Not sure why you bought the cases either. It adds a bit of cost, and probably greatly increases temps; cold winters or not.
Pretty impressive setup otherwise. Any houses for sale near that hydroelectric dam?

As for my rig.. heh, well, I dont have canadian winter here, nor your electricity prices, so my setup is way more humble,  and since Im anal about low temps, it also looks a bit different



Anyone else mining on a 5870 @900 MHz  @ 41C?

Hey goat, I want in too!
Either you bet me on bets of bitcoin where you have a 7/5.25 odds advantage, or, if for some reason you dont want that, Ill take those 10-1 2-1 staggered odds with Inaba too. Just tell me how much you want to bet.

Not only that, because of a weird side effect from I *think* the caching algorithm (dont quote me on that), decreasing the memory clocks actually increases performance. Only marginally, but measurably. Saw a nice chart with the effect a while ago, Ill see if I can find it.


That is because the memory controller is on the GPU. The memory controller will draw less power if clocked lower, hence lowering your temperatures a bit.

Im not disputing it. However at stock core speed and 300 Mhz memory you may get better MH/W depending on the power draw of the rest of the machine.  Its almost certain if you combine it with undervolting (most 5850s seem happy at 1v for stock speed).


Thats mostly true when you have bad voltage regulation (power spikes) and for mechanical parts. But indeed, heat cycling can in fact cause "mechanical" failures with gpu's: hair cracks in soldering or underfill, something nVidia chips had large problems with some years ago, a problem that could often be solved by baking the card (reflowing the solder).

However, electromigration is your biggest enemy here, and heat cycling plays no role there. Think of electromigration as internal wear. Many people witness this when they notice their maximum stable overclocks reduce over time. With a new cpu they can overclock, say, by 1 GHz, and after a year or two its only 800 Mhz. Thats electromigration. Its irreversible and will over time kill any chip, slowly or suddenly. And it will happen a lot sooner with high temps and voltages, though if you are lucky, that means 10 years instead of  500. If you are less lucky, its 1 month instead of 10 years.

Thats amazing. So why dont you just buy 100.000 BTC on the market and turn it into twice that? Dont tell me you actually still work for a living when you have this secret way to print money?


Sure, but why do you buy your bitcoins for $7 in electricity when you can buy them for $3 on Mt GoX?  You can buy mine for $6 if you want You can still keep them as long as you want if you are in for the long run.


No argument there. If you have fun doing it, who I am I to comment. Just dont kid yourself or anyone else that you are making profits, because you are not. You are losing money hand over fist. Well you would be if you were paying for your electricity, but clearly you arent. Just please refrain from advising others how to mine as "profitably" as you.


Be your own judge. Reread your posts, and tell me you come across like someone who knows what he is doing.

You spent something ike an average month salary on hardware thinking every pence you spent on electricity would turn in to 7 pence worth of bitcoin while in reality its 10x less and it generates even less than what you pay in electricity. When I pointed that out, you revised your estimate to 3 pence worth of bitcoins for every pence spent and only now, after who knows how many months of mining, do you realize your are not even getting once pence worth of bitcoins. Barely half of it.

Knowingly mining at a loss doesnt make one stupid, getting your costs wrong by a factor 10x, well...
Oh, and get a spell checker. saying you are " blatently" not stupid doesn't exactly make your case .

Just be clear: the powerconsumption of your GPU will scale linearly with clockspeed (and exponentially with voltage). So overclocking does very little to improve effciency/w. It mostly helps efficiency/$. But the higher you clock it, the more power it will consume and vice versa. With constant voltage, the power efficiency of your GPU remains the same at the higher clock, in fact, it might even drop a tiny bit because it will run hotter, and hotter gpus draw more power (and require higher fan speed, which is also power, however little). That effect is marginal though, and probably more than offset by the constant power consumption of the CPU, motherboard, ram etc, which becomes a lower % of your power as you increase hash rate of the gpu's.


True that. Moreover, videocards were designed for gaming. So for a few hours per day at medium load. They werent specced to run 24/7 with a power virus.

Tesla's and Firestream cards are designed and specced for that, and no surprise, they are generally clocked much more conservatively  despite insanely high prices...

There is no compulsory exchange rate for bitcoin. Its value floats, as does that fiat currency. Gresham's law applies to gold or silver coins with a face value that doesnt match its intrinsic value, or it applies to currencies with artificially fixed exchange rates that dont match market supply/demand. It doesnt apply to bitcoin or dollars/euro's.

No offense, but clearly you do not know what you are doing. You just posted you mine 1 day per week to cover electricity cost when it turns out your bitcoins dont even cover half the electricity costs. So you dont pay less for them, you are paying almost twice the market rate, just in electricity alone, never mind the hardware costs. Your utility provider loves you for it.

Click the order button and read on:

Performance: Each pre-order unit performs 1.05 Giga Hash / second (plus or minus 10% running variant) for a two step SHA256 process using Bitcoin block mining as a common measure of reference. Power consumption reaches a maximum of 19.8w under full load.

Maybe you should reread that law.

If your card is still covered by warranty for a long time to come (and you think you can get away with voiding it), well, even then probably not. Depends on your luck on how much of an increase the overvolt nets you. Keep in mind it may take weeks before you get an RMA, thats a lot of lost revenue to make up with slightly higher hash rates.

At least if you do want to overvolt, do minor increases and keep your temps low, not 80+C. I killed a 5850 by accidentally overvolting it through the roof and hitting the thermal throttles of 100C GPU and 130VRM. It took about 30 minutes to kill it dead. Heat is bad, overvolt is bad, the combination is stupid.

Are you that thick? No, I bet on http://betsofbitco.in/item?id=141 where you currently get far better odds than the 10-1. In fact odds are in your favor, so cough up those coin or just shut up. BTW, its hysterical how you have moved from wanting to bet $10.000 no matter the odds to haggling over single digit bitcoin amounts now.

i already bet 5btc on a bet i didnt formulate and didnt want to make . waiting for your 50

maybe you should have done that thinking before you spent all that money on hardware? But seriously,  you can not possibly pay your utility bill without noticing

i do understand your overclocking logic better now though; any card that fails saves you money !

LOL. Math isnt your strong suit is it.

3 KW/H * 0.16 pound per Kwh = ~11.5 pound per day in electricity.
That nets you (optimistically) 3.5  BTC or 7 pound per day. But you think you make money for 6 days out of 7?
You are losing 4 pound every day. Well done, overclock them some more, maybe you'll make profit at 1.4v

You were so sure you would bet $10.000 but you dont like odds for betting 10BTC

blah blah blah

Because its time weighted. Early bets weigh much more than late bets.


Of course the assumption is wrong. If you bet 1000 bitcoins, I and others will bet against that, even people who might be skeptical. The final ratio is a good indication of what people actually expect.  So why dont you start with 100 BTC, heck, just 10BTC and see what happens?

The statement is BFL will NOT ship a card that matches their specs in time.  6 BTC  disagree with that, so the majority so far believes they will ship in time and meet specs.

I dont believe that for a second. Some googling suggests electricity in the UK costs ~10p/KWh or $0.16.
Even if you would achieve 2MH per watt (which you wont with those cards at those voltages, youd be lucky to get 1.5-1.7MH/W) that works out around $3 in electricity cost per bitcoin. You are saying it costs you $0.4 per bitcoin in electricity. I call BS on that.

Naysayers put your money where your mouth is.  Bets so far are clearly in favor of BFL not only being legit, but even going to live up to all its claims and in time. Stop saying you are sure its a scam unless you put some coins on it:

http://betsofbitco.in/item?id=141

My bet was rejected because "scam" is apparently a curse word or something  
But someone else started a bet, which should be a lot safer for the naysayers than I what I was willing to bet, since it includes the 8 week deadline and no margin on the specs. Anyone thinking this is a scam can make a few bitcoins by betting this:

http://betsofbitco.in/item?id=141

Goat, Im sure you will put 100 or more BTCs in there, there is no way you could lose that bet.