Path: utzoo!utgpu!news-server.csri.toronto.edu!dgp.toronto.edu!jonah
Newsgroups: comp.arch
From: jonah@dgp.toronto.edu (Jeff Lee)
Subject: Re: 386 Clones [really: IEEE floating point & various approaches; long]
Message-ID: <1990Nov1.232508.18287@jarvis.csri.toronto.edu>
References: <1990Oct26.015244.586@amd.com> <8464@scolex.sco.COM> <42597@mips.mips.COM> <4174@goanna.cs.rmit.oz.au> <42618@mips.mips.COM>
Date: 2 Nov 90 04:25:09 GMT
Lines: 13

mash@mips.COM (John Mashey) writes:
>Now, how about input from people who actually develop FP programs:
>a) Do you use traps for SIGFPE, or not?

I'm not a numerical analyst, but I did recently take a graduate course
on numerical software and the advice given by the professor was to
avoid traps and make explicit tests for +INF, -INF, NaN, and 0 where
appropriate.  Usually only one or two are appropriate in any given
case and explicitly checking makes sure that you *think* about the
possibilities for overflow/underflow, how to avoid it (if possible),
and how to retain accuracy.  The opinion on traps was that they were
generally more trouble than they were worth, especially if the trap
handler overhead was any significant amount.