Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!think.com!mintaka!bloom-beacon!eru!hagbard!sunic!mcsun!ukc!tcdcs!vax1.tcd.ie!hughesmp
From: hughesmp@vax1.tcd.ie
Newsgroups: comp.sys.acorn
Subject: Re: Proposed ARM enhancements
Message-ID: <1991Apr22.013410.1@vax1.tcd.ie>
Date: 22 Apr 91 01:34:10 GMT
Sender: news@cs.tcd.ie
Organization: Trinity College Dublin
Lines: 14
Nntp-Posting-Host: vax1

It's all very nice talking about loads of new instructions for future
ARMs, but they would only be of use to home users, who wrote their
own code, because anything commercially released using the instructions
wouldn't work on lesser chips. The alternative would be to write a 'ARM3'
patch module, and so on, for each chip down the range, as each new chip was
released, which would trap undefined instrucions and do them in software...
Not that nice really. I'd prefer to see basic speed, video, cache, and memc
improvements first... Backwards compatibility is of prime importance in
developing further chips in the series.

(I'm not saying the ARM3 is slow, or has a bad cache, I just don't
have one, but we can always do with more speed...)

Merlin.