Level 3 of 12 · about 2 weeks
CPU scheduling
How an operating system chooses which task runs next: the measures, the classic policies, shares, several processors and deadlines, and how Linux does it.
Modules
Core: everyone needs it. Common: most jobs use it. Specialist: optional depth, safe to skip on a first read.
1
What scheduling is and how to judge it
Bursts, the scheduler, the dispatcher, and the numbers used to compare policies
Core 6 parts · about 4 hours
2
First come first served and shortest job first
The two simplest policies, a trace of each, and why shorter jobs first lowers the waiting time
Core 7 parts · about 4 hours
3
Round robin and the time quantum
Taking turns with a timer, and how the length of a turn changes response time and overhead
Core 6 parts · about 4 hours
4
Priorities and multilevel feedback queues
Letting important work go first, preventing starvation, and learning which jobs are short
Core 7 parts · about 5 hours
5
Fair-share scheduling, lottery, stride and Linux
Giving each process a stated share of the processor, and how Linux does it with virtual runtime
Core 7 parts · about 5 hours
6
Multiprocessor and real-time scheduling
Several processors, balancing the load, keeping a task near its cache, and meeting deadlines
Core 7 parts · about 5 hours
7
Scheduling in Linux in practice
Looking at the scheduler, changing policies and priorities, and diagnosing a busy machine
Core 7 parts · about 5 hours