Any thoughts about updating this and making it a Google doc?
No pressure of course.
Topic: Estimated long-term costs of owning and running different ASICs (Read 6369 times)
August 10, 2013, 10:43:06 PM
Should you upgrade MegaBitPower's 400GH to $8000 now? That's the October delivery price, and August deliver is sold out.
How does multiplying B10*B3 (BTC/Ghs * # to get to 1 TH/s) give you total power draw? Maybe I'm reading your spreadsheet incorrectly or the written formula is just in the wrong spot?
Edit: Nevermind, the B10 should just be B11 and then the math works out just fine. Thanks for the spreadsheet!
Edit: Nevermind, the B10 should just be B11 and then the math works out just fine. Thanks for the spreadsheet!
Some time ago I tried to find a link for the spreadsheet and didn't see it. Now even picture seems gone 
Ya Postimage was giving me issues trying to link the full-rez pic - it would work, and then not. I've reuploaded it to tinypic, so it should work better now. I've also made some changes, including removing the USB sticks, the K64, the BFL Little Single, and I updated the 5GH/s and the 10GH/s ASICMiner blades, including the new prices.
Real men post links to the spreadsheet, not pictures of it.
Some time ago I tried to find a link for the spreadsheet and didn't see it. Now even picture seems gone
Ya Postimage was giving me issues trying to link the full-rez pic - it would work, and then not. I've reuploaded it to tinypic, so it should work better now. I've also made some changes, including removing the USB sticks, the K64, the BFL Little Single, and I updated the 5GH/s and the 10GH/s ASICMiner blades, including the new prices.
Some time ago I tried to find a link for the spreadsheet and didn't see it. Now even picture seems gone
Updated OP chart with the new, "low end" KNC miner, as well as miners with BitFury chips being sold in the US. For me, that's a big deal. Also, can't understand how they got the power numbers they did. BFL thought they'd get 1GHs/W too, but looks like BitFury was actually able to do it.
As the chart gets bigger, I'll prolly take off some of the more useless ones, such as the USB Block Erupters and the K1s. The point is pretty clear that they're terrible for a large scale operation, and that doesn't even include the added cost of hundreds of USB hubs. I'll also prolly take off the BFL Little Single, as it's almost exactly the same numbers as the SC Single. I'd also like to move more towards complete miners, which would give a number that represents the final cost. This means the K16 and K64 units are gone, and resellers of those in complete packages are in (such as TerraHash).
As the chart gets bigger, I'll prolly take off some of the more useless ones, such as the USB Block Erupters and the K1s. The point is pretty clear that they're terrible for a large scale operation, and that doesn't even include the added cost of hundreds of USB hubs. I'll also prolly take off the BFL Little Single, as it's almost exactly the same numbers as the SC Single. I'd also like to move more towards complete miners, which would give a number that represents the final cost. This means the K16 and K64 units are gone, and resellers of those in complete packages are in (such as TerraHash).
Partially I am trying to justify to myself my modest B3 (^/v) and Bitfury (v/^^^) orders.
Take a look above. I really wasn't comfortable putting KNC and Bitfury's options up there, as we don't know what the final numbers are. We still haven't seen any working products from either of them. Specifically with BitFury, I think his power numbers are way off. The numbers are supposedly in the range of ~1W/GH system draw.
There should be plenty of bitfury information in the next couple of weeks as 100TH starts deploying their mine (50- 75TH in phase 1 I believe), and I think it was mentioned that they are not overclocking but instead going for high density boards... I guess they got a good deal on chips
Nice spread sheet. If that is a google doc, you can share it read only and post the link here.
If it isn't, you could import it then share it read only. Might be easier to read for some.
If it isn't, you could import it then share it read only. Might be easier to read for some.
Partially I am trying to justify to myself my modest B3 (^/v) and Bitfury (v/^^^) orders.
Take a look above. I really wasn't comfortable putting KNC and Bitfury's options up there, as we don't know what the final numbers are. We still haven't seen any working products from either of them. Specifically with BitFury, I think his power numbers are way off. TLDR; Most power efficient miners are the best for long haul mining.
You must not have looked at the chart very long. Compare those Block Erupter Blades (ASICMiner), and the TerraHash DX. They get almost the SAME power efficiency: ~133 vs 138 GHs/kW. Yet the initial investment of the ASICMiner Blade is so high that it jacks up the overall cost by a HUGE margin. You really should be looking at line 7, which is the cost for just the hardware.And as I said before, this isn't a profitability calc or a return estimate. This is simply taking what hardware is available today, and putting realistic numbers to the cost it will be to run them.
Correct... as stated... I looked at AM's availability (^) and then their BTC/GHs (vvvv) and then they get taken off my list. Partially I am trying to justify to myself my modest B3 (^/v) and Bitfury (v/^^^) orders.
TLDR; Most power efficient miners are the best for long haul mining.
You must not have looked at the chart very long. Compare those Block Erupter Blades (ASICMiner), and the TerraHash DX. They get almost the SAME power efficiency: ~133 vs 138 GHs/kW. Yet the initial investment of the ASICMiner Blade is so high that it jacks up the overall cost by a HUGE margin. You really should be looking at line 7, which is the cost for just the hardware.And as I said before, this isn't a profitability calc or a return estimate. This is simply taking what hardware is available today, and putting realistic numbers to the cost it will be to run them.
TLDR; Most power efficient miners are the best for long haul mining.
This however is only true if the rate of return is constant... which it isn't.
I don't think people plan on running their equipment longer than it is profitable. The biggest return while difficulty is growing is up front. So.. I would posit that the most important metric is time to market followed by hashrate/btc, screw power efficiency.
1 TH/s isn't going to be all that profitable next year with Tytus' 100TH mine, AsicMiner's 1PHs mine, etc. 100 GH/s today is a different story.
This however is only true if the rate of return is constant... which it isn't.
I don't think people plan on running their equipment longer than it is profitable. The biggest return while difficulty is growing is up front. So.. I would posit that the most important metric is time to market followed by hashrate/btc, screw power efficiency.
1 TH/s isn't going to be all that profitable next year with Tytus' 100TH mine, AsicMiner's 1PHs mine, etc. 100 GH/s today is a different story.
Thank you crazyates.
A lot of work.
Very interesting.
A lot of work.
Very interesting.
Oh I forgot they are much faster, but still ~90 celsius seems a lot
Where did you see 90C? Most of the one's I've seen (online and in person (Jalapeno and 60GH Single)) have been in the 65-75C range. If one's over 80C, it's considered pretty damn hot. 
One thing that surprises me, BFL claims to have 65nm and its amazing that the chip runs hotter than 110nm avalon. Are they overclocking that much to fit in that box? Sounds like a timebomb...
BFL consumes more power (heat) per chip because it's a physically larger chip (56mm2) that does 4GH/s, rather than the tiny little chip (15mm2) that does 300MH/s.If BFL wanted to make a chip that was comparable to an Avalon or ASICMiner chip, it would be 3.5mm2 (yes, less than 2mm x 2mm), have a single engine (rather than the current 16), hash a little higher than 250MH/s, and consume 0.2W. But that's kind of a stupid design for a chip, so obv that wouldn't make any sense.
Oh I forgot they are much faster, but still ~90 celsius seems a lot
One thing that surprises me, BFL claims to have 65nm and its amazing that the chip runs hotter than 110nm avalon. Are they overclocking that much to fit in that box? Sounds like a timebomb...
BFL consumes more power (heat) per chip because it's a physically larger chip (56mm2) that does 4GH/s, rather than the tiny little chip (15mm2) that does 300MH/s.If BFL wanted to make a chip that was comparable to an Avalon or ASICMiner chip, it would be 3.5mm2 (yes, less than 2mm x 2mm), have a single engine (rather than the current 16), hash a little higher than 250MH/s, and consume 0.2W. But that's kind of a stupid design for a chip, so obv that wouldn't make any sense.
One thing that surprises me, BFL claims to have 65nm and its amazing that the chip runs hotter than 110nm avalon. Are they overclocking that much to fit in that box? Sounds like a timebomb...
I'm also going to post another revision, with 2 additions: Jalapenos that are flashed to 8GH/s, and Bitfurys ASICs.
Updated OP with those 2 revisions. How does BitFury plan on getting 3x the power efficiency as BFL at the same 65nm? He's claiming even 50% more efficient than KNC, which is supposed to be using a 28nm process. Weird. I'm also going to post another revision, with 2 additions: Jalapenos that are flashed to 8GH/s, and Bitfurys ASICs.
Updated OP with those 2 revisions. How does BitFury plan on getting 3x the power efficiency as BFL at the same 65nm? He's claiming even 50% more efficient than KNC, which is supposed to be using a 28nm process. Weird. Jump to: