[En-Nut-Discussion] RTOS Comparison

Nathan Moore nategoose at gmail.com
Tue Mar 23 20:45:58 CET 2010


Thanks Michael!

If in addition to these threads you only had the idle thread and
possibly one other
inactive thread then that should have been just about best case for
the currentthread
swapping code in Nut as I understand it. Especially since the only
timer used is from
the sleep and shouldn't be in the timer list at the time of the
context switch you were
measuring.
Processor wake-up time also shouldn't come into play because the high
priority thread
only wakes up during the time when the low priority thread has already
ensured that the
processor is awake.

If you changed your timing calculation to (Tcd - Tab)/Tab then you
would take the
clock speed out of the picture (for the most part), which might be interesting.


Nathan



More information about the En-Nut-Discussion mailing list