I'm a C++ programmer working on various projects which run under Windows but which require accurate timing.
Traditionally these have tended to run better on Intel CPU's (because for a multi-core device, Intel CPU's tend to use a single high-precision timer which gets shared by all the cores). AMD CPU's on the other hand have traditionally used different timers for each core (and AFAIK they're not synchronised). Therefore when the OS switches a task onto a different core, the timing can go utterly haywire (which defeats the object of it being "high precision")..
Is that still the case with the latest AMD processors (e.g. ThreadRipper)? Or do the different cores finally share a common, high-precision timer?