ULE/SCHED_SMP diff for 7.0, buildkernel & thanks.
Jeff Roberson
jroberson at chesapeake.net
Tue Jul 17 19:59:48 UTC 2007
On Tue, 17 Jul 2007, youshi10 at u.washington.edu wrote:
>
> Unfortunately those results are still based on 2.6.20, not 2.6.22 (2 minor
> patch revision difference).
>
> I assume that that's for a vanilla Linux kernel?
Look more closely; The green line is linux-2.6.21 with the new glibc. I
sent my kernel config to some linux hackers to look at. We removed some
minor debugging code to get these results. I have also tested with 2.6.22
with no real change. Although many other people saw great improvements.
I will update to fedora core 7 eventually although I supposedly have the
relevant fixes.
In my mind I hope that linux addresses their issues and that really isn't
my primary concern. My primary concern is that freebsd is now becoming
competitive on higher-end server class hardware for a variety of
workloads where it was not before. I benchmarked linux just to see where
we are at as they are generally considered fast and scalable.
The credit for the great improvements we've seen from 6 to 7 should go to
the many developers who have put a lot of hard work into locking
individual subsystems and the primitives. The scheduler is only able to
do better now because there is less contention over all.
Thank you,
Jeff
>
> -Garrett
>
> _______________________________________________
> freebsd-current at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe at freebsd.org"
>
More information about the freebsd-current
mailing list