*BSD News Article 73828


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.mel.connect.com.au!news.mira.net.au!news.vbc.net!alpha.sky.net!news.aimnet.com!ossi.com!agate!howland.reston.ans.net!newsxfer.itd.umich.edu!news.mathworks.com!enews.sgi.com!fido.asd.sgi.com!neteng!lm
From: lm@neteng.engr.sgi.com (Larry McVoy)
Newsgroups: comp.os.linux.networking,comp.unix.bsd.freebsd.misc
Subject: Re: TCP latency
Followup-To: comp.os.linux.networking,comp.unix.bsd.freebsd.misc
Date: 15 Jul 1996 21:36:44 GMT
Organization: Silicon Graphics Inc., Mountain View, CA
Lines: 54
Message-ID: <4sedlc$f0l@fido.asd.sgi.com>
References: <4paedl$4bm@engnews2.Eng.Sun.COM> <4s8cuq$ljd@bertrand.ccs.carleton.ca> <31E7C0DD.41C67EA6@dyson.iquest.net> <4s8tcn$jsh@fido.asd.sgi.com> <DuI083.FH3@kithrup.com>
Reply-To: lm@slovax.engr.sgi.com
NNTP-Posting-Host: neteng.engr.sgi.com
X-Newsreader: TIN [version 1.2 PL2]
Xref: euryale.cc.adfa.oz.au comp.os.linux.networking:45340 comp.unix.bsd.freebsd.misc:23677

Sean Eric Fagan (sef@kithrup.com) wrote:
: In article <4s8tcn$jsh@fido.asd.sgi.com>,
: Larry McVoy <lm@slovax.engr.sgi.com> wrote:
: >Nobody said that Linux' TCP latency under load is faster,

: Part of the reason John is so upset, I think, is because Linux users
: (including Larry) used to point out how wonderful Linux' context switching
: numbers were -- wonderfully low, making things so fast, right?  But it
: bombed under load, meaning that anyone who had a heavily-used machine (such

Umm, this is old, old news.  The context switch alg was rewritten back
around 1.3.28, well before the Usenix paper & conference, it it does
not bomb at all.  In fact, I'll venture a guess that it is the best
performing scheduler for multi user, I/O style (your typical Unix
workload) in existance.  

: That doesn't, however, tell the whole story.  Sure, you might get a context
: switch number from lmbench of three nanoseconds, but how well does that
: scale with having twelve million processes, half of which are ftp processes
: over your 4GB/s networking interface, and the other half are doing a build
: of the entire OS from scratch?  (Okay, so I'm exaggerating a bit, and
: assuming that hardware will continue to improve ;).)

It's a good point, and I hope to start addressing that in 2.0.

: >BSD guys:  "your benchmark sucks!  your numbers are wrong!  you mislead the
: >	    world!  you suck!  whine!"

: I could point out that you, Larry, have made it obvious to some people
: (including myself) that if we didn't want to discuss Linux with you, we
: weren't welcome to discuss *anything* with you.  

That's unfair.  The full context of that was that I wasn't interested in
hiring you and allowing you to work on free software if the free software 
was BSD based.  If there's money involved, and I have some say over that
money, then, yeah, I want you to work on Linux, I happen to think it is
a better use of your time.  If you want to get paid, then you have to be 
willing to play by your employer's rules.  

: However, I don't know if
: you have taken that approach with lmbench; a couple of comments I've heard
: have indicated that that might be the case, but I haven't heard or seen
: anything from you about it.

lmbench is OS neutral.  The only "stacking" I do is that people that
are fun to talk to generally have more influence on my work than those
who scream at me (hey, I'm human just like anyone else - I like talking
to people that are fun to talk to.  Surprise, surprise.)  But I'm very
fair about what gets stuck in the benchmark and I routinely work with
folks from Intel, Sun, and SGI, as well as people like Linus.  If the
FreeBSD crowd wanted to be part of the process, that's fine with me.
--
---
Larry McVoy     lm@sgi.com     http://reality.sgi.com/lm     (415) 933-1804