That article's based on data from Anandtech, where the conclusion was a bit more mixed:
http://www.anandtech.com/show/9766/the-apple-ipad-pro-review/4
Last table on that page is the one to look at. Basically the positioning of A9X vs m3 varies depending on the work load, but on average the Intel chip is still faster. That's really impressive, but I don't think it's fair to make the blanket statement that A9X is faster than M3 when it's still losing in overall performance.
Note that I said some Achips were faster than Mchips. I have no concerns with anything you stated. Actually very well said. Back to the original point - can Apple actually port macOS to the A chip? My point was YES and that they probably already have it in the lab. But to be fair to your point, the question is if Apple will ramp up the A chip to be able to compete with the higher end M chips or even the I chips from intel. Actually intel is already creating some ARM chips and I believe there already exists ARM based servers in the wild. So my thought is that the ARM architecture can support performances suitable for higher end PCs. Its up to Apple to leverage the architecture to create a competitive and performing chip. For the iphone the A9 is super fast. For the iPad Pro, it can already use a little more power. Let's see what the A10 brings. But if they grow the chip up to compete (and I believe they can) then there is no reason why I couldn't run macOS on the ARM other than marketing/company-politics.