That or something like it definitely still applies to my R7 1700. If Windows is configured to ever let it drop to its minimum state, it gets intermittently slower even when it should be maxed out (full load and anything reading frequency directly shows max).
Captaintrav, what's your CPU?
My observations are on Skylake and Kaby Lake R laptops. I had heard a little about Ryzen parts, but this is new to me. What's more is we're not talking about a 8% difference, more like %50 difference in one case, and in a situation where you'd think the system would be more I/o bound - applying a Windows image from a .wim file. Although I am seeing smaller, but still significant differences in other apps. One situation was that users are complaining certain CPU intensive tasks were slower once they got Windows 10, when the hardware was either the same or faster. The weird thing is the CPU is clocking up to 3.6Ghz or so, just like it should (single threaded workload). Really puzzling why the power settings would have that much effect.