The fact that a simple OS update can make the CPU up to 13% more effective makes me wonder how much performance difference there is between Windows and Linux.
Since I have dual boot, I'll check later if I can find any benchmark tool that works on both so I can compare.
I don't watch videos, but I'm going to blindly guess that this is fixing borked Windows scheduling that pegged SMT's fake cores before real cores.
If that's the case, Linux was basically just that much better (though you could "fix" it on Windows by forcibly disabling SMT in bios). But most of the time the performance is pretty similar (especially in the real world where you're aiming to be GPU limited).
Does this mean review/benchmarking sites need to reevaluate all chips after os updates or at least publish the versions used while testing?
pretty much yeah as the performance gains are significant enough to impact a user's buying decision.
To not update previous benchmarks is to do buyers a disservice as they would no longer be able to make their best informed buying decisions, the only ones who would benefit from a situation like this would be the corporations for no reason