Re: volanoMark regression with kernel 2.6.26-rc1

[Dhaval Giani]

On Thu, May 08, 2008 at 01:53:33PM +0800, Zhang, Yanmin wrote:
> 
> On Thu, 2008-05-08 at 11:10 +0530, Dhaval Giani wrote:
> > On Thu, May 08, 2008 at 01:18:34PM +0800, Zhang, Yanmin wrote:
> > > 
> > > On Wed, 2008-05-07 at 23:03 +0530, Dhaval Giani wrote:
> > > > On Tue, May 06, 2008 at 05:22:07PM +0530, Dhaval Giani wrote:
> > > > > On Tue, May 06, 2008 at 10:06:30AM +0800, Zhang, Yanmin wrote:
> > > > > > Comparing with 2.6.25, volanoMark has big regression with kernel 2.6.26-rc1.
> > > > > > It's about 50% on my 8-core stoakley, 16-core tigerton, and Itanium Montecito.
> > > > > > 
> > > > > > With bisect, I located below patch.
> > > > > > 
> > > > > > 18d95a2832c1392a2d63227a7a6d433cb9f2037e is first bad commit
> > > > > > commit 18d95a2832c1392a2d63227a7a6d433cb9f2037e
> > > > > > Author: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
> > > > > > Date:   Sat Apr 19 19:45:00 2008 +0200
> > > > > > 
> > > > > >     sched: fair-group: SMP-nice for group scheduling
> > > > > >     
> > > > > >     Implement SMP nice support for the full group hierarchy.
> > > > > > 
> > > > > > If I reverse the patch with resolving some conflictions, volanoMark result could
> > > > > > be restored completely.
> > > > > > 
> > > > > 
> > > > > ok, that's bad. Let's get vatsa and Ingo also involved.
> > > > > 
> > > > 
> > > > Just to confirm, do you still have a performance regression with
> > > > !group_sched?
> > > I just tried it with CONFIG_GROUP_SCHED=n a moment ago. The regression becomes less than 3%.
> > > 
> > 
> > One more thing if you can try out, please set the shares for other users
> > to 2 except for the one which is running the benchmark. You can set it
> > at /sys/kernel/uids/<uid>/cpu_share
> I might try. There are only 2 users active in my system, root for background processes and mine
> for the testing. In the other hand, I kill most backgroud services when starting testing. So
> it might not have help.
> 

Even if one service is running as root, it is guaranteed 2/3 the CPU
(since root's shares are double that of other users by default). It
would cause issues for sure.

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