This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Loop optimizer issues
- From: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>
- To: tm_gccmail at mail dot kloo dot net
- Cc: Diego Novillo <dnovillo at redhat dot com>,"gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>,Richard Henderson <rth at redhat dot com>, Jason Merrill <jason at redhat dot com>,pop at gauvain dot u-strasbg dot fr, Jan Hubicka <jh at suse dot cz>,Daniel Berlin <dberlin at dberlin dot org>
- Date: Tue, 29 Jul 2003 22:33:02 +0200
- Subject: Re: Loop optimizer issues
- References: <1059492808.3164.61.camel@frodo.toronto.redhat.com> <Pine.LNX.4.21.0307291316140.17919-100000@mail.kloo.net>
Hello,
> > I've been thinking about this proposal and I am not convinced that it
> > would be a good idea to merge the two branches.
> >
> > * The tree-ssa branch is primarily trying to replace RTL passes.
> > This may include the loop optimizer. So far, it is not clear
> > whether we will actually need to have a sophisticated loop
> > optimizer at the RTL level (if at all). The only thing that
> > comes to mind that may needed from RTL are:
> > * instruction lengths/costs to guide the unroller
> > * prefetching capabilities and memory costs to guide
> > software pipelining and memory hierarchy optimizations.
>
> We do BIV-to-GIV conversion and some addressing mode selection in the RTL
> loop optimizer. This is critical for targets such as the SH.
this is done by the new rtl-level optimizer. The support probably is
not flawless, but if you encounter some problems and provide me with
more details, I am fairly confident I am able to fix them quickly.
Zdenek