On Sun, 21 Apr 2002, Felix Karpfen wrote:
> Thank you for the prompt reply.
>
> That was one of my guesses. However, apart from the unexplained sudden
> massive increases in (apparent) CPU usage - which remain unaffected by
> closing and reloading fvwm2 and only disappear on a reboot of the system
> - there appears to be little correlation with the graphed display and
> the output of top.
The display is of the "load factor", which is more or less the
number of "runnable" processes. This is not *exactly* the same
as CPU load, since CPU load cannot exceed 100%; load factor has
no real upper limit. Moreover, the graph's display is a time-average
over a sample interval. (Every time it scrolls is one interval).
> It is a very trivial problem and does not appear to make the slightest
> difference on the way programs actually run.
Heh. The genius of Unix. This is probably because the programs
are IO bound. I am currently compiling three kernels in parallel
on one machine, and its load factor is running around 5.5 to 6.
But the most top shows for any process is 2-3% cpu. Compiling is
mostly IO work. With good disks, you don't really notice a slow down
in things like keyboard response or process initialization.
(This discussion based on BSD, I dunno what Linux or Solaris do.)
When I want to see a high load factor, I simply fetch a bunch of mail
from a popserver. The various sendmails are actually trivial tasks,
but they stay runnable for a long time doing dns lookups and so on.
> So if there is no obvious answer, please ignore.
For even more graphs, try fetching and installing xperfmon or
one of the other performance monitors.
Dave
--
The man who reads nothing at all is better educated
than the man who reads nothing but newspapers.
-- Thomas Jefferson
--
Visit the official FVWM web page at <URL: http://www.fvwm.org/>.
To unsubscribe from the list, send "unsubscribe fvwm" in the body of a
message to majordomo_at_fvwm.org.
To report problems, send mail to fvwm-owner_at_fvwm.org.
Received on Tue May 14 2002 - 04:03:15 BST