Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Speculate on what will become of all the mining farms when the Crypto craze finally comes to an end. - Page 2
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Speculate on what will become of all the mining farms when the Crypto craze finally comes to an end.

24

Comments

  • jsgjsg Member, Resident Benchmarker
    edited June 2018

    @Abdussamad said:
    @jsg the bitcoin core team for one is very capable. they've found bugs in openssl for instance.

    That's an indicator of capability but not more. There are indicators pointing in the other direction, too.

    @mksh said:

    @jsg said:
    Crypto currency? Thanks no, keep it.

    Well, shouldn't that rather be:

    Anything written in C? Thanks no, keep it.

    Theoretically, yes. Practically though we need an OS, don't we? Another very major factor at least for me is that e.g. Unix and Windows were created in merrily ignorant old times. So, Unices for example can be forgiven for not throwing away all their work and starting fresh at least in my opinion.

    The crypto currencies though are youngster and where created when it was already very well known how problematic C and C++ (and frankly most commonly used languages) are. So I feel it's alright to judge them in a much harder way than say Unices. I have btw. (a) looked at both some crypto currency code and designs and (b) worked in crypto heavy security projects myself and have a pretty good understanding of what I'm talking about. Most didn't give a damn about safety. Marketing bla bla, yes, but MUCH harder work in the coding workshop, nope.

    It's just the good old C dilemma. Great language but awfully easy to shoot yourself in the foot with. While i agree math guys might produce exceptionally shoddy code in my opinion anyone who proclaims he can write any siezable amount of C code and be reasonably sure to not introduce any kind of critical bug should get a head check. As for avoiding C, well, that would basically translate to avoiding computers in general.

    For a start the code of the math guys is usually not exceptionally shoddy. It's just not up to par considering the field. Btw. it's usually not the crypto guys writing the code themselves. Pretty much all the well established ones are at least associate professors and can hand the coding work to e.g. someone doing his master grad. Typically those guys aren't very experienced but they are smart and they are more or less good coders.

    Also keep in mind that they are not free to choose their language. Typically the competition rules ask for C which makes sense because in that field you need a language that is relatively close to the iron plus omnipresent on all architectures.

  • randvegeta said: I'm not convinced in the long term future of Crypto, so this question is obviously assuming that Cyrpto mining will eventually come to an end.

    Cryptocurrency will outlive your hosting aspirations

  • mkshmksh Member

    @jsg said:

    @mksh said:

    @jsg said:
    Crypto currency? Thanks no, keep it.

    Well, shouldn't that rather be:

    Anything written in C? Thanks no, keep it.

    Theoretically, yes. Practically though we need an OS, don't we?

    That's pretty much what i was hinting at with avoiding C would translate to avoiding computers.

    Another very major factor at least for me is that e.g. Unix and Windows were created in merrily ignorant old times. So, Unices for example can be forgiven for not throwing away all their work and starting fresh at least in my opinion.

    The crypto currencies though are youngster and where created when it was already very well known how problematic C and C++ (and frankly most commonly used languages) are. So I feel it's alright to judge them in a much harder way than say Unices. I have btw. (a) looked at both some crypto currency code and designs and (b) worked in crypto heavy security projects myself and have a pretty good understanding of what I'm talking about. Most didn't give a damn about safety. Marketing bla bla, yes, but MUCH harder work in the coding workshop, nope.

    I see where you are coming from and while i might not be all that concerned with using C in general i wholeheartedly have to agree on the last part. A lot of crypto projects are nothing more than 105% marketing garbage. Sometimes to the point of not even being an actual project but just a couple of colorful advertizing graphics. Expecting any kind of quality control there is what i'd call wishful thinking at best.

  • randvegetarandvegeta Member, Host Rep

    @doughmanes said:

    randvegeta said: I'm not convinced in the long term future of Crypto, so this question is obviously assuming that Cyrpto mining will eventually come to an end.

    Cryptocurrency will outlive your hosting aspirations

    What aspirations?

  • raindog308raindog308 Administrator, Veteran

    irm said: Wouldn't be difficult at all to accomplish this for anyone looking to convert a space to a grow.

    As you point out, though, there are already plenty of purpose-built highly efficient operations.

    I'm only familiar with North America, but it's not like the marijuana market is a land of opportunity right now. I live in a state where it's legalized (which borders two other states where it's legal) and marijuana is silly cheap here - like $3/gm plus specials plus loyalty rewards plus someone's always going out of business. On my 12-mile commute to work I pass more than 40 marijuana retailers, and that's mostly along one road. Many of these retail spaces have had multiple tenants because everyone opens a pot shop and then goes out of business in six months. Most of these places over hundreds of varieties, and they're all 80-90% pure THC. Or pure CBD if you prefer.

    Etc. There is a TON of high-grade, dirt cheap marijuana in the western US. Which means there's a ton of high-grade, dirt cheap marijuana everywhere because if you can't make a profit selling to local dispensaries, you can make a fortune selling underground. It's grown by the acre now.

    Good luck turning your DogeCoin farms into grow farms...

  • joepie91joepie91 Member, Patron Provider

    sunnyg said: With Bitcoin, how much would the seller fee be? Zero... How much would the transaction fee be for the customer to send the BTC/ETH payment? £0.50 at best...

    Until the transaction congestions are fully solved, you should expect transaction fees that are at least an order of magnitude higher than that, if not more.

  • randvegetarandvegeta Member, Host Rep

    @joepie91 said:

    sunnyg said: With Bitcoin, how much would the seller fee be? Zero... How much would the transaction fee be for the customer to send the BTC/ETH payment? £0.50 at best...

    Until the transaction congestions are fully solved, you should expect transaction fees that are at least an order of magnitude higher than that, if not more.

    Solutions already exist. Just need to be adopted.

  • sunnygsunnyg Member

    @joepie91 said:

    sunnyg said: With Bitcoin, how much would the seller fee be? Zero... How much would the transaction fee be for the customer to send the BTC/ETH payment? £0.50 at best...

    Until the transaction congestions are fully solved, you should expect transaction fees that are at least an order of magnitude higher than that, if not more.

    My 20sat/byte payments (£0.15) get mined in the very next block. If that is supposed to be on a congested network, I'm excited to see what Bitcoin has in store next :)

  • jhjh Member

    If it's anything like deserted petrol stations, my guess is hand car washes.

  • joepie91joepie91 Member, Patron Provider
    edited June 2018

    @randvegeta said:

    @joepie91 said:

    sunnyg said: With Bitcoin, how much would the seller fee be? Zero... How much would the transaction fee be for the customer to send the BTC/ETH payment? £0.50 at best...

    Until the transaction congestions are fully solved, you should expect transaction fees that are at least an order of magnitude higher than that, if not more.

    Solutions already exist. Just need to be adopted.

    No, they don't. All the existing solutions are subject to specific operational limitations; they either only work in specific circumstances, or they require a particular 'user investment' (eg. escrow sum) that makes it unsuitable for many cases, or they require a persistent connection to the network, or they provide reduced security, or...

    The fundamental underlying problem, namely unspent balance (UTXO) bloat, has not been solved. That problem is why the block size limit exists, and even that limit is just a temporary mitigation that doesn't solve the long-term storage requirements issues.

    @sunnyg said:

    @joepie91 said:

    sunnyg said: With Bitcoin, how much would the seller fee be? Zero... How much would the transaction fee be for the customer to send the BTC/ETH payment? £0.50 at best...

    Until the transaction congestions are fully solved, you should expect transaction fees that are at least an order of magnitude higher than that, if not more.

    My 20sat/byte payments (£0.15) get mined in the very next block. If that is supposed to be on a congested network, I'm excited to see what Bitcoin has in store next :)

    The current congestion isn't what I'm talking about. The inherent congestion problems in the current design, are.

  • randvegetarandvegeta Member, Host Rep

    joepie91 said: No, they don't.

    Yes they do. Move small transactions off the block chain and you need not worry about congesting the network.

    There are no blockchain solutions. Blockchain does not scale, no matter what you do. The only solution is to move transactions off the blockchain and accept semi centralization. This is not a big deal IMO. Though blockchains in themselves are 'trustless', you still need to trust the 3rd party (or parties) you are dealing with. If you imagine a solution where almost all sub $50 transactions are done off-chain, then the risk is minimal, cost is tiny, and transaction speed is near instant (much faster than block time).

    It's really quite a simple concept, and solves BTC's scalability problem entirely.

  • AidanAidan Member

    randvegeta said: What becomes of these spaces, kitted out with massive amounts of power and ventilation/cooling?

    Most would qualify as semi-decent warehouses & would likely be rebranded as such.

    Seeing as we're talking about a couple of dozen large buildings, there'll be no net effect.

  • randvegetarandvegeta Member, Host Rep

    Aidan said: Most would qualify as semi-decent warehouses & would likely be rebranded as such.

    I suspect alot of places started out as warehouses. Power infra is also not used at all for warehousing.

  • AidanAidan Member

    I suspect alot of places started out as warehouses. Power infra is also not used at all for warehousing.

    Italian loaves are stored at negative 60'C, which uses a fair amount of power.

    The world's a big place, I'm sure other industries would gladly absorb the warehousing space.

  • @sunnyg said:
    I hope Bitcoin/Ethereum prevail and go mainstream for payment processing. I just received £1800 payment on Paypal, and guess how much the seller fee was? £88... And I still run the risk of being disputed over the course of the next 6 months or charge-backed over the next 12 months. With Bitcoin, how much would the seller fee be? Zero... How much would the transaction fee be for the customer to send the BTC/ETH payment? £0.50 at best... Personally I could care less about the price of the actual coin, I was surprised back in 2013 that Bitcoin had 100x the value of a Dollar, now it is 7000x and it feels a little less ridiculous some how.

    I am afraid, that is not a valid comparison. On the other hand, if you accepted cash, you'd have 0 costs.

    You are comparing a commercial service vs block-chain (kinda like open source).

    Of course with Paypal, the buyer gets protection, would the same buyer be willing to pay the same price if all protection is removed? Just like buying a mobile phone from grey market vs official outlet, you pay less but you have no warrant.

    PS: £88 on £1,800 is nearly 4.88%, which is very high!

  • joepie91joepie91 Member, Patron Provider

    @randvegeta said:

    joepie91 said: No, they don't.

    Yes they do. Move small transactions off the block chain and you need not worry about congesting the network.

    There are no blockchain solutions. Blockchain does not scale, no matter what you do. The only solution is to move transactions off the blockchain and accept semi centralization. This is not a big deal IMO. Though blockchains in themselves are 'trustless', you still need to trust the 3rd party (or parties) you are dealing with. If you imagine a solution where almost all sub $50 transactions are done off-chain, then the risk is minimal, cost is tiny, and transaction speed is near instant (much faster than block time).

    It's really quite a simple concept, and solves BTC's scalability problem entirely.

    I've seen exactly zero off-chain transaction mechanisms that could replace on-chain transactions in full, without limitations or additional requirements. If you know of one, I'll gladly look at it...

  • WilliamWilliam Member
    edited June 2018

    Why speculate? This happens daily. Power costs go up and miners move.

    90% of these are just industrial warehouses, much like Cocius datacenters in Romania, they either will be used for storage again or stay, as probably before, empty.

    randvegeta said: I suspect alot of places started out as warehouses. Power infra is also not used at all for warehousing.

    A normal house in Europe has at least 100A on 1 phase, if not 3 phase (way more common). My flat has 100A.

    I've never seen a warehouse with less than at least a single 100A 380V 3 phase feed - it simply makes no sense wire less, you will have at least one 380V device and thus get 3 phase and practically many hundred A more if desired easily.

    Thanked by 1Aidan
  • MikeAMikeA Member, Patron Provider

    So who wants to sell me a used GTX 1080 for cheap?

    Thanked by 1TriJetScud
  • MikeA said: So who wants to sell me a used GTX 1080 for cheap?

    No, but i have tons of 470s/480s with near dead memory. They would be useful if AMD had something like Cuda, Vulkan is more of game obviously and OpenGL is useless.

    Nvidia cards have only lost resale value since they have stock again available, same as AMD, so you probably pay now the same or similar for an expensive used to a cheap new one. A cheap used one will probably rare to find, mining is still profitable.

    Thanked by 1vimalware
  • williewillie Member

    AMD has HIP (ROCm) which is like Cuda but it's not ready for prime time yet. They are waking up to the necessity for this now, so it will probably work better soon. New hardware from them is also coming.

  • WilliamWilliam Member
    edited June 2018

    Yea, but if it only works on GCN5 (Vega) it will be useless for miner cards (mostly GCN4/GCN3) as the HBM based do not really suffer from memory death anyway.

  • defaultdefault Veteran
    edited June 2018

    @William said:

    MikeA said: So who wants to sell me a used GTX 1080 for cheap?

    >

    A cheap used one will probably rare to find, mining is still profitable.

    I concur. Old and expensive (but used) mining GPU's have not flooded the auctions market yet, which means mining with those is still profitable.

  • AbdussamadAbdussamad Member
    edited June 2018

    joepie91 said: I've seen exactly zero off-chain transaction mechanisms that could replace on-chain transactions in full, without limitations or additional requirements. If you know of one, I'll gladly look at it...

    Have you heard of lightning? It is the future of bitcoin at least.

    @jsg they need control over memory which is why they use c++. decrypted private keys cannot be swapped to disk by the OS for example. they must stay in memory.

    edit: actually here's what one core dev has to say about it.

  • mkshmksh Member
    edited June 2018

    @Abdussamad said:
    @jsg they need control over memory which is why they use c++. decrypted private keys cannot be swapped to disk by the OS for example. they must stay in memory.

    While i guess this can be done (at least on linux) with mlock i have my doubts the average reference implementation makes use of it. Short of this managing memory yourself doesn't mean you get to control swapping and afaik C/C++ doesn't have a standardized way to control it. It's OS dependent.

  • jsgjsg Member, Resident Benchmarker
    edited June 2018

    @Abdussamad said:
    @jsg they need control over memory which is why they use c++. decrypted private keys cannot be swapped to disk by the OS for example. they must stay in memory.

    edit: actually here's what one core dev has to say about it.

    That's, Pardon me, a typical example of the "know-how illusion" in crypto currency circles.

    For a start the way you put it is quite vague but I guess you are talking about garbage collected languages vs. C++ where memory management is in the hands of the developer.

    What you say is wrong. Plain and simple wrong. Whether memory gets swapped to disk or not is nothing to do with C++ vs. a GC language. Plus in virtually any modern OS memory is virtual in one way or the other anyway. In some OSs one can, however, at least have some control over memory being swapped out or not.

    But note that that issue is not even critical in the realm of security because there we always assume a worst case (Dolev-Yao) Eve. Those who thought that anything in memory is somehow more safe than on disk should have learned their lesson with Meltdown/Spectre at the latest.

    Re your edit: That guy seems to be somewhat concerned about safety and security (which is laudable) but stays rather vague and seems to have only intermediate knowledge in the field. He pretty much just repeats typical intermediate level beliefs and standard remarks.

    Thanked by 1mksh
  • mkshmksh Member

    @jsg said:
    But note that that issue is not even critical in the realm of security because there we always assume a worst case (Dolev-Yao) Eve. Those who thought that anything in memory is somehow more safe than in disk should have learned their lesson with Meltdown/Spectre at the latest.

    Not to say that (at least on linux again) setting up an encrypted swap is as easy as it gets.

  • AbdussamadAbdussamad Member
    edited June 2018

    jsg said:

    Re your edit: That guy seems to be somewhat concerned about safety and security (which is laudable) but stays rather vague and seems to have only intermediate knowledge in the field. He pretty much just repeats typical intermediate level beliefs and standard remarks.

    he's one of the top developers of bitcoin core and a cryptologist.

    well i guess you know more than any of them. there's a lot of money to be made from hacking bitcoin. some call it the biggest bug bounty ever.

  • btw doesn't this prevent swapping?

  • mkshmksh Member
    edited June 2018

    @Abdussamad said:
    btw doesn't this prevent swapping?

    See my post above. It's a linux syscall. Not related to C/C++. Other OSs likely have something similar but if not, well, tough luck. C/C++ itself have no mechanism to control swapping.

    Edit: I see mlock actually has some POSIX support. Still not the whole world is POSIX (does full compliance even exist?) and i would'nt count on it to be available across the board.

  • jsgjsg Member, Resident Benchmarker

    @Abdussamad said:
    he's one of the top developers of bitcoin core and a cryptologist.

    well i guess you know more than any of them. there's a lot of money to be made from hacking bitcoin. some call it the biggest bug bounty ever.

    The fact that you use authority and personal attacks on me instead of arguments doesn't make you look strong, quite the contrary.

    I don't care about "top developer" of project xyz. I care about facts and real security.

    The real issue here is that you played too big and tried to make yourself look smart when in fact you are clueless and now you try to bring up (presumed) authorities and to attack me personally to save face. Simple as that.

    I suggest you let it go and try to learn something.

Sign In or Register to comment.