Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!solace!nntp.se.dataphone.net!nntp.uio.no!uninett.no!news1.best.com!bofh.noc.best.net!not-for-mail
From: rone@bofh.noc.best.net (Ron Echeverri)
Newsgroups: comp.unix.bsd.freebsd.misc,comp.unix.bsd.bsdi.misc,comp.sys.sgi.misc
Subject: Re: no such thing as a "general user community"
Date: 2 Apr 1997 16:02:46 -0800
Organization: fidgety systems administrators gmbh
Lines: 34
Message-ID: <5hus36$3ev$1@bofh.noc.best.net>
References: <331BB7DD.28EC@net5.net> <5hui15$8ru@fido.asd.sgi.com> <5hul1c$o4s$1@bofh.noc.best.net> <5hun4a$eav@fido.asd.sgi.com>
NNTP-Posting-Host: bofh.noc.best.net
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:38319 comp.unix.bsd.bsdi.misc:6556 comp.sys.sgi.misc:29623
In article <5hun4a$eav@fido.asd.sgi.com>,
Larry McVoy <lm@slovax.engr.sgi.com> wrote:
>: >Sure they are. The numbers don't represent any science. The code that
>: >generated the numbers, on the other hand, certainly does represent science.
>: >While people have pointed out a few bugs in lmbench 1.1, the code has
>: >proven to be quite accurate and robust. Is that not science?
>: Certainly. And running a benchmark on two operating systems, one
>: running on faster hardware, the other running on slower hardware, is
>: not science.
>It's not? What about Spec, TPC, LADDIS, Netperf, etc.? They do exactly
>what you just claimed is not science. So what's wrong with them? Are you
>saying that all benchmarks are valid only if they are run on identical
>hardware? Makes it kind of hard to compare vendor A with vendor B,
>doesn't it?
That depends on whether you are benchmarking the operating system or
the hardware, which are, indeed, two entirely different things. You
can either benchmark two different operating systems on the same type
of hardware, or you can benchmark two different machines running the
same program. Benchmarking operating system A on machine I and
operating system B on machine J is an exercise in masturbatory
numbercrunching unless you also benchmark A on J and B on I.
Of course, if i'm wrong, please excuse my ignorance; it's just this
"common sense" thing that keeps getting in my line of sight.
rone
--
Ron Echeverri Best Internet Usenet Administration rone@best.net
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Your Operating System Sucks! For a pamphlet with more info send $95 to
666 Millisoft Ln #666, REDmond WA 98666-0666