Re: signifanctly slowdown of FreeBSD 8.0-CURRENT/amd64

--- On Sat, 5/30/09, Kip Macy <kmacy_at_freebsd.org> wrote:

> From: Kip Macy <kmacy_at_freebsd.org>
> Subject: Re: signifanctly slowdown of FreeBSD 8.0-CURRENT/amd64
> To: "bf" <bf2006a_at_yahoo.com>
> Cc: freebsd-current_at_freebsd.org, attilio_at_freebsd.org, ohartman_at_mail.zedat.fu-berlin.de
> Date: Saturday, May 30, 2009, 11:55 PM
> >
> > I'm running the r193133 amd64 with a custom kernel and
> all debugging off
> > on an AMD Athlon64 3400+ single-core, and I haven't
> noticed any significant
> > slowing, although I haven't been doing any systematic
> benchmarking.
> >
> > What would be the penalties of running an SMP -CURRENT
> kernel on
> > single-core hardware with no hyperthreading? Can
> anyone quantify the
> > typical added overhead?  Or, counterintuitively,
> would an SMP kernel
> > be better in some ways?
> >
> 
> He is trying to diagnose if the problem was introduced by
> enabling
> adaptive spinning on sx locks. They're only enabled on SMP
> kernels.
> 
> Cheers,
> Kip
> 

So I inferred from his request to have Oliver (the original poster)
revert r193011.  But it seems unlikely that this is solely or even mostly
responsible for the slowdown, as Oliver reported that it first began
to occur several weeks ago, and Attilio only committed r193011 on Friday
-- unless Oliver independently added ADAPTIVE_SX to his custom kernel
a few weeks ago.  In any event, I'm still interested in any reports of
the relative performance of SMP vs. UP kernels on UP hardware that 
anyone is able to share.

Thanks,
         b.