Topic: what's the most efficient mining hardware available? (bachelor thesis) (Read 5813 times)

sorry, yuup - there are 4 blades you can't see in the pic doh. can't count. but, those 4 are not spaced so I could have it running cooler (or will have when i do a bit of feng shui) at night - when I need to close the office room door, so the cat doesn't go investigating, the window is still open and the redundant hard drives take heat from the top shelf, so the psu is more efficient, during the day, office windows are wide and door open to let the lo-vely warmth fill the house (this saves on heating bills). so I'm getting around 148gh thru the night (11pm-6am) and the rest of the time it's around 152-155GH. I remain sure that I can push it to 160GH when I get those 4 blades on copper risers, anyone got any spare please?

  

oh... and it's all sitting on a cold stone slab (an 1800's fireplace stone) - and that is super-chilly, check the PIPOTAC link in my sig to see where i live

With overclocking, you could improve the hashrate a bit, but definitely not that significant.

I'm not sure how the cooling makes 150 GH/s mine faster than 150 GH/s.  At current difficulty (2,193,847,870.17) mining rate should be around 0.000229 BTC per day for each GH/s (based on Eligius pool) so with 150 GH/s you would be clearing €0.016 per minute before power TODAY.  Maybe you haven't re-calculated in a few months.

Well that's a problem, because pre-order hardware is an order of magnitude more efficient than hardware available right now. It's certainly possible to profit from mining even with this huge increase in difficulty, but not if you don't take a risk with pre-orders.

yup, you're not far off. good cooling is the key. open window behind the rig.


there are 3blades from what you can see above - the first of the three is showing ^

How is this supposed to be possible?
Mining Calculator calculates per Day 0.03438524 BTC = $27.42  with 150gh and current difficulty

i've 150gh that's currently mining an average of €0.047 per minute (1/2 - ish)
happy.

Haha. No I'm doing payment rows and calculate the net present value. (including capital costs)

So you plan to forecast profitability on mining an output that fluctuates wildly? You might consider changing topics to analysis of blindfolded darts for picking options strategies. Sorry, I'm being a dick. But seriously I don't get it, as a research paper or a writing piece for an economics class maybe, maaaaybe if you are talking about supply an demand forces, and decentralized markets, secondary markets, etc... but a thesis, never. Good luck if you decide to do it, hopefully your professor isn't anything like me.

Edit: ugh, please, PLEASE tell me you aren't using "ROI" in a forecasting context... ROI is a measure of historical activity, not to be used for forecasting, this place is full of crap information btw and that term is molested here on an hourly basis. This is a good place to talk with tech folks, sadly not many Econ nerds participate here. Feel free to dig through my old posts, on a few occasions I went on a rip about ROI and other misrepresented bullshit, now I'm beat down from trying to spread knowledge that nobody gives a shit about, "ya can't save 'em all!" lol... https://bitcointalksearch.org/topic/m.3699883

+1

Yeah, I noticed that too.
I didn't want to get started on cex.io, which in my opinion is the single largest treat to bitcoin that exist, Santoshi warned about companies like them. 

lol.
Anyway, it is always fun to see guys spamming their cex.io referral links, and then lure newbies to the site.
You can find lots of them here. https://bitcointalksearch.org/topic/for-newbies-is-cexio-a-good-investment-387280

The currently most interesting mining project is DZ coop r17:
https://bitcointalksearch.org/topic/closedr17x-black-arrow-prospero-x-3-dzmc-exclusive-130-40ghs-335145

Shares was sold for $110 per 40GH/s in November, the hardware is due to be delivered in ultimo February.
Back in November it was extremely difficult to predict the difficult for late February.

Now it is possible to make a much more certain prediction, and ROI is definitely possible, even with weeks of delay from the fab.

In case you don't already know here are some great calculators (any miners most important tool):
http://mining.thegenesisblock.com/
http://bitcoinwisdom.com/bitcoin/calculator
http://bitcoinwisdom.com/bitcoin/difficulty

Yes that's in. I explained the hardware evolution from cpu's to asic's and made 2 calculations with an actual asic-device. In the first calculation the device is delivered in time and the miner makes a huge profit, in the second one the device is delivered with 2 months delay and the miner will never reach break even. That's how I show how critical the delivery time is. (my underlying assumptions are: difficulty rises every 11 days about 28% - that's the average from the last half year) (unfortunately I'll have to calculate this new because I calculated with a device from xtrememiners, which is certainly scam.. :-( )


Yeah I'm thinking about something like that too, but actually I don't know how to do this. I'd have to make a lot of uncertain assumptions and therefore it could happen that it becomes not scientific enough for scientific paper.


I'm writing for an audience with economical background and it was really not easy to explain the technical stuff in an easy to understand manner. If someone with technical background and or knowledge about bitcoin read this he will for sure shake his head :-D


This sounds very interesting, I'll work through that, thanks!


Yeah, cex.io is one reason why I think many or most miners don't act rational. If they would the price would go down, but it don't. Another reason are those usb-mining-sticks (usb-miner) which were sold like crazy and the next reason is the insane prices people are paying on Ebay for mining hardware.

Interestingly it is really considerable to buy some GH on cex.io, because if the prices keep on not going down you can sell them after a period and had the mining for free.

Thanks guys for all you replies yet :-)

As for your question on what hardware cost to use - naturally hardware that can be purchased right now isn't such a great deal (or everyone would purchase it until it wasn't available).  The recent bitmain usb chips are priced at $50 for 2 GH/s (overclocked) and cooling would need to be applied and the bitmain S1 is $1700 for 180 GH/s

I don't think using those prices ($10/GH + electricty) are fair for your analysis.  You can find cheaper options with a high chance of delivery if you are willing to wait until April / May - but no one knows what difficulty will be then.  If you were going to buy mining equipment today I think you'd need to get in line on one of this and gamble on the difficulty or continually monitor for the next generation (more efficient) pre-order and again gamble.

I'm interested in how this chapter will fit in - what is the real thesis topic?


In any case for the chapter you should consider doing a review of bitcoin mining profitability in the past since it greatly influenced the current state of the bitcoin economy - especially since it widely distributed coins to those that were willing to turn their cpu and then later gpu cycles to it.  You could discuss the current state where it requires dedicated ASIC (application-specific integrated circuits) to have profitability and the hardware that is profitable generally requires a pre-order gamble to acquire in a timely manner.  Then I suggest you turn an eye towards the future.  The total computational power of the network has grown exponentially - fueled in my opinion somewhat by increased bitcoin value and somewhat by technology that is more efficient - can this continue for the long term?  While moore's law has stunned some by holding true with exponential growth in computing power over more than 40 years - the limited pool of resources for bitcoin (the remaining coins to be minted + transaction fees) will likely mean mining power growth will likely become linear and logarithmic after a time.  But you certainly shouldn't try to predict that timeframe too accurately - the error bars are probably measured in tens of months.  Not to mention the exit of the home miner - soon typical household electric circuits will be sufficient to host the required juice to keep relevant miners powered.

You could look at the cost of such a large bitcoin computation network.  Both the social cost in current electric consumption which has some environmental implications and in terms of how much investment at current prices would be required to increase difficulty.  If everyone adding computation to the network is paying the pre-order retail price of $3 per GH/s how much capital was spent acquiring minng hardware last month?  Is that sustainable?

Lots of great avenues for topics, I hope you've carved out a good one.  I also encourage you to use analogies to relate the at times tediously technical details of bitcoin operation.  Things like What is difficulty?  What is one bitcoin?  Why is it secure?  Analogy is a technique often leaned upon by the great physicists to translate the technical into the everyday - think http://en.wikipedia.org/wiki/Schr%C3%B6dinger%27s_cat

you might want to catch up with this site too: http://btcinfo.co.nf

Writing a theory on rising prices down sounds like an interesting task, I must say. Maybe you should do some more reading.

Ironically, cex.io has a built-in calculator showing that people will never get their investment back (not even close), but it seems no one cares about it.

most irrational.

 

I am sure there are irrational miners out there, but I don't most miners are irrational.
However, I am quite sure it is not just a few single cases.

For example, if you take a look at cex.io, you will find people buying 1 GH/s mining contract for 0.043+ btc.
And, the GH/s price is pretty stable even after the difficulty readjustment.