Topic: [1050 TH] BitMinter.com [1% PPLNS,Pays TxFees +MergedMining,Stratum,GBT,vardiff] - page 363. (Read 837180 times)

OK, you convinced me. I'll keep testing Bitminter a few days more, even if I suspect that with a so low Hashrate it tends to be very unlucky most of the time (BTW: how about a measure of the pool's luck in the statistics?). On the other hand, I love the app and the interface.

I dont see why.  It shouldnt make any difference.  The average of the last 80 hours wont be different from the past 24h. So your payment wouldnt change either.
In fact, you will likely get underpaid in any  pool that does not use PPLNS, because those few pools that still havent adopted it, attract all the pool hoppers, and pool hoppers gain a disproportionate amount of the payouts (which means, you lose).

My puters are up about 12-16 hours per day. So maybe 53 hours every 80. In any case I would be underpaid in comparison to any other pool with regular proportional payment.

I kinda doubt it would make a difference. The average of the last 10 shifts.. a shift is taking over 8 hours. So its your average contribution over the last 80 hours. Unless by part time you mean you turn the rig on once every two weeks, I dont see the problem.

But you made equally pointless to mine with your pool for part-time miners like me :-(
Good bye

As a result, difficulty is going down too. I just keep mining, no matter the price. The way I see it, either bitcoin takes off in the next years, and Id bang my head against the wall for not mining now, or they dont, and Id have lost, frankly, not much. That the price is low, and the difficulty dropping accordingly makes it a great time to mine. Yes, its speculative, not that different from buying coins, but where is the fun in that . Besides, for me prices would have to drop quite a bit still to make it not operationally profitable even at the current difficulty. Its still cheaper than buying coins.

wtf is going on with the price?  i might have to shut down......

Dont break your head over it. It seems to be some problem on my side. Hours later, Bitcoin.cz also stopped accepting or rejecting shares. Last share accepted was 3 or 4 hours ago, ever since I do get messages about new blocks, I get the MH rate, but its stuck on 6 accepted shares. No idea whats causing it is, but I suspect it will be solved by switching to an AMD card and diablo.

Yes, something is seriously wrong there. What amazes me is that it keeps working doing hashes all this time, but no accepted or rejected shares. And also no news from long polling.

On one machine I am running a GeForce 9400 GT at 3 MH/s with the bitminter miner and everything works fine. So it's not because your card is slow - this one is only 10% of your speed and still works.

I am confused with this too. I guess bitcoin.cz must be doing something different, that keeps Phoenix happy. I just wish I could reproduce the problem, then I could find a way to fix it. Phoenix runs just fine on my main mining machine. Well, maybe I should try it with the old GeForce.

I pointed my rat rig to bitcoin.cz pool, and got some results that may shed more light on it, as for some reason its more verbose.:



It does get disconnected, but reconnects with the next block? Since that pool is much larger than bitminter,  that happens every x minutes as compared to every x days on bitminter, so it keeps working. Not sure. The last disconnection seems have been fixed without a change in block.

my two bitminter rigs have 36000 works/89 rejects in 1 day and 3:49hrs/3 cards, this morning around 9 am i had 600(wasnt paying attention, just woke up) rejects on my guiminer, reset it and its been good all day

Restarted the rat rig this morning, and again after a (short) while, its  not getting any shares confirmed:



Note this screenshot was taken after 10:00 am, so an hour and a half after the last confirmed share. It hasnt crashed or anything, MH rate is fluctuating, gpu load 100%, just not getting any accepted (or rejected) shares. Im confused.
Ill try another pool and see what happens. Maybe the nv card is buggered?

oops,  I used bitcoin.cz

Cant find any such setting in my router. But Im using the same router with my main mining rig which has extremely few stales. Then again, maybe it is related to the speed.. if it takes the nvidia card so long to finish one share, maybe the connection times out where with my main rig it does not, as it never takes 5 minutes to complete a share with my 5850.

For me its not worth bothering. But in the stats I see some with pretty slow machines (unless they turn them off frequently), perhaps you could check the stats and see if anyone is really suffering from many stales. As for how to solve it.. dont ask me

I thought slush's pool didn't even have long polling? Or maybe it does now?

I see in the screenshot that you had some longpoll (LP) messages, the latest at 19:54, then nothing all the way down to 21:39. This tells me long poll has failed. You do have some rejected shares over that timerange, and that's probably the places where longpoll should have been giving you new (non-stale) work.

Maybe you have a router that kills idle connections? You could look in its configuration if there is a setting for how long TCP connections are allowed to be idle before they get killed off. This is the reason I made the bitminter miner restart the longpoll connection every 5 minutes.

Perhaps I should kill long poll connections also from the server-side after 5 minutes? I am working on getting TCP keep-alive in there though, much better solution than restarting connections all the time.


The server is just fast with long polling, that is correct. I don't really see another way the server can reduce stales.


Sure, I want all client software to work, and cgminer seems like a popular one nowadays.

Part of the reason you get such ultra-low stales is also cgminer I think. I heard somewhere that if it is about to submit a share and sees it is stale, it won't submit it. Something like that. And it makes sense, why submit shares you know are stale. I'll probably put something like this in the BitMinter program soon.

I have found bitminter to be one of the best pools when it comes to stales.



Ignore the 0MHz I am accessing the desktop remotely through RDP and that seems to confuse cgminer.

~15 hours,  19,462 shares and 5 rejected = 0.03% stale rate.

Is bitminter doing something special?  I am guessing if it is ultra fast at reporting new block via longpoll that would be the only thing that would bring down stale rate right?  Weirdly I have noticed my rejected shares are lower since I started watercooling but that can't be right because rejected shares should only depend on network latency right?

The fact that bitminter is closed source is still a concern of mine but bitminter the pool seems to be doing very well using 3rd party minter (cgminer).

No, it was with phoenix. But when I ran the same miner on the same machine on slush pool for a while, I didnt have any rejects. If it helps, here is a screenshot:




I actually did put a fire detector/alarm in there And Im still gonna turn it off for the night. There is a supply of firewood in there, and its right below my bedroom.  Illl probably also add a deskfan of sorts,  as despite being so open, heat does get trapped in it. Couldnt have used a worse CPU, 3.4 GHz prescott and a motherboard that doesnt let me underclock it. And with nvidia openCL its constantly running at 100%. Thats what you get when you use scrap lol. Think I have something lower power somewhere or at least a motherboard that will let me underclock it.

Over 10% rejected is ridiculous. It could be a problem with long polling. Long polling should normally keep the rejected numbers way lower than that.

I hope this wasn't with the bitminter miner. In the last version I thought I made long polling pretty safe even with the routers that are most aggressive with killing idle connections.


Cool Just keep it open like on the picture or I think a fire will break out.

Well, as I mentioned in the edit, for whatever reason it does seem to work now, only,  Im getting suspiciously high rejection rates on the slow nvidia/phoenix machine. On my main rig I have 2 rejects out of 6000+ (which is awesome). On the slow nv machine Im now looking at  4 rejected out of 27. Is that just terrible luck or because its so slow?

as for the fastest cuda miner; no matter, I was simply testing the 'case' and internet connection, running it headless etc, as the machine is tucked in a kewlox cupboard in a side building of my house:


[/URL]

Since it seems to work and no fire has broken out yet, so I have just ordered a 5850 to replace the nvidia card.

That's actually on my list already. I just have to find the time to work through that TODO list!


Hmm, yeah, that would be useful. Added to the list.


Hmm, yeah, I suppose I could make the miner report in the devices used and make some statistics on the server. It would be cool stats for sure. Just too bad it wouldn't work with other miners. Anyway - I put it on the list.


Yeah, everyone seems to want a bigger top hashrates list. The way you put it sounds like a good solution. Maybe a button to switch between 10/20/50. Added to list.


I don't think so. I run multiple workers all the time, and I see in miner stats that we have a few Phoenix users. I just tried Phoenix now and it worked fine.


Well, that does sound bad. Anyone else have similar problems?


You probably have "manual vectors" on with the bitminter miner. On nvidia things go much better with them off.

That said, I think there's a CUDA miner that might be fastest on nvidia. Can't remember what it's called though.. cudaminer?

edit seems to be okay now 

/edit

Is there a problem with multiple workers and/or phoenix?

Ive just brought an old rig online as test, and created a new worker for it, and I would get between 1 - 3 accepted shares fairly quickly, and then nothing for hours, while the machine was still claiming a whopping 30MH.

Ive deleted and recreated another worker and reconfigured the miner, and had the same thing. First share was accepted, then nothing.

Now Ive tried it on a different pool, and I have no problems there, currently looking at 10 accepted shared in about as many minutes?

Im using phoenix btw, as its a really old machine (pentium 4 with a geforce 8800) and I got way "better" performance with phoenix than bitminter