Not on an exchange = worthless at best, or scam.

 R9 290 thermal limits at 94, on the reference-board design.
 NOT a good idea to let them get that hot though, they're a lot happier for the long term if you can keep them down to around 80.

 290 is still a good ETH mining card, though I'm not sure that's the BEST option.

Even before the last week of price collapse (which I do NOT anticipate recovery from for a while, unless the Chinese Government turns it's apparent current policy around and specifically states "domestic cryptocoin exchanges are in fact legal"), I was anticipating DASH mining profitability collapsing to an extremely lower level by the end of October, and even more so in November and December, JUST from the very large number of Antminer D3 units that have been sold.

 Then add in all of the OTHER folks selling "current generation" miners like the Innosilicon A5, the new Pinidea, the newly-announced Baikal X10, the DM11G and DM22G.....

 I don't *THINK* we're tallking "5 cent electric or lose money" territory by turn of the year, but I am NOT sure that will be wrong.


 (Edit) then the Bitcoin Price Collapse due to China getting "anti-Bitcoin" hits, now I suspect 5 cent electric might not be enough for a D3 to be profitable by end of December.

Interesting that they have the 19 PCI-E slots but specify that they support "up to 16" cards.

 I do like the multi power supply connections, should make coordinating the multiple power supplys such a high GPU count rig would need trivial.

 If I ever started building riser rigs, I'd likely go with this board JUST for that reason.

I wish folks would do a SEARCH before starting more threads about this scam.

 This is AT LEAST the 4'th thread about this SAME SCAM COMPANY AND MINER.

Did you set up xorg's config so that the drivers know where to FIND the card?

If you have more than one or 2 miners, garage is probably a better idea - usually a lot easier to get ventilation going in a garage.

 A basement that has some of those "sunken" below ground level windows, and some way to be able to OPEN those windows, might be viable.

And Bob stated in at least one interview that it was a paraphrase of a long-running "economist" comment to the same effect.

 TANSTAAFL is the actual quote - "There Ain't No Such Thing As A Free Lunch" - it's been a long-running Science Fiction meme.

Keep in mind that most BIG farms are paying 3c / kwh OR LESS for electric.

 Even a lot of us SMALL farms can get into that ballpark, by being in the right area.

 You are now starting to understand the reason I moved last summer when I decided to go "pro" as a miner.

 As I already pointed out, it's often a case that the WIRING won't support Gigabit, and in THOSE cases the cost would be a ton higher and it's impossible to justify the extra cost for a router that will likely never run at Gigabit speeds.

 The other factor is that a lot of the time good 100MB gear can be had for PENNIES ON THE DOLLAR - in fact I was recently GIVEN a Cisco 48-port 100MB switch (it also has 2 Gigabit "backbone" ports) that the school district a friend was working for was going to just toss in the trash during "infrastructure upgrade work" they were doing this summer.
 My TOTAL cost for 48 ports of 100MB + 2 Gigabit ports? Well, it DID need a power cable - standard C13 that I have a TON of spares for sitting around - so ZERO.


 For reference, up to middle of last year I was still running machines on 10Base-2 (COAX) - the only reason I stopped is that all of the cards I had were ISA and I retired all of my old ISA machines when I moved cross-country last summer (they weren't worth loading up for the trip).
 10Base was severe overkill for what they were doing - but when I found a bunch of 3Com "trio" port cards at a Hamfest for $2 EACH in the early 1990s, and already had a bunch of BNC "T" connectors and patch cords, the price was impossible to pass up and I got something like 15-20 YEARS of usage out of most of those cards (I did have a couple of the 20ish I bought die over the years).


 But as I already pointed out, for MOST folks there is no reason to NOT go Gigabit, and pay the small premium for futureproofing.

 When the latency in both cases is measured in MICROseconds as a general rule, it's NOT AN ISSUE.

 The 14/16 nm generation is the FIRST that is the case for though - aided and abeted by the massive jump in Bitcoin pricing during most of it's existance.

 It probably WILL be the case for any future miner generations though, since miners have caught up to semiconductor state-of-the-art for the most part.

 I don't see it either.

 As I recall Apple is using Intel foundry space for at least some of their chips.

Just use them as individual drives, no need to do anything fancy like JBOD or such.

 I don't see any reason why 10TB (or bigger) drives won't work, as long as your MB supports them and they're formatted properly under an OS that supports BIG drives.

 10 at this time however is NOT cost effective, unless you already HAVE the drive(s) - 8TB seems to be the current sweet spot on capacity per $.


 So far, I'm batting 50/50 on the Seagate Expansion external drives I have bought - one was an Archive, the other was a Compute.

 I suspect some already have - they're NOT profitable unless you have SUPER cheap electric.

The probability of folks being able to get mining chips made on a 7nm process next year is pretty much ZERO.
 2019 - perhaps, but even THAT is looking very iffy.

 For perspective - Intel MIGHT be shipping 10nm late THIS year, and they've been working on THAT node a lot longer than anyone has been working on 7nm (and Intel is USUALLY the first folks to market with usable chips on a given node).

 Keep in mind some history - 14/16nm was first ANNOUNCED as being available almost 2 YEARS before the first actual products hit retail using THAT node.

 10nm has proven to be MUCH more difficult, and preliminary indications are that 7nm (in the cases that will actually BE 7nm) is going to be much much tougher yet.


 Mining gear hasn't so much "stagnated" as it finally caught up to current state-of-the-art on semiconductor technology, so it doesn't have the OPTION of "move to a better node every year or less" any more, it has to wait on the INDUSTRY to be able to move - which has commonly been a 4-6 year cycle and seems likely to get WORSE as semi tech has gotten well into the range where quantum effects and the limits of silicon are bringing Moore's Law to it's knees.

 Intel has said flat out that "10nm is the end of the road for silicon", IBM has already announced that it's work on the 7nm node is using HYBRID Si/Ge wafers because pure silicon couldn't scale any more (but Germanium isn't going to be able to scale much further than Silicon)....


 I'm not going to be shocked if 7nm is the END of "standard" semiconductor technology, at best I'd guessing the next node after THAT will be 10 years more-or-less down the road.
 It might be time for alternative tech to start getting itself OUT of the lab, like carbon nanotube electronics....

There is no "benefit" to using PCI-E generation 1 - but for mining, it's overkill enough that there's no reason to NOT use an older motherboard that only supports that.

 It BEEN DONE a few times - most if not all of the "mining specific" LINUX distros are intended to be usable from USB flash drives and are Ubuntu derived, and I have set up a few machines to do the same on my own.

 Linux does NOT have "less hash rate" unless the mining software isn't being developed for both for a specific algo.
 Linux DOES have inferior tools for fan control and overclocking (no Afterburner or even close), but it CAN be done just not as easily.

Given the current rumblings out of China, we might be looking at THE END of the profitability bubble by the end of the month, or more likely sometime next month.

http://fortune.com/2017/09/11/bitcoin-price-value-usd-china/ among other reports.