I recall this blog post[1], discussed here[2] at the time, discussing how a SATA chip driver prevented the CPU from going to sleep state C6, instead keeping it

I recall this blog post[1], discussed here[2] at the time, discussing how a SATA chip driver prevented the CPU from going to sleep state C6, instead keeping it at C3, leading to higher power consumption. Switching the SATA card to one with a different chip fixed that. There were other surprising factors which affected power draw. My takeaway was that as in real life, sleeping can be hard. [1]: https://mattgadient.com/7-watts-idle-on-intel-12th-13th-gen-the-foundation-for-building-a-low-power-server-nas/ https://mattgadient.com/7-watts-idle-on-intel-12th-13th-gen-... [2]: https://news.ycombinator.com/item?id=38823514 https://news.ycombinator.com/item?id=38823514