Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!bloom-beacon!gatech!rutgers!bellcore!faline!ulysses!sfmag!sfsup!sfmin!lmg
From: lmg@sfmin.UUCP (L.M.Geary)
Newsgroups: comp.sys.ibm.pc,comp.arch,comp.unix.wizards
Subject: Re: 386 unix (Dhrystones)
Message-ID: <846@sfmin.UUCP>
Date: Fri, 18-Sep-87 10:17:04 EDT
Article-I.D.: sfmin.846
Posted: Fri Sep 18 10:17:04 1987
Date-Received: Sun, 20-Sep-87 16:04:46 EDT
References: <125@snark.UUCP> <299@nuchat.UUCP> <358@netxcom.UUCP> <306@nuchat.UUCP> <306@kksys.UUCP> <1337@van-bc.UUCP>
Organization: AT&T Information Systems, Summit N.J.
Lines: 16
Xref: mnetor comp.sys.ibm.pc:8014 comp.arch:2257 comp.unix.wizards:4351

[ Various Bell Technologies UNIX System info deleted. ]

> Dhrystone with registers was 2184.
                               ^^^^
I checked out a pre-release version of the raw Intel 386 port on
a Compaq 386 (2Mb RAM) and got similar Dhrystone numbers. However, I got
around 3350 Dhrystones using Microport V/AT 1.3.6 (the 286 binaries).
Turbo C gave around 3500 Dhrystones.

What's going on here? Earlier posted Dhrystone results showed 386 boxes
running the 386 Intel port at > 5000 Dhrystones. Now I'm seeing a 40%
performance DECREASE over the 286 versions of the system! And it isn't
just benchmarks; the system feels slower, too. Something is wrong.

					Larry Geary
					ihnp4!attunix!lmg