Date: Mon, 10 Mar 1997 21:31:34 +0100 (MET) From: Luigi Rizzo <luigi@labinfo.iet.unipi.it> To: bde@zeta.org.au (Bruce Evans) Cc: dg@root.com, bag@sinbin.demos.su, hackers@freebsd.org Subject: Re: Q about 100Mb ether cards Message-ID: <199703102031.VAA27621@labinfo.iet.unipi.it> In-Reply-To: <199703102014.HAA11658@godzilla.zeta.org.au> from "Bruce Evans" at Mar 11, 97 07:14:28 am
next in thread | previous in thread | raw e-mail | index | archive | help
> >> been able to get it to work. The Pro/100B is well supported; it's what I'm > >> using in wcarchive and I haven't had any problems. It consumes about half > >> the CPU time that the de driver does and has about the same level of > > > >that's curious though, since both do DMA from/to mbufs. What CPU > >overhead are we talking about, 1% or 10% ? > > On a P5/133 with a 21040 (driven by de0), total network overhead for > ttcp -t is 10-15%. The idle overhead is about 11% (systat reports 6-7% > Sys and 3-4% Intr, and a process that doesn't do any memory accesses > slows down by about 11% when ttcp is running). A process that does lots > of memory accesses slows down by 15%. I am not sure on how to interpret these data. I guess the 6-7% is common to all drivers (the 21040 is a 10Mbit/s thing, right ?) and the remaining 3-4% changes from board to board. I guess I'll have to determine numbers for 100 Mbit/s boards and the 21140... can you tell me more details on your test setup ? Luigi -----------------------------+-------------------------------------- Luigi Rizzo | Dip. di Ingegneria dell'Informazione email: luigi@iet.unipi.it | Universita' di Pisa tel: +39-50-568533 | via Diotisalvi 2, 56126 PISA (Italy) fax: +39-50-568522 | http://www.iet.unipi.it/~luigi/ _____________________________|______________________________________
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199703102031.VAA27621>