[Darryl Ramm]

Are you comparing past behavior of an identical Pro Tools versions with identical plugins versions on the same Windows version? What did that do with CPU use as you decreased the buffer size? "not much better than" the old system... based on what a CPU meter is doing? Or other stuff? What exact CPU meter/metic are you looking at? What are you doing with that VI in the session? How many voices? How do you have NI configured/optimized itself?

So an OEM install of Windows, full of usual crapware? I would not assume that is a good idea to start with, just as a matter of principle I would get rid of that and do your own clean install. Ideally from a Microsoft installer, not an OEM recovery volume etc.

Get rid of/do not install any AV software. All plugins and apps need to be reinstalled from scratch using correct versions of installers, do not try to copy stuff off your old PC.

What sample rate are you running at? High CPU use is unlikely to be your new hardware and very likely just a software issue. Maybe just behaving as a particular plugin does on any CPU at that IO buffer size... have you tested the same on your old PC? I have no experience with NI strings...

Are you getting any AAE errors? Make sure ignore errors is not checked.

What problem were you trying to solve by upgrading this PC? Ignoring CPU meters does it do that?

And ultimately how small a IO buffer you can run a VI at is not simply about CPU performance, its very dependent on lots of stuff. Needs very careful optimization and setup, tuning of the VI, especially streaming vs caching samples etc. and you always have workflow choices, including freezing/committing tracks. And for systems optimization there make sure you get **everything** including BIOS c-state settings. Try with or without hypethreading. Try CPU priority tweaks discussed on DUC... but they are more for solving actual AAE errors, not for high apparent CPU burn.

And as others have commented spinning HDD for sessions and samples is not the best choice nowadays. You would even be better off getting a big single M.2 boot drive and putting sessions and samples on that fast one drive, than on much slow slower dedicated HDD. For now while testing make sure the session is fully cached, with disk cache set to a large enough size, not "normal" and the disk cache meter goes green before you start testing.