*BSD News Article 39729


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!agate!howland.reston.ans.net!EU.net!news.eunet.fi!news.spb.su!KremlSun!satisfy.kiae.su!demos!dnews-server
From: Ivan Popov <pin@kremvax.demos.su>
Newsgroups: comp.os.386bsd.misc
Subject: Re: How fast? [was: ... slugish ...]
Date: Mon, 19 Dec 1994 09:50:35 +0300
Organization: Demos, Moscow, Russia
Lines: 23
Sender: news-server@news.demos.su
Distribution: su
Message-ID: <LAxqIzkaT1@kremvax.demos.su>
References: <1994Nov28.194617.18912@system9.unisys.com>
    <3bf6ou$pm7@wup-gate.wup.de>
    <MICHAELV.94Dec9223853@MindBender.HeadCandy.com>
    <199412190232.UAA24259@dworkin.wustl.edu>
Reply-To: pin@kremvax.demos.su
NNTP-Posting-Host: news.demos.su
X-mailer: Mail [v2.1 SunOS]

In comp.os.386bsd.misc article <199412190232.UAA24259@dworkin.wustl.edu>
    you write:

>Ivan -
>    Your experiment is incomplete.   You haven't determined why
>there was a difference in performance and where the bottleneck is.

Well, I haven't that goal at all :)
I only needed to compile the new version of the package under both systems...
The only thing I added was "time makescript".
I know those systems are _very_ different, one 16-bit while the other 32,
one real mode, other - protected and so on.
Nevertheless, both are Unices and run on the same hardware.

I see that for relatively simple appllications
(say, alphanumeric without huge amounts of data - this is the case
when we are compiling reasonably sized C-programs)
the progress in systems' architecture didn't make any performance
improvements  :-(

Yours,
Ivan <pin@demos.su>