On Mon, 1 Oct 2001, Lorenzo Allegrucci wrote:
> Disclaimer:
> I don't know if this "benchmark" is meaningful or not, but anyhow..
I'm not sure either, since qsort doesn't really have much
locality of reference but just walks all over the place.
This is direct contrast with the basic assumption on which
VM and CPU caches are built ;)
I wonder how eg. merge sort would perform ...
> Below are linux-2.4.10 results
> real 4m54.728s
>
> kswapd CPU time: 3 seconds
> qs RSS always on 238-240M, very stable never below 235M.
> .. and 2.4.10-ac2 results
> real 6m2.139s
>
> kswapd CPU time: 20 seconds
> qs RSS never above 204M, average value 150M.
The RSS thing is just a side effect of how swap is allocated
and should have little or no influence on which pages are
kept in memory.
One thing which could make 2.4.10 faster for this single case
is the fact that it doesn't keep any page aging info, so IO
clustering won't be confused by the process accessing its
pages ;)
cheers,
Rik
-- IA64: a worthy successor to i860.http://www.surriel.com/ http://distro.conectiva.com/
Send all your spam to aardvark@nl.linux.org (spam digging piggy)
- 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 : Sun Oct 07 2001 - 21:00:16 EST