Topic: Gigantic difficulty jump of the last few days (speculation) - page 2. (Read 4640 times)

it's actually the opposite, increase in price = increase in diff

we have a better profit with today diff than what it was a 230 per btc, 93B diffculty is better with 460 in value, than 230 inv alue and 60B diff(the previous stable diff)

the current value also, will permit a good margin of profit even after the halving and even without the new efficient machine, currenty efficiency is very good, only a consumption of $45 per month with a single s7

Dilute, hide, and broadly unfairly distribute costs.  Sounds like everyday business to me.

I disagree.  Quite a bit....  you can never remove the person that has cheap or close to zero cost power.

What you can do is lower  his hash to network hash percentage but if he has really cheap power you can not beat him.

My friend leases a floor in a building and he then sublets he keeps 20% of the floor and pays 20% of the power bill.   Normal rates are 10cents  but he is paying 20% of that or 2 cents.  he can give me 2 s-7's
So I will mine s-7's for years there.  A second benefit is his place is cold in the winter as he does not control heat he used to use 2 1500 watt heaters for 5 months  a year.  he won't be using them.

So basically he has perfect mining setup for 2 s-7's  no one can beat him  he is cheaper then free power as 5 months a year he is gettng better then free power and the other 7 months he gets 2 cent power.

Granted he is limited but he is not the only person  with these conditions.
You will have small mining as they are the only miners  that can have 'free' power setups.

How many of these are sold every year?
Sooner or later  this will come about  that a company like Delonghi will cut a deal with a company link bitfury

are you implying that difficulty and price have forever been synchronized even if not in immediate fashion, ergo increasing difficulty will yield increasing price?

secondly, and possibly a crude iteration of what you said,  I think that mining will essentially cease at the halving because only those with 16 or whatever nm architecture will even attempt to survive and technology won't meet demand for 4 or 8nm...at least nowhere near the halving time.

As it largely now is,  two or three companies will control everything after the halving and even they will be forced to 'pan' for their pay. If correct, will the diff after halving maintain the synchronization with price that I interpret you've suggested?

i don't know why people are worried about diff increase all the time, the ratio between the profit and the diff will always be the same, so if many will join and it will be not-profitable, other will simply leave

it's not like everyone out there has free electricity, that would be scary, because they will dump until btc is zero in value, but even then there is the halving...

Not really.... 21 said they were very into this.  And what did they give us? A custom RPI board... they say the above but they yet to do it. 

I would still buy a bitfury lightbulb for fun if they ever do sell them.

I love the light bulb idea but was wondering how they might get it connected. wifi? usb doggle?

1 is possible. It is mainly caused by S7 coming to the market. Bitmain is going to build more s7. So the difficulty will rise again.

Think of people with old school light bulbs burning 60, 75 or 100 watts a blub.  An led bulb burning 12 watts with a chip mining 5 watts same light at 17 watts and  some coin back.  I could see this happening.

This is the vision of 21 Inc. ^

I see some interesting things happening:

1. If Bitfury will be selling chips to anyone, then they are positioning themselves to be a.k.a Intel of bitcoin, then someone (maybe 21 inc) will be a software (mostly) arm-with their emphasis on programming solutions; Coinbase will be a bank, etc, etc.
Sooner or later a whole ecosystem will be fully fleshed out.

2. with 0.05 J/Gh, some dual use devices actually make more sense. For example, a dual use lamp w/mining chip.
Light ON-40W to light, 20 W to hashing, light switched off-all 60w to hashing.
Light on-hashing 400Gh (20W), lights off-up to 1200Gh-basically an S5 in a lamp!
How many lamps do you have? A few dozens easily.

as to your below quote copied from above



--------

 Sfards had that SF100 unit, but it appears to have been a very poor board-level design AS USUAL from the successor to Gridseed.

 Innosilicon A4 - sometime next year likely first half.
 Alcheminer is also working on a next-gen Scrypt chip, but doesn't seem nearly as well funded.


--------


they both are gonna be like 500mh or so units ...ASSUMING you can actually get some of them in 1 unit batches from somebody ...in a approprate amount of time (dubious)

The alcheminer II in 4-6 months if I remember right it 510mh 1150 watts at $4950 usd

I have Titan(s) that have ROI'd so I'm hoping these folk know something I don't on expected BTC/LTC prices etc

because 1 Titan is 350mh 1250 wats and again ROI'd so I can run with this pack in 6 months...depending on how data hall nuts they go on these

so yeah..only real option if you are CRAZY (I don't recommend it buy crypto in BTC/LTC instead) is buy used scrypt ...but essentially the prices are TOO high on ebay etc imho
and not only that you probably could only pull this off with the Alcheminer 256 (with new firmware 3rd party patch fixes on litecointalk.org) or Titans at 350 mh. Would have to be used equip OVER 200mh as a unit to pull this OFF and remember Alchimer 256 1st batch use (i think) 2x the watts of a Titan 350mh...so that also plays into all this...at least at my 13c kwh rates or more thereof.....anything below would kill you on ROI at these used prices...only the BIG ass beasts would pay (used) but they are SOOOOO high priced imho...but what the hell you could get lucky or have cheap ass under 5c kwh electric...just a note with 2 titans I make at 645mh ave speed about 450 net after elec so 1 Titan is 1/2 that ....so these units in 6 motnhs are NOT impressive at my elec rate indeed

me at 13c kwh just can run the Titan's out....it especially (likely) will not be worth me chucking out say $4950 usd for a 510mh unit in 4-6 months when you on ebay NOW can get a 350mh Titan (used) for around 3,250 usd or so ....and with new equip in these asic units both btc and scrypt with only 90 day warranties ...your odds are not too bad on a used unit via say a paypal purchase on ebay (45 days) so wtf ..if you are that nuts...just grab a used Titan now and mine like hell imho why wait for these newer units vs difficulty now vs then and price now vs then it looks at worse a wash at best you are ahead with used equip NOW

but again I like the mnfg optimism that they can sell alcheminer II's at 501mh for 1150 watts 4-6 months from now (and they are always 1 month or two late imho) and you have a shot at ROI....or so the mnfg must be thinking to push these at that mh and price range...if they are correct then we will be looking at 7 buck LTC imho to pull that off!

anyway don't like the odds/ the somewhat modest increase on MH / the price / nor the 4-6 month *supposed* delivery time ....add to that a big 90 day warranty and it just plain don't make a lot of sense..

now 1 year warranty on parts/labor etc .....510mh for say 3,250 usd ...can buy 1 unit at at time for that price and speed ....well that would correspond AT THAT POINT IN TIME 6 MONTHS FROM NOW SAY nicely with what you can get on ebay now for big assed miners..but again 90 day warranty for 5k or so of equip at these speeds /watts ...its a 'lead duck' imho


   

 Sfards had that SF100 unit, but it appears to have been a very poor board-level design AS USUAL from the successor to Gridseed.

 Innosilicon A4 - sometime next year likely first half.
 Alcheminer is also working on a next-gen Scrypt chip, but doesn't seem nearly as well funded.


 
 I've been saying all along that probability of RoI on the S7 was pretty much zero unless your electric was VERY VERY cheap.
 Bad timing on the release and WAY too high of a price.

 Avalon 6, even worse - lower performance for MORE money.

 B-Eleven - going to have to be VERY cheap to have a prayer.

 There is hope on the horizon though.....

 Well KNC it is looking like 20mw is going up soon?

http://bitcoinist.net/new-kncminer-data-center-will-running-within-four-weeks/


 Also bitfury 40mw up this coming week?

 http://www.businesswire.com/news/home/20151211005837/en/BitFury-Launch-Energy-Efficient-Immersion-Cooling-Data


 Someone should have come out with bigger units in scrypt imho before this ..they would have killed.....as an exmple difficulty of the network is up like 21% or some
such from when I started with Titan (2nd titan march 2015) orig anyway nov 8th 2014.....

 BTC difficulty on the network has gone up 23% or some such since Oct 1st ....(did this this the other day these are ballpark) for the btc network

So man me choosing (lucky) to switch my Neptune btc miner this last spring to a knc titan scrypt miner (and march used one) paid off and continues to do so

but BTC difficulty wise......imho ...it would still be 'chancy' but get a used scrypt unit even at ebay prices..mine LTC convert when the ratio to btc is around 0.01 ltc to btc
as I've done...the odds are damn long with that now trying to get over priced scrypt equipment..but the odds are better (as a home miner for sure or even a small data hall operation of btc etc) then when these data halls hit (the ones we know about ...god knows what is going on in china data halls)

Just saying.....when the Elephants decide to get on the dance floor and do the Polka ..its time for us Mice to frigging scatter away and get some fruit punch!

I see no way to get a home or even modest A7 10 pack for BTC miners with frigging almost FREE electric keeping up with these data halls (the ones we know about above and probably the ones we don't China) all hitting within the next 3 months...looks like a perfect storm imho...gonna be very very ugly...the only thing the could (I hope) make me eat my words would be a MASSIVE BTC  price spike say up to $600 bucks due to adoption/china etc...then maybe ..just maybe ......

anyway hope I'm wrong but if I won the lottery and had the itch to get a data hall type asic operation going..it would not be in BTC at this time nor even looking at 6 month time line

well we will see ...just saying (not that I know squat at one time drank the bfl kool aid..got my refund 1 yr 20 days later due to luck...but consider that as the author of the msg posted here)

Just in time when the new batches of S7s shipped. Too bad, this is gonna take longer to ROI.

IMHO, we are riding in a car (at 0.25J/Gh) while an airplane just flew over.
Someone is deploying tremendous amounts and we (amateur miners) are falling by the wayside at 20% an adjustment period.
Nobody sells us a competitive product, unfortunately. There is no Intel equivalent in bitcoin.
If Bitfury is anywhere close to deploying up to 100mW at 0.06 or even 0.08J/Gh, it means about 1exahash (1000 petahashes) just from one company.
http://insidebitcoins.com/news/bitfury-ceo-talks-bip-100-and-announces-16nm-bitcoin-mining-asic-chip-tapeout/34664

I think that i bought my last largish miner.
It is kind of sad to realize that the era is almost over, apart from fun mining with sticks, pods, etc.

EDIT: yep, it is coming..
https://twitter.com/bitfurygeorge/status/677132522074017794

Some comments.  

1) The 50 Megawatts is presumably the amount of power available to the Bitfury site. In practice electrical systems are almost never run more than 80% loaded. There also need to be power available for fans, lights and so on.  So I suspect a 50 Megawatts site would be able to run say 30 Megawatts of active miners max.

2) The network hash rate over short periods, saw one day or two, can only be very approximately estimated.  The estimation is just based on the block solve rate.  There can be statistically variation swings of 10-20% in hash rates over one day and even the 2016 block adjustment each fortnight is only accurate to a couple of percent.

For the technically mined Bitcoin blocks are found every 600 second on average with a poisson distribution. The Standard Deviation over N blocks in seconds is 600/Sqrt(N).  The greater the number of blocks used the more accurate the network hash estimate.

3)  KNC and Bitfury 16nm chips running under 0.1 J/GH seems totally believable. I am surprised it is questioned. It is the sort of improvement you should expect with the move from 28nm (or 20nm) to 16nm chips.  The companies have no reason to lie about their chip power consumption as it is very easy for potential bulk purchasers or investors to check.

Another example, the 28nm Pickaxe ASICs chips used in the SP50 claim 0.15W/GH/s.  So I would be very surprised if 16nm cannot do considerably better.

The mystery spike was about 200 ph if your numbers are try they did not use all they can do.

So we will soon know if we do 200 or 300 ph

OK, so I get the "BitFury testing" hypothesis. But then why did it "go away"? Unless there was some kind of huge failure, wouldn't you let it run for the most part and fix what was broken? Even if things came in too high in terms of J/GH, or maybe the immersion cooling really won't do what's expected, why not "throttle back"  in terms of frequency or whatever and still hash with what you have? I could understand testing prior to shipment of a product, but it's all internal to BitFury isn't it? It's not like ByFury has to collect a 100PH monster order and test it prior to shipment, do they?

I also take the coindesk articles with a grain of salt. They are derived entirely by what KNC or BitFury tell them. They have no way way to independently verify anything they write about KNC or BitFury, do they? We've all seen ASIC manufactures come up short on their efficiency estimates before. I see no reason to think that KNC and BitFury are immune to that possibility.

It is likely new Bitfury chips run well below .1 watt a gh (or Joules/Gigahash).

Their press release and site claims "16nm chip that is expected to achieve 0.06 J/GH this year.".  See http://www.coindesk.com/bitfury-completion-16nm-bitcoin-mining-asic/

KNC claim 0.07 J/GH. See http://www.coindesk.com/kncminer-deploys-next-generation-16nm-bitcoin-asic/

So it is safe to assume the latest miners run below 0.1 J/GH.   Or 2 or 3 times more energy efficiently that an S7.