On Fri, 16 Jun 2000, Kees Bakker wrote:
> >>>>> "Mike" == Mike Galbraith <mikeg@weiden.de> writes:
>
> Mike> On Thu, 15 Jun 2000, Kees Bakker wrote:
> >> With test1-ac17 I still have very bad performance, because kswapd eats
> >> almost all of the CPU cycles. The bare test1 was a little better, but not
> >> much.
>
> Mike> Hmm, plain test1 gives MUCH (>100%) better performance here (128mb.. not
> Mike> tested under ram-impoverished config) Have you tried 99-pre5 on this box?..
> Mike> pre5 was the last kernel which gave me nice all-around (vm) preformance.
>
> No I haven't tried 99-pre5. 2.4.0-test is the first kernel I tried on this
> machine. Kswapd uses about 86-90% CPU. Then I grabbed ac17, but now kswapd
> takes even more CPU.
I suggest giving pre5 a trial run. I suspect that it will work quite
well on your low ram box. If this suspicion proves accurate, that bit
of info could be useful to the vm wizards. [mvho: if it takes your box
20 minutes to an hour to boot with >=test1 kernels, further testing
of these kernels on your hardware is a pointless waste of your time]
-Mike
-
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/
This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:12 EST