*BSD News Article 30925


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!howland.reston.ans.net!gatech!prism!prism!not-for-mail
From: gt8134b@prism.gatech.edu (Robert Sanders)
Newsgroups: comp.os.386bsd.questions
Subject: Re: Linux or FreeBSD?
Date: 28 May 1994 20:07:19 -0400
Organization: Georgia Institute of Technology
Lines: 117
Sender: gt8134b@prism.gatech.edu
Message-ID: <2s8mbn$e1o@acmex.gatech.edu>
References: <CqH2z7.29E@dit.upm.es> <2s618a$34t@pdq.coe.montana.edu> <2s86fj$cn4@acmex.gatech.edu> <hastyCqJBED.9yz@netcom.com>
NNTP-Posting-Host: acmex.gatech.edu

hasty@netcom.com (Amancio Hasty Jr) writes:

>In article <2s86fj$cn4@acmex.gatech.edu> gt8134b@prism.gatech.edu (Robert Sanders) writes:
>>nate@bsd.coe.montana.edu (Nate Williams) writes:
>>
>>>In article <CqH2z7.29E@dit.upm.es>,
>>>GARCIA VALDEARENAS <cdt94001@oasis.dit.upm.es> wrote:
>>>> Linux is faster than FreeBSD, but has a very poor network support. If

Okay, before I have to fend off Amancio's hostile remarks, let me emphasize
that all the "performance" issues I brought up were simply to help explain
why several people who have used Linux and *BSD have remarked that Linux
was "faster."  I didn't want to start yet another penis-size war, just
explain that design decisions made by both sides will affect perceived
speed.  Sheesh.

>>meta-data updates (I turned that option for Linux on and the slowdown is 

>** I have never lost a file system on any of my FreeBSD systems... **

Er, ok.  I never said you did, and I never said that the synchronous updates
were a bad thing.  For your information, I've lost filesystems twice, both
times due to bad disks; Linux's ext2fs has proven very reliable for me.
The fact is that synchronous updates are merely one small measure against
filesystem corruption; recently-updated files may have a better chance
against corruption than on a non-synchronous system, but serious damage
won't care either way.  I'd rather trust fsck when I have problems than suffer 
throughput loss all the time for this unlikely occurrence.

If you want to impress me with filesystem stability, show me BSD 4.4's
log-structured filesystem running under FreeBSD.  Then I'll agree that
*BSD has a definite advantage over Linux in that area.

>>similar), the shared libraries, whilc more conceptually "cleaner", were
>>slower than Linux's (or, at least, executable loading seemed slower), and

>Does Linux have Dynamic Shared Libraries?
>Because we do ....

I'm very well aware of the advantages of both implementations, thank you.
FreeBSD's implementation is inherently slower, but the tradeoff is a "cleaner"
system.  What do you mean by "Dynamic Shared Libraries"?  Linux has shared 
libraries, and symbols in the executable can override those in the shared
library.  But is the linking done at runtime?  No.  It's done at compile-time.
And we save a lot of CPU and page-dirtying for it.

>>the base system was very spare (= didn't include perl, or much of any non-
>>BSD utility).                                    ^^^^

>If you look just a tiny bit at any freebsd ftp site you will
>definitely fine ported packages for sound, lang, etc...

Yes, thank you.  We did do that, but we were rather surprised that the "base" 
system didn't include many programs we had come to expect.  Please read
my posts more closely before mouthing off.

>>Another problem was that the sound driver was apparently an old version and 
>>produced poor results when used with GMOD on our Gravis Ultrasounds.

>Well if you get an old kernel you are liable to get poor results.
>GMOD works fine over here...

It was FreeBSD 1.1 Gamma.  At the time, there was no newer kernel except
one culled from the -current tree.  I thought you *BSD people were so
uppity about not having to play patch-of-the-day like Linuxers supposedly
do?

>Also, Jim Lowe (for FreeBSD) modified the sound driver to support VAT, Van
>Jacobsen's visual audio tool. What is vat? well it is currently being
>used in the internet for voice conferencing, radio broadcast, or just
>to chat with someone on the internet.
>	The *sources* for vat are not available!

You keep harping on this in a thousand newsgroups.  Frankly, Amancio, if I
want radio I will turn on my *radio*.  I'm not interested in some bandwidth-
sucking audio net.geek IRC that Van Jacobsen was too high-and-mighty to 
release the sources to.  And if I were, it wouldn't be too terribly
difficult to get BSD binary compatibility working under Linux.  We both
know how low a demand there is for that.

>vat also uses IP Multicasting not sure if IP multicasting is available
>for Linux.

Last I heard it wasn't officially available for FreeBSD, either, and you were
causing hackles to be raised by insisting that it was.  Linux's device drivers
support multicast lists, but the actual protocol stack isn't ready.  Seems
there isn't much demand for that, either.

>>A point I hesitate to bring up is that FreeBSD didn't seem to work with Linux's
>>NFS server.  Knowing that no *BSD will admit that it got things wrong, and not
>>knowing that Linux's NFS server got it right, all I'll say is that Solaris,
>>an MS-DOS client, and Chameleon NFS/32 for NT all worked perfectly with it, as
>>did Linux's client.

>How odd, that Sun OS  NFS seems to work with FreeBSD...

In which direction?  And, if you wish to exchange childish snotty remarks,
how odd that SunOS's NFS worked with Linux (Linux->SunOS and SunOS->Linux).

If anyone's curious, FreeBSD would work well as a NFS client, but when
reading large files, the reading program would die with a "protocol not
supported" error in the middle.  As I said, I'm not blaming anyone for
this, but it was one reason my roommate switched back to Linux.

I'd like to know why it didn't work: if it's Linux's problem, I'd like 
to get it fixed.  If it's FreeBSD's problem, I'm sure someone on the
core team would like to get it fixed.  If it's an interoperability
problem that nobody but the spec can be blamed for, then perhaps
Linux (or FreeBSD) ought to be changed to conform to the "reference"
implementations.  And don't say FreeBSD *is* the reference implementation,
because that code was rewritten for Net/2.

-- 
 _g,  '96 --->>>>>>>>>>   gt8134b@prism.gatech.edu  <<<<<<<<<---  CompSci  ,g_
W@@@W__        |-\      ^        | disclaimer:  <---> "Bow before ZOD!" __W@@@W
W@@@@**~~~'  ro|-<ert s/_\ nders |   who am I???  ^  from Superman  '~~~**@@@@W
`*MV' hi,ocie! |-/ad! /   \ss!!  | ooga ooga!!    |    II (cool)!         `VW*'