[c-nsp] 7200 / NPE-G2

[Rodney Dunn]

On Wed, Apr 18, 2007 at 03:23:40PM +0200, Gert Doering wrote:
> Hi,
> 
> On Wed, Apr 18, 2007 at 08:41:42AM -0400, Rodney Dunn wrote:
> > While the graph of CPU vs. pps goes up pretty closely inline with
> > what a G1 will do the G2 will forward close to double the G1 in
> > packets per second approaching the no drop rate (NDR) the G2 can handle.
> 
> So what sort of magic happens when both G1 and G2 hits 100%, and the G2 then
> goes up to forward twice this "100% CPU" pps rate?

It's not magic. It has to do with the way CPU accounting is done on
the box in conjunction with a bit different memory architecture. I'd
have to go back and read the engineering documents to go real deep in to
it. It's been a few weeks since I read it.

Just trust me. ;)


> 
> gert
> -- 
> USENET is *not* the non-clickable part of WWW!
>                                                            //www.muc.de/~gert/
> Gert Doering - Munich, Germany                             gert at greenie.muc.de
> fax: +49-89-35655025                        gert at net.informatik.tu-muenchen.de