This video (link bellow) started to be very interesting when Paul got to the details about what PCIe devices will be disabled with lowest level CPUs in X299 platform.
Basically that previously obvious (usually) choice between that second PCIe slot or two more SATA slots, is now even more "exciting".
For me such combo (those CPUs in X299) looks like a really drunk twister game with computer components (not that it is any surprise).
TL;DW: Kaby-X is bullshit.
Repair shops will only need to stock a $250 CPU for testing purposes, instead of a $500 CPU. Apart from that, I can't think of a single use case for the X299 16 lane processors.
This seems like nothing more than a cash grab. Folks trying to build a budget X299 system will be pissing away $250 if they fall into this trap. Hopefully, the major retailers will refuse to stock this junk.
WTF Intel!!!
The chip is supposedly marketed to Extreme overclockers. A disables iGPU and increased voltage is what Intel believe will help the chip get high record breaking clocks...
That doesn't work if you need to test the memory banks though 
They still need a quad channel CPU to be able to test all memory banks.
"Supposed" is probably the good word. As one thing I remember from YT videos that SkyLake-Xs have poor TIM material. That contrasts the "Extreme Overclockers" idea.
The only possible reason i could think of from intel doing it,
is because the want to have the cpu with the highest clockspeeds and ipc.
But other then that, those 7640X and 7740X chips dont make any sense to exists at all.
Its just pointles, because you can't use half the feutures of the more expensive X299 boards.
It just makes no single sense what so ever.
You can get the same performance from the mainstream Z270 platform,
for way less cash.
Intel does not care about poor TIM on chips they know Overclockers will use LN2 or other exotic forms of cooling.
My thought was for diagnosing non-booting systems, but you are entirely correct. For a comprehensive diag, you still need the big $ CPU.
Its going to be very interesting as to the pricing versus AMD x399's, in theory just the fact that you must support 60 PCIE lanes, you must support quad memory and you must have 4 lanes for chipset should equate to the cost being spread out over more boards. Also the microcoders have a far simpler task versus their x299 brothers and sisters. And yeah, 250 buck processor and 500 dollar motherboard which is gimped to hell as "high end", thanks Intel.
I think you're onto something here. Intel know full well that in comparison to the mainstream i7 CPUs, their 10+ core CPUs will have abysmal clock frequencies. While overclockers may, or may not adopt X299, that is irrelevant. Intel will still be able to make outrageous single thread performance claims for the platform ... never mind that no informed consumer will ever build/use such a machine in production. Obviously, the exception to the rule are those wacky competitive CS:GO players who think that 350fps isn't good enough and that a X299 rig may get them to 400+fps.
The motherboards for X299 may not even be up to snuff due to intel's rushed launch.
Do you know how much butthurt from the blue side that video is causing? It's hilariously fun.
That guy has been overclocking CPUs for a while now, I wouldn't wanna try to flex my keyboard warrior skills if I was them.
Well you know, they obviously know better..as someone said VRM was just copy and paste from other models of motherboard.
Not sure what you are trying to say with this.
Not every motherboard has the same vrm implementation as far as that is concerned.
That was one of the comments about the VRM issues with the new motherboards.
Nah, it was just that some of the entry level X299 motherboards showed similar issues.
But that doesnt mean that they have the same vrm implementation atall.
It really depends on which particular mosfets / powerstages that are used,
and in which kind of configuration.
In this case there seems to be nothing wrong with the vrm atall.
It just seems that there might be some issues with the heat dissipation of the used mosfet heatsink.
This could be a matter of a bad contact, due not the best thermal pads or what not.
As someone who diagnoses and fixes PCs in a professional work environment on a daily basis. Thats completely wrong. If the chip can't test full functionality its not good enough to test a board. Test if it boots? Sure but that's basically the end of its usefulness