This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Insert on edge comment
- From: Andrew MacLeod <amacleod at redhat dot com>
- To: Jeff Law <law at redhat dot com>
- Cc: Chris Lattner <sabre at nondot dot org>, vmakarov at redhat dot com,matz at suse dot de, gcc mailing list <gcc at gcc dot gnu dot org>
- Date: 02 Jun 2003 11:13:31 -0400
- Subject: Re: [tree-ssa] Insert on edge comment
- References: <200306010202.h5121wcB024816@speedy.slc.redhat.com>
On Sat, 2003-05-31 at 22:01, law@redhat.com wrote:
> In message <Pine.LNX.4.44.0305311132300.22731-100000@nondot.org>, Chris Lattner
> writes:
> >IMO, the real problem in tree-ssa is that it uses a nested control flow
> >structure instead of a linearized one...
> Amen. It's a model I'd like to move away from, it introduces a fair
> amount of complexity. However, it's pretty far down on the list of
> things to do right now, but it's definitely on my list.
>
Well, the longer it left, the more impossible it becomes to remove...
So if there is a plan or thought to change it, it ought to be sooner
rather than later...
There are positives and negatives, we probably ought to enumerate them
so we know what we are looking at. Up to this point, the amount of
hassle for me directly hasn't been significant... but then we haven't
done any heavy duty loop work yet.. Thats where the tale is really going
to be told. We probably need a basic loop optimization written shortly
so we have an idea where we stand...
Andrew
Andrew