*BSD News Article 39715


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!uwm.edu!spool.mu.edu!caen!usenet.coe.montana.edu!bsd.coe.montana.edu!nate
From: nate@bsd.coe.montana.edu (Nate Williams)
Newsgroups: comp.os.386bsd.misc
Subject: Re: How fast? [was: ... slugish ...]
Date: 19 Dec 1994 01:15:09 GMT
Organization: Montana State University, Bozeman  Montana
Lines: 42
Message-ID: <3d2mqt$iuq@pdq.coe.montana.edu>
References: <1994Nov28.194617.18912@system9.unisys.com> <3bf6ou$pm7@wup-gate.wup.de> <MICHAELV.94Dec9223853@MindBender.HeadCandy.com> <3cs7rd$k1u@news.demos.su>
NNTP-Posting-Host: bsd.coe.montana.edu

In article <3cs7rd$k1u@news.demos.su>, Ivan Popov <pin@demos.su> wrote:
>alexd@system9.unisys.com (Alex  Dumitru) writes:

>My remark about the compilation speed:
>
>I tried to compile same package (screen control libraries)
>on the same computer (386/25, 4Mb RAM)
>under FreeBSD-2.0 (generic kernel)
>and under Demos-86 (real mode UNIX for i8086 [sic], using 640K mem
>                    with /tmp in extended memory).
>
>The result: _user_ times used are almost the same (about 17 min)
>            _real_ times used are: 50 min under Demos-86,
>                                   90 min under FreeBSD-2.0
>In both cases other activity in the system was reduced to the minimum -
>no users, no activity of daemons.
>The difference between user and real times was due to I/O and swapping.
>(BTW in the case of Demos-86 all passes of the compiler and make itself
>fit into 640K, hence almost no swapping)

Port that compiler to FreeBSD and try it out.  GCC is a *pig* (and
that's being nice) and the major reason why development and performance
sucks on 4MB machines.  Between it and X makes running any of the free
OS's diffuclt to run well on low-memory hardware.  Now, if we could get
lcc and mgr to become standardized...... *grin*

If you didn't have to swap/page then I bet FreeBSD would have 'similar'
performance to you real mode UNIX.  However, give me access to your
real-mode box, and I'll bet I can get root in less than 24 hours. ;-)

This statement assumes that your real-mode unix was written and designed
for all x86's, and does not take advantage of the HW protection features
of the 386 or better processors and that I have access to the
c-compiler.


Nate
-- 
nate@bsd.coe.montana.edu     |  FreeBSD dude and all around tech.
nate@cs.montana.edu          |  weenie.
work #: (406) 994-5980       |  Unemployed, looking for permanant work in
home #: (406) 586-0579       |  CS/EE field.