I just spent two weeks trying to convince a new intel Zenbook laptop to have decent battery life. It would eat the battery both awake and asleep. Went through the Arch wiki on suspend issues. Discovered that the bios has a broken vestigial S3 suspend (which more and more vendors are shipping); the modern suspend mode is now S0ix (s2idle). Found that my system was only getting into C2 and C3 out of C10 levels of S0ix power-saving-state nirvana.

Somehow, I lucked upon finding that the Intel Rapid Storage/VMD setting in bios was what kept the processor from ever going to lower power states. Once I disabled that, nearly everything else fell into place. The cpu ran cooler at normal use, battery lasted longer, and power burn during sleep went from 4% an hour to negligible.

This was fun. Not one tool successfully pointed me at the real problem. It took one random dell support post to set me on the right path. https://bugzilla.kernel.org/show_bug.cgi?id=211879. I spent two weeks chasing the same problem that somebody else had in 2021. Linux doesn't have a [WARNING] for detecting a damned VMD, and it doesn't have a means to tell the VMD to fuck off? The stupid hardware doesn't have the sense to not fuck up the processor if it isn't attached to its Windows-only driver? I don't understand how anybody has been able to use an intel for the last couple of generations if this is how they work.

In conclusion - battery life is actually pretty great now. But it was a bloody nightmare to get here.