In this article on MSDN about mutitasking the author mentions that each thread on a Windows system runs for the duration of a "time slice" and that one of these time slices is around 20 milliseconds. In another article I read the author referes to time slices as "quantums" and wrote that they are usually around 15 milliseconds on a Windows system.
Now, what I don't understand is that when I run a program and use a timer (like QPC/QPF or timeGetTime) to continuously spit out time-stamps it seems my program is indeed not suspended every 15-20 milliseconds. In fact, it looks like it isn't suspended at all.
Also, when I ran the program there was a total of 618 threads running on my system which means that if each of them was granted a 20 milliseconds time slice each thread would have to wait no less than 12.36 seconds every time it had worked for 20 milliseconds, which obviously isn't the case. I know a lot of these threads are sleeping most of the time, but even when I had Left 4 Dead 2 running which has 45 threads (during gameplay) and uses quite a lot of CPU resources it still didn't seem to affect my program ... ?
