Get the top HN stories in your inbox every day.
xenocyon
drran
Definition of efficiency (from Google): the ratio of the useful work performed by a machine or in a process to the total energy expended or heat taken in.
Heat pumps are >100% efficient.
Usually, we cannot convert all input energy into useful work, so efficiency is always <100%, because of energy losses, which are producing waste heat. However, in case of heater, heat is the "useful work", so electrical heaters have "impossible" 100% efficiency, while heat pumps have even more "impossible" >100% efficiency.
piracy1
I consider personal enrichment as a sort of useful work as it stands to replace my own 'useful work'
coldtea
Perhaps parent meant useful with the caveat that "it also wont fuck up the climate, and assist in huge destruction and killing millions in the process of making me rich".
Which is so selfish of him. Why should the safety of billions matter over me becoming rich? /s
OJFord
> or heat taken in.
drran
Ambient heat is not a "heat taken in", because you don't need to supply it. Usually, heat is taken out from the system and dissipate as ambient heat, so a heat pump can use the same heat again and again, without exhausting of ambient heat.
fomine3
Totally agreed but I wonder is it applied to extremely cold situation. Possibly outside unit produce heat but not all heat supplied to room.
seniorsassycat
A heat pump in a −273 °C space will convert electricity to heat at a 1:1 ratio.
If you count the heat taken from the space around the condenser, which you should based on Google's definition, heat pumps are only 100% efficient.
Calling heat pumps >100% efficient is like calling my solar water heater >100% efficient. True if you only count the electrical input, and not other sources of energy.
CobrastanJorji
You will be pleased to know that the majority of heat pumps are installed in homes whose exteriors stay above 0 kelvin year round. Also, in what I assume is an efficiency feature, most heat pumps do not allow operation in a zero Kelvin environment.
Also, to nit pick: Kelvins are not degrees and do not use a ° sign.
stopping
This is incorrect. We measure the "efficiency" of a heat pump by the amount of heat moved into a space divided by the work energy provided to the pump (electricity in this case). This is called the "coefficient of performance". Typical heat pumps have a CoP of 3 or more. Meaning, you can spend 100W of electricity to pump 300W of heat into your home. A purely resistive heater only has a CoP of 1.
yarcob
OP is in Austria where it often gets below freezing.
Air/air heat pumps and ventilation heat exchangers tend to freeze up when it gets too cold. So heat pumps often need to run in reverse (to melt the ice on the condenser) and heat exchangers need to pre-heat air with resistive heating to prevent ice build up.
If he can pre-heat the air with a crypto miner that would be better than either of the above.
_0w8t
Consumer-grade heat pumps do not approach the theoretical thermodynamic efficiency limit especially when the outside temperature is significantly below freezing temperature as that complicates design. Plus they do not last forever and it takes a lot of energy to make them. So a hybrid design with an electrical pre-heater for really cold weather can be the least consuming in the total balance of energy. But then one can just as well mine crypto currencies in the preheater just as the article described.
ramraj07
Looks like even OP with their fancy own home didn't actually install a heat pump the way which will make it more than 100% efficient, what about the large horde of people who rent and can't do a thing about how their house is heated? It's trivial to find out if your heating system is actually efficient and if it's not why shouldnt you run crypto mining?
bigiain
> didn't actually install a heat pump the way which will make it more than 100% efficient
Or, more likely, they installed a heat pump of a suitable capacity for 80+% of the expected conditions, with a pragmatic "old school" pre-heater to make up when conditions are out near the end of the bell curve.
> why shouldn't you run crypto mining?
I think the key takeaway here is that any time you're using resistive electrical heating for air, crypto mining is a suitable alternative to just making resistive nichrome wire glow red. Whether the cost of 1000W worth of GPU cards compared to the cost of a cheap 1000W fan/column heater makes sense for you is a different question. Whether the payout in whatever crypto you're mining ever comes close to breakeven on the cost of the mining rig and it's depreciation as it becomes more and more out of date is another question. If, like the article's author, you can build an effectively zero cost mining rig from parts on hand, then it's almost certainly worthwhile (modulo what you value your time building/configuring/maintaining it...)
(And there's the deeper ethical question of whether participating in a Proof Of Work cryptocurrency at all is just outrageously leveraging your personal heating requirements into a planet burning speculative ponzi scheme "asset" trashfire...)
the_jeremy
I am renting a house and can't do a thing about how my house is heated. How do I find out if my heating system is efficient?
lalaland1125
An alternative to using heat pumps is to use gas heating. Gas heating lets you achieve >100% "efficiency" relative to electric heating because you don't have to pay the price of converting the gas into electricity.
bigiain
This section of this book:
https://www.withouthotair.com/c21/page_150.shtml
Has some calculations that show that using gas to generate electricity (in a modern gas fired power station) which you then use to run heat pumps for heating - is overall more efficient than just using the gas for heat given real world numbers. (It does go on to show the limitations of how much ground-source heat pumping you can do without freezing the ground - his ballpark numbers suggest maybe only 25% of the heating requirements for a typical British suburb could be met with just ground-sourced heat pumps before you'd freeze the available ground...)
(That whole chapter starting back on page 140, and the whole book - are well worth reading in my opinion.)
_0w8t
There is another non-trivial advantage of natural gas. Energy losses from sending it over pipes are smaller that electricity losses in transmission lines, https://www.eng-tips.com/viewthread.cfm?qid=76881
perl4ever
People sure talk a lot about heat pumps, but from what I read, they're best suited to around 10 C, which in my climate is a brief interval between the seasons changing and hardly requires heat at all. It may be two or three times as efficient, in a sense, but if it's not usable all winter, it can't save much energy, can it?
m463
I think that applies to air source heat pumps, but if you have another medium that might be better.
Sohcahtoa82
I did this in an apartment I lived in years ago, and do it now in my house.
My computer room in the house is the lowest point in the house, so naturally, it tends to be a couple degrees cooler than the upstairs. This means that if I set the thermostat so that the computer room is comfortable, the bedroom is too warm. If I run the house circulation fan constantly, it's not an issue, but that consumes a decent amount of power.
If I'm going to consume that much power, I might as well mine crypto and make a few dollars. Mining for NiceHash, a service that lets people rent hashing power from miners, I'll average $100-400/month worth of Bitcoin on my RTX 3080, depending on current hashing prices, and how much of my time I spend gaming, since mining is effectively paused while gaming. The mining happens while I use my computer with no noticeable loss in performance, and it puts a few watts of energy into the room to heat it up a bit, evening it out with the rest of the house.
gkfasdfasdf
I would really like to try NiceHash, but the fact that I have to add it as an exception to Windows Defender antivirus, and that one of the founders has previously been convicted [0] and served time for writing botnet malware, really gives me pause.
Sohcahtoa82
Windows Defender flags all crypto miners. It is not unique to NiceHash. You don't have to add an exception for the whole NiceHash directory for it to work, but it makes it easier. NiceHash updates the miners occasionally, and if you don't add an exception for the whole directory, then every time a miner gets an update, it's disabled by Windows Defender until you manually allow it.
gkfasdfasdf
Seems like there should be a way to have NiceHash sign their binaries or otherwise vet with Microsoft so Defender can automatically allow it. Setting an exception seems like it makes it a target for other malware to hijack or piggyback on.
kiddico
I'd bet that writing botnet software was pretty good training for writing distributed processing software like NiceHash, though I think your concern is still valid.
gkfasdfasdf
Yes absolutely, I'm sure writing a botnet is not trivial so I respect the technical ability!
mtone
Nicehash (the app) is a convenience tool to manage miner executables -- it's mostly those that are flagged by AVs. You can connect most miner executables to Nicehash (the service) or other pools directly, and some are open-source.
For instance I GPU mine with https://github.com/ethereum-mining/ethminer and CPU mine with https://github.com/xmrig/xmrig, both built from source and I don't use Nicehash (the app). But I haven't looked at the source, so at some point I'm still trusting someone.
gkfasdfasdf
This is interesting, so I could run the NiceHash app (which doesn't require an AV exception) and have it mine using open source miners which I have built myself?
lostmsu
Self-ad: check mine (which is also conveniently called Mine). https://losttech.software/Downloads/Mine/
I am considering to add a feature, that would pause when external sensor reports high temperature. But you can already hack around that by creating a background window titled "DON'T MINE" and setting the tool to stop when it sees it.
lunatuna
I'll take a bit of your NiceHash and add some water [0] to that and make you something tasty. A bit more turnkey and a middleman if you don't mind.
xienze
NiceHash is on GitHub and you could always build it yourself if you’re worried. I think the reason it flags as malware is because virus scanners flag anything crypto mining-related as malware by default.
bseidensticker
You're not actually mining Bitcoin with your GPU. I think nicehash automatically picks the most profitable coin to mine at the time and just pays out in bitcoin.
It hasn't been profitable to mine bitcoin with a GPU since 2013 or 2014.
randomhodler84
No, nicehash pays you in bitcoin for the specific hash rate that you sell. It’s a hash marketplace denoted in BTC. Nicehash doesn’t mine, it’s customers do.
bseidensticker
Sure, but nobody is buying time on your GPU to mine Bitcoin. They are buying time on your GPU to mine some other cryptocurrency. One that uses a different hashing algorithm.
qvrjuec
Aren't you concerned with hardware degradation with your GPU running at 100% most of the time?
barkingcat
This is usually a non-concern. It's likely that the GPU will become obsolete in the time that it fails due to being run at 100% of the time.
The other question to ask, if you compile programs, write webpages, and edit photoshop images: are you concerned with hardware degradation with your CPU if you run at 100% all time time, for example, compiling chrome that takes 3+ hours, applying photoshop filters to an image to get a production quality output, or for rendering a 3 hour long animated movie?
Do you worry about the GPU when you are retraining a GPT-2 or GPT-3 scale AI framework?
These questions are kind of ... dependent on each person. Presumably you buy a computer in order to perform computations. That is its purpose.
Just because it's crunching bitcoins/cryptocurrencies doesn't degrade it any more than ... say running an AI framework on a GPU for 3 months continuously to generate a self driving model ...
balls187
The flipside, if you have a CPU and you can't keep it pegged 100% of the time, are you wasting resources?
Sohcahtoa82
No.
My GPU hovers at around 55-60 C while mining with the fans at 77%.
Years ago, I was running an AMD R9 290. THAT ran HOT. Even with fans at 100%, it would hover around 95 C and constantly fight with thermal throttling.
EDIT: And as someone else said, at these mining rates, it will pay for itself in 3-4 months.
cwkoss
At those rates, the GPU will likely pay for itself before it fails.
coralreef
The only thing that really degrades is fans.
Running a GPU for ~3 years at 75-82c
Bayart
The main causes of degradation in electronics are usually material failure due to thermal stress (cycles of expansion and contractor weaken the electrical pathways, resistance goes up, that causes more stress etc.) and electromigration due to over-voltage (the flux of electrons induces a shift in the position of the atoms it goes through).
Considering mining is a constant load (not much thermal stress) and improving efficiency requires lowering the voltage those aren't really concerns. There's far less risk than overclocking a GPU for gaming. The only electrical part that may be of concern would be the voltage regulation, but that's still an outlier as long as cooling is adequate.
Fans are know to fail because they're ran at high speeds all the time, but they're a commodity.
tullianus
Electromigration can happen because of constant high temperatures as well!
xienze
I was quite surprised by this, but when running NiceHash on my 2070S with no overclocking etc. it stays at ~60C, fans aren’t even audible, etc. And it generates the equivalent of $5 of BTC every day. Can’t complain.
polyterative
just limit power to maximum efficiency. My card mines eth 52MH/s@100% and 49MH/s@55% power
alasdair_
I have a 3090 and cheap-ish electricity. Is it actually worth doing GPU mining again? I thought that died out long ago.
deepakhj
The 3090 runs super hot. You probably want to look into a heatsink/fan for your vram.
https://www.reddit.com/r/EtherMining/comments/loxagu/improve...
https://www.tomshardware.com/how-to/optimize-your-gpu-for-et...
bigiain
Not for Bitcoin.
Other crypto currencies can work better though.
https://www.nicehash.com/profitability-calculator/nvidia-rtx...
Claims you can earn ~$17 per day if you can get electricity to 10c/kWhr running a single RTX3090 but that's all mining Etherium (DaggerHashimoto) not BTC.
abiogenesis
Nvidia even offers dedicated mining GPUs: https://www.nvidia.com/en-us/cmp/
In related news: https://news.ycombinator.com/item?id=26192201
bseidensticker
Mining bitcoin no, but other cryptocurrencies yes.
Sounds like with the run crypto is on right now, you can make some pretty decent money.
tablespoon
> If I'm going to consume that much power, I might as well mine crypto and make a few dollars. Mining for NiceHash, a service that lets people rent hashing power from miners
What's the benefit of this vs. just participating in a mining pool?
mschout
I'm not sure I get it either. Maybe just simplicity / path of least resistance in that NiceHash provides a quick way to get up and running potentially without needing to install mining software and decide on a pool to participate in etc?
bolasanibk
https://www.nicehash.com/ got the HN hug of death!
lupire
What is the $ cost of the power?
Sohcahtoa82
About $20-25/month.
But it's not as if the cost of power is wasted. It's generating heat which I desire. If my GPU is consuming 300W, then that's 300W of heat that my heater doesn't have to generate.
ericd
Electric resistive heating is the worst form of heating, though - heat pumps are something like 3-4x the heat per kwh used. So not totally wasted, but a much worse way to get heat into your house, if that's your goal.
salawat
You're still maximizing power generation demand. That makes environmental issues worse even if it does make fiscal sense. You aren't "using power tgat would go to waste anyway".
All power is JIT. It makes a difference.
srcmap
Is there any cost analyze for using solar panel power for Crypto?
I am installing Solar from Tesla, wonder if I should just use the external power generated for crypto than to send them back to PG&E.
sgc
Since you can make money mining crypto paying for power, and the power company pays you less than they charge, or course using it for crypto makes financial sense.
dreen
Years ago when mining with a single gaming card was more common and viable, I lived in a rented room in a house owned by a live-in landlord who controlled the heating, and wouldn't turn it on for very long in the winter. Since electricity was included in my (fixed) rent, I was able to keep my room a few degrees higher than the rest of the house.
However I must point out that you shouldn't do this in most circumstances, there are environmentally friendly ways of heating your house.
arnaudsm
In countries like France [1], electricity heating is the most environmentally friendly, at 40 g Co2/kWh, while gas (the cheapest and most common) is >250.
hokkos
If the average carbon intensity of french grid is in the 40g, if you take into account when heating is used it is more in the 80gCO2/kWh, anyway a heat pump is always better.
http://www.carbone4.com/wp-content/uploads/2020/06/Publicati...
hnuser123456
France's electricity is largely generated by nuclear power which has minimal CO2 emissions.
adrianN
Even in France a heat pump will give you at least twice the heat for the gram of CO2.
arnaudsm
Heat pumps need energy to operate. Hence at low temperatures the best combo is electricity+heat-pump. So cryptomining+heat-pump is exactly the same, minus the e-waste.
legulere
At least in Germany you can also buy electricity that is 100% renewable at 0 g CO2/kWh. It also does not really cost much more (though we generally have pretty high electricity prices because of taxes on electricity).
Aachen
They don't run an extra cable though, they just buy credits and if you're lucky the profit doesn't go to the parent fossil energy company. It's still better than old energy firms but it's not as if the coal plants run any less hard when you turn on your bitcoin farm with a green contract instead of a fossil one. It's just a matter of where profits flow, and also some public perception. (E.g. Climeworks says that individual customers help show investors that that's a market for it, which helps them do more good than purely the CO2 you pay them to remove.)
Regarding this "fossil parent" thing btw, list of checked companies: https://www.robinwood.de/oekostromreport (I'm with Green City Power because they don't only go the easy hydro route but also build out solar -- iirc, it has been a while). Switching is a matter of signing up with the new provider. Nobody needs to come by, power doesn't go out at all, you just tell them what the meter says on date X and all is good.
tgb
I know that's what you're paying for, but I wonder what the marginal effects are. I.e. am I paying for a 100% renewable energy that otherwise would have been sold to a customer not on a renewable plan? If I use more of the 100% renewable energy, does that increase the amount of non-renewable energy that must be generated for other people's consumption? I have no idea how grids work.
berdario
I'm in a similar situation, and I'm just using an electric heater... No fancy mining setup.
But I'm also thinking, maybe actually this is for the better? (Not only for the stingy landlord, but for the environment).
I mean, even if a heat pump is much better than an electric heater... Having to heat up only the rooms that need it, instead of kitchen, bathrooms, living room (plus the rooms of other flatmates who might not be bothered) might actually use up less energy total, than using a more efficient mechanism, which otoh would be used for the whole home.
cjrp
I'm not sure if they're common where you are, but this is what Thermostatic Radiator Valves are used for. You can set an approximate room temperature, and the valve will open/close based on that. There are also smart ones now which can have a schedule applied to them, e.g. don't heat my bedroom during the day.
dreen
Despite my precarious housing situation at the time, I also own a big house in the country over 150 years old. The main form of heating it originally were 2m tall ceramic furnaces, always placed so that they could heat two rooms at the time. You put material to burn in from either side, and after getting hot the furnace would keep heating the room for hours, and all it takes is a few logs of wood (that you later have to regrow anyway). I wonder how many bitcoins is that in term of environmental impact.
jiofih
0.5kg of CO2 per kWh is the average for electric grids around the world.
1kg of wood produces around 2kg of CO2, every 400g of firewood gives you 1 kWh, so 0.8kg CO2/kWh.
That’s slightly worse than mining. Bitcoin is wasteful in its total energy consumption which is not reused, but 99% of the power you put into that GPU will be released as heat.
adrianmonk
The concept makes sense, but I'm not sure the numbers work out. Heat pumps are pretty efficient, something like 300 to 400%. Are you really going to use less than 25 to 33% of your living area and also not provide any heat at all to the other areas?
Also, with forced air heat pumps, you can usually adjust the registers (air vents) to reduce the flow of air to unused rooms.
dmos62
> However I must point out that you shouldn't do this in most circumstances, there are environmentally friendly ways of heating your house.
What aspect do you think could be friendlier to the environment? Rare metals and production emissions in high tech electronics? Or the fact that it's using electricity (as opposed to more isolation)?
IgorPartola
Your electricity likely comes from burning coal. Coal pollutes quite badly.
Other ways to heat a dwelling are more efficient. For example natural gas heaters burn cleaner and energy losses are much smaller compared to all the conversion losses (by the time you produce a BTU of heat from electricity, several BTUs of energy have been lost due to conversion losses, while burning gas is much closer to perfect efficiency though obviously not perfectly efficient). Or if you can use co-generation, you essentially produce no noticeable environmental impact at all. Or you could go the other route and invest in insulation or a molten salt wall for passive heating.
dmos62
But electricity isn't inherently dirty, right? It's probably better to have the network shift to greener sources than have every consumer do that individually.
dreen
Apart from what others mentioned, proper insulation is the best way to keep your house warm, it is missing from many houses.
celticninja
I mean the OP found an environmentally friendly and economic way of heating his house.
undefined
undefined
cgufus
I'm not sure I understand the principle.
A heat pump achieves a COP (coefficient of performance) of approx. 3-4 (e.g. by investing 1 kWh of electricity, a heat pump generates 3-4 kWh of heat by extracting from the surroundings, air or sole, 2-3 kWh). In this example, by pre-heating the air, you supply the heat pump with ~0.9 kWh of thermal energy (the miner will convert 900 watt directly to heat I would assume). So instead of 1 kWh of electricity consumption from the heatpump, you have 1 kWh of electricity for the heat pump, plus 0.9 kWh for the mining, and you end up with a bit more than 3-4 kWh (since the COP of a heat pump increases if the source temperature is higher).
So in a nuthsell: before: in 1 kWh, out 3-4 kWh after: in 1.9 kWh, out 3.5 - 4.5 kWh
so you lose 0.4 kWh?
shoo_pl
On principle, yes. There are quirks - the COP is different for lower temperatures, and becomes 1:1 at around 5F (-15C). So preheating the air could improve the effectiveness at low temperatures, and I am guessing it might look like this:
- before: 1kWh in, 1kWh out - after: 1.9kWh in, > 1.9kWh out
However, if it was that simple, I suppose heat pump manufactures would include a pre-heating as built-in feature.
It's very likely that he reduced the energy consumption of heat pump by 50%, but at the same time he uses more than those 50% for mining and has a negative total result that is being offset by the profit from mining itself. Which probably nice for him, but not really for environment :)
throw0101a
> There are quirks - the COP is different for lower temperatures, and becomes 1:1 at around 5F (-15C).
It depends on the unit. Some (Mitsubishi FE12NA) have a COP of 1.75 even at -10F / -20C. See Table 6:
cgufus
The used coolant mainly infleunces these numbers.
check chart on: https://www.researchgate.net/figure/Coefficient-of-performan...
bdcs
IIRC my thermodynamics classes correctly, the heater would be optimally placed on the hot effluent out of the heat exchanger (HE) going into the house. This is because the COP is improved (similarly to heating the cold side) because the hot side of the HE doesn't need to be as hot to get to the same T, but also the HE doesn't need to move the heat through it, increasing efficiency. (COP decreases with increasing heat flux [Q] in practice.) For well-mixed air in a house (a poor assumption), this is the same as throwing the miners in a closet. I would suggest to the author to move the miners to the hot side the HE going into the house's rooms. Simulation or measurements (over the course of a week, not just instantaneous measurements) would be helpful here .
If I were a HVAC company with WiFi thermostats, I would look into including miners in heating solutions.
elgaard
Heat pumps have thermostats so they do not run at max all the time.
So if the heat pump without the miner use 1 kW to heat to room temperature it would need less than 1 kW with the miner to heat to the same room temperature. Plus as you say the COP increases.
So it could be that 5 degC to 22 degC requires 3 kW but the heat pump can do it with 1 kW (COP=3).
The miner use 0.9 kW to heat the the outside air to 11 degC. It would take 2.1 Kw to heat from that to 22 degC. But the heat pump now has a COP of 4 so it can do it with 0.5 kW.
So you lose 0.4 kW but gain bitcoins.
Or you could say that you get 2.25 as many BTC for the same electricity cost.
lmilcin
I wonder if we could make it even better and make smart heaters that instead of wasting calculations actually do something useful, something like folding proteins or some other computing load that doesn't require huge network load and can be easily distributed (something like SETI@home but actually useful).
In densely populated cities we could bury huge datacenters underground and use that energy for heating directly, and use excess to produce electricity.
zirkonit
Another one: Yandex (disclosure: the company I work for) is using one of its datacenters to heat the entire town it is in.
Proof: https://helsinkismart.fi/case/waste-not-want-not-data-center...
420codebro
Interesting. Never ran across someone who worked for Yandex. Overall are you happy with your employer?
zirkonit
Absolutely! By any definition, the pinnacle place to work for in Russia.
mrweasel
Currently a Facebook datacenter in Denmark is supplying around 10.000 homes with heating.
You can argue that Facebook isn't actually doing anything useful, but still, it's better than to waste the generated heat.
mensetmanusman
I don’t think any rational person argues that Facebook isn’t doing anything useful at all, lots of people value the connectivity and ability to share photos during lockdown.
Facebook has a good and bad side just like humans do. For example, if Facebook got rid of the ability to post news, it would be a much better place.
nr2x
The questions are twofold: does Facebook do more good than bad? Could the good be accomplished without much of the bad? The answers to both questions are arguably bad for Facebook.
pmontra
Or completely remove posts and keep only Messenger and Whatsapp. I've been using almost only Whatsapp and Telegram to share stuff with friends and groups of friends for a few years. If I want to read news I look for them either on Google News or on the very web sites that publish them, the ones I trust.
If Google and Facebook would block news on my country I wouldn't notice much.
MeinBlutIstBlau
Also, Facebook created apis that other developers utilize for many AppStore apps. So it's really nothing you can control.
fiftyacorn
I think Amazon does the same with a datacenter and powers a refuge
trungdq88
Fascinating! Is this true? Do you have a link? Thanks!
mrweasel
Most articles are in Danish, the best I could find quickly is: https://www.datacentremagazine.com/data-centres/facebook-exp...
Edit: It's should be noted that most Danish cities already have a remote heating infrastructure in place. Aside from the regulatory issues, it's mostly a question of hooking up datacenters and other heat producing industries to that infrastructure. In most places utilising the remote heating isn't voluntary, if it's available where you live, your home has to be connected.
Things like datacenters are slowly replacing coal fired heating plants, because most of those plants where made to generate electricity, but that's now supplied by more and more renewable energy. So cities need to find other sources of heat, to replace the volume no longer coming from the power plants. Where I live that's datacenters, waste incinerators and heavy industry.
As a new thing, remote cooling is now also attempted by using cold water from limepits.
Delk
There have also been plans to do this in Finland, but most of what I can find is either marketing material or news articles discussing plans [1, 2] rather than actual achievements, so I'm not sure what actually became of it.
There's a brochure from the national innovation fund that mentions a town actually covering about half of its heating needs with heat from a data center, though. [3]
Swedish telco Telia also has had similar plans for a data center in Helsinki, Finland, and their website says their "goal is to recover and reuse all the heat produced" [4], but I'm not sure how much weight to give that since proclaiming a goal only costs a few words. It would be nicer if they said what they're actually doing at the moment even if it were much less than "all of it".
[1] https://www.zdnet.com/article/from-deep-underground-data-cen...
[2] https://www.theguardian.com/environment/2010/jul/20/helsinki...
[3] https://www.sitra.fi/en/cases/district-heating-from-data-cen...
[4] https://www.telia.fi/business/telia-helsinki-data-center
Aachen
I don't doubt it, they'd be silly not to do it: good PR from a waste product? Perhaps you can even charge for it? Amazing deal, especially the former and especially for Facebook. It's also very common; a school I went to was heated by the data center across the road.
bayindirh
Welcome to Qarnot.
neilalexander
> The computing-heater warms buildings ecologically and for free, thanks to the waste heat released by embedded microprocessors. By performing complex IT operations...
... which are what, exactly?
ryankrage77
Going by the vaguness of the site, I wouldn't be suprised if they're mining bitcoin for themselves while you pay for the electricity.
Edit: Dug around some more, looks like they're building a BOINC-like service? https://computing.qarnot.com/en/
bryanrasmussen
looking through their tech stack and the hello world example and stuff I think you can put your own stuff i n it, but probably the most common stuff is 3d rendering.
on edit: I got a downvote so someone must think I'm wrong? the basis of my idea was - in the FAQ
https://computing.qarnot.com/en/FAQ
3D:
Blender, Maya, V-Ray, Guerilla
IA / ML / Big data or simulation:
Code Saturne FreeFem OpenFOAM PyTorch TensorFlow SickitLearn Spark
If there is a Docker image, we support the software! You can either bring your own or choose an existing one on Docker Hub. You can also ask our experts for help!
I expected the 3D stuff put up top made it the most used, but could be wrong in that. Obviously also some data crunching, ML tasks, but at any rate if my answer was wrong and so off base as to get a downvote maybe you could also just say why I'm wrong and what it's generally used for?
on second edit: developer documentation https://computing.qarnot.com/en/developers/overview/qarnot-c... made me think that maybe if you have one you can get your own api token and put your stuff on it, obviously you would have to pay them for that so not sure how it would work.
acomjean
Im not sure what it’s doing but “folding@home” is doing biology simulations as it’s useful distributed computing, that can probably generate some warmth.
BenoitEssiambre
I don't know much about this company but I love the name Qarnot which is suggestive of thermodynamically optimal computing. Carnot (https://en.wikipedia.org/wiki/Nicolas_L%C3%A9onard_Sadi_Carn...) was the father of thermodynamic efficiency. He came up with thermodynamics to optimize steam engines, maybe making him the most steampunk of scientists. Thermodynamically optimizing bits is an echo of that for the age of computing.
agumonkey
happy to know they're still on
dvdbloc
For years in an apartment I generated heat using surplus servers folding proteins. Some servers that are still quite fast and only a few years old are pretty cheap on eBay. Dockerize the whole thing and it’s easy to start and stop. One time I woke up however and it was especially cold, I realized I had a network issue and my servers could not get any more work units...
exdsq
'Proof of Useful Work', as opposed to proof of work, is a thing!
monkeydust
Any coins / tokens that implement this or something similar?
Nextgrid
The proof of work in tokens has to be a net waste of energy; the idea is that the proof of work is costly enough to prevent a 51% attack.
Making the proof of work do “useful” things would lower the cost of said work thus lowering the barrier of entry to an attack.
Uberphallus
Gridcoin.
vladvasiliu
I think that could be interesting. As others have said, whatever the computations are, if they are being done anyway, might as well use the generated heat for something useful instead of letting it go to waste.
But I wonder what would happen with such a system in the summer? For example in most on France it gets pretty cold for long enough every year that having a proper heating system and good insulation makes financial sense. But during the summer it's pretty hot, especially in cities. While the heating can be turned off between March and October (give or take), Facebook & co would probably like their DC to keep on working, so to keep on heating, year round.
galangalalgol
How hot is hot? A few days ago it was -18 here and now its 27. In a few months we will hit 38. I'm sure I dont want to be doing computations then.
tuukkah
District heating is used year round to get warm water.
rypskar
There are several useful BOINC-projects, for protein folding you have https://foldingathome.org/, the projects mine computers use most time on are the projects at https://www.worldcommunitygrid.org. You can find lists of projects at https://boinc.berkeley.edu/projects.php and https://www.boincstats.com/page/projectPopularity.
Interesting idea to combine it with larger heating than to heat a room or two. Many will probably argue that it is an ineffective way to create heat without also looking at the benefits from the work done
ryankrage77
> Many will probably argue that it is an ineffective way to create heat.
Electrical heating is as close to 100% efficient as you can get. Every watt your computer uses ends up as heat.
Generating those watts from non-renewable sources is much less efficient though.
I wonder if it's possible to calculate when the benefit of contributing to BOINC projects outweighs the CO2 generated.
hectormalot
Residential sized heat pumps can do '400-500%' efficiency (i.e. 4-5kWh of heat for every kWh of electricity), so electrical heating at 100% efficiency is indeed inefficient. If the calculations are really valuable it could still be worth it though.
Also, what would you do in the summer?
vbezhenar
Heat pumps have >100% efficiency. It's better to use AC to heat your house as long as outside temperature is not too cold.
undefined
eurasiantiger
Can climate change be solved by crunching more numbers?
medstrom
Produce electricity? I assume electricity was used to do the computation in the first place, turning into let's say 80% work and 20% heat. The idea was to use that 20% to heat our homes. The original electricity still needs to come from elsewhere.
oefrha
Doing computations hardly stores any energy anywhere (the energy stored in, say, magnetization in a magnetic disk is pretty much negligible), so almost 100% of electricity is turned into heat.
The problem with this ~100% efficiency is that, if your goal is heating, you can move way more than 100% heat with 100% electrical energy if you use, say, a heat pump.
owlmirror
Heat pumps are not viable in a lot of cases, in an urban setting, it's probably the vast majority. So for the many cases were the alternative for heating would be burning fossil fuels, using electricity, ideally produced via renewable energy/nuclear, could be a superior alternative.
lmilcin
When I said Data Center can produce electricity I did not mean that it can produce as much or more than it takes in. I thought this obvious enough that it did not have to be said.
But alas, it has to be said.
No, you can't create perpetuum mobile ie. build a machine that given a supply of energy produces as much or even more energy.
innocenat
It is still not viable.
Assume your datacenter runs at 50C (122F) and the temperature outside is -40C (-40F), using the datacenter heat to generate electricity has theoretical maximum efficiency of 27.85%. If your datacenter is at 23C (73F) and room temp is 0C (32F), then the theoretical maximum is 7.77%. Note the word 'theoretical maximum', in reality probably at least 10 times worse than that.
virgo_eye
No, essentially 100% of energy used for computation will turn into 'waste' heat.
If you calculate 1000 digits of pi, those digits will not embody any energy.
medstrom
Just turning chaos into order takes thermodynamic work, but I was wildly off with my percentages. Thinking in terms of an idealized computer that doesn't create heat, but didn't realize how far we were from that.
elwell
> Some crypto currencies (don't call them "crypto", that's lame and wrong)
"Cryptocurrency" is more than twice as long in letters and syllables. I think we've reached the point where "crypto" is disambiguated in most contexts from "cryptography", or did the author have a different reason for calling that abbreviation "lame and wrong"?
blueline
Crypto/cryptography is perfect nerd bait. Easy way to listen to yourself talk while being technically correct about something.
andy_ppp
Crypto is an abbreviation of cryptography, by trying to force the abbreviation to mean a specific subset of cryptography you limit all future uses of the word crypto to mean crypto currencies which seems excessive. We might have other words that follow crypto in future.
akvadrako
I think blockchain or cryptocoin is also acceptable.
pedrocr
The explanation of needing to pre-heat the air into the heat pump is strange. Modern A/C systems can pump heat at 500%+ efficiency even with very cold outside temperatures, no pre-heating needed. Replacing that with resistances is a 5x or more reduction in efficiency. That heat pump doesn't seem to have an outside unit so may be less efficient but it would need to be very poor to go down to only 100%. Installing a proper A/C system would provide both much higher efficiency heating in winter but also add the capability to cool in summer when the solar panels are actually providing a lot of energy.
Haemm0r
"Modern A/C systems can pump heat at 500%+ efficiency even with very cold outside temperatures, no pre-heating needed."
Name a source please.
pedrocr
Here's a current R32 system by Daikin:
https://www.daikin.co.uk/content/dam/dauk/document-library/d...
SCOP is at 5.15 and that's a seasonally adjusted value. So over the entirety of the winter the unit is expected to deliver 5.15 units of heat for each unit of energy. The unadjusted COP is 6.27. For this unit operational limits are listed at -15C/5F and I think there are units that go quite a bit lower. I haven't looked at that much as -15C is plenty for us. And this is with air to air. With air to ground much higher efficiencies are possible.
For cooling it's even better, with the equivalent SEER at 8.75. The improvements have been so good that the efficiency scale is already at A+++ because the A standard has been so exceeded.
Haemm0r
If you do not have temps below 0-5°C that might be true. Below the freezing point efficiency drops quite a bit.
klmadfejno
I did this for a while. It worked. Too well. We had to open the windows amidst a boston winter. In net it was barely profitable, especially with the recoverable cost of the computers. Fun and kind of a weird art installation in a way. Ultimately wasteful.
joosters
My mining rig will stay profitable until the ETH price is at ~900$
I don't think this is at all true. Even if the price of the crypto remained stable, the mining difficulty will keep on increasing as more miners join. So after a time, his mining rigs will consume more money in electricity than they generate in coins.
3np
Well, no, at a certain point (and indeed during normal market conditions, which right now isn't) there's an equilibrium. As global hashrate goes up, as you not, profitability goes down, which makes miners switch to other chains, or, eventually, power down their rigs.
I think your initial statement is correct, but from the wrong direction; if the ETH price is at 900$, the incentive to mine is less, some of current hashpower will already have left the mining of ETH, ergo OP is actually likely to still be profitable, although less so.
joosters
I didn't specify a direction, so I'm not sure how I can be describing the wrong one? My point is that the price of ethereum doesn't have to move at all and yet his mining rig will become unprofitable after a time. The hash rate will increase and/or more efficient miners will come online, leaving his equipment needing to mine faster or use less electricity in order to generate the same returns.
3np
I misinterpreted your comment then, apologies!
But, it's not a given that we'll have a significant enough increase of miners for that to be the case and time will tell (unless you happen to have knowledge I don't!)
progfix
The author did not mention the key factor in this: Sufficient thermal insulation. Looking at the photo of his house, this type of modern house has at least 16 cm thermal isolation on the facade and at least 20 cm on the roof.
pstrateman
This makes sense unless you have the option of natural gas heating, then not so much.
blacksmith_tb
How so? He says "on sunny days the miner and whole heat pump are running fully on solar energy collected on my roof", hard to beat that with gas. I would say that moving more houses to all-electric heating, water heaters, clothes dryers etc. is future proofing - right now, a lot of that energy would be from fossil fuels, but swapping in renewables is easy.
oconnore
Natural gas heating contributes significantly to climate change, and if you're in an area with significant renewable electric (or if you expect you will eventually over the lifetime of your system), you can lower your carbon footprint by heating with electric heat pumps.
shawnz
You could run a gas generator to power the miners and all of the waste heat could still heat your home.
rblatz
Most electric heaters are heat pumps which are significantly more efficient than heating via resistive or mining.
In fact a heat pump's efficiency is about 300%.
Aachen
Just to be sure, you're not talking of those €20-50 electric heaters right? Because you say "most" and afaik those are pretty much the only ones people have unless they're in some fancy new building or have a fancy AC and paid extra for that option.
lgessler
GP's referring to HVAC systems you'd find eg in a relatively new (1980s+) single family home in America
noxer
Someone clearly doesn't understand thermodynamics at all if he seriously thinks putting that thing outside and let the warm air be sucked in, is more efficient that running it inside. for obviously reasons (isolation loses/heat radiation) more energy (heat) is lost outside than inside (inside has zero loses because even the heat that doesn't go where you want it, still is inside the house.)
dang
Ok, but can you please post to HN without supercilious disses? The information here is good but we don't want a culture of putdowns here.
yarcob
- He mentioned that mining is more efficient when the GPUs are cooler, so using them to pre-heat outside air is better than using them to heat up room temperature air.
- since he would be forcing a lot of air through the miner, almost all heat would be transferred into the house using convection. The miner would stay cool and there would be very little radiation loss.
- finally, his ventilation system probably needs pre-heating on very cold days (often with a resistive pre-heater). Using the miner for that is smart.
noxer
>He mentioned that mining is more efficient when the GPUs are cooler...
Its simply not true. GPUs dont work fast or slower based on temp as long as the temp is in the rated range. It could potentially save energy by spinning the fans slower if the air around is super cool but the energy savings from that would be incredible small.
>since he would be forcing a lot of air through the miner, almost all heat would be transferred into the house using convection. The miner would stay cool and there would be very little radiation loss.
That's also not how thermodynamics work. The greater the temp differences are the more heat will be lost. The air need time to pick up the heat so you have a partially open box full of warm air outside that will heat up everything around it and possibly radiate heat directly in the sky unless it is extensively isolated. But even then you cant effectively isolation the air intake.
>- finally, his ventilation system probably needs pre-heating on very cold days (often with a resistive pre-heater). Using the miner for that is smart.
Sure, using the miner _outside_ is the not smart part.
yarcob
> GPUs dont work fast or slower based on temp as long as the temp is in the rated range
Correct me if I'm wrong, but I assumed that mining would quickly cause the GPU to reach the maximum temperature, at which point it would automatically throttle. The performance mainly depends on the cooling power. And air at -10°C will cool a lot more efficiently than air at 20°C. (I'm assuming the ventilation would move a roughly similar amount of air through the miner as the normal fans, which may or may not be a sensible assumption)
> The air need time to pick up the heat so you have a partially open box full of warm air outside
As far as I understood he's planning to pull the cold outside air through the miner. The warm air inside the miner would immediately be pulled inside the house.
I assumed that the amount of air used by the ventilation system of the whole house is enough to keep the miner pretty cool. The GPUs would get hot, but since a lot of air is moving through the outside of the miner would not get very warm at all.
I mean, yes, depending on where he places the miner and how long the hose to the air intake is, there will obviously be some lost energy. But I don't think it's as stupid as you say.
Especially since there is probably not a lot of space to put it inside the house, as typically the heat exchanger is very close to the air intake, to avoid long ducts with cold air inside the house.
Get the top HN stories in your inbox every day.
1. At an individual/hobbyist level this is a fine thing to do, if one is already a cryptocurrency enthusiast.
2. At a social/planning level it is important to remember that opportunity costs matter, and this should not be recommended as public policy. One kWh of electricity could either be used for mining bitcoin (creating at most one kWh of indoor heat), or it could be used for running a heat pump[*], yielding multiple kWh of indoor heat due to the magic of heat pumps and their >100% "efficiency"[**].
[*] Yes, I realize the OP is already using a heat pump, but assisting it with preheat is inherently less efficient than plowing the same energy into running the heat pump itself.
[**] Yes, heat pumps aren't >100% efficient per a physics textbook, but what most people care about is that you can use them to get >1 joule of indoor heat from 1 joule of energy input, which feels magical when compared to electric heaters, gas furnaces etc.