On pricing, AMD hardware is marked up about 20% for the next month or so. It’s an early adopter tax.
On multithreading, parallelism is limited by how parallel the problem is. I suspect that’s also related to why the 30 series was only so-so at low resolution, despite its very high theoretical performance; it depended to much on parallel execution, and could only really fill it’s pipes when there were a lot of pixels.
The very first problem is the"business logic" part. I do not know of a single flight simulator that is not very single core heavy. That may mean no-one has solved the puzzle yet. It may also mean that there are critical things than cannot be parallelized. We just do not know.
As I understand it, DX11 itself was designed before multi threading was even a thing, so it has some intrinsic bottlenecks in it, as well as limitations that likely his flight sims hard. As I understand, DX11 has a fixed limit on the number of draw calls be cycle, so things like flight sims, where you are looking at a vastly larger space than any other game genre, likely smack into that limit hard. That may also be why flight sims seem to be as cache hungry as database servers are too, but that’s pure speculation on my part.
DX12/Vulkan are very new ways of drawing frames, and it will take a while for developers to really get the best out of them, especially in areas that have no cross-over with popular genres.
So, when building a PC for flight sims, focus more on single thread performance, and CPU cache/memory performs performance, and consider multi thread capacity only in terms of the rest of the things you may be doing with your PC.