> > that only 0.3% CPU time is spent in the kernel. (readprofile shows
> > that
>
> This is bogus. [...]
oops, right. Anyway, vmstat shows that about 200 reschedules per second
happen with 200 running processes:
r b w swpd free buff cache si so bi bo in cs us sy id
200 0 0 0 156248 274668 26340 0 0 0 0 109 215 99 1 0
200 0 0 0 156248 274668 26340 0 0 0 0 109 215 99 1 0
200 0 0 0 156248 274668 26340 0 0 0 0 108 215 99 1 0
200 0 0 0 156248 274668 26340 0 0 0 0 109 217 99 1 0
200 0 0 0 156248 274668 26340 0 0 0 0 108 215 99 1 0
200 0 0 0 156248 274668 26340 0 0 0 0 109 215 99 1 0
200 0 0 0 156248 274668 26340 0 0 0 0 108 215 99 1 0
a reschedule every 5 msecs is not a big overhead at all, considering the
_ridiculous_ CPU load.
> [...] Scheduler is highly unlikely to measure _its own_
> overhead. Basically, do not trust this meters this much. Take a look
> at how scheduling is done - it is in sync with timers which measure
> who eats time -> it is highly unreliable.
you are right, i just knew it's not much overhead and quoted the first
number available, mea culpa :)
-- mingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/