*BSD News Article 62319


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!news.gmi.edu!zombie.ncsc.mil!news.mathworks.com!newsfeed.internetmci.com!inet-nntp-gw-1.us.oracle.com!news.caldera.com!news.cc.utah.edu!park.uvsc.edu!usenet
From: Terry Lambert <terry@lambert.org>
Newsgroups: comp.unix.bsd.freebsd.misc,comp.os.linux.development.system
Subject: Re: The better (more suitable)Unix?? FreeBSD or Linux
Date: 16 Feb 1996 23:27:53 GMT
Organization: Utah Valley State College, Orem, Utah
Lines: 47
Message-ID: <4g33tp$esr@park.uvsc.edu>
References: <4er9hp$5ng@orb.direct.ca> <4fjodg$o8k@venger.snds.com> <311DA774.167EB0E7@FreeBSD.org> <31220995.C4C54C1@acm.org> <4g0sam$r6p@agate.berkeley.edu>
NNTP-Posting-Host: hecate.artisoft.com
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:14486 comp.os.linux.development.system:18180

nickkral@america.CS.Berkeley.EDU (Nick Kralevich) wrote:
] In the extended abstract, they claim "On small file workloads, LInux is 
] order of magnitude faster [10 times] than the other systems [FreeBSD 
] and Solaris]".  (items in brackets were added by me).

From the same paper:

"Given that we measured the average non-cached seek time of these
 systems to be 14ms (Figure 11), Linux is clearly not accessing
 the disk during this benchmark.  This is because the Linux file
 system, ext2fs, uses and asynchronus metadata update policy,
 unlike the FreeBSD and Solaris file systems.  While this gives
 Linux a performance advantage, it could result in losing more
 data after a system crash."

Also:

"FreeBSD writes files of less than eight megabytes 50% faster
 than Solaris or Linux.  Linux maintains less than half the
 write bandwidth of FreeBSD or Solaris for almost all file sizes."

[ ... more benchmarks ... ]

The benchmarks described are against the 2.0 Alpha release.  They
are prior to the buffer cache unification.  The Lai/Baker paper
gives a better set of numbers.

Note that the Lai/Baker paper did not compare NFSv3 vs. NFSv2
performance, nor did it operate with a variety of NFS server
platforms (as noted in the paper itself).

It should also be noted that the pipe benchmark section has been
totally invalidated in -current by John Dyson's rewrite of the
pipe code.  I thought the pipe benchmark to be a rather bogus
test in any case, but John has "fixed" it anyway (it outperforms
Linux).


Not that I'm terribly interested in discussing non-file system
issues anyway.


                                        Terry Lambert
                                        terry@cs.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.