Date: Thu, 20 Jun 1996 13:33:54 -0500 (EST) From: John Dyson <dyson@dyson.iquest.net> To: jkh@time.cdrom.com (Jordan K. Hubbard) Cc: hackers@FreeBSD.ORG Subject: Re: In case anyone's interested in P5 vs P6 benchmarks. Message-ID: <199606201833.NAA09813@dyson.iquest.net> In-Reply-To: <15977.835291316@time.cdrom.com> from "Jordan K. Hubbard" at Jun 20, 96 10:21:56 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > I ran the full "suite" of benchmarks we have in /usr/ports/benchmarks > against a quiet P5/133 and a P6/200 of otherwise similar attributes > (same disk controller, similar drives, etc). > > Results in: > > ftp://time.cdrom.com/pub/ > -rw-rw-r-- 1 jkh bin 28297 Jun 20 10:20 p5.time.times > -rw-rw-r-- 1 jkh bin 27762 Jun 20 10:20 p6.calweb.times > > The p5 was my own machine, time.cdrom.com. The p6 is owned by CalWEB. > > Jordan I just looked at the results, and have a few comments. The results are as expected, but can be confusing. The p6 times were done on 2.1. Relative to 2.1, 2.2 is approx 2x faster on Execl throughput, pipe throughput, shell scripts and fork times. The 2.2 code gives you 'almost p6 on 2.1' VM performance on a P5. If you look at the 2.2-current numbers, and assume the P5 processor speed is 10x the reference, we scale very well. Unfortunately, on 2.1, we did not fare quite as well in some areas. It would be *really nice* to see results for 2.2-current (as of today, with all of the VM bugfixes) on a fast p6!!! (Of course, also compile with all of the gcc options like -fomit-frame-pointer -O2, etc -- just like the 'other' OS does.) But anyway, thank you, Jordan for some valuable numbers, and hopefully, there will be more to come!!! John
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199606201833.NAA09813>