From owner-freebsd-hackers Sun Feb 18 13:37:40 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id NAA05790 for hackers-outgoing; Sun, 18 Feb 1996 13:37:40 -0800 (PST) Received: from brasil.moneng.mei.com (brasil.moneng.mei.com [151.186.109.160]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id NAA05785 for ; Sun, 18 Feb 1996 13:37:37 -0800 (PST) Received: (from jgreco@localhost) by brasil.moneng.mei.com (8.7.Beta.1/8.7.Beta.1) id PAA08031; Sun, 18 Feb 1996 15:35:12 -0600 From: Joe Greco Message-Id: <199602182135.PAA08031@brasil.moneng.mei.com> Subject: Re: BSD/OS 2.1 To: hsu@clinet.fi (Heikki Suonsivu) Date: Sun, 18 Feb 1996 15:35:11 -0600 (CST) Cc: dennis@etinc.com, freebsd-hackers@FreeBSD.ORG In-Reply-To: <199602172006.WAA11931@cantina.clinet.fi> from "Heikki Suonsivu" at Feb 17, 96 10:06:35 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Sender: owner-hackers@FreeBSD.ORG Precedence: bulk > In article <199602151844.NAA00492@etinc.com> dennis@etinc.com (dennis) writes: > I've finished installing the new release of BSD/OS....which by the way is rather > ... > I'd like to benchmark it against freebsd and perhaps publish the results (if > they're > favorable, of course )...can anyone point me at some good, portable, widely > accepted benchmark utilities. Obviously network-related stuff is of interest > > Allow 50-100 users to log in and use anything they want, including > installing new software (non-root) and run it, and tell us how long the > thing stays up. For large site use, it is the one and only benchmark. IMHO that is silly. That is not a benchmark, in any way, shape, or form. Users do not produce repeatable results. This is fine as a stress test, a reliability test, but it is not a benchmark. A benchmark is taking a test and repeating it under controlled conditions on controlled hardware. Typically you try to retain as much similarity between the variables as possible (i.e. FreeBSD vs. Linux, but both on the same hardware platform). If you do not do this, the results are meaningless (i.e. FreeBSD vs. Linux, Linux on a P90 and FreeBSD on a 386sx/16). The benchmark can be arbitrarily complex (i.e. run a Web server on a router) as long as there is some useful conclusion to be drawn and the benchmarks are run in a meaningful way. For what it's worth, as often as I tend to have differences of opinions with Dennis, I think it's a good thing that somebody is LOOKING at BSD/OS and seeing what's going on. I for one would be very interested in any relative comparisons. (Thanks in advance for any results, Dennis!) Mind you, I'm not saying somebody shouldn't stress test it, I'm just saying that that's not a benchmark at all.. ... Joe ------------------------------------------------------------------------------- Joe Greco - Systems Administrator jgreco@ns.sol.net Solaria Public Access UNIX - Milwaukee, WI 414/546-7968