If you're using resistive heating, a computer is about as efficient as a space heater.

I have not recently done the math, but I'm pretty sure in my area its still more economically efficient to heat my home with natural gas than resistive heat. Obviously, emissions are a concern, future price uncertainty of fossil fuels, etc. Not the best way forward.

However, heat pumps in my area can make a ton of sense. They're far more efficient than resistive heating as long as it doesn't get practically arctic cold outside. When in decent climates they're about as economically efficient than a gas furnace.

In other words, mining crypto for heat isn't necessarily the most economical thing to do. It may just be better for you to use a heat pump and put your savings into buying crypto if ultimately having crypto and being warm is your goal.

I use NatGas to heat my house here in Alaska (in-floor radiant heat, quite efficient). But I also have wood burning stoves and am surrounded by forest. If the price of NatGas increases to the point it becomes unaffordable, trust me when I say my carbon footprint will go up dramatically. NOT heating my house is not an option in the winter here.

If you're suggesting you'd burn trees, that's carbon neutral (very nearly, maybe a tiny bit used to fell and move them.)

> I have not recently done the math

What calculations are these? How do I figure out if it's worth it?

Pretty easy.

Electric heat is 100% efficient. Natural gas is anywhere between 80-97% efficient depending on your model of furnace. Convert therms of gas[0] and kWh[1] to BTUs, and multiply by efficiency and price to determine which is cheapest[2]. Usually natural gas wins.

For emissions it’s more complicated. You need to look up the energy source for your grid[3], it will vary a bit seasonally, and factor in transmission losses for electric heat. You can then calculate the emissions per BTU of natural gas heat vs. the emissions of the grid to give you a BTU of electric heat. For most grids, natural gas still wins.

Where this gets complicated is if you use a heat pump, either a “mini split” (common in Europe and some new apartments) or a ground source (“geothermal”) heat pump. These have a seasonally adjusted efficiency rating[4], usually between 2 and 4. This means per unit of electrical energy the unit is moving 2-4 units of heat (depending on model and exterior conditions). You can take the electric heating number from above and divide by this number to figure out the cost and emissions for a heat pump. Usually a heat pump will win on emissions and lose on price (including installation costs) because natural gas is very very cheap.

0 - A “therm” of gas is the traditional measuring and billing unit of natural gas in the United States, it’s equivalent to roughly 100,000 BTUs. Your bill will probably be in therms, but cubic feet are trivially convertible to BTUs too.

1 - 1 kWh is 3,412 BTUs

2 - You can either compare per BTU costs, or your estimated monthly cost based on your heating need. The latter is important if you have tiered or seasonal electricity pricing. To figure this out take a sample monthly gas bill from winter, subtract your summer gas bill (if you have gas hot water) and use the calculations above to determine your heating needs in BTUs. This lets you calculate your electricity needs in kWh and therefore costs.

3 - It’s important that this is a seasonally correct. Most grids get dirtier as more polluting and less efficient power plants are turned on for summer AC, often coal power plants in most regions. Therefore the emissions per kWh will vary on a month by month basis for a given grid.

4 - If we’re being very precise, a non seasonally adjusted number would be best here. An air sourde Heat pump will struggle to extract heat from very cold exterior air, reducing its efficiency. A ground source one will be more efficient, given that ground water is warmer than the air typically.

Assuming you already have a gas furnace (I do, was in the house when I bought it), and the option is to either use resistive heat (space heaters, computers, inefficient lightbulbs) or the natural gas furnace, you would need to look at two main things. The cost of electricity per BTU and the cost of natural gas per BTU.

On a somewhat recent decently high gas usage month that emulated a cold winter (forgot to turn off the pool heater from spa mode on the whole pool circulation mode for a few days...oof) my $/CCF (Centi-Cubic Foot) was 1.22. Converting $/CCF to $/MMBTU, divide by 1.037 to get ~$1.176/MMBTU.

Electricity is usually billed at kWh pricing. 1MMBTU = ~293kWh. I pay ~$0.086/kWh, so 293 * .086 = $25.198/MMBTU.

Obviously, you're not getting 100% efficiency with the natural gas furnace. I think the one I have in my attic which is almost a decade old is rated at like 85% efficient. 90% efficient furnaces are available, some of them don't even require a metal chimney as the resulting air isn't even that hot.

But even at only 85% efficiency, you're still taking more than 10x less dollars per heat.

This math changes when thinking about buying new equipment and other factors. I haven't looked at equipment pricing so it would be hard for me to say how expensive it would be compared to alternatives. This can also vary a good bit from your local market. Energy is stupid cheap in my area, natural gas is a good bit cheaper here than a good chunk of the US while electricity prices are about median, maybe a bit lower.

Oof, misnamed a unit there. CCF = centa cubic foot. 100, not 1/100.

> Computers also generate heat. Wish I could use cryptocurrency miners for my heating needs. Why burn fuel or pass current through resistances for that?

Realistically? Cost. If your goal is to heat a space, it's much cheaper and more reliable to NOT use thousands of dollars of silicon. Doing calculations based on the site below, you might be able to heat a smallish bedroom with a very beefy PC. I can get a space heater that puts out that much heat for less than $50 and is far less likely to stop working (e.g. heatsink gets clogged and system overheats).

https://www.stelpro.com/en/expert-advices/calculating-the-ri...

I mean, you laugh, but during winter in my office at home whenever I get cold I just start running Folding@Home on my pc, it helps cure cancer and keeps me nice and warm. As far as I can tell it's cheaper than turning the heating on in the whole house too.

I already have an rtx gpu. It nets about 8$ of eth a day. Uses 250w. So that's around 1$ of power a day. And enough wattage to heat my office. So I can get paid to keep my room at a nice temperature. This is why I mine now

> I already have an rtx gpu. It nets about 8$ of eth a day. Uses 250w. So that's around 1$ of power a day. And enough wattage to heat my office. So I can get paid to keep my room at a nice temperature. This is why I mine now

Unless your office is the size of a closet, that doesn't sound like it's enough to provide all of its heating needs (at least according to the formula of the website I linked).

It's certainly practical to provide to exploit heat generated by computers to partially heat a space, but I don't think it's practical to use silicon as the main heat source.

Using adsorption chillers [0] you can use the heat to also cool down your office.

[0] https://www.technologyreview.com/2011/03/30/196054/using-hea...

My heater doesn't become less effective each year, requiring replacement. While ASIC concretely implies this, GPU is not excluded if you want to remain competitive (read: generating profit)

You might say that a regular electric heater is 100% efficient, converting every unit of electricity into exactly one unit of equivalent heat.

Crypto mining hardware is >100% efficient, as it's exactly as efficient as an electric heater, but also generates revenue. Over time, others make more efficient crypto mining hardware, reducing the yield from yours. This causes your "efficiency" to gradually reduce down to 100%, but no lower.

However, nowadays we have heat pumps, which are also more efficient than our 100% benchmark, and produce additional useful heat.

> Crypto mining hardware is >100% efficient

It also makes the hardware depreciate (simply with the passage of time, but also because of the particular usage), which in practice effects the balance and can easily turn the monetary efficiency into negative.

It doesn't have to be competitive.

4+ kWh electric showers are ubiquitous in my country. Literally a dumb piece of metal resistance immersed in water. Anything is better than the current situation. What if it was a processor instead?

What if a bunch of home servers could heat up stored water through their heat instead of just dissipating that energy uselessly into the air?

We'd be getting the heat we need and we'd be helping build a decentralized cash network. We were gonna spend energy heating up stuff anyway. Might as well get some cryptocurrency out of that heat.

This is what WiseMining.io is addressing with their Sato “bitcoin boiler.” It uses hashcards (daughterboards populated with mining ASICs) immersed in coolant to augment a typical heat pump setup.

That's interesting. My search for products in this category turned up the Qarnot company which also has a "digital boiler". Apparently, it's a general purpose Linux server that can be used to mine any coin or perform any type of computation.

https://qarnot.com/en/heating-water/

Not sure if it's even available outside Europe. I wonder if it's possible to somehow replicate this with a home server. How is it pumping the heat out of the components and into the water system?

A datacenter that doubles as district heating seems like a pretty good idea.

Crypto miners are pretty inefficient at heating your home.

A heat pump is more than twice as efficient as direct heat.

Maybe so, but they generate crypto, which heat pumps don’t.

Yup. Also mining rigs produce a lot of noise, are more likely to break down & need an internet connection 24/7. Not your average heater.

Immersion cooling makes it quiet and efficient, totally suitable for in-home use.

Cooling for your heater, genius idea!

Can you elaborate? I've seen products that do this, it must be possible.

Resistive heaters use energy to create heat. Heat pumps use energy to move heat.

For a resistive heater, its taking 1 energy unit from the wall and putting 1 energy unit into the air.

A heat pump will use one unit of energy from the wall to take 3-5 units of energy from one side of the system and dump that 3-5 units of energy to the other side of the system.

I see. Heat pumps must move existing heat from somewhere though. Won't those come from burning fuels?

I have a lot of solar panels in my home. I'm generating more energy than I'm using. I don't consider efficiency to be a significant concern at the moment.

There's plenty of available heat outside, even when its somewhat cold outside. This breaks down when it gets really cold, like arctic kind of cold, but most places where lots of people live don't get that cold for very long. Its the same principle as how even when its hot outside your AC unit can still manage to dump some heat out there. Its because of the latent heat of the refrigerant and the phase change induced by the changes in pressure from the compressor. Its kind of hacking the physics of phase changes to move energy from one place to another.

> I'm generating more energy than I'm using. I don't consider efficiency to be a significant concern at the moment.

I also produce a surplus and I sell it (for more than the local utility charges to boot). More efficiency is more money in my pocket, is it not for you?

> I also produce a surplus and I sell it

> More efficiency is more money in my pocket, is it not for you?

Unfortunately, it isn't. I also inject surplus energy into the local power distribution network. In my country, I don't actually get paid for that. They give me "kWh credits" that I'm supposed to spend whenever I generare less energy than I consume. These credits will expire if unused, I can't accumulate them.

So it is simply not viable for any home or business to generate more power than they consume. I have space for more solar panels but I'd be giving the power companies free energy if I installed them.

Cryprocurrency is the perfect solution for my problem. Instead of giving them my energy surplus, I sink it into my computers instead so they can mine cryptocurrency. If the waste heat could be reused somehow, it would be even better...

It moves existing heat from the outside to the inside of the house. It works exactly like an air conditioner, but in the other direction. (And some heat pumps are reversible so they can be used as both)

You drill very deep shafts and move refrigerant up and down. It’s efficient but capital costs are huge. Usually it does not make sense economically without government regulations.

You don't even need that. Up to circa -20C you can use a regular AC with an inverter and it would still work, albeit at reduced efficiency.

This explains it far better than I can. https://www.popularmechanics.com/home/outdoor-projects/a2586...

There was a video I've seen someone, a year ago, setup a 'pre-heater' for their home heating system with a mining rig, so that the home heater doesn't have to use nearly as much fuel/electricity to warm the home. I forget how efficient it was, but the summary was that the net decrease in cost to run the heater was observable but not significant enough to justify everyone doing so.

That sounds freaking awesome. Do you know how to find it? I found a company called Qarnot that's offering products along those lines but they're probably extremely expensive. A home project would be ideal.

While I don't remember which video it was - I think this is a pretty close estimate to it[0].

[0]https://www.youtube.com/watch?v=uLO73bbIJtU

The premise was to only operate the miner in cold months, when the outside air could easily cool the system. I didn't fully re-watch this video but it does give a good setup and explanation for it.