This is after a full night:


Results are much closer now, and much more believable, though bitminter is still faster by a non trivial amount.
cpu speed is a non issue clearly, but we could have guessed that.

I also notice diablo timing out a few times contacting the pool. 6x in the past 10 hours. Not sure if that is really an issue here, as a few seconds later it always gets through, but  bitminter  (machine is on the same network) never timed out once.

lastly, I have 1 hardware error and 4 rejects on diablo, and zero of either on the bitminter machine, but that may not be related to the software, Im not sure.

started a new run. Created 2 new workers, booted both machine in to identical ubuntu installs this time, so same drivers, same sdk same everything. Im running bitminter on the old P4 now, and diablo on the Core2q. Ill post results tomorrow if neither machine crashes (the c2q is a rebuild thats not quite done yet. Like the CPU fan is not even attached its merely balancing on the CPU lol. And no, its not overheating at all, dont worry).

update: before I go to bed, this is what I have after 30 minutes or so:



Even with a pentium 4 bitminter seems to be pulling way ahead. Yes, its way too early, but still... color me surprised. I expected them to be close.  So am I doing something wrong with the diablo settings? Im using this:

./DiabloMiner-Linux.sh --url http://P4man_diablocore2:xxx@mint.bitminter.com:8332 -v 2 -w 128

Ive experimented with 196 and 256 but the results were inconclusive. So I kept it at 128, same as Im using with bitminter.

Does anyone know how to overclock a non reference 5850 beyond 775 Mhz on linux? Its an MSI twin frozr II. I flashed the bios of my 5850 reference card which unlocked all speeds, but I havent found a bios for this one...

No temperature issue, both cards are below 55C. The ubuntu/diablo machine does have a pathetic CPU by comparison (single core Pentium 4 vs Core2 quad) but I dont think that makes any difference.

The bitminter one is the one that produced more hashes. You must either compare the live apps (so 301 vs 255) or the ones shows on the website from a few minutes earlier (288 vs 253)

Ill do a longer sample later, but Im using my windows machine for other things it doesnt mine 24/7 like the ubuntu rig. When I do it, Ill reverse it. Ill run bitminter on the ubuntu/pentium 4 machine and diablo on the windows/C2Q.

Regarding the allegations of potential cheating with the reported MH, I decided to a little (admittedly, limited and somewhat flawed) test now that I have 2 5850s. One machine is running diablo (on ubuntu) the other is running bitminter (on windows). Both clocked at 775/300, similar optimal settings, same 2.5 version of the SDK. I started them simultaneously let them run for a few hours.

Bitminer and diablo both report very similar MH/s, around 310 (taking the screenshot makes it drop). Diablo is a bit more consistent, bitminter, definitely peaks higher, but if anything might average a bit lower in what it reports, hard to say. Both systems have had zero rejected shares in this run. Anyway, this is what the pool reports:



bit=bitminter
dia=diablo

Perhaps its too small a dataset to be conclusive, but from the very start bitminter has produced noticeably more shares than diablo. If Dr Haribo is cheating, it seems he is paying for it .

Its the AMD drivers (though nvidia drivers are not better, at least on linux)
Im now running an 5850  on the same old Pentium 4 and my CPU load is pretty much zero with bitminter. Im using (I think) catalyst 11.3. This is on ubuntu, but the problem is the same as on windows, so try grabbing some old ati drivers.

Where are you guys getting your news from I wonder... is it Fox spreading that sort of BS?

Yep; its a great pool, great tools and stats, and bonus coins on top of that. Now if only we'd have some more luck

And have you seen whats happened since december and may? Hint:
http://bitcoin.sipa.be/speed-lin.png

With the current difficulty it will take you on average 6459 days before you find your next block. Yes, almost 20 years.

no one said liquid cooling was a recent idea.

Card is up and running. Works great. Thanks a lot.
+1 rep to seller.

Its really not a question of OR, but AND. You can have the best VGA cooler in the world, if your case becomes a sauna, your card will still run hot. This is even more true for most aftermarket coolers that do not exhaust their heat through a PCI slot, but dump it inside the case. This requires good case ventilation.

That said, I do have an accelero twin turbo on my 5850 and its awesome, BUT.. well, there are serious buts.
Its awesome in the sense that it dropped my GPU temperatures from high 80C to 50C. 55C max when overclocked in a hot ambient room. Its that good. BUT, the accelero only cools the GPU, not the memory, and most importantly, not the VRMs. They include sticky alu fins for that, but they dont stick and are worthless. Throw them away. Instead I used the base plate from the stock cooler. That works fairly well, but if you have a reference card, you will have to modify things a  bit. I had to dremel the baseplate a bit to make room for the heatpipes, otherwise the heatpipes would push on the baseplate. You also need to bend a few fins of the accelero to make room for the DVI block. Neither is hard to do, and with a 5870 you dont need to do either I think.

Last point; VRM cooling. Im still not happy with my VRM temps. @850 Mhz the VRMs will heat to 80+C. This is well within spec, but still higher than I like, so I ordered a thermaltake VRM R5 cooler specifically for the VRM. Dont have it yet, but reviews look promising (although I will have to cut my stock cooler baseplate in 2).

Many cards dont support VRM temp monitoring, and I suspect this is the n1 cause for premature failures. With the stock cooler my VRMs went near 100C at stock speed. If you cant monitor your VRM temps, dont assume they are ok.

FWIW, Im not unhappy with my reference design powercolor 5850. I did add an aftermarket cooler (accelero twin turbo), and Ill soon add a VRM cooler, but with the accelero and at stock voltage it seems happy up to ~900 Mhz stock voltage. GPU temps arent a problem with the accelero (55C max) but VRMs get smoking hot at those speeds (90-100C). I hope the thermal take VRM R5 will take care of that, only then ill try some overvolting.

Frankly, I think all reference design cards are the same, regardless of brand.

And it arrived this morning
Havent tested it yet, but Im sure it will be fine. Thanks for the quick delivery. Thumbs up.

Not even sure what version of the drivers Im using, whatever is nvidia-current on ubuntu. But my 8800 has been working stably for a day now, using phoenix and connected to bitminter's pool. 

I think my problem was related to a little utility I downloaded to manage nvidia fan speeds when running headless on ubuntu; nvidia-settings requires a running x-server, and X wont start without a monitor. (Much as I appreciate's nvidia's usual level of support for linux, its pretty lame their smi command line utilities only work with quadro's and tesla's).

Regardless, another 5850 just arrived 5 minutes ago and the 8800 will soon be relegated to its previous role of being spare part.

Rather than forking, perhaps an alternative inflationary chain could be developed that derives its value from bitcoin. Akin to the gold standard, where the gold becomes the bitcoins and the inflationary currency is linked to it. In the far future, I could even imagine printed money like our current currency, but backed by bitcoins. Bitcoins would become a storage of wealth (similar to gold), the other currency or even currencies,  a method to conduct financial transactions.

Gimme a break

Only a year ago, a pizza would have cost you about that many bitcoins. So what? Rather than using bitcoin, you'd typically use "KiloCoin" or whatever, and you pay for your pizza with 40KC. I dont see the problem with that as a trade instrument.

This potential donor still is . It sounds too scary, too radical. I suspect most people protesting there wouldnt even back an abolishment of the federal reserve, let alone the dollar or the entire system of fractional reserve banking. They want jobs, reforms, both economic and political, but not "anarchy". End the banks sounds Utopian, unrealistic and frankly scary. What happens to our money when you end banks?

So I think "Its our money" is a better slogan, although perhaps more to the people who already know bitcoin and understand the double or triple meaning.


My 2 cents.

That doesnt make sense . A low combined hashrate does lead to larger variability, but statistically that has equal chances of working out in your favor or against it. In the long run, the average will not be different from anyone else.

 (BTW: how about a measure of the pool's luck in the statistics?).

Look here:
https://bitminter.com/blocks

Its the CDF that is what you are looking for.  Under 50% is more than average lucky, more than 50% is unlucky.