Code-compile and (de)compression are (kinda) valid android benchmarks scenarios now.
I've been using Youtube ReVanced for a while now and noticed the great patches for other apps also available in ReVanced manager. If you are unfamiliar ReVanced can patch other apps with additional functionality such as blocking ads and correcting other annoyances like re-enabling taking screenshots.
The whole process involves downloading, upacking, patching, repacking, and installing an APK file and it has shown the slightly aging side of my phones SD845 chip, as it takes a few minutes to complete with larger apps. Its not something that comes up too often, but when it does, it could go a bit smoother.
I wouldn't mind seeing some numbers for this workflow when choosing my next phone.
Yes, from what I can tell you there is quite the difference. I for example have noticed up to 25% higher geekbench scores when I place my phone a cold metal plate.