Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!srcsip!klemmer!vestal
From: vestal@SRC.Honeywell.COM (Steve Vestal)
Newsgroups: comp.arch
Subject: Re: Integer Multiply/Divide on Sparc
Message-ID: <52075@srcsip.UUCP>
Date: 2 Jan 90 17:58:52 GMT
References: <84768@linus.UUCP> <8840004@hpfcso.HP.COM> <1804@l.cc.purdue.edu> <1535@cbnewsi.ATT.COM> <28548@amdcad.AMD.COM> <1989Dec30.161619.22893@phri.nyu.edu> <1296@quintus.UUCP>
Sender: news@src.honeywell.COM
Lines: 17
In-reply-to: ok@quintus.UUCP's message of 1 Jan 90 01:24:04 GMT

In article <1296@quintus.UUCP> ok@quintus.UUCP (Richard A. O'Keefe) writes:
>   Fortran 8X and Ada also allow arrays with dynamic bounds.
>   [...]
>	   real array a[1:P,1:Q,1:R];

I believe that induction variable optimizations, which replace multiplications
by additions for certain array address computations within loops, work just
fine even with dynamic bounds.  My limited experience evaluating such
optimizations has been that they are surprisingly widely applicable for
matrix/vector manipulation code.  Personally, I suspect explicit
multiplies/divides/mods/rems in the application code might be the hazard to
watch out for rather than integer multiplications used in addressing
computations.

Steve Vestal
Mail: Honeywell S&RC MN65-2100, 3660 Technology Drive, Minneapolis MN 55418 
Phone: (612) 782-7049                    Internet: vestal@src.honeywell.com