[rtl] What causes poor scheduling accuracy?

[Robert Merithew <merithew@uiuc.edu>]

If I run the scheduling accuracy tester under rtl1.2 (kernel 2.0.36), I
find an occasional error of up to 15,000 ticks (for example, if I run
'startx', which requires lots of disk activity, perhaps page faults).

If the rtl kernel is emulating interrupt-masking, what is preventing the
scheduler from getting control for so long after the programmed interrupt?

I'm using all the default parameters of the 'rt_process' module (i.e.
one-shot mode, minimize_jitter=1 ...)

please copy any replies to me, as I'm not yet on the list.

thanks,
--
  Robert Merithew (merithew@uiuc.edu)           _______________    __o
  University of Illinois at Urbana-Champaign   ______________     -\<,
  Dept. of Physics                                          ......O/ O



--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail majordomo@rtlinux.cs.nmt.edu OR
echo "unsubscribe rtl <Your_email>" | mail majordomo@rtlinux.cs.nmt.edu
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/