Topic: Hardware powering the network (Read 963 times)

Nowhere in my post did I say it was profitable. Quite the contrary. It is for speculative and entertainment purposes only. But if you're lucky, sometimes speculation on alt coins can be more profitable than straight bitcoin mining. Regardless, I find it highly entertaining especially now with btc in the doldrums. It's great fun to watch the price of some spunky new coin gyrate up and down (and hopefully up again) on the charts.

Reselling obsolete gear is hardly worth the effort these days. Unfortunately I missed that boat by miles.


Btw, I'm highly allergic to buying coins, bitcoin or otherwise. That's why I mine. 

With the proviso that mining at a loss may be rational, if you have difficulties accessing other ways of obtaining BTC.  Consider, if China continues to make it hard to deposit money to exchanges, then many Chinese miners might quite rationally continue to operate their older ASICs, even if they are effectively paying a substantial premium for their coins, because it may still be cheaper and easier than the alternatives (e.g. an in person transaction might involve travelling to another town, which obviously has its own costs associated with it).

roy

if mining bitcoins with old equipment is now unprofitable, then mining alt-coins that are even less profitable is an even worse thing to do with them.  unless, as you say, its for pure entertainment.  but there is a cost for that entertainment in electricity use... so it may be better, if you want those alt-coins, to just buy them and not waste the electricity.  and sell off the old miners to someone less capable of calculating their profitability or open them up and recover the valuable parts (power supplies, some cables) and send the rest for recycling.

An interesting way to extend the life of old miners is to mine alt coin with them. It's more like speculation since it isn't really profitable to mine them. But it's fun to have other coins to hold or trade without having to spend our beloved bitcoins.

I'm "mining" some PoS coins right now with some old Singles but I just shutdown my AM Blades today. I used "mining" in quotes because you can't actually mine proof-of-stake coins. But there are some pools that mine other alt coins and exchange them for PoS coins. I know it's kind of silly but it gives me cheap entertainment.

ha! my BFL singles were still making £20 profit per month and i've powered them down just today.
The fact that my rig is now running at my folks place - in their garage, something to do with this fact... I know exactly how much power draw my SP10 is pulling (without the need for a kill-o-watt) and can work out exactly to the penny how much money I need to give my mum each month to run it. I'm rambling now...

I'd say that we could bring that down to 3 months, which I'm sure would still give plenty headway to take into account, those running at a loss.

I think a decent way to remove old asics is to just remove them after 6 months of running on the minus. I can understand people keeping them running while paying more on electricity than what they earn in BTC, but keeping them for more than 6 months like that is close to stupid and we can ignore them.

Disclosure: I think this May will be my first month of running burnin's (gen1 avalon chips) boards on the minus. Let's see for how long.

Of course you are right, i was thinking slightly idiosyncratically and not taking into account the fact that the old 130nm rigs are indeed, many times smaller - once those are all powered down, it would only be a small reduction in the network hashrate. I do also think that you are right regarding the sub 0.5w/GH technology, and I find great interest in watching the manufacturer markets, trying to predict who will break through first with a 16nm rig. I think even 20nm is not going to cut the mustard after only a short time, as these chips are just a stepping stone to the more efficient 16nm. I was actually astounded at the efficiency of the spondoolies gen1 - which my new rig is, performs at around 0.85w/GH, i'm keen to see what they can do with a simple die shrink on their gen2 machines and how far that goes in an overall sense.

Anyway, going slightly off topic there now. Apologies - you can probably tell I enjoy my new machine very much!

So, how do you see the collection of the data occurring? It would be foolhardy to simply ask the manufacturers, because they will just tell us what they want us to hear, and what percentage of people would we equate to be those 'die hard' miners who keep their 130nm running in speculation of a rise in BTC price?
It's an interesting endeavour, to say the least! my hat is off to you if you intend on heading this up, and i'll watch with great admiration if you can show something close to an actual figure. It would be a great thing, if we can also monitor the changes as time goes on.

i think we might be able to group the systems in easy ways.. by their TCO - total cost of ownership... i.e., purchase price of the mining system, AND its power consumption, because those two factors ultimately determine their on-going profitability, and over time the latter (cost of operation) will become more important, as the network TH's increase yet the cost of power doesn't decrease.

the older systems (the bfl 65nm's, older avalons & asicminers etc) were all built using chips that consume multiple watts per GH (some as high as 6w/gh), are all being turned off or phased out (even as we speak) as they're no longer power efficient and the cost of electricity and price of bitcoin make them uneconomic to run.  BUT the sum total of all of those 'multi watts per GH' systems ever sold, is still in the low single digit PHs (lets be generous, and say 5-10 PH's, and i think closer to 5 is likely)

thus even when people decide to switch off every single 'multi watt per GH' bitcoin miner, its will hardly make a dent in the network size because the growth today is already more, in one month, than the sum total of all 'multi watt per gh' systems that exist.

What you're left with (when all the older/inefficient miners are switched off) are all the newer systems, as sold by BitFury, KnCMiner, Cointerra, AntMiner, Innosilicon, etc... that are all currently operating around 1Watt/GH system power (give or take)... and thus are much more efficient than the previous generation, and also much lower cost to purchase thus there's many more of them and they're dominating the network...   and once they've been purchased and deployed, few people will have an interest in turning those systems off until the price of bitcoin and the price of electricity make them uneconomic...

We'll wait til the next generation of lower powered miners comes out that may consume - perhaps <0.25 Watt/GH  - that might emerge next year (or perhaps late this year)

i'm thinking it needs to be in the ballpark of 0.25 watt/gh rather than <0.5w/gh, because people need a big incentive to change behaviour so there may need to be a significant change in the cost of operation (power consumption) - or the cost of purchase  - rather than a simple halving the power consumption like the newer systems coming this year might at best achieve).

i think it will be very difficult to ascertain, for example; my latest addition to my rig arrived last week, and since powering it up, i've now powered down other parts - AM, BFL and a few garden blades... sensible people switch off when it's no longer worthwhile, only the die hard will keep all their equipment running regardless of BTC price. A good indicator might be to watch ebay to see what people are selling off... I doubt we'll ever be able to ascertain true values on the network, one might assume lots of things, and manufacturers providing numbers is merely that, their own assumption that all of their customers are still running all of their ASICs. I would very much doubt that BitFury have anywhere near the amount of 40%. possibly still 30%, but i'm just pulling numbers from thin air, in that.

we don't know, but thats why doing this openly with crowd sourced data is the way to weed out what sounds truthful, what can be corroborated, etc..

Interesting idea, but how do you know if people are telling you the truth about their hardware (which may or may not exist)?

it would be useful to crowd source some information to figure out which hardware is powering the network.  This is a different statistic to which mining pools are operating, since they are often multi-vendor.

This isn't a forecast, post mortem, or pool statistic.  Its simply trying to figure out what it is going on Right Now and which hardware is being used by miners of every category (home, enterprise, private mines, etc) to power the network and is regardless of which pool they choose to hash on.

Current Network (April 26th) is approx 60 PH.

From information we have available it breaks down like this :-

BitFury 20 PH
KnCMiner 12 PH
Cointerra 8 PH
AntMiner 8 PH
ButterFlyLabs 7 PH ?
HashFast 2 PH ?
Innosilicon/Bitmine 4 PH ?
Avalon 1.5 PH ?
AsicMiner 1.5 PH ?

Need help filling in the details.

The first four are public statements so have some basis in reality - BitFury claims 40% of the network as recently as last week (source: themselves), KnCMiner has 8 PH in their own private mine (source: organofcorti), and sold 4 PH of Jupiters (over $40m in sales, Andreas announced on a panel at Inside Bitcoins in Berlin), CoinTerra announced 5,000 systems (15% of the network when it was 50 PH) delivered by mid April (at 1.6 TH/s), and AntMiner claimed 20% of the network (i.e.: 8 PH when it was 40 PH)

The top four stats are reliant on each company's claims being true.  The guys that claim a % of the network without referencing actual hardware units shipped are likely to be less accurate since they can only be true at a moment in time and if this time was known it was factored in.

The next bunch, BFL, Innosilicon/BitMine, HashFast, Avalon and AsicMiner are needing more accurate estimations, and are probably very inaccurate at this point.

Does anyone have any better information that can improve the accuracy of this list?   When we have something more accurate, its probably worth charting it over time.