Erm. This thread, with its 'internet engineers' and 'self-proclaimed hardware alpha testers' turning out to be *14* yrs old is making me feel rather the dinosaur.
However, whilst it's historically made more sense to sell spades to miners than to dig oneself, I've always been attracted to the shiny stuff. The great thing about cryptocurrencies is that they are intangible, so instead of getting all Sméagol about some chunks of polished high-density pretty-coloured metallic elements, it's actually more fun to assemble a big rack (erm, well only 25 of them) of little circuit boards and watch the blinkenlights.
Ahem. I'm sure I'll get a YouTube style 14-word-max 'witty' putdown by the kids for writing complicated multi-clause sentences but, hell, blame it on multiple generational differences.
On to the sausage. I've always just merge-mined and traded the LTC for BTC. I ran a VERY small bitcoin mining operation - 5 GH/s worth of Ztex FPGAs, and at max 7.5 GH/s of various GPUs - my frame rigs have always been entertaining builds but 5 kW power consumption 24/7/365 in south east England didn't make much sense at £7 bitcoins. So the FPGAs are still chopping away, and I've let the GPUs fade out through natural wastage (running overclocked at max fan speed since 2011 is 'survival of the fittest' at the extremes).
I've just checked out the bitcoin price and it's £102.
err wut?
Now I've got money to invest in this, but after over £5k worth of FPGAs (and too much time designing and integrating a thermodynamically efficient assembly and power / control harness - though it was ACE fun), I decided to hold off and wait for the ASICs. Well, having lost quite a bit of money to 'bASIC Tom' (and I'm a consultant to the investment management industry in the 'real' world - having a track record delivering his FPGA product, compared to BFL's zero coupon, priced at par, junk bond 'pre-order' - I'd say that my due diligence was appropriately scaled for the investment and was genuinely surprised when he lost it...), I'm now resigned to having picked the wrong horse and having missed the boat, as well as liberally mixing metaphors right left and centre.
For the youtubers - I'm still keen to maintain my tiny mining operation, but picked the wrong ASIC, and giving money to someone like BFL (which failed my risk analysis first time round) is guaranteed to leave me as a 'late adopter' and making no profit. There are a lot of screaming creditors who have already given BFL their money, and it's only fair that *if* BFL deliver an ASIC then they should get their units first. The evidence trickling out suggests that BFL only have a certain number of viable chips actually manufactured, so the number of working *devices* will be limited. It is very possible that BFL already know they can't fulfil all the orders on the books.
With the smaller ASICs claiming a few tens of gigahash *per device*, other mining businesses who *do* get an ASIC or ten will be sitting on hundreds to thousands of GH/sec - my 5 GH/s of FPGAs will still be efficient, and pure profit, but just bugger all pure profit. In order to accumulate 1000 GH/s worth of my FPGAs, I'm looking at £100,000 capital investment and probably a couple of months just to put them all together - so that's the same again in lost consultancy (obviously it'd make sense to outsource, but building mad electronic assemblies is a bit like 'do-it-yourself supercomputer' assembly and really good fun
)...
So what to do? Bitcoin is getting expensive now - and whilst the Litecoin 'silver to Bitcoin's gold' didn't make much sense when you needed 140,000 LTC to buy a slice of pizza, it is looking a damn sight more reasonable now that you can get a gram of reasonable cocaine for less than a single bitcoin on the Silk Road these days (not that I'm a connoisseur - it's just a high-price product that has always cost multiple BTC, until recently...). Taking SR as an example, their website is geared up to pricing product in Bitcoins and Bitcents. This was never a problem, but now that the smallest 0.01 BTC figure seen throughout their merchant system (yeah, I know BTC is more divisible, but I don't know if SR's developers would need significant rewrites and concomitant change management risk to increase precision) is around £1, the cheaper products - even illegal drugs - are getting close to the minimum visible granularity. It's already annoying that postal charges are cheaper than the minimum the vendor can charge, etc. etc. etc.
Let's say one *did* want drugs. In general, a retail customer wants to buy *small* quantities if mail-order, since larger quantities risk disproportionate penalties (possession vs. supply, etc. - at least here in the UK). The price of Bitcoin (and transaction costs in and out, if one doesn't mine) encourages purchases of a few BTC, but now that's a few hundred quid's worth of your chosen 'little luxury' and looking like 'distribution', especially for cheaper products.
All the Silk Road needs to do is to bung some more zeros on their web shop, but again there's the peculiar psychological factor here. The majority of what I've written above will be dismissed as irrelevant nonsense by the subset of mathematicians who suffer from narrow-mindedness, because (correctly) everyone concerned could just think in ten thousandths of BTC and there's no problem. But humans don't think like that.
Cryptocurrencies, after using them in the real world (tangible purchases - my first Bitcoin purchase wasn't an online bet or similar intangible or service, but actually a nice Canadian one ounce Silver Maple coin - whereas from most of my mad rambling you'd bet on drugs... heh), become *psychologically* associated as *money*. My BTC hasn't gone on drugs - I bought lots of (metal) coins and my ASIC 'purchase' was also BTC. So for me, even more so. Bitcoin and nice chunky bullion coins seem interchangeable (a whole new topic of discussion!).
And it's *this* psychological aspect that makes LTC possibly viable. Technically, one only needs BTC and that great tool called 'mathematics' to divide them up. But it's getting similar to the idea of buying a nice bottle of wine with a Krugerrand. Let's say it's a nice enough bottle of plonk and the same price as one ounce of silver, a single chunky coin. The same chunky ounce *gold* coin would also buy the wine but would leave 98% of the value of the coin (based on £20 silver / £1000 gold). I could, as per the SR drugs analogy, take my Krugerrand and instead order 50 bottles of the wine, but it'd have to be damn good stuff and last a long time because I don't drink that often.
So, back to psychological economics, it just *feels* right to buy that nice bottle of wine with a integer unit of an appropriately-valued coin, rather than using 49/50 of the value of a unitary coin worth much more. The original 'silver to Bitcoin's gold' does, in fact, make sense from the view of human behaviour (even though technically it's somewhat redundant).
However, for this to work, Litecoin needs to be worth about the same ratio-wise. With BTC at £103 now, one LTC would need to be worth £1 or so - and it's not. So there's a potential early-adopter bandwagon here for those who didn't win the Avalon lottery. Selling LTC 'shovels' could turn out to be a sensible business plan, because investment activity with Bitcoin is already turning to gold-like behaviour (speculation / store of value) leaving a space for *transaction* currency (the silver analogy again, with industrial usage and old coinage).
I grant that BTC could be in a big speculative bubble here, but with competitive devaluation in Western FX markets, some *are* genuinely using BTC as an inflation hedge and a means of capital preservation (albeit a risky one). And as a store of value, money velocity slows, saving is preferred over spending.
So I'd be very interested in the Litecoin mining gear - especially FPGA kit that consumes very little power. I have the risk appetite for this game as I've already shown, but the ASIC situation has me very wary of allocating further capital at Bitcoin mining equipment. I'm bored of GPUs and having my main house sweltering through winter at over 30 deg C with the windows open in winter (!!!!), and whilst a fancy rack of tens of GPUs can be used *right now*, I've also paid the electricity bills for a couple of quarters of constant use of such devices. They have very very substantial operating costs. A fancy FPGA assembly with 25-100 modular boards, whilst fun in terms of geek-appeal, also has the big benefit of not costing significant amounts of money to run.
Finally (phew...) - is there something genuinely innovative and unique to your approach that gives you no competition, so that I should be seriously starting the process of considering allocation in your direction? What is the barrier preventing my existing cluster of Spartan-6 FPGAs being repurposed with cheap plug-in auxiliary boards and a new assembly plus bitstream (I won't play Internet Engineer, I'll simply ask my old college mate who does this for real, at a senior level - but I assume there's *some* hack possible to give additional working memory to my Ztex 1.15x boards... my 1.15d boards already have more resources to begin with though).
Would your boards have fallback value too? Litecoin aren't worth much right now, so if your approach resembles the BTC experience where the FPGA units have around the same performance as a mid-range GPU but consume less than 1/10th the power, but cost more than the GPU up front, as a business case the time horizon is quite far away. This risk becomes more acceptable if the devices have intrinsic value (can be repurposed, or worst case, the chips recycled). Even better would be the ability to mine Bitcoin at low power, when transaction fees are the main mining income from *that* particular currency, if LTC doesn't work out.
I once had a couple of thousand LTC floating around before my pool arranged auto-trade for BTC, and now the pool has stopped merged mining I've renewed my interest in LTC and have set up a few test rigs for 400-500khash. However, with the price of electricity, I want to be 100% FPGA and ASIC by the end of Q2. It's only a small operation after all, and I don't want the significant ongoing cost, heat and noise of GPU rigs.
Seems though I have a choice - stick to Bitcoin, big risk, 'pre-order' nonsense with ASICs... or consider the first application of FPGA tech for the LTC blockchain. The fact that I've already lost on an ASIC 'pre-order' makes the BFL option rather unpalatable. But I'd like to keep mining - on a socioeconomic level, cryptocurrencies are a big deal to me philosophically, and part of me (less rationally) likes the 'silver to BTC gold' analogy of LTC. I'll be keeping an eye on this - when you're ready to take beta-level investors then I'd be interested. I'm no VHDL guru but I ported Stefan's (Ztex) toolchain to Mac OS X 64-bit 10.6.8 a while back, so am competent hacking the software side of things.
Same interest here. Small BTC Miner of Ztex boards (counts 18 Startan's; 3,9GH/s) with the luck not to have invested in false ASIC (i was 99% sure to go with bASIC!) and totally mised the Avalon preorders
