I'll BET it has 2304 CUDA cores or if it gets priced at $449 or $459 MSRP - otherwise it's gonna eat too many 1080 sales.
If it IS in fact priced around $400 (which is the SAME or nearly so as the 1070) it'll have to drop the core count down to 2048 to avoid eating 1070 sales.

 Card is about 2 months too late to be a good ETH mining card - RX series AMD have dropped back down too far on price - but it might be a very good ZEC miner for the price.

 $250 represents quite a bit more than one month of GRC income for me at current pricing.

 2000 GRC (somewhat less than 2 WEEKS of my current income) gets me a state right now in slightly less than 2 weeks based on recent staking intervals.
 For someone with 80 magnitude of production (which would put them very close to or IN the Top 200 Gridcoin producers), they would have to keep over 3 MONTHS of income around to stake every other week.

 For smaller folks, this a totally INSANE requirement and makes Gridcoin pretty much WORTHLESS to them.

 I don't know of ANY other "project out there" that ties up so much of your earnings just to get paid your non-staking earnings - in fact, I'm not away of ANY other project that does so except for STAKING SPECIFIC type coins that don't have anything resembling a "proof of work" requirement.

I would expect they are using a mining-specific Ubuntu variant and booting via PXE with no drive at all - but that's really not a needed option for a small farm.

 Ubuntu is perfectly capable of being installed to a USB key with no need for Windows at all.
 The way I set that up is only ONE option - I created a "base" install on a HD, with a total partition size of small enough to fit on a USB key (don't use GPD/UEFI large drive support, use the old style partitions) - reset FSTAB to use STANDARD partion names instead of that stupid UUID garbage, then used DD to copy the entire drive over (usually "dd if=/dev/sda of=/dev/sdb bs=16M conv=notrunc,noerror" for the command line on my machines but the specific drive naming might vary).

 As long as the motherboard puts the GPUs at the same PCI-E addresses, you can also use this to CLONE an existing setup.
 If you use static IP addressing, you will then need to change that address on the "cloned" machine to it's own address.
 It's best to change the hosname and the "worker name" your mining program uses so you can keep track of each machine seprately.

 If you close to a motherboard that does NOT address the GPUs to the same PCI-E addresses as on the original machine, you'll also need to update xorg.conf with the correct PCI-E addresses.
 You may need to modify whatever startup scrypting file you use to set overclock/fanspeed/TDP limit for the right card addressing (if you use all the SAME model card, this can be ignored).

 For a pure mining rig, you PROBABLY can cram all you need onto an 8GB pen drive - but since Claymore at least wants a minimum of 16GB virtual memory (even when you have more RAM in the machine, his software STILL insists on that) best to figure on a 32 GB pen drive and 17GB or so of it allocated to swap space.


 For reference, one of my FOLDING rigs (folding has a good bit higher requirements for CPU and MB support than mining does) was running on a pen drive - booted up a lot slower but the actual PPD once the machine was running was almost identical to another machine I had with the same hardware running on an actual HD.

GTX 1080 hasn't made $3 a day gross for a month or so on a reliable basis no matter WHAT you are doing on it. 2 months ago, MAYBE if you were doing merged folding.

GTX 1080ti has been very close to that even the last few days, with proper tuning for max hashrate instead of tuning it for max efficiency (which costs net income if you are in a low electric cost area like where I am at).
https://whattomine.com/coins?utf8=%E2%9C%93&adapt_q_280x=3&adapt_q_380=0&adapt_q_fury=0&adapt_q_470=02&adapt_q_480=0&adapt_q_570=0&adapt_q_580=0&adapt_q_750Ti=0&adapt_q_10606=0&adapt_q_1070=19&adapt_q_1080=6&adapt_q_1080Ti=1&eth=true&factor%5Beth_hr%5D=35.0&factor%5Beth_p%5D=140.0&grof=true&factor%5Bgro_hr%5D=58.0&factor%5Bgro_p%5D=210.0&x11gf=true&factor%5Bx11g_hr%5D=19.5&factor%5Bx11g_p%5D=170.0&cn=true&factor%5Bcn_hr%5D=830.0&factor%5Bcn_p%5D=140.0&eq=true&factor%5Beq_hr%5D=760.0&factor%5Beq_p%5D=190.0&lre=true&factor%5Blrev2_hr%5D=64000.0&factor%5Blrev2_p%5D=190.0&ns=true&factor%5Bns_hr%5D=1400.0&factor%5Bns_p%5D=190.0&lbry=true&factor%5Blbry_hr%5D=460.0&factor%5Blbry_p%5D=190.0&bk2bf=true&factor%5Bbk2b_hr%5D=2800.0&factor%5Bbk2b_p%5D=190.0&bk14=true&factor%5Bbk14_hr%5D=4350.0&factor%5Bbk14_p%5D=210.0&pas=true&factor%5Bpas_hr%5D=1700.0&factor%5Bpas_p%5D=210.0&skh=true&factor%5Bskh_hr%5D=47.5&factor%5Bskh_p%5D=190.0&factor%5Bl2z_hr%5D=420.0&factor%5Bl2z_p%5D=300.0&factor%5Bcost%5D=0.046&sort=Profitability24&volume=0&revenue=24h&factor%5Bexchanges%5D%5B%5D=&factor%5Bexchanges%5D%5B%5D=bittrex&dataset=&commit=Calculate

 Shows $3.11 gross and very close to $3 net after electric for yesterday on my Aorus, for a specific example.

 And no, I'm NOT using Nicehash figures for that, I'm using the actual amounts I've earned with the cards on what I've actually been DOING on them (currently ZEC mining).

 HDD mining is profitable - just that the profit level is VERY LOW compared to the investment vs GPU mining, unless you only do HDD mining on HDs you were going to get ANYWAY for new rigs so that the actual investment cost is low or zero.

 If you are in an area where electric cost is in the 20 cent/kwh or higher (large parts of all of California by report, for example), HDD mining might be your ONLY profitable option and has a much higher probability of being your highest return per $ invested option even if GPU mining IS profitable.

In my experience, XP was MORE stable than 10 is - once it got enough service packs under the hood.

 Still wasn't anywhere close to 2000 stability, much less NT4 SP6 or NT 3.51 SP5.

 For that matter, *7* seems to have turned out eventually to be able to match or SLIGHTLY exceed XP stability - and 10 is nowhere NEAR 7's poor stability level.

 Then there was the combination of Windows for Workgroups 3.11 with QEMM 7.54 - which wasn't NT level stability but could argue with 2k and definitely beat 10.


 Still waiting on a "how long they have stayed up at a time" mention.

 Perfect uptime for a couple months at a time is NOTHING.

 There is no point to buying a managed switch for a mining situation (unless you get a used one DIRT cheap).

 I've been thinking about getting some of the Quanta Windmill servers they have on surplus.

 As long as my A2 units stay profitable though, I'm not in a rush - don't have a lot of 220 available in the current space and the A2 make more per watt used at this point.

 100% uptime over how many years?
 "100% uptime" by itself is a MEANINGLESS claim.

 I have owned and maintained a LOT of LINUX machines (some Ubuntu but mostly Slackware and some Redhat) that achieved uptime measured in YEARS between power outages, physical moves of the machines, HARDWARE failures (usually the hard drive), and such - and some of them might have exceeded a DECADE had I stayed put in the same place long enough and had no power outages that exceeded the runtime of the UPS some of those systems were on.
 Most of these have been doing Crypto work (keep in mind that at it's root, Cryptocoin Mining is crypto work) on the distributed.net RC5-72 project.
 This is the SAME sort of workload Cryptocoin mining puts on a GPU, and very similar stress levels - the Dnet client used to be widely-used AS a "stress test" tool by review websites for both CPUs and GPUs.

 I have seen exactly ONE Windows machine (it was running NT 3.51 SP 5) that managed ONE year of uptime under the same conditions - and nothing since that could even get to the one year mark without a crash.


 Win 10? I gave up on it entirely due to the crashyness - even on the one mining machine I tried it on it never managed one MONTH of uptime without a bluescreen, and that was running the SAME AMD driver version (15.12) I used on the same hardware under LINUX with long uptime.
 Don't get me started on the "crashy" level of Win10 on my gaming rig - it was SO bad I reverted to Win7 despite the performance hit of "no more DirectX 12" in the highest-end game I usually play.
 I can't remember if I ever had more than a 3 day period without a crash - no matter WHAT driver version I tried.


 I do agree that Windows is more user friendly (and that's said as a 25+ year LINUX veteran), but it's not even in the same COUNTRY when it comes to stability, much less the same ballpark.


 As far as I know, ALL of the "mining specific OS" are based on Ubuntu.

 I didn't say it was comparable - you're arguing MY case that it's not - unless you have a VERY HIGH electric cost that makes GPU mining prohibitively low on profitability.


 With that said, the nitpicking 9-)

 You DID forget to factor in the electric usage of the hard drives, 120 TB of Seagate Externals = 16 drives (they FORMAT to 7.27 GB not to 8GB on both the Archive "Backup" version and on the Compute "Expansion" version) probably average 8-10 watts each (peak usage is about 12 watts per drive as I recall the specs but they're not accessing most of the time which will drop that - but the platters keep spinning which uses a lot of the power they consume on 12VDC).
 128 watts at MY power cost however is only pennies a day - but it IS barely into the double-digit pennies a day - while 160 would still be under a quarter per day FOR ME.

 BURST calculators are notorious for being optimistic, which seems to be due to how the "network TB" figure is calculated.
 Pools seem to only count the "machine that most recently submitted a share to an account" on the total TB of that account, there are times the pool I uses says I have well under 1 TB of hard drive mining when my total is well over 20TB.
 In my exprience, you need to DISCOUNT the "estimated return" any BUST calculator comes up with to a little over HALF, perhaps as high as 60% of it's estimate, to come up with a REAL figure of any accuracy.

 You also should keep in mind that electric COST varies a lot - mine is some of the lowest in the USA, nowhere near the Nicehash 10cents/kwh default - which is a plus for GPU mining where I'm at.
 You probably deducted more for electric usage than I actually have to pay.

 Nicehash also badly underestimates the Equihash (ZEC) hashrate of the GTX 1080 ti at 630 sol/s.
 I'm seeing more like 760-770 right now out of my Aorus on the same EWBF 3.4b miner Nicehash uses in it's recent Legacy versions.
 Oddly enough they DO seem to use 250 watts as the power usage for the card, which is where mine is set at  - a 1080ti pulling 630 sol/s is probably only using 150-180 watts of power.

 Probably because they got used to ETH where it IS needed for max hashrates, and on a FEW cards it might actually help with ZEC hashrates as well.
 ZEC isn't nearly as memory-intensive as ETH though - a lot of my cards could max out ETH hashrate by maxing memory clock but running at stock OR LOWER core clock (which made them more efficient with NO loss of hashrate).

 I am one of the Top 20 research magnitude producers in Gridcoin (top 16 if you exclude the 3 GRCPOOL entries that aren't single producers, and sometimes in recent times Top 15 COUNTING them).
 In order to stake at least MONTHLY on a semi-reliable basis, I have to keep about a month worth of coin in the wallet - and I HATE that they're just sitting there USELESSLY instead of my being able to use them to add more hardware.
 (don't even MENTION "staking income", I get better return on that out of a SAVINGS ACCOUNT out of my credit union - BEFORE I have to factor in the GRC price drop of the last couple-three months).

 For folks with small research numbers, they'd have to keep YEARS of coin on hand to get payouts at all (don't forget the 6 months "lose it if you don't stake by then" limit).

 I probably should move my wallet to a LINUX machine - though the Win7 machine it's on now is fairly reliable, generally only needs a reboot every couple months or so since I don't DO a lot on it.
 However, I went with NO wallet crash/reboots for MONTHS prior to the recent wallet changes, the last two-three weeks I have been seeing a wallet crash/reboot on average every other DAY.

 Cisco has had a few turkey products over the years mostly in their "consumer-grade" gear, but overall in routers they are generally the best (their "consumer-grade" stuff though isn't nearly as good as their "pro-grade" stuff).
 For switches, they WERE at least matched by 3Com (I'd actually argue the 3Com SWITCHES were better but it was very close) and are still CURRENTLY matched by Netgear pro-level hardware.

 Cisco DOES get argued with on the high end for routers - but it's by other high-end companies most folks have never heard of because they ONLY make high-end gear.

 Actually, on Monero the highest-hashrate stuff is still high-end CPUs like the AMD Epyc, Threadripper, and some of the top-end Intel many-core server CPUs.

 The only GPU that can get into the 1000 hash/sec range is the Vega, per ONE report from ONE person - everything else that manages over 1000 hash/sec is a CPU - and the highest single-CPU report I've seen to date was an Epyc pulling over 3000 hash/sec vs the Vega at 1800.


 The main issue is that nothing in cryptography that is in widespread usage uses anything but integer operations - which GPUs are perfectly capable of doing JUST as efficiently per clock cycle as any CPU can, but have a lot more "cores" to do the work with than any CPU does.
 Even though GPUs clock a lot lower than CPUs (as a general rule on same-timeframe hardware), they have so many more "cores" that they can do a lot more work in total on anything that can be processed in a widely parallel way of work.

 This also applies to ASIC hardware - the only reason you don't see a lot more algorithms with ASIC available is that most of the applicable coins would not generate enough income to make an ASIC unit able to pay off in a reasonable timeframe, coupled with the VERY HIGH COST to design, debug, and put an ASIC chip for mining into production that would be able to beat GPUs by a wide enough margin to matter.

 This is the main reason there is no ASIC hardware for ETH mining - the uncertainty about "when is it going to go POS" has kept ASIC making companies uncertain that their miners would SELL - even though ETH would certainly make enough income to support ASIC if one COULD count on it being mineable for long enough.
 The other reason is that ETH is hard on memory access - and ASIC *CAN'T* do much if any better on that than GPUs can do.
 Wider RAM doesn't always help on that - otherwise the Fury and the Vega lines would be by far dominate ETH mining, instead of being high-end but NOT crazy-outstanding.

 Newegg - but I'm in the US, you EU types seem to get shafted on pricing due to your generally very high taxes like VAT.


 My Aorus in the gaming machine runs at 65-66C under full ZEC mining load, not "about 70" - but it's in a case with good cooling in an air-conditioned environment.
 I DO have it set up for 100% TDP, a mild core overclock, and my usual fan profile in Afterburner.

 The Zotac 1080 and 1070 mini cards run reasonably cool, but they're at a lot lower TDP than a 1080ti has to deal with and all 3 models appear to use the SAME identical cooler.
 I'd expect the 1080ti model to run quite a bit hotter.

ZEN is sometimes a little more profitable than ZEC - used the same mining software, just change the pool you point at and the payout address.

 It's similar to the ETH / ETC divide - same algorithm, different blockchains.

 GPU mining WAS profitable in all of those months - if you had low enough electric costs.

 There is nothing inherently wrong with buying via eBay - Newegg in particular sometimes offers better deals there than on their own website (though you can sometimes browbeat them into MATCHING the deals via their own website via their "price match" policy).


 ZEC is only one option - I spent a few MONTHS making quite a bit more (about DOUBLE at some points) on the Nvidia 10xx based part of my farm working "merge folding" of Curecoin and FoldingCoin than those cards would have made on ZCash.
 There have been other times that they made more on other stuff like DGB-Groetsl or DGB-Skein and GameCoin, or even the ZCash spinoffs like ZEN.

 Don't lock yourself into a single coin and refuse to consider anything else.


 The profitability in June, July, and August was quite a bit higher than now - you're NOT going to see GPU rigs 75% paid off over the NEXT 3 months unless we see another major price spike - but if you are willing to tolerate much longer payoff timeframes AND have cheap enough electric, you'll pay off EVENTUALLY.

 Wierd card then, EVERY card I've worked with that had 2 connectors would flat out refuse to run FOR SAFETY TO THE POWER REGULATION CIRCUITRY if it only has one of the 2 connectors hooked up properly.

 Don't try to mine on a laptop.
 They are NOT designed for 24/7 operation at high load levels like mining puts on a machine.

 The only VIABLE option is to use your laptop as a USB hub controller for a USB-based miner, as that doesn't put a lot of load on the laptop itself.

 EVGA G2 1300 have 8 GPU connectors in stock form - all 6+2 - and if you get some extra of their "2 GPU connector" cables they can get to 12.

 Dunno where you are getting that "they only have 6 connectors" thing from, unless you think the PS end of the cables is the same as the GPU end (they are NOT).

 Not per card, but on an entire 4 card rig it seems LIKELY as that would only be a bit over 10% of TDP per card.

Dual mining DOES eat noticeably more power, as it is designed to stress more of the card than single-algo mining does.


 Nicehash isn't always "throwing your money away" - sometimes it's MORE profitable due to how their market works.