Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!yetti!geac!daveb
From: daveb@geac.UUCP (Brown)
Newsgroups: comp.sys.mac
Subject: Re: Mac C Compilers, Benchmarks, Stupidity
Message-ID: <1179@geac.UUCP>
Date: Wed, 19-Aug-87 13:29:42 EDT
Article-I.D.: geac.1179
Posted: Wed Aug 19 13:29:42 1987
Date-Received: Fri, 21-Aug-87 02:28:10 EDT
References: <3560@cit-vax.Caltech.Edu> <3676@sdcsvax.UCSD.EDU>
Reply-To: daveb@geac.UUCP (Dave Collier-Brown)
Organization: The little blue rock next to that twinkly star.
Lines: 25

In article <3676@sdcsvax.UCSD.EDU> jww@sdcsvax.UCSD.EDU (Joel West) writes:
>While I might disagree with the tone, I certainly agree that
>the existing 68000 compilers for the Mac are disappointing
>in their use of peephole optimization.  The 68000 is no VAX,
>but it has many peephole optimizations that can be done;
>while every compiler does not remove needless code
>(like ignoring the function result) is beyond me.

It is interesting to note that the Pascal compilers don't seem to be
as bad.  In a time-critical loop, I found that writing an explicit
array addressing calculation in MPW Pascal produced code only one
instruction away from my optimal assembler (a register transfer). I
promptly switched to using the Pascal, since it didn't have the
parameter passing overhead of the quasi-inline assembler inserter.

Can someone with MPW please dump and annotate the inefficencies
in the Pascal compiler, please? pretty please?

--dave (I am _NOT_ a DCB) collier-brown

-- 
 David Collier-Brown.                 {mnetor|yetti|utgpu}!geac!daveb
 Geac Computers International Inc.,   |  Computer Science loses its
 350 Steelcase Road,Markham, Ontario, |  memory (if not its mind)
 CANADA, L3R 1B3 (416) 475-0525 x3279 |  every 6 months.