FPGA devices! A big thanks to Fefox for lending me his BFL singles for testing and his participation in the testing!  If you have one or more such devices, please help test this beta. It has only been tested on 64 bit Windows. That leaves 32 bit Windows, 32/64 bit Linux, 32/64 bit Mac OS X. How to use it: It doesn't probe the serial ports by default. There is a new menu "devices" with two possible actions: checking all serial ports for BitFORCE devices, or a specific one. Also there is a new FPGA tab in the options (Settings -> Options in the menus). Here you can set how to detect FPGAs on startup and whether to scan for added/removed devices at intervals. There are also some settings you can change on each device. Click the wrenches to change these settings. Please try it out:  If the device reaches max temp it will stop. Click the wrench to change the max temp. Is 75 C a good default max temperature? Any feedback appreciated. |
|
|
|
Good day to be mining! 5 bitcoin blocks so far... can we get one more?  |
|
|
|
There is no list of payouts for the block 177274 (the one before the one marked as stale).
Does it mean that now the unconfirmed rewards won't be listed on the block details page unless we activate the new prepay perk?
This was caused by the bug mentioned above. Now fixed and everything is rolling again. The bug was causing some operations to abort, for instance the pay splitting of block income, as you noticed. After I fixed the bug all those operations went through. Sorry for the confusion and the delay in block payouts. On the upside, nice to see some lucky blocks again. |
|
|
|
|
The latest BTC block is not stale despite what the block list shows at the moment. There's a bug, I'm working on it.
|
|
|
|
|
Quick bugfix this morning. Account details was still showing unconfirmed amounts after they were already paid out with the prepay perk. Should now be correct.
|
|
|
|
Stats->blocks: "all" filter seems to only show BTC, otherwise looks good.
Oops! Thanks for letting me know. Fixed now. I guess sorting the blocks by height wasn't such a good idea - it put all the namecoin blocks at the end.  Me and my 6 gpu farm are back, I lost faith with p2pools luck (12% missing since forever). Any big changes that could require my immediate attention in the last 2 months ?
Nice to see you back on the team. Like P4man said the luck was not good. A lot of users left and we are only now back to the same number of active users. You can be glad you missed the block that took 11 million proofs of work. But we did have a block with only 282 proofs of work after that, which partly makes up for it. New on the feature side is the donation and perk system. I have been working on other things too, but it's mostly behind-the-scenes stuff that you won't notice yet. There is more stuff on the way, though. |
|
|
|
So how about GCN (HD7970)? Is this supported/faster/...?
Yes, it is supported and from the reports I get the BitMinter client seems to run very fast on 7970. Of course you can also use DiabloMiner, cgminer or any other miner in the pool. I don't have a 7970 so I haven't been able to speed test the miners properly against each other. |
|
|
|
New "prepay" perk is now available (and default for new users). With this on you will get paid when a block is created. No need to wait for 120 confirmations, which on average would take 20 hours. Hopefully this makes the pool even more appealing, especially for beginners trying out mining for the first time. Full update details: https://bitcointalk.org/index.php?topic=27062.msg867354#msg867354 |
|
|
|
Another website update. Here are the news:
- Perk support completed. One perk added so far:
- Prepay perk. If you have this perk activated you get paid when the block is created (unless stale). No waiting for 120 blocks, and you keep the coins if it later becomes orphaned and worthless.
- Start off new users with the prepay perk activated for both BTC and NMC.
- Stale blocks now listed as "stale" (previously not shown). These are blocks that have never been part of the longest chain. There is no pay for these, even with the prepay perk.
- Small bugfix on login page: switch keyboard focus to input field after you click an OpenID provider logo (was not working).
- Don't extrapolate stales in live stat shift display (forgotten in the last update).
- Site navigation and info/error messages (if any) stick to the top when scrolling down. This was done so you can see new messages appearing as you interact with something far down on a page. Having navigation always available also seems handy.
- Fix luck statistics. Since a backend update a few days ago it's been displaying blocks we didn't make (with 0% CDF).
You may have to clear your browser cache before the navbar & messages will stick when scrolling. As always, please let me know ASAP if you experience any difficulties. |
|
|
|
Dr, I think the luck page is a little broke. All but a couple of points show 0% CDF for me.
Oops! Sorry about that. I'll try and get a fix in with some other updates tonight. So far, the fixed 5870 is holding up. We shall see how they go through this night.
Looking good! |
|
|
|
Nice summary about block withholding, DnT. Should perhaps go on the bitcoin stackexchange? Since there is no economic value to withholding blocks the attackers motivation is non-economical. The minting bonus doesn't really materially change the cost.
Yeah, the hope would just be to discourage it by making it as expensive as possible. I think at least that PPLNS has an advantage over PPS-based payout models in this. If you hurt the pool 10% then you hurt yourself 10% too. But you are probably right that adding a minting bonus to this wouldn't be a big deal. If someone is willing to lose a considerable income from their mining just to hurt your pool, then an extra 1% loss is not going to stop them. Sadly the "real fix" would be blind hashing (miner doesn't know if the hash solves a block or not) and that simply isn't possible with Bitcoin without a breaking protocol change. Maybe the "next" blockchain will incorporate ideas like that.
Yeah, perhaps something for the altchains to think about. But I have the feeling most of them just want to make a quick buck, not contribute something useful. Nice job on your solved block Doc!
Thanks! I got two blocks only 3 days apart - quite happy with that. |
|
|
|
Another option is a new donation cause ('Minting bonus' or something like that?). The payout for each block could be whatever is in the donation box at the time. Or it could be CDF percent of the donations. So if you get an easy 10% block, you get 10% of the current donations. For a block like today 99.9% of the available donated funds. To be fair one thing you may wish to do is make the "minting bonus" only payable to those who enable "minting bonus" donation. Could make it something like 2% max donation. If you donate the max and solve a block you get the payout. If you donate say 1% instead you get half the payout, 0.5% gets you 1/4 the payout. The payout is based on the optional minting donation/fee set by the miner at the time the block is found. That way the minting bonus money circulates among "players". More likely to encourage donations that way. Good idea. I like the thought that if you contribute something you get something in return. Only downside is that it doesn't help against block withholding. But with PPLNS miners at least have some incentive to not withhold blocks. So maybe it's not needed. |
|
|
|
|
Yes, interesting. Namecoins have more than doubled in value lately. Something going on?
|
|
|
|
|
Show Posts
