Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!mcvax!enea!tut!santra!jmunkki
From: jmunkki@santra.UUCP (Juri Antti Munkki)
Newsgroups: comp.sys.mac
Subject: Macsbug 5.5 and SANE in Mac II
Message-ID: <7384@santra.UUCP>
Date: Thu, 10-Sep-87 14:16:33 EDT
Article-I.D.: santra.7384
Posted: Thu Sep 10 14:16:33 1987
Date-Received: Sat, 12-Sep-87 18:36:33 EDT
Organization: Helsinki University of Technology
Lines: 20
Keywords: Macsbug, SANE, Macintosh, Macintosh II, benchmarks

While trying some benchmarks on my Mac II I noted that the results changed if
I booted without MacsBug. After some thinking I traced the problem to floating
point operations.

Trying the Mandelbrot programs posted in spring gave me some figures for proof.

Macsbug & Mandel881		5432
Mandel 881 without Macsbug	6653
Macsbug & MandelSane		1576
Mandelsane without Macsbug	2187
(I only tried the program for a minute or so per run, so these numbers may
 not be very accurate, but they clearly demonstrate the problem.)

Maybe someone would be interested in finding out what MacsBug does to slow
down the 68881...

Juri Munkki
mcvax!santra!jmunkki

"Using Hypercard isn't programming, but it's almost as much fun..."