patch for merging graphite branch (before tuplification)

[Jack Howarth]

Tobi,
    I am finding that the polyhedron tests aermod, capacita, channel,
gas_dyn, nf, protein, rnflow and testfpu all fail to compile with 
'-floop-strip-mine -floop-interchange -floop-block -fgraphite' which produces
errors of the form...

aermod.f90: In function ‘anyavg’:
aermod.f90:21764: internal compiler error: in get_first_matching_sign_row_index, at graphite.c:3853
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Is this another tuplification error that is already fixed in graphite
branch?
                   Jack


On Sat, Aug 09, 2008 at 10:55:09PM +0200, Richard Guenther wrote:
> On Sat, Aug 9, 2008 at 10:53 PM, Jack Howarth <howarth@bromo.msbb.uc.edu> wrote:
> > Richard,
> >   What are those new optimization options? The graphite wiki doesn't go
> > into those. Also are they already in gcc trunk or part of the currently
> > proposed graphite patch for gcc trunk?
> 
> -floop-block -floop-interchange -floop-strip-mine
> 
> Richard.