This is a good revelation but is a technique that is old as the hills. My first post university programming job was working for CAE Electronics, a Flight Simulator company. The simulations ran on a variety minicomputers of the era, primarily DEC Vax 780 and successors or Gould SEL during my time there.
If memory serves me correctly the simulation ran like a super loop on a microcontroller every 50ms or 33ms. On each loop the dispatcher would run a selection of the core simulation functions some every loop, some at every 2nd, 4th or more loops. The system was called banding. The dispatch table was redone to optimize the needs of that particular simulator.
There were various optimization outside of the main cpu as the peripheral devices did computation as well. Also as the systems got faster there was time left over for the normal OS to run low priority function with non real-time requirements.
The technique was not new in my time and had been used on the previous generation of Xerox systems about a decade earlier.
Interesting trivia: Do you know that VAX 780 from 1978 is still supported by CURRENT(2026!) version of NetBSD? :D
First modern 32 bit CPU with MMU.
I worked on machines of the articles vintage - I would expect this technique to be common the further you go back due to the lack of power. Just a variant of CPU slicing.