This content has been marked as final. Show 1 reply
I don't quite understand the question. If you run on a non-real-time Linux then the non-preemptiveness of the kernel will add latencies all over the place. In addition there will be priority inversions all over the place. There may also be differences in how basic real-time priority scheduling is performed.
In short you have lost determinism so the timing properties of your application will be unknown and likely vary from run to run.
Edited by: davidholmes on Jan 22, 2009 8:44 PM