Real-Time in general is not about speed, it's about guaranteeing, that a specific function will be executed within a specified time window.
Imagine a car with some sort of fieldbus. You want the brakes to be working at any time within a time limit (say, 1ms). That's Real-Time.
Windows for example uses some sort of scheduler to distribute CPU time to the different threads with a granularity of - i think 1 ms. So each thread gets 1 ms, then it is interrupted and the next thread gets its 1 ms. Of course there are priorities which thread gets a little more time (or more "slots") and which thread gets less. BUT: You cannot tell in Windows, that your application will execute in a specified time window - at any time, it can be interrupted and therefore delayed and theres nothing you can do about it.
Real Time OS make sure, that - if some specified event happens - it will be executed within a specified time window.
