Newsgroups:	comp.os.msdos.djgpp
Date:	Thu, 6 Mar 1997 13:58:18 -0500 (EST)
From:	Peter Berdeklis <peter AT atmosp DOT physics DOT utoronto DOT ca>
To:	Leath Muller <leathm AT solwarra DOT gbrmpa DOT gov DOT au>
cc:	djgpp AT delorie DOT com
Subject:	Re: Netlib code [was Re: flops...]
In-Reply-To:	<199703060354.NAA20358@solwarra.gbrmpa.gov.au>
Message-ID:	<Pine.SGI.3.91.970306134622.3600A-100000@atmosp.physics.utoronto.ca>
MIME-Version:	1.0

On Thu, 6 Mar 1997, Leath Muller wrote:
> > As to the problem of gcc reloading memory locations already in registers, 
> > this is not just a problem of FPU code.  I used the -S option to generate 
> > asm code for a interrupt handler that I wrote in C.  The asm code had _a 
> > lot_ of unnecessary register reloads that I had to eliminate.  I think 
> > that this is a deficiency in the gcc optimizer.  (Considering the 
> > performance of DJGPP relative to other compilers, I assume that the 
> > problem is not unique.)
> 
> Are the official gcc dudes (FSF is it?) looking at a pentium optimization
> for gcc (and thus DJGPP) in the near future? Is there a way to find out,
> or even make suggestions? (Well, I know there is, I just dont know where
> to go... :)

There is already a pentium patch for gcc.  I don't know where to find it, 
but it was discussed a month or two ago on the newsgroup.  Check DJ's 
mail archives on the web site.

Note that the optimization problem I describe has nothing to do with the
Pentium (I'm using 486's and a 386!:).  This seems to be a general problem
with gcc's optimizer tracking register variables in its object code.  I
would assume this problem is also present for other architectures,
although I don't know any other asm language to check, e.g. using gcc on
the SGI at work.  Any one more knowledgable willing to take a look?

---------------
Peter Berdeklis
Dept. of Physics, Univ. of Toronto

- Raw text -