This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Use of build_binary_op
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Pop Sébastian <pop at gauvain dot u-strasbg dot fr>
- Cc: Steven Bosscher <s dot bosscher at student dot tudelft dot nl>,Jason Merrill <jason at redhat dot com>, Paul Brook <paul at nowt dot org>,<gcc at gcc dot gnu dot org>
- Date: Thu, 26 Sep 2002 13:18:07 -0400 (EDT)
- Subject: Re: [tree-ssa] Use of build_binary_op
On Thu, 26 Sep 2002, [iso-8859-1] Pop Sébastian wrote:
> On Thu, Sep 26, 2002 at 04:08:42PM +0200, Steven Bosscher wrote:
> >
> > Think "legacy Fortran 77 code" here, with GOTO's everywhere. Eliminating
> > all those GOTO's would create way too many "on/off" variables (ref.
> > simple-goto-elim.c), most of which can probably never be eliminated
> > during optimization (if there's a lot of overlap for example).
> >
> > In general, I do not believe that "whole function" goto/break
> > elimination is a very useful transformation in most cases. Doing the
> > transformations on big trees is expensive, and sometimes it does more
> > harm than good.
> > (ref. http://www.sable.mcgill.ca/~hendren/ftp/erosa/thesis.ps.gz).
> >
> Yes but we could do something as in Open64: just put a limit on
> the number of gotos to be eliminated.
> The code in simple-[goto|break]-elim.c is just experimental, and I have to
> work a little on it in order to make it work again on GIMPLE.
>
>
If you want my patches to greatly speed it up again, let me know.
They never were committed by you or anything.
--Dan