Revisiting Linux CPU scheduling

The post revisits the concept of Linux CPU scheduling, focusing on features like `autogroup` introduced in Linux 2.6.38, which groups processes by terminal windows, impacting their CPU time allocation. It also explores how `nice` values affect process prioritization within these groups and highlights the dominance of `cgroups` under systemd, which overrides `autogroup` settings. The author shares insights and practical examples, discussing how to manipulate CPU scheduling with `systemd` units, slices, and their hierarchy.