4 ms·

Look at GPU hash rates, CPUs can't compete at hash rates that much is clear.

IkmoIkmo 12y ago
Even GPU rates are absolute crap compared to ASICs, and that's the top-end cards used for gaming. You generally get a few hundred MH/S at a few hundred dollars, for example a GTX580 at $200 used gives you about 150 MH/S. So 0.75 MH/S/$. Compare that to ASICs that can get 3000 MH/S/$. Meanwhile the graphics card is a huge power hog, at 244 watts the GTX580 will cost you $320 in electricity if ran 365/24/7 at an OECD average. So let's plug it into today's hashrate calculator: $0.15c, if ran for a year 100% maxed out 24/7. Imagine we lower that to 25% usage, 8 hours a day, and you get roughly 1 cent. Oh and these hashrates are growing exponentially so it's likely much, much lower. So even if you do this year-round, you'd blow your user's graphics card in a year, generate $320 in extra electricity costs and make 15c (very optimistic estimate), and completely lock up 100% usage of his graphics card 24/7. And that's for a decent gaming card that you can buy at $200 used today. Don't mind the millions of torrenters who use laptops with on-board shitty cards who just want to pirate Netflix. In any case, I'm not even sure if it's relevant as I just checked the CPU rates because users in the thread mentioned their CPU went up while idle that hadn't happened before, indicating they're CPU mining as opposed to GPU mining (which is probably easier than configuring for different cards, and allows mining on computers with a decent CPU and a shitty on-board video card for non-gaming.) So I still think it's the dumbest idea ever.
- squeaky-clean 12y ago
  The problem with this is you keep factoring the costs of a computer, or electricity or an ASIC. It's expensive or the user. But for uTorrent, that computer is free. You can't trick people into installing bitcoin mining crapware on their ASIC. There was a similar scandal[0] in 2013 where the ESEA client (an e-sports league, and the client is required to play in ESEA matches) "infected" about 14,000 users. They were able to generate a few thousand dollars worth of Bitcoins (worth more now) over only two weeks until the miner was noticed by the community. They have an advantage because most people who maintain a paid subscription to an e-sports league will have powerful GPUs. But compared to the size of uTorrents userbase, 14,000 gaming PCs is nothing. In this case though, it looks like EpicScale is only using the CPU. And probably not mining bitcoins, as they never mention them specifically. Their terms[1] list "[...] various purposes, including, but not limited to, cryptocurrency mining, data processing, data analysis and/or scientific research and development." [0] http://www.wired.com/2013/11/e-sports/ http://www.wired.com/2013/11/e-sports/ [1] http://www.epicscale.com/terms http://www.epicscale.com/terms
  - IkmoIkmo 12y ago
    > The problem with this is you keep factoring the costs of a computer, or electricity or an ASIC. It's expensive or the user. But for uTorrent, that computer is free. How is that a problem in my analysis? It's important. For example say you bought a Macbook Air and came with mining software, used 100% of your CPU and GPU and made your computer unusable. And it generated 1 penny per year in return. And it'd cost you $100 in electricity. Would you consider it normal to say 'But for Apple it's 1 penny, for free! Your analysis is flawed because all those costs, who cares, Apple isn't paying for them, for them it's a free penny'. Of course not. It's ridiculous. Screwing over your user is not without consequence, there is a cost involved. A moral one, for one, which if you can't express financially still matters. That is, locking up 25% or 50% of a CPU, reducing the lifetime of their CPU and costing them $20 in electricity per year is screwing over your users. But secondly there is also an economic cost, just not directly. But indirectly in the long run, doing things like this will lead people to uninstall and avoid your product. In the exaggerated example about Apple above, it's obvious nobody would ever buy that Macbook Air again. A similar thing will happen for uTorrent on a smaller scale, and that's a cost. And if your little scheme generates peanuts in revenue, then that cost of losing users is likely bigger than your revenue if you'd built a less intrusive scheme. As for your example, it's from spring 2013. Spring 2013 the hashrate was about 50k, today it's 350 million. So literally, the same infected computers today would earn $1.17 in the same period, or about $28 per year. And the electricity costs would be thousands of dollars. Even if you didn't have to pay for it, a business model where you screw with your user's hardware performance, generate thousands of dollars in costs for them to make $28 is ridiculous. Even if you scale that up from 14 thousand to 14 million users, not with heavy GPUs but average users using their CPU, it's absolute peanuts. I showed that already. I'm not making this up. This is exactly my point, as I mentioned, a few years ago this would've been a brilliant idea, hell even spring 2013, today, not so much.
    - edwintorok 12y ago
      Apparently it was supposed to be opt-in and their website says they only use CPU so you were right in comparing CPU with ASICs, it doesn't make sense: https://www.reddit.com/r/techsupport/comments/2y0g43/epicscale_compute_is_this_legit/cp64ida https://www.reddit.com/r/techsupport/comments/2y0g43/epicsca... https://www.reddit.com/r/techsupport/comments/2y0g43/epicscale_compute_is_this_legit/cp5v423 https://www.reddit.com/r/techsupport/comments/2y0g43/epicsca... They should drop the cryptocurrency mining part, it just looks shady to do that when there is no real benefit as you showed. And they should make sure that opt-in really is opt-in...

Even GPU rates are absolute crap compared to ASICs, and that's the top-end cards used for gaming. You generally get a few hundred MH/S at a few hundred dollars, for example a GTX580 at $200 used gives you about 150 MH/S. So 0.75 MH/S/$. Compare that to ASICs that can get 3000 MH/S/$. Meanwhile the graphics card is a huge power hog, at 244 watts the GTX580 will cost you $320 in electricity if ran 365/24/7 at an OECD average. So let's plug it into today's hashrate calculator: $0.15c, if ran for a year 100% maxed out 24/7. Imagine we lower that to 25% usage, 8 hours a day, and you get roughly 1 cent. Oh and these hashrates are growing exponentially so it's likely much, much lower. So even if you do this year-round, you'd blow your user's graphics card in a year, generate $320 in extra electricity costs and make 15c (very optimistic estimate), and completely lock up 100% usage of his graphics card 24/7. And that's for a decent gaming card that you can buy at $200 used today. Don't mind the millions of torrenters who use laptops with on-board shitty cards who just want to pirate Netflix. In any case, I'm not even sure if it's relevant as I just checked the CPU rates because users in the thread mentioned their CPU went up while idle that hadn't happened before, indicating they're CPU mining as opposed to GPU mining (which is probably easier than configuring for different cards, and allows mining on computers with a decent CPU and a shitty on-board video card for non-gaming.) So I still think it's the dumbest idea ever.

squeaky-clean 12y ago
The problem with this is you keep factoring the costs of a computer, or electricity or an ASIC. It's expensive or the user. But for uTorrent, that computer is free. You can't trick people into installing bitcoin mining crapware on their ASIC. There was a similar scandal[0] in 2013 where the ESEA client (an e-sports league, and the client is required to play in ESEA matches) "infected" about 14,000 users. They were able to generate a few thousand dollars worth of Bitcoins (worth more now) over only two weeks until the miner was noticed by the community. They have an advantage because most people who maintain a paid subscription to an e-sports league will have powerful GPUs. But compared to the size of uTorrents userbase, 14,000 gaming PCs is nothing. In this case though, it looks like EpicScale is only using the CPU. And probably not mining bitcoins, as they never mention them specifically. Their terms[1] list "[...] various purposes, including, but not limited to, cryptocurrency mining, data processing, data analysis and/or scientific research and development." [0] http://www.wired.com/2013/11/e-sports/ http://www.wired.com/2013/11/e-sports/ [1] http://www.epicscale.com/terms http://www.epicscale.com/terms
- IkmoIkmo 12y ago
  > The problem with this is you keep factoring the costs of a computer, or electricity or an ASIC. It's expensive or the user. But for uTorrent, that computer is free. How is that a problem in my analysis? It's important. For example say you bought a Macbook Air and came with mining software, used 100% of your CPU and GPU and made your computer unusable. And it generated 1 penny per year in return. And it'd cost you $100 in electricity. Would you consider it normal to say 'But for Apple it's 1 penny, for free! Your analysis is flawed because all those costs, who cares, Apple isn't paying for them, for them it's a free penny'. Of course not. It's ridiculous. Screwing over your user is not without consequence, there is a cost involved. A moral one, for one, which if you can't express financially still matters. That is, locking up 25% or 50% of a CPU, reducing the lifetime of their CPU and costing them $20 in electricity per year is screwing over your users. But secondly there is also an economic cost, just not directly. But indirectly in the long run, doing things like this will lead people to uninstall and avoid your product. In the exaggerated example about Apple above, it's obvious nobody would ever buy that Macbook Air again. A similar thing will happen for uTorrent on a smaller scale, and that's a cost. And if your little scheme generates peanuts in revenue, then that cost of losing users is likely bigger than your revenue if you'd built a less intrusive scheme. As for your example, it's from spring 2013. Spring 2013 the hashrate was about 50k, today it's 350 million. So literally, the same infected computers today would earn $1.17 in the same period, or about $28 per year. And the electricity costs would be thousands of dollars. Even if you didn't have to pay for it, a business model where you screw with your user's hardware performance, generate thousands of dollars in costs for them to make $28 is ridiculous. Even if you scale that up from 14 thousand to 14 million users, not with heavy GPUs but average users using their CPU, it's absolute peanuts. I showed that already. I'm not making this up. This is exactly my point, as I mentioned, a few years ago this would've been a brilliant idea, hell even spring 2013, today, not so much.
  - edwintorok 12y ago
    Apparently it was supposed to be opt-in and their website says they only use CPU so you were right in comparing CPU with ASICs, it doesn't make sense: https://www.reddit.com/r/techsupport/comments/2y0g43/epicscale_compute_is_this_legit/cp64ida https://www.reddit.com/r/techsupport/comments/2y0g43/epicsca... https://www.reddit.com/r/techsupport/comments/2y0g43/epicscale_compute_is_this_legit/cp5v423 https://www.reddit.com/r/techsupport/comments/2y0g43/epicsca... They should drop the cryptocurrency mining part, it just looks shady to do that when there is no real benefit as you showed. And they should make sure that opt-in really is opt-in...