> Sorry, I misread mingo's posting, my error! But this means that with
> 2000 processes, assuming the run queue has 1 entry (as mingo points
> out), we will recalculate 5 times/second on i386. Each time we will
> use 1ms which means we use 0.5% cpu in the scheduler. This is on a
> very fast computer (pII/300MHz). On an alpha, it could easily be 5%
> since it does 10x as many schedule() calls. Are there any numbers on
> this?
On both Alpha and other platforms the timeslice length
is the same (20*HZ/100 == 200ms).
Therefor, the number of schedule() calls on Alpha will
be similar to that on eg. PPC or SPARC.
Rik.
+-------------------------------------------+--------------------------+
| Linux: - LinuxHQ MM-patches page | Scouting webmaster |
| - kswapd ask-him & complain-to guy | Vries cubscout leader |
| http://www.phys.uu.nl/~riel/ | <H.H.vanRiel@phys.uu.nl> |
+-------------------------------------------+--------------------------+
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu