I noticed this a few months ago, and after consulting this list was told
that it was an artifact of the CPU utilization monitoring process. (I
was using xosview). I wasn't sure that I bought it though, so I set up
some cpu-bound test programs that monitored their scheduling via the
proc filesystem and sure enough, they were *not* really bouncing back
and forth from CPU to CPU. In about 50 test runs it only realy bounced
around on one of them, and I assume that some other process was doing
wierd things during that run, rather than it being a scheduling problem.
John
Marc MERLIN wrote:
>
> I have dual CPU machine on which I run two niced crack processes.
>
> If I send SIGSTOP to one of the two processes, the remaining crack process
> keeps bounces quite between the two CPUs (as shown by the CPU0_Nice and
> CPU1_Nice bars in procmeter3), so as to give almost the same load on each
> CPU
>
> I thought that because of CPU affinity, my remaining process was going to be
> assigned to the same CPU (at least as long as there isn't enough non niced
> work to keep the other CPU busy)
>
> Is this expected behavior for the scheduler?
> (I am running on 2.2.7)
>
> Thanks,
> Marc
> --
> Microsoft is to software what McDonalds is to gourmet cooking
>
> Home page: http://marc.merlins.org/ (friendly to non IE browsers)
> Finger marc_f@merlins.org for PGP key and other contact information
>
> -
> 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/
-
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/