• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Analysis: Mt Gox uncertainty & Where to put miners & shipped scrypt ASICs

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

I.M.O.G.

Glorious Leader
Joined
Nov 12, 2002
Location
Rootstown, OH
Analysis: Mt Gox uncertainty & Where to put miners & shipped scrypt ASICs

Many of you are on automated pools and don't concern yourself with where to put your miners, however I figured it'd be worth sharing some of my current thoughts anyway. Also, Mt Gox is a good topic to be aware of right now as it could be seen as the next most likely drastic change in the markets.

I've recently done some mining tests involving, LTC, Doge, and PTS. My experience with Doge mining was very positive (profitable) due to timing (low difficulty and exceptionally good exchange rates) - I discussed it here. My experience with PTS mining was alright due to timing (slightly more profitable than LTC as it was struggling, a good amount less profitable than Doge because it was surging) - I discussed it here.

So currently, LTC mining is in position to become much more competitive due to a decent exchange rate and decreasing difficulty, while Doge is looking like it might become less dominant due to a decrease in exchange rate and slightly higher difficulty (both LTC and Doge look like they are could get into alignment around $20/daily with 2.2MH/s hashrate - this assumes no large market movement boom/crash, which is a BIG if)... The two coming together in the middle should be expected regardless though as the gap was large and it should naturally even out.

Meanwhile however, PTS exchange rate is hanging out around .0137 PTS/BTC, and it remains relatively stable - which makes the picture for PTS brighter than it was on the 24th when I did my last test (the exchange rate had dipped at the time), and it could likely be the best place to put my 2MH/s of CPU miners at ypool, as my estimates put income at around $25/daily (about $1.25/4770K).

That concludes my recent analysis of where to put miners to maximize returns.

Gox... Hm. It continues to detach further from other BTC markets. It's current spread is $160, and I've seen it as high as $190 personally - I'm not sure where the spread has peaked. Another problem with Gox is volume - no one can get USD out, and people are giving up en masse. So the only way to get USD out of Gox is to pay the $160+ spread to buy into BTC and send the BTC somewhere that USD withdraws are possible. Due to these things, many people think Gox will fail because in this situation no one wants to have USD on Gox or put USD into Gox. If Gox fails, it will hurt people there hugely, but worse yet will be the mass media news as many mainstream places still view Gox as a market leader despite what we in the trenches know - Gox and its inflated prices are irrelevant, and getting more so everyday. But the bad news in the media is likely to have big impact on the markets regardless.

I'm not in Fiat now, holdings are 40/60 BTC/LTC. But I feel like I should go Fiat, in light of the Gox uncertainty and the threat it poses - but there's just no way to know when things will hit a critical point, and unless the market does crash, being in Fiat doesn't look good. If BTC goes, LTC tends to follow so holding either is likely to play out about the same... Not sure what to do, but I'm hoping I make the right decisions at the right time. If not, plan B is to just hold, and after things pass over I should come out even... Ideally though I hope to be in Fiat at the right time, the crash comes, and I buy back in on a ton of cheap coin.

So finally, no commentary here, but Dualminer (70KH/s, 2.5W, $99) and DigInForce (300KH/s, 8W, $1300) have shipped scrypt ASICs. I don't have any more information than that, but they have shipped product - I don't believe they are ready to ship in mass quantity, but reports are that working miners are actually in the wild but haven't been reviewed/confirmed yet. DigInForce doesn't look any good on paper as its just too expensive, however Dualminer could look alright on paper - Get 1000 of them, and you could do 70,000KH/s for $100,000 at only 2.5KW. This wouldn't be unreasonable for a small time institutional/business miner.
 
Last edited:
I don't know how long you have been into bitcoin litecoin etc., but, Mt. Gox's downfall has been "coming soon" for at least a year. Though, I haven't looked at recent analysis or speculation. I don't know who uses it, perhaps larger sellers that can arrange a special withdrawal method (in person, etc.) I think there is more going on there, than what is outwardly visible.

I think litecoin value will come up more in the future. I prefer xpm to pts - I think it is kind of tied to ppc and will surge in the future. I mine it and hold for now, though only on one cpu.
 
I've been following casually for a bit, but I'm mostly basing my current concern with Mt Gox on the growing spread... Since I started watching closely in November, the spread has become a much bigger concern, while the volume has gotten much worse as well. Historically people had bashed on Mt Gox because their user support was terrible, and because you couldn't withdraw USD easily - but things would go through eventually, and even if they didn't it wasn't a big deal to buy into BTC for an extra $50/coin or so and transfer to a different exchange. However when you have to buy into BTC for an extra $200/coin, to me that makes it clear its a growing problem to a point that won't be sustainable... The unknown is what is behind the problem, and that's the real issue, but it doesn't really matter - the behavior of the market in separation from other markets is the smoking gun, we just don't know the crime.

I haven't looked closely at XPM, because it wasn't competitive on current profitability against PTS. One or the other may be better in the long run, kind of anyone's guess - but right now I'm taking an immediate profitability approach, trying to stay on top of whatever fairly well-known coin shows advantage for a matter of days at least and exchanges well to increase BTC/LTC holdings.

I think LTC will come up more in the future as well, however I'm not sure if that will be before or after it goes takes another dump. (same goes for BTC)
 
dualminer does look decent, but still, 10 of them will give same khs as a single 280x and cost nearly 3 times .. not agreeable to me, id rather have gpus, a bit more on power, but also a good device in hands which i can sell if worst comes to worst
rather buy a sha asic box lol

gox is a scary thought, but that exchange should be wiped out, it is nothing but trouble, just hope i know when :)

about cpu cpins, xpm is still better, used to be they were close by, now its better thanks to a lighter diff increase, imo that is
 
dualminer does look decent, but still, 10 of them will give same khs as a single 280x and cost nearly 3 times .. not agreeable to me, id rather have gpus, a bit more on power, but also a good device in hands which i can sell if worst comes to worst
rather buy a sha asic box lol

I think that's the most reasonable way to look at it from a personal mining standpoint. GPUs are undoubtedly the best option for a personal miner right now and for the foreseeable future.

However I think there is a bigger message this sends, that many people may not appreciate - ASICs that don't make sense for personal miners to buy can have less obvious consequences.

Here's an example to demonstrate:

Once it is possible to receive a large shipment of Dualminer 1st gen ASICs, a business (or a rich guy) could get 70MH/s for $100/each totalling $100K at only 2.5KWh of power (likely with much better hashrate per sqft than GPUs). If that business pays $.12/KWh for power, then it would cost them $216/month to run ($2592/year).

Without the ASIC option, to get 70MH/s they would need about 100 280X's for $400/each totaling $40,000 at 22,500KWh (without adding in the power from non-GPU components, not adding in the cost of mobos/CPUs/powersupplies/cases/racks/cables, and ignoring the cost of dissipating the heat output - considerable stuff even for a small scale business farm). If that business pays $.12KWh for power, then it would cost them $1944/month to run ($23,652/year).

Bottom line - ASICs open up the opportunity for businesses to go BIG in scrypt mining with huge advantages of economy of scale. ASIC solutions can be setup in a form factor that optimizes hashrate per sqft, they directly save thousands of dollars per month in power costs, they cost less to setup compared to GPU solutions that require many other components, and they cost less in facilities charges (Cooling, power to run the cooling, cabling, etc).

At the current market rate and difficulty, 70MH/s would earn $184K/year - this would vary wildly based on difficulty and exchange rates of course, but it would be a very reasonable bet on a phenomenal rate of return on investment.

Also, this is an arbitrary 70MH/s example with a meager $100K investment... A $500K or multi-million dollar investment would be far more realistic for a business investment. 70MH/s with 20-25 rigs with 4/5x GPUs wouldn't be that complex, take up that much space, etc... But 100-200 would present much greater challenges. For a $1 Million investment, you could get 700MH/s running 10,000 ASIC chips, which isn't out of the realm of possibility, but running 200 PCs with 5x 280X's each is pretty far fetched in regards to what is possible - plus the facility/power cost gets really insane at that point.

When that happens, individuals getting 700KH/s on a single card will quit mining, because it costs them money. The businesses will all race to get as much hash rate as possible, as the biggest and fastest are the ones who will get the return on investment while many of those slower and smaller will lose money... This is why, virtually overnight, BTC became unprofitable to mine without ASICs. And the same will be true for Litecoin, once a solution makes sense for businesses...

I don't know if Dualminer is good enough to trigger this, and it doesn't matter until they can ship in large quantities (1000/10,000/100,000), but when the time does get ripe it will be the end of our scrypt fun. :-/
 
Last edited:
but then someone will just make a new gpu only type of coin.

God let's hope so.

I'm still the smallest miner (Kh wise) on OCF from what I gather. It would not take much to knock me off the mining bandwagon... :(
 
but then someone will just make a new gpu only type of coin.

This is quite possible. Hard to get a new GPU coin to catch on, and also possible for litecoin/doge or other scrypt coins to require more than 128KB per thread which would further deter ASICs that would need redesigned.

I'm hopeful GPU mining will remain a thing into 2015. Its fun and makes money. :) A lot like folding or seti, but rather than donating to power/time to their projects, by helping others here with mining we donate to them in a way.
 
God let's hope so.

I'm still the smallest miner (Kh wise) on OCF from what I gather. It would not take much to knock me off the mining bandwagon... :(

You are no smaller than me... I just have borrowed KH/s which I won't have in a month or two. I look forward to being right there with you though, and I'll still be scratching along for any coin I can get. :)
 
This is quite possible. Hard to get a new GPU coin to catch on, and also possible for litecoin/doge or other scrypt coins to require more than 128KB per thread which would further deter ASICs that would need redesigned.

I'm hopeful GPU mining will remain a thing into 2015. Its fun and makes money. :) A lot like folding or seti, but rather than donating to power/time to their projects, by helping others here with mining we donate to them in a way.

maybe hard to get a new gpu coin to catch on, but if everyone who was gpu mining suddenly it isnt worth it for them anymore, thats a lot of people who can make a new gpu coin catch on!
 
So finally, no commentary here, but Dualminer (70KH/s, 2.5W, $99) and DigInForce (300KH/s, 8W, $1300) have shipped scrypt ASICs. I don't have any more information than that, but they have shipped product - I don't believe they are ready to ship in mass quantity, but reports are that working miners are actually in the wild but haven't been reviewed/confirmed yet. DigInForce doesn't look any good on paper as its just too expensive, however Dualminer could look alright on paper - Get 1000 of them, and you could do 70,000KH/s for $100,000 at only 2.5KW. This wouldn't be unreasonable for a small time institutional/business miner.

Viper

>=25000kh/s @ <=300w for around 9000 USD + shipping IIRC, This I am watching to see if it actually happens.
 
You are no smaller than me... I just have borrowed KH/s which I won't have in a month or two. I look forward to being right there with you though, and I'll still be scratching along for any coin I can get. :)
:grouphug: :D
Viper

>=25000kh/s @ <=300w for around 9000 USD + shipping IIRC, This I am watching to see if it actually happens.

Let's (I'm) hoping this is just fantasy...... :-/
 
face it. if scrypt asics come out en masse, someone/everyone wont go and dump all their mining hardware, especially if they spend a good amount of time/money on them. There will be a new algo , that most/all miners move to. then lets hope LTC prices skyrocket. :)
 
Viper

>=25000kh/s @ <=300w for around 9000 USD + shipping IIRC, This I am watching to see if it actually happens.

Worth mentioning, so I appreciate that... However the other two I noted are actually real physical things right now. From the specs, it sounds like the viper isn't targeted as a first to market product... I have read the alpha tech stuff some, so I believe it will happen, but not that it has shipped a working product to anyone yet (dualminer and diginforce have reportedly).
 
face it. if scrypt asics come out en masse, someone/everyone wont go and dump all their mining hardware, especially if they spend a good amount of time/money on them. There will be a new algo , that most/all miners move to. then lets hope LTC prices skyrocket. :)

+1 for sky high LTC prices :)
 
Viper

>=25000kh/s @ <=300w for around 9000 USD + shipping IIRC, This I am watching to see if it actually happens.

take this with a grain of salt. they are projecting 3x the power efficiency and 4x the cost efficiency of the DualMiner ASIC. im hard pressed to believe that.

from what i had seen, they are only projecting those numbers based on very small scale. (like the performance of a single chip). im not inclined to believe that it will scale linearly.
 
Back