Re: [patch] "HT scheduler", sched-2.5.63-B3

From: Mike Galbraith (efault@gmx.de)
Date: Fri Mar 07 2003 - 03:09:32 EST


At 08:45 AM 3/7/2003 +0100, Ingo Molnar wrote:

>On Fri, 7 Mar 2003, Mike Galbraith wrote:
>
> > Weeeeell, FWIW my box (p3-500/128mb/rage128) disagrees.
>
>UP box, right?

Yes, p3/500 w. 128mb ram and rage128.

> > I can still barely control the box when a make -j5 bzImage is running
> > under X/KDE in one terminal and a vmstat (SCHED_RR) in another. I'm not
> > swapping, though a bit of idle junk does page out. IOW, I think I'm
> > seeing serious cpu starvation.
>
>which precise scheduler are you using, BK-curr? And to see whether the
>problem is related to scheduling at all, could you try to renice X to -10?
>[this is not a solution, this is just a test to see the problem is
>scheduling related.]

I plugged your combo patch into virgin .64.

>is there any way we could exclude/isolate the VM as the source of
>interactivity problems? Eg. any chance to get more RAM into that box?

I can (could with your earlier patches anyway) eliminate the X stalls by
setting X junk to SCHED_FIFO. I don't have ram to plug in, but I'm as
certain as I can be without actually doing so that it's not ram shortage.

Best would be for other testers to run some tests. With the make -j30
weirdness, I _suspect_ that other oddities (hmm... multi-client db load...
query service time) will show.

         -Mike

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Mar 07 2003 - 22:00:36 EST