These things are still for sale.  Surely even at 1 per account they must have sold more than 5k of them by now.

/\                                             /\
|  This is the explanation right here |
|                                               |

Windows defender is intended for noobs so it pretty much tags all miners as a virus and quarantines them.  So if you are running windows defender you need to create exclusions.

The AV I use most is Sophos and it doesn't usually do that.  It flags them all as PUPs (potentially unwanted program) so you can decide what to do with it.  That classification is more honest, it recognizes the fact that maybe it's there because you wanted it there or maybe it's there because of malware but the AV doesn't know for sure.  I still have to make an exclusion to stop it from popping up with warnings but it doesn't usually quarantine them straight away like windows defender does (although truthfully it is a bit inconsistent and occasionally some of them do go straight to quarantine).

Newer AV's are moving towards behavioral scanning techniques so they can spot programs that aren't in the virus definitions so if they see something peg the CPU at 100% for an extended period that could certainly look suspicious.

Last time I sent them (myrig) a hashboard for repair they let me know right away that they had received it and then crickets for over a week.  Then I started pinging them every few days and either got no response or time wasting responses like "I'll ask the tech" or "I'll ask the tech again".  In the end they had the thing for a month, said I had run it in a too hot environment (suspiciously the exact same thing they have told me all three times I have used them even though I have always run all of my asics in climate controlled environments) and then finally said it couldn't be fixed and sent me back a board that was missing various heatsinks. Was it my original board or just some other random board they sent to make me leave them alone?  I have no idea.  Not a fan!  

Sorry I have a hard time keeping up, what with the long time intervals here, but is this dumbshit Kermy actually Carlos the drunkard then?  Someone clarify for me please?

Yep I had the same problem as OP using the 2 x molex to PCIe adapter.  Rig needed regular reboots, it was freezing.  Finally one day it wouldn't work at all.  Opened the case and one of the molex connectors to the PCIe adapter was melted and burnt.  This was because I built that miner using a cheapo PSU that I had laying around.  I try and use quality modular PSU's when possible but sometimes I give in to the attraction of building a cheap miner (all I had to buy was the GPU's, everything else came my way for free).
 

I probably am talking about an APU for the machine from work (I had to search what APU meant in this context).  I am not aware of any current AMD CPU with integrated GPU.  AMD may still have them but if so I hadn't heard about them, and as I say my experience with the one I got some years back was not good at all.  But what I have been seeing is stuff like this about the Intel CPU/AMD GPU integration.
https://arstechnica.com/gadgets/2018/01/kaby-lake-g-unveiled-intel-cpu-amd-gpu-nvidia-beating-performance/

This is already looking like a successful collaboration on the high-end processors so why not continue it on down the line for the lower-end processors?  A lower end version with i3's and lower-end Radeon chips would likely be a huge improvement for the average HP, Dell, Lenovo desktop lines that pretty much all come with Intel HD Graphics chipsets.

Honestly I don't know what's so different about this than what AMD was doing in the past with their own GPU integrated on the CPU thing except that this version sounds like it's good whereas the old AMD version wasn't.

Hmmm.  Lots of confusing signals here.  For starters Intel already makes GPU's, just not high end ones.  But pretty much 90% (or maybe even more) of the mobo integrated GPU's are Intel.  The exception, AMD CPU boards with AMD integrated GPU's are arguably even worse.  I bought an AMD CPU\GPU integrated PC for work and even though the GPU model was decent as a stand-alone (it's a few years old now but back when I got it, the stand-alone card was decent), in the integrated version that GPU sucks really bad.  I regretted it immediately.

Also, the recent Intel\AMD partnership with the integrated AMD GPU has been regarded everywhere as a huge success, even this article talks about it in very positive terms.  Why would Intel go in the opposite direction after that?

You're welcome.  BTW, where did you buy your miner?  It seems rare these days that anyone actually delivers a new asic that makes that much when it arrives.  Usually they announce it after they have been mining the hell out of it and then by the time they deliver it's a 6 month ROI (which means by the time you actually ROI it's been a year since the difficulty goes up all along).

So just for the record and in case you didn't know this already the pool doesn't know how many hashes any of the miners are doing.  When you look at your miner interface (I am unfamiliar with this particular miner but this is true for every miner) but when you look at the miner interface, it is reporting to you the number of hashes the miner is actually able to compute, so a real-time hashrate.  The pool on the other hand is only getting occasional submitted shares from the miners that are connected to the pool.  Vardiff (variable difficulty) which is what pretty much every stratum implementation is using now days, varies the difficulty based on how many shares a miner turns in with a specific goal in mind, usually something like 4 or 5 shares per minute.  So the more shares a miner submits, the higher the difficulty the pool will set for the miner until it reaches the point where it is getting 4 to 5 shares per minute from the miner.  The entire time, the pool is calculating the hashrate of the miner based on the number of shares it submits and the difficulty it has set for the miner.

So a miner that has a very high hashrate, will be able to find 4-5 shares per minute for a higher difficulty than a miner with a lower hashrate.  But there is variability and luck.  So any miner could in theory get a lucky streak and hit more shares of a given difficulty in a period of time than the hashrate would normally expect, in which case it will appear to be a higher hashrate than it is to the pool.  Bad luck also happens, which makes a miner appear to be lower hashrate than it is.

So the point is this, the hashrate you see on the poolside, for any pool and any miner, is a calculated value and that calculation is subject to variance and luck.  Over time though, the average on the poolside should be fairly close to what the miner interface reports.  So it's totally normal to see some fluctuation in what the pool says your hashrate is, what you want to pay attention to is the average.

So for example here is my hashrate according to the pool for 3 S9's (one with a dead board):  That 10 minute average of 39 is higher than my actual hashrate.  That 1-day average of 36 is very close to it.  The actual hashrate is  8.9 + 13.15 + 13.85 = 35.9

I just use MSI afterburner for that on all of my GPU rigs.   Then you don't have to learn how to handle it with each different software miner if you switch miners or coins.

If by "shut off" you mean it would just power off without warning then the first thing that comes to mind for me is possible CPU overheating.  Double check your CPU heatsink.  A CPU overheat issue will do exactly that. 

Well there is bitfury.  There was those guys in Europe who turned full scammers, KNC.  There were those guys from Israel, Spondoolies, but they sold out or went out of business or something.  Of course there was BFL, fuckers.

That top miner at 1400 Kh would be something like 1500-2500 high-end GPU's.  A bit hard to believe for a single farm.  Not to mention if you had a farm that big, why not solo mine?  It's 13% of the entire pool.  If I was a gambler I would bet on a botnet.  But seriously hoping it isn't asics.  I don't really know enough about the v7 changes to make any sort of intelligent comment about the feasibility of modifying the existing asics to work on the new version.

You're definitely right though about profitability being no good yesterday whereas right now it's looking pretty good.

Biggest miner on MPH was 200 Kh yesterday, now there is one at 1,400 and another at 400.  I'm guessing either the botnets are getting back in the game or someone figured out how to get their asics going on this version already.

Ah geez, difficulty is going up.  I was thinking that this might happen.  All of these slow blocks are being treated as outliers in the difficulty retarget so they aren't being included.  We gotta get over the 20% outlier hump so that the retarget will start to include the slow blocks.  Monero devs should look at this issue for future forks.  They might want to figure out something that will help difficulty retarget faster right after a fork.  The 720 block minus 20% outlier thing is actually a decent formula for when hashrate isn't changing a lot in a short time frame but this current hashrate drop and associated slow block time thing is going to be made even more painful with this short term increase in difficulty.

Not that long.  I am unclear as to how the edge case blocks are selected and then disregarded in terms of how monero retargets difficulty, but after the first 144 blocks (at the most) we should definitely begin to see the difficulty go down.  We are 44 blocks of the way through that 144.  It may well start to head down sooner than that though, on the other hand, it may head up some more as well.  As I say I am unclear on how it determines which blocks are considered the edge case ones that don't get counted towards the next difficulty.  

At any rate, we should see the other side of the first 144 blocks within 24 hours and see a definite reduction in difficulty begin (if it hasn't started before that).  As heavyarms1912 says above, the full extent of what network difficulty will look like will take 720 blocks but once it starts going down, blocks should begin to be a lot faster again so hard to say how long it will take to get those 720.  I would guess that 48-60 hours would be long enough for us to see something like normalcy in terms of block time on the network again, although it could certainly be less time than that.

The monero POW v7 changes started at block 1546000 and we are at 1546024 now so that already happened.  I don't know about any monero-wanna-be coins that may have forked off at that point or may be forking off later. 

I hope we see some difficulty adjustment before 1546144, this is pretty slow...  So far though the difficulty hasn't really budged.  In fact it has actually crept upwards a bit which is not helping.

Did you try xmrig? https://github.com/xmrig/xmrig

Seems like either xmr-stak or xmrig would be your best bet for now.  You may have to figure out the best settings to use.

Also, my understanding is that the new algo change does reduce hashrate somewhat compared to the older version.

Blocks are coming very slowly.  In case anyone is interested (as I was) there is a good explanation of monero difficulty adjustment in the answer on this page: https://monero.stackexchange.com/questions/987/how-is-the-monero-network-hashrate-calculated-and-how-is-it-different-than-bitco/1010

So I guess we will have to wait a bit to see the difficulty start to come down but it sounds like it should probably start happening within a day or so.

Assuming that explanation is correct then I would expect all of the current blocks to be in the 20% of outliers that are discarded.  If so, then 20% of the rolling 720 blocks used to calculate difficulty would be 144 blocks.  So theoretically it may not be until after block 1546144 that we start to see a real difficulty drop and faster blocks again.  Which right now means another 126 blocks.

I suspect that a good chunk of network hash is coming from botnets as well.  I am hoping some of those will be unable to update their hosts\victims quickly and we'll lose some hashrate there as well.   It'll be interesting to see what happens.