I recently upgrade my CPU from a 1950X Threadripper to an Intel 12900KS, in the hopes of speeding up my physics sims (and generally speeding up the viewport when there is a lot of geometry). I have not noticed any improvement in speed, despite the significant upgrade. (I also have 64MB RAM, and a lightning fast M.2 drive, so there should be no data bottlenecks on my system.)
I ran some tests and when I’m running a complicated physics bake (which is painfully slow) my CPU is only processing at about 1-3% power, despite being overclocked to 5.6 Ghz.
Now I know that Blender is mostly optimized to calculate physics using single core, but taking that as a given, why is it only using a faction of the processing power that a single core is capable of? Is there a reason for this?