because ARM's weaker memory model than x86 is a major challenge for translating multi-threaded code with low overhead
As someone else here has pointed out, Windows on ARM emulation ignores this by default and just accepts the possibility of wrong results or app crashes.
So, you're likely not going to see much performance difference, but...
As far as I know, current Windows on ARM ”cheats” by pretending memory order is not a problem.
Indeed, as your link confirms:
These settings change the number of memory barriers used to synchronize memory accesses between cores in apps during emulation. Fast is the default mode, but the strict and very strict options will increase the number of barriers. This slows down the app, but reduces the risk of app errors. The single-core option removes all barriers but forces all app threads to run on a single core.
Very good catch!
Yeah but 2700X to 5800X(3D) can be done on the same motherboard with the same RAM!
The incremental cost of the 7xxx series is a lot larger.
"Tensor Units" are just low-precision matrix multiplication units.
Completely pointless to compare IMHO. A bit like comparing "core counts".