Newsgroups: comp.arch
Path: utzoo!utgpu!watserv1!maytag!watdragon!watsol.waterloo.edu!tbray
From: tbray@watsol.waterloo.edu (Tim Bray)
Subject: F.P. vs. arbitrary-precision
Message-ID: <1990Sep12.220536.7986@watdragon.waterloo.edu>
Sender: daemon@watdragon.waterloo.edu (Owner of Many System Processes)
Organization: University of Waterloo
References: <3755@osc.COM> <4513@taux01.nsc.com> <119244@linus.mitre.org> <6837.26e7ee92@vax1.tcd.ie>
Date: Wed, 12 Sep 90 22:05:36 GMT
Lines: 17

Some time ago, rwallace@vax1.tcd.ie wrote:
>There is practically no processing done which depends on integer * and / being
>fast (accessing an array of structures doesn't count...

First off, it is VERY DANGEROUS to make statements like that without a LARGE
suite of benchmarks and end-user surveys to back them up.  Go ask the guys
who designed the sparc.

Second, I think this statement is wrong.  Associative lookup, i.e. symbol
table management, is fundamental to many modern applications - compilers,
interpreters, very high level languages, computer algebra systems,
PostScript.  Symbol tables are usually done with hashing.  Hashing functions
very often involve multiplication (but note the neat algorithm for string
hashing in the recent CACM).

Just another data point,
Tim Bray, Open Text Systems, Waterloo, Ont.
