Linux 6.8 Merges Fix For Recent Performance Regression Spotted By Linus Torvalds

[aufkrawall]

Can you elaborate on this?
Did amd_pstate=active + performance mode caused stuttering?

Had it wrong in my mind: It is amd_pstate=active that doesn't allow selecting schedutil. At least I'm deducing that from the fact than I can switch to schedutil with =guided.
performance governor never caused stutter for me with any scaling driver, it was schedutil.

[creative]

I don't know how that guy puts up with what he puts up with especially considering how absolutely obnoxious fTPM and TPM have been to work through. Even there stuff is still not ironed out.

This regression appears to be a fairly much more simple fix but I don't think I could ever handle being a software engineer.

[leledumbo]

Test stutter/frame times in games/video/browser with different loads, it always caused me some noticeable impact at some point.

This is my daily workflow, everything is smooth.

With some amd_pstate driver (afair it was active that allowed it) it even caused higher power draw vs. performance. For now, I ended up with pstate=guided and powersave governor for the 5700X. It still shows lowest idle clocks for cores while not causing any apparent issues.

This could be the reason, I think I have amd_pstate disabled totally but let me check again later. For context, I'm on mobile 6800H.