Hallo,
I have problem to understand the “ticksize”. I read, when you have a
ticksize from
10 ms than you have a max jitter from >10ms. Is that right?
My next problem is, in the Neutrino-book is standing, that the smallest
ticksize is
0,5 millisec. That is very slow.
The ticksize is for the process request, isn`t it?
And in which relation are standing the ticksize and the cycletime?
Hallo,
I have problem to understand the “ticksize”. I read, when you have a
ticksize from
10 ms than you have a max jitter from >10ms. Is that right?
Yes, ticksize is time precission of the kernel. Any timer should give you time equal or more than
you asked for…
My next problem is, in the Neutrino-book is standing, that the smallest
ticksize is
0,5 millisec. That is very slow.
It was for QNX6.0-6.1A. Now, in QNX6.2, smallest ticksize should be 10uS (microsec)
The ticksize is for the process request, isn`t it?
It’s timeslice between resheduling processes in system. You could read Robert Krten’s book for
details
And in which relation are standing the ticksize and the cycletime?
I’m sorry for ignorance, what is cycletime? Is it CPU frequency? Than there is no any relation (it
is different oscillators . Except only one thing: slow CPU is unable to serve the system with
small ticksize. All the time you will have resheduling, but not actual work