Topic: Idea to go GPU and not FPGA (Read 2760 times)

I don't see how you couldn't come out a head. Farmers that built their windmill here and had professionals hook it up have had their windmill running for over 10 years. My Uncle which owns land out in the country has had his for over 20 years and looking at his bill he averages 1,100KW a month. That is roughly 2,000 a year at .15 cents a KW. He built his system as something to do after retiring and spent under 10k. All he has done over the years is add and replace batteries. With the other hardware it is like buying a PSU. The higher quality and proper sized unit will save you money in the long run.

This is super wrong. You need to have consumption large enough to amortize the cost of the solar system over the life of the system (including service, batteries, and inverters) and still come out cheaper than grid rates. Below a certain threshold, that ain't happening.

LOL. Priceless.

I don't consider the green energy investment part of the cost. Why? Because it would serve as taking me off the power grid after mining is over for good, and going green is more than saving money on your bill. It is or should be about helping the earth. Not to mention depending where you live you can actually get paid for your excess energy production.

I assume in the dead of July you had AC running. 

AC = transfers thermal output of cards to the outside to prevent rise in temps.
Open Window in February  = transfers thermal output of cards to the outside to prevent rise in temps.

Ironically in the winter time the better your house is built/insulated the better it is going to keep that miner waste heat inside and raise the temp of the room (and thus cards). 

FPGA also are less tolerent of thermal stress (heat).  You can get away with a GPU hitting 110C+ for a short period of time.  I wouldn't recommend it but it is unlikely you will kill it.  An FPGA hitting same temps is toast.  

FPGA are precision equipment generally used in industrial or scientific applications by qualified personnel.  GPU are used by complete idiots.  GPU are engineered to be as idiot proof as technology will allow.

Highly unlikely. GPUs are extremely reliable. They have several temperature monitors and will down-clock when they detect overheating. FPGAs? Nada. If you have any cooling issues (fan stops, etc) with an FPGA it'll fry itself before you know it.

36 x 150W = 5400W
And unless you plan to mine only 5 hours/day you'll need at least 3x that solar capacity and batteries. This gets extremely expensive very quickly. The cost of your rigs would be the small part in this setup. Overall this would be much more than the cost of using FPGAs.

Last time I checked 65W panels ran around $400. So a quick rough calculation gives me 250 panels or about $100,000. Actually, when I plug 5400W into an online solar calc it says I should have 25920 W solar capacity! Or about 400 solar panels. So right off,  even without batteries (which are awfully costly and not so green either) I can say without any doubt that I'd prefer to spend the $200,000 on FPGAs here.

You base this on what?  The short warranties or the short period of time any FPGA miners have been in the field.

WARNING WARNING WARNING.  Logical failure in progress.  Someone trying to have his cake and eat it too.

If you get 10x the hashing power at the same electrical load where do you think that energy goes.
Same electrical load = same heat.
Same electrical load = same amount of cfm required = same amount of fans = same amount of noise.

Now you can say you could get the SAME hashing power with
1/10th the electrical load and thus 1/10th the heat and 1/10 the cooling requirements.

Still even that isn't accurate.  Good GPU rig can get 2.6 MH/W.  There are no 26 MH/W FPGA boards (yet).  ztek is close w/ 22 MH/W but the payoff period isn't 5 months and unless you buy a lot it isn't 10 months too.

Solar and/or wind power is not free.  I don't know why people think it is.

Power is power.

If solar power costs less than your grid power then it is economically viable alternative form of power however for most people it isn't cheaper so you are simply paying more to get the same power.

Yeah. Considering that FPGA are less likely to die than your GPUs and you got a clear winner. FPGA FTW !

There's other reasons NOT to go green. Besides the fact that it is way more expensive to get the green energy in the first place, and that it has an undetermined ROI, you should also consider that FPGAs have a HUGE benefit when compared to GPUs:
HEAT!

While going green sounds nice for smaller arrays, I have to keep the windows open for my 5 GH/s rig(this is january, -5° C), otherwise 2 of the cards hit a 100C. Besides, GPUs make lots of noise, which is another excellent reason to switch to FPGAs. My point is, that you can easily deploy FPGA's with about 10x the hashing power at the same electrical cost, without most of the heat, without almost all of the noise(since FPGAs can be cooled with lower-rpm fans that do make less noise), and at a ROI of 5-10 months. after that point you make decent profits.

You have to be pretty aggressive to get a ROI on a solar panel or some wind power generator in less than a year, and green energy, that isnt that green btw(considering that the CO2 bilance for producing a solar panel alone exceeds the CO2 balance for creating that same energy with coal or gas), doesnt solve the essential problems of mining and living in coexistance: Heat, maintenance and noise.

I have to admit though, that we did never turn on even one of our heaters since we installed the last 2 GH/s.

For me it is because FPGA route will not give me any fun.

1. I know a little bit about FPGA, but it is not enough for actively engaging a FPGA mining project. So, with FPGA route, basically I just bought some black boxes and then connect power supply to the black boxes and then run the miners. Period. Do you think it is fun?

2. Upfront cost is too high. I just don't have such a big deal of money to burn.

3. With GPU route I can play with them for many hours! I build the frame (metalwork is my hobby), setup the rig, playing around with the settings, overclock, underclock, write control scripts...Countless activities.

4. It is not risky, I can sell my GPU at a good price

As long as the numbers add up...

This is like some miners considering their electricity "free" since the parents/school/landlord pays. You can question the morality of it (as all subsidies, it's picking from somebody else's pockets) but you cannot question it's a sound financial plan   as long as the numbers add up.

If going for solar power makes financial sense for mining, it makes financial sense without mining and you should do it anyway.
The electricity you consume for mining is no different that for any other use, so you should separate the issues.

That seems like a mighty big if though.  And I certainly cant believe it would pay itself back + your mining hardware in less than a year. No way. Unless of course you ignore the solar investment and wrongly assume your solar electricity to be free for mining.

I been looking at FPGA mining because of the cost of electricity and Dec 2012. Well I figure why not just keep going the GPU route? Why not power your mining rigs with wind/solar power? Sure there is a nice little start up cost; however, going green for your home means you qualify for tax credit. So for us with a set goal in mind for our desired hashing power wind/solar makes sense. Instead of investing in FPGA the offset cost between GPU and FPGA can be spent on going green. So you get same energy cost if you were running FPGA and you get the cheap hashing power of GPU.

So assuming 36 6870 for 10Gh/s
11,500 cost of PC hardware (not solar/wind hardware)
5 dollars a coins

By Dec 2012 You could very well be in profit assuming you sale coins above 5 dollars. Also after you're ready to stop mining or bitcoin goes under you still have green energy to keep your bill down, PCs, GPUs that can be sold, no trying to find a use for your FPGAs, and the warm feeling of going green.