When running single-threaded benchmarks / applications, windows (win 10 - 1903) schedules processes/threads as if it doesn't know about the "fastest core" in my 3700x. It always seems to prefer Core-0 in the system (marked in ryzen master as "second fastest core", whereas Core-3 ("fastest core") is always idle for low-thread workloads.
I wouldn't mind, however the first core only turbos to 4.3ghz and I've read windows-10 1903 should actually be able pefer the core which can be clocked highest.
My system consists of a 3700x on a phantom gaming 4 asrock board running the latest BIOS (1.0.3aab) + latest chipset drivers (ryzen performance power plan).
Should this actually work, or am I hoping for too much?
Best regards, Gerhard
I wonder - is this working for everybody else as expected (or is the behaviour I expect to see expected at all?) or does just nobody care? This is the second forum where I ask this question - yet the post always remains completly silent.
I have exactly same problem. my fastest core is 4 and 8 but Windows scheduler prioritizes second fastest core 1. But i noticed that changes if you enable manual settings under PBO. It starts prioritizing core 9 instead of 1 but still not the fastest in the chip.
Happens with me too. Windows always prioritizes CCD0/CCX0 for all main tasks, regardless of the fastest core of CCD0 being in CCX1, Core 4. Tho, I wouldnt worry about it too much. Ive confirmed all cores of my CCD0/CCX0 can do the rated boost speed regardless.
The issue you two are describing is within the Windows Power Plan ("AMD Ryzen Balanced") not using "performant cores". You can use POWERCFG.EXE to remedy this. Ensure you've got all the doodads pertaining to CPPC enabled in your BIOS/UEFI enabled too.