This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Basic block renumbering removal
- From: Jan Hubicka <jh at suse dot cz>
- To: Richard Henderson <rth at redhat dot com>, Jan Hubicka <jh at suse dot cz>,gcc-pdo at atrey dot karlin dot mff dot cuni dot cz, gcc-patches at gcc dot gnu dot org
- Date: Fri, 17 May 2002 13:48:11 +0200
- Subject: Re: Basic block renumbering removal
- References: <20020429151942.D11370@redhat.com> <20020430160511.GK18000@atrey.karlin.mff.cuni.cz> <20020430163200.A3211@redhat.com> <20020514091823.GN6514@atrey.karlin.mff.cuni.cz> <20020514172803.GC1535@atrey.karlin.mff.cuni.cz> <20020515072607.GH4292@atrey.karlin.mff.cuni.cz> <20020515151110.GA24680@atrey.karlin.mff.cuni.cz> <20020515162535.D25141@redhat.com> <20020516062436.GA1786@atrey.karlin.mff.cuni.cz> <20020516103755.B25733@redhat.com>
> On Thu, May 16, 2002 at 08:24:36AM +0200, Zdenek Dvorak wrote:
> > I have removed as much as I found (but surely some of them remains;
> > I really am not able to find all in 16000 lines of patch, sorry).
>
> Thanks.
>
> > > Second, this test exists only to cut down the number of redundant
> > > tests. Thus we are not interested in the actual ordering of the
> > > blocks, only that every block has a unique index. Thus the original
> > > test is still correct.
> >
> > I had two reasons for this:
> > 1) I wanted to keep semantics the same as before, in order to make bughunting
> > easier.
> > 2) It did not work other way (in cfg-branch; I didn't check this after
> > adapting the patch for mainline).
>
> Hum. I guess we can leave this for now, but I'm certain this
> will turn into a quadratic performance problem with some test
> case. Please figure out how to address this.
>
> I suspect that both here and in back_edge_of_syntactic_loop_p
> you can arrange for the block indicies to be monotonicaly
> increasing.
I am happy to address these issues by incremental patch later.
I plan to do couple of similar fixes/cleanups and tweeks for
different part.
> Since I don't want to read through a patch of this size again,
> and none of the points above is critical to correctness, I've
> committed the patch as-is.
Great :)
Honza
>
>
> r~