From owner-freebsd-questions Sat Feb 14 20:40:47 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA22100 for freebsd-questions-outgoing; Sat, 14 Feb 1998 20:40:47 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from alexanderwohl.complete.org (ict61.southwind.net [206.53.100.61]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA22057 for ; Sat, 14 Feb 1998 20:40:26 -0800 (PST) (envelope-from jgoerzen@alexanderwohl.complete.org) Received: from localhost (localhost [127.0.0.1]) by alexanderwohl.complete.org (8.8.7/8.8.7) with SMTP id WAA00332; Sat, 14 Feb 1998 22:38:12 -0600 (CST) (envelope-from jgoerzen@alexanderwohl.complete.org) Date: Sat, 14 Feb 1998 22:38:12 -0600 (CST) From: John Goerzen X-Sender: jgoerzen@alexanderwohl To: Konrad Heuer cc: Vincent Defert , freebsd-questions@FreeBSD.ORG Subject: Re: FreeBSD vs Linux In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Thu, 12 Feb 1998, Konrad Heuer wrote: > 1. The Linux scheduler which is very different from other UNIX > schedulers (and thus the FreeBSD scheduler) behaves very poor when the > system is heavily loaded (no fair scheduling!). In comparing Linux to SunOS and Solaris in heavliy-loaded systems, I can say that Linux performed much better. However, I have not compared it directly to FreeBSD. > 2. The Linux NFS implementation doesn't compare to the FreeBSD > implementation. It's neither Version 3 as in FreeBSD nor does it support > write-behind by the nfsiod daemons. So for NFS clients which need write > access Linux is a bad choice (only about 1/3 of the FreeBSD performance). This is correct, although I haven't directly compared the speed. Again, though, it easily outperformed SunOS. > 3. Since the Linux 2nd Extended File Systems by default also buffers inode > and comparable data it's faster in operations like unpacking tar files. Anything that touches a lot of inodes. This includes rm -f's, compiles, etc. However, I wouldn't really give Linux an edge here since it is just an issue of semantics -- what is the default. Just because it isn't enabled in FreeBSD by default doesn't mean that FreeBSD should be faulted for it. > With some risk one can mount a FreeBSD Fast File System with an async > option but then the dirty buffers containing critical data will be flushed > only in 30 second intervals. Linux runs a special bdflush daemon with a 5 > second interval for critical data which is more reliable. Both operating systems let the user tweak these values. You could set FreeBSD's default of 30 seconds down to 5 seconds or set Linux's default of 5 seconds up to 30. On my laptop, when testing either OS, I set it to an hour. I also set the noatime option in both systems. This let the hard drive spin down a lot more. And since a laptop has a battery, and both kernels are quite solid, I have no stability problems by doing that. > On the other hand I found the sequential writes and (much more important) > reads of larger files are about 30%..50% faster with FreeBSD and the FFS. I have found no large difference here. > > Last, non-technical point: > > For people like me who are accustomed to UNIX for years FreeBSD is very > pleasing since it is in fact *UNIX* although it doesn't wear the > trademark. Linux is Linux and no UNIX - it's a reimplementation with a lot > of more or less perceptible small differences. This, I think, is a matter of personal preferences. I prefer the approach of Debian to that of FreeBSD. However, as I said in a different message, you cannot in cases like this compare Linux directly to FreeBSD. For instance, Unifix Linux is certified POSIX compliant, etc. John To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message