Capital cost.

[rograz]

I don't get how some of you can build such inefficient rigs, especially those of you in the US with your hardware prices :/ My absolutely worst rig is in the 0.9$/MH and I have to deal with the hardware prices in sweden/europe (25% VAT). When the 5850 Xtreme was launched here for example they were in the 175-200$ area and I've yet to see the 5830 Xtreme below 140$. People just seem to be spending way to much on the rest of the hardware rather than on cards where it matters.

[syb3ria]

$130 for single 5770. I should get some more since i don't seek mh/j efficiency.

[V4Vendettas]

Woot messing about in trixx I just made it to about $0.92 
Well happy with this its my first one \o/

Once those PCI risers make it over from Hong Kong I should beable to address some heat issuse's and pust it down even more.

[Reckman]

3200$ vs 4020 mhash, 12 5850s, 3 systems with risers

[rograz]

$130 for single 5770. I should get some more since i don't seek mh/j efficiency.

5770 is a horrible $/MH card since each pci-e slot comes with a cost as well, unless you are just filling up slots in your current systems then I would agree it's a decent deal. Also 5770 is actually a pretty decent MH/W card once you downclock the memory.

[SlipperySlope]

When I built my rigs the 5770 was the best card available - as 5830s were gone.

[CydeWeys]

I've spent $1660 on ~2370 MHash/s, or about 70 cents per MHash/s.  Not too shabby, especially in comparison to some of the other numbers on here!  It helps that $310 of the $1660 was to purchase two 5850s that went into otherwise already-bought-and-paid-for computers.  I only have a rig and a half that was purchased specifically for mining.

[KarlSpaat]

im at 12 5850 with riser cards on 3 MBs. Spent 3250 $ for about 4,5 GHash/s. These configs need 2,1 KW and my kitchen became pretty warm

[1bitc0inplz]

When I built my rigs the 5770 was the best card available - as 5830s were gone.

Not a bad choice considering the 5830s weren't available.

[syb3ria]

$130 for single 5770. I should get some more since i don't seek mh/j efficiency.

5770 is a horrible $/MH card since each pci-e slot comes with a cost as well, unless you are just filling up slots in your current systems then I would agree it's a decent deal. Also 5770 is actually a pretty decent MH/W card once you downclock the memory.

Yep, filling up slots on existing systems that run 24/7. And i plan putting 2x5830 or 5830 + 5770. And im not after MH/W

[evlew]

I'm roughly around $3000 for 3600mhps = $0.83/mhps or 1.2mhps/$

first rig was 4x 5830's at 109 + free ship (wish i had to balls at the time to go all out and just get 12)
Then after cashing out my first coin I ordered 2 more rigs of 4x 6870's.

MSI 890FX GD70's
900watt & 1000watt PSU's

[jasonk]

I already had 1 desktop, so that kept my costs down.

Capital: ~$1100

M hashes: ~2000

One rig already had 1x 5830, just added a 2nd 5830 card @ $90.  Does about 580M hashes.

My 2nd rig I built from scratch.  It has 2x 5850's and 1x 5830 running off PCI-E 1x.  That rig cost me $600.  Does about 900M hashes

My 3rd rig my friend and I went 50/50 on.  It has 4x 5830's.  Cost us about $800, but we split the cost and hash rate.  So thats +$400 for about 550M hashes

In 2 months I paid for all hardware investment and put $600 in my pocket.

[stapler117]

Between me and my friends, we had 2 unused motherboards with 3 pcie slots total. All we needed were better power supplies and gpus. We put one of them in my friend's desktop as an upgrade to his older graphics card so it's not running as fast as the other 3, but it brings our total up to 4 5830's.

$650 invested
1000 Mhash

1.54 MH/$ or $0.65/MH

^^proud of that.

[grod]

3x sapphire 5830, $129 each = 387
3x pcie risers, $18
Any board with 3 PCIe slots of any sort, cheapest CPU - $60
Antec 750 watt gamer series, $55 (+2) = 57 AR
1 Gb USB memory stick, $1

Total: $533, should pull nearly 900 Mhash and pull about 600 watts out of the wall.  All of this may still be possible today, was definitely possible last week on newegg.  Even without rebates there are plenty of 700-800 watt PSUs capable of feeding 3 overclocked 5830s under $100, so call it $600 for 900 Mhash or 1.5 Mhash/$ or .66 $/Mhash.

Assuming fixed difficulty this gets nearly half a bitcoin a day.  Let's call it $6 after power costs, or 180 a month.  3 and change month ROI.

OP's builds require more than 6 months of no difficulty increases and $15/BTC exchange rates.  Good luck with that.  I'd send that crap back and eat the restocking fee if I was him.

[Jack of Diamonds]

I only got into mining because I got bulk 5830's, 5850's and 6990's cheap (-30-40% under market price).

For example, 8x 5830's for $792 (2400mhash/s). 5x 6990 for $2600, (4000mhash/s). Bunch of sapphire 5850's xtremes at $99 each.  There's of course PSU and mobo costs too, but IMO if you can get great deals, mining is still worth it.

Don't just go to Newegg and buy the most expensive thing. Look around, make phone calls and see if importers have some cheap ones.

[grod]

Here's another way to look at the OP's investment.  He could have spent $8000 on bitcoins.  I'm sure they'll be $13 again this weekend, so I'll use that as the price.

8000/13 = 615

At 7 gigahash the OP should be mining around 3.5 bitcoins a day, at a power cost of at least .5 btc/day.  Now assuming no difficulty increases we're talking 200 days to earn that many BTC.

Look at prices and difficulties right around the end of last year (200 days ago) to guestimate how likely it is your hardware will *ever* generate anywhere near 600 bitcoins over its lifetime.  You've just bet 8000 dollars no ASIC of FPGA miners get introduced until well into next year, and the 7 series radeons won't be very good at mining -- both very poor bets.

[mike678]

Here's another way to look at the OP's investment.  He could have spent $8000 on bitcoins.  I'm sure they'll be $13 again this weekend, so I'll use that as the price.

8000/13 = 615

At 7 gigahash the OP should be mining around 3.5 bitcoins a day, at a power cost of at least .5 btc/day.  Now assuming no difficulty increases we're talking 200 days to earn that many BTC.

Look at prices and difficulties right around the end of last year (200 days ago) to guestimate how likely it is your hardware will *ever* generate anywhere near 600 bitcoins over its lifetime.  You've just bet 8000 dollars no ASIC of FPGA miners get introduced until well into next year, and the 7 series radeons won't be very good at mining -- both very poor bets.

the op made a sloppy investment. You should be able to get at least 1.3 megahash per dollar easily. He gets .875
$8000*1.3=10.4 gigahash

Under current difficulty that's 6.693 BTC I assume that .5 btc is still acceptable because some of the slower more expensive cards consume more power.

So 615/6.193 a day =99 days

So in about 3 months he will have the same amount after that he will continue to gain more bitcoins then the original possible investment.

edit: If for example he went with the most extreme case he could get 1.58 megahash per total cost with 5830's at 110.
$8000*1.58=12.64 gigahash
thats 8.134-.5=7.634

615/7.634= 80 days

[grod]

Oh, I agree an efficient builder can still make the case for mining.  The OP is far from an efficient builder.  In fact, it'd be hard to build less efficiently unless you buy 6850s at ebay markups to go with $250 blinged out PSUs, $200 gamer cases and $400 gamer motherboards in addition to getting hard drives and windows licenses.  The OP is a classic case for buying efficiently mined BTCs instead of overpriced mining gear as a BTC investment.

Also, your 3 month timeframe for an efficient builder (see: my post above) ignores the possibility of hardware failure, brownouts, downtime due to DoSed pools, network outage, and a wildly spiking difficulty.

[mike678]

Oh, I agree an efficient builder can still make the case for mining.  The OP is far from an efficient builder.  In fact, it'd be hard to build less efficiently unless you buy 6850s at ebay markups to go with $250 blinged out PSUs, $200 gamer cases and $400 gamer motherboards in addition to getting hard drives and windows licenses.  The OP is a classic case for buying efficiently mined BTCs instead of overpriced mining gear as a BTC investment.

Also, your 3 month timeframe for an efficient builder (see: my post above) ignores the possibility of hardware failure, brownouts, downtime due to DoSed pools, network outage, and a wildly spiking difficulty.

True I'm sure as you gain more hardware there will be more failures but there isn't really a way to calculate that. Also I doubt all pools will be ddosed at the same time and worse case at that power he could still mine independently. Also the difficulty shouldn't be unpredictable as long as the price doesn't go sky high.

[muyoso]

Here's another way to look at the OP's investment.  He could have spent $8000 on bitcoins.  I'm sure they'll be $13 again this weekend, so I'll use that as the price.

8000/13 = 615

At 7 gigahash the OP should be mining around 3.5 bitcoins a day, at a power cost of at least .5 btc/day.  Now assuming no difficulty increases we're talking 200 days to earn that many BTC.

Look at prices and difficulties right around the end of last year (200 days ago) to guestimate how likely it is your hardware will *ever* generate anywhere near 600 bitcoins over its lifetime.  You've just bet 8000 dollars no ASIC of FPGA miners get introduced until well into next year, and the 7 series radeons won't be very good at mining -- both very poor bets.

Not to mention that he could sell those 615 bitcoins for 16 dollars within 1-3 days of buying them at 13 making him $1840 within a week.  He could probably do this multiple times.  If he bought again using all $9840 at 13 he would have 757 coins and could cash out at $12112 if it hit 16 again.  Thus making more than he could ever hope to mine in the matter of 1-2 weeks.