Date: Wed, 09 Sep 1998 17:10:06 -0700 From: Mike Smith <mike@smith.net.au> To: current@FreeBSD.ORG Subject: ELF worldstone (etc.) Message-ID: <199809100010.RAA02290@dingo.cdrom.com>
next in thread | raw e-mail | index | archive | help
Well, I guess it had to be fairly obvious that I had an ulterior motive
asking whether anyone had done any ELF worldstone testing. So here are
the numbers that you're looking to beat.
Benchmark: mm:ss Notes
--------------------------------------
GENERICstone 00:42 (1)
ELF-worldstone 42:35 (2)
ELF-NOAOUT-worldstone 32:01 (2)(3)
(1) /usr/src on a single Quantum Atlas II, building GENERIC.
(2) /tmp, /usr/src and /usr/obj on a single 650MB MFS filesystem.
Splitting /usr/src and /usr/obj onto two separate Atlas IIs
increased the ELF-worldstone time by about 4 minutes
(async, not using softupdates).
(3) there was a CVSup fetch running for some of this.
/etc/make.conf had 'CFLAGS+= -O -pipe' enabled, no part of the tree was
turned off, all builds were with -j8. Perhaps significantly, mean CPU
utilisation was only a little over 200% for both worldstones.
Hardware for this exercise was generously loaned to FreeBSD Test Labs by
Intel, and software support to make it work came from Tor Egge and Steve
Passe, with contributions from other subscribers to the FreeBSD-smp
list.
System configuration summary:
CPU - 4x400MHz Intel Pentium II Xeon, 1MB L2 cache
Motherboard - Intel B0 'Bear', 450NX chipset
Memory - 1GB
Disk - 2x9GB Quantum Atlas II, Onboard Adaptec aic7880
--
\\ Sometimes you're ahead, \\ Mike Smith
\\ sometimes you're behind. \\ mike@smith.net.au
\\ The race is long, and in the \\ msmith@freebsd.org
\\ end it's only with yourself. \\ msmith@cdrom.com
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199809100010.RAA02290>