Date: Mon, 26 Jul 2004 20:00:10 -0700 (PDT) From: Kelly Yancey <kbyanc@posi.net> To: Don Bowman <don@sandvine.com> Cc: freebsd-net@freebsd.org Subject: RE: device polling takes more CPU hits?? Message-ID: <20040726195807.R76990@gateway.posi.net> In-Reply-To: <FE045D4D9F7AED4CBFF1B3B813C85337051D9445@mail.sandvine.com> References: <FE045D4D9F7AED4CBFF1B3B813C85337051D9445@mail.sandvine.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 26 Jul 2004, Don Bowman wrote:
> From: Luigi Rizzo [mailto:rizzo@icir.org]
> > On Mon, Jul 26, 2004 at 01:18:46PM -0700, Kelly Yancey wrote:
> > ...
> > > Out of curiousity, what sort of testing did you do to
> > arrive at these
> > > settings? I did some testing a while back with a SmartBits
> > box pumping
> > > packets through a FreeBSD 2.8Ghz box configured to route
> > between two em
> > > gigabit interfaces; I found that changing the burst_max and
> > each_burst
> > > parameters had almost no effect on throughput (maximum 1%
> > difference).
> >
> > fast boxes are pci-bus limited, not CPU limited(*) so
> > changing the burst
> > size (which basically amortizes some CPU costs) has little if any
> > effect.
>
> The PCI-X bus will probably be 64-bit 133MHz in this case,
> the limit moves up to the P64H2 hub for large packets,
> to the CPU for small packets. Polling becomes quite
> critical to prevent livelock.
>
Sorry, I should be been more clear. Polling certainly stopped livelock
under extreme load, however I never found much difference whether the
burst size was small or large. I was wondering if it was just the nature
of my test and if in other environments the burst_max and each_burst knobs
have a greater affect.
Kelly
--
Kelly Yancey - kbyanc@{posi.net,FreeBSD.org} - kelly@nttmcl.com
FreeBSD, The Power To Serve: http://www.freebsd.org/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040726195807.R76990>