Path: utzoo!lsuc!ncrcan!ziebmef!cks
From: cks@ziebmef.uucp (Chris Siebenmann)
Newsgroups: comp.sys.amiga
Subject: Re: C vs. Assembler
Message-ID: <1988Jun22.190630.4151@ziebmef.uucp>
Date: 22 Jun 88 23:06:27 GMT
References: <2008@ssc-vax.UUCP>
Reply-To: cks@ziebmef.UUCP (Chris Siebenmann)
Organization: Ziebmef Public Access BBS/Unix
Lines: 22

In article <2008@ssc-vax.UUCP> dmg@ssc-vax.UUCP (David Geary) writes:
...
>...............  Sure, there are some assembly gurus out there who
>could take the assembly output by a C compiler, and optomize it for
>more speed, but I think our current compilers on the Amiga are good
>enough so that the gains made by optomization of the assembly are
>rarely justifiable, and/or rarely make a significant difference in speed.

 Unfortunately, Aztec C usually produces noticeably sub-optimal code,
as did old version of Lattice (I haven't seen the output of Lattice
4.0; anyone care to comment on its code quality?). Atzec C is a good
base-level optimizing compiler; it gets most of the basic
optimizations like common subexpression elimination (I won't call
constant folding an 'optimization'; I call compilers without it
'brain-damaged' instead), but compared to GNU C or Greenhills C it has
a long way to go.

-- 
	But he said leave me alone I'm a family man
	And my bark is much worse than my bite
Chris Siebenmann		uunet!utgpu!{ontmoh!moore,ncrcan}!ziebmef!cks
cks@ziebmef.UUCP	     or	.....!utgpu!{,ontmoh!,ncrcan!brambo!}cks