This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [lno, tree-ssa] Fix kill_redundant_phi_nodes
- From: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Jeff Law <law at redhat dot com>,"gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 20 May 2004 20:22:24 +0200
- Subject: Re: [lno, tree-ssa] Fix kill_redundant_phi_nodes
- References: <20040413203933.GA19867@atrey.karlin.mff.cuni.cz> <1085076806.23047.55.camel@localhost.localdomain>
Hello,
> > * tree-ssa.c (raise_value): Removed.
> > (get_eq_name, check_phi_redundancy): New.
> > (kill_redundant_phi_nodes): Use standard algorithm.
> >
> I've been looking at kill_redundant_phi_nodes and I've now half
> convinced myself that we don't actually need it. I talked with Jeff
> about it and I will be looking at a reorganization of DOM to make it a
> collection of parts rather than one monolithic module.
>
> The thing is that DOM already has the propagation machinery needed
yes, most of the time. It is not hard to create an example where the
propagation done by DOM won't be sufficient, but they are a bit
artificial. Removing the iteration from kill_redundant_phi_nodes it
probably would not hurt much (but given that it does not cost us almost
anything anyway, I don't see much reasons for it).
Zdenek