On Saturday, March 30, 2013 7:42:38 PM UTC+1, leif wrote:
>
> Jean-Pierre Flori wrote:
> > On Saturday, March 30, 2013 6:25:37 PM UTC+1, Jean-Pierre Flori wrote:
> > On Saturday, March 30, 2013 2:46:56 PM UTC+1, leif wrote:
> >
> > Jean-Pierre Flori wrote:
> > > On Saturday, March 30, 2013 7:15:10 AM UTC+1, leif wrote:
> > > Just for the record, I get ICEs with the GCC 4.7.2.p0
> > spkg (with Sun as
> > > and ld) on Solaris SPARC (32-bit) when compiling GAP and
> > libGAP with
> > > -O3; -O2 works in both cases. (I didn't get them with
> > the system-wide
> > > GCC 4.7.0 on {mark,mark2}.)
> > >
> > > I cannot reproduce these on the Solaris I have access to
> > (same as the
> > > one with singular/restrict problem).
> > > With Sage 5.7 and the GCC 4.7.2.p1 spkg (not that it should
> > not make any
> > > difference), I can build the GAP and libgap spkgs both with
> > CFLAGS unset
> > > and with CFLAGS="-O3".
> >
> > I also built GCC itself with -O3 (and -mcpu=ultrasparc3
> > -mtune=ultrasparc3); not yet sure whether that matters.
> >
> > I rebuilt GCC with CFLAGS="-O3 -mcpu=niagara2 -mtune=niagara2".
> > And ... tada! It breaks when building GAP:
> > {{{
> > ../../src/blister.c: In function 'InitKernel':
> > ../../src/blister.c:2743:1: internal compiler error: in
> > simplify_immed_subreg, at simplify-rtx.c:5262
> > }}}
>
> Exactly.
>
> > three times.
> >
> > No problems with the same GCC built with CFLAGS="-O3 -mcpu=niagara2
> > -mtune=niagara2" if I build GAP with CFLAGS="-O2".
>
> Yep. Although I didn't get them with GCC 4.7.0 and '-mcpu=...
> -mtune=... -O3', so this might be a regression in 4.7.1 or 4.7.2.
>
> Did you search GCC bugzilla? (I haven't yet.)
>
Nope, I rather tried to finish building Sage.
Here is the result of "make ptestlong":
{{{
sage -t --long devel/sage/sage/crypto/mq/sr.py # 1 doctest failed
sage -t --long devel/sage/sage/interfaces/expect.py # 2 doctests failed
sage -t --long devel/sage/sage/matrix/matrix_modn_dense.pyx # 5 doctests
failed
sage -t --long devel/sage/sage/matrix/matrix_modn_dense_template.pxi # 5
doctes
ts failed
sage -t --long devel/sage/sage/combinat/backtrack.py # Time out
sage -t --long
devel/sage/sage/combinat/cluster_algebra_quiver/cluster_seed.py
# Time out
sage -t --long
devel/sage/sage/combinat/cluster_algebra_quiver/mutation_type.py
# Time out
sage -t --long devel/sage/sage/combinat/cluster_algebra_quiver/quiver.py #
Time
out
sage -t --long devel/sage/sage/rings/polynomial/polynomial_element.pyx # 1
doct
est failed
sage -t --long devel/sage/sage/interfaces/ecm.py # Time out
sage -t --long
devel/sage/sage/quadratic_forms/quadratic_form__automorphisms.py
# Time out
sage -t --long
devel/sage/sage/schemes/elliptic_curves/ell_rational_field.py #
Time out
}}}
The non-timeout errors are worrying:
* sage.crypto.mq.sr.test_consistency
{{{
File "multi_polynomial_ideal_libsingular.pyx", line 250, in
sage.rings.polynomial.multi_polynomial_ideal_libsingular.slimgb_libsingular
(sage/rings/polynomial/multi_polynomial_ideal_libsingular.cpp:3568)
RuntimeError: Bus error
}}}
* sage.interface.expect: Singular crashes
* sage.matrix.matrix_modn_dense.Matrix_modn_dense._unpickle: unpickle
failures
{{{
File "matrix_modn_dense_template.pxi", line 640, in
sage.matrix.matrix_modn_dense_double.Matrix_modn_dense_template._pickle
(sage/matrix/matrix_modn_dense_double.cpp:6719)
RuntimeError: Bus error
}}}
* matrix.matrix_modn_dense_template.Matrix_modn_dense_template._unpickle:
unpickle failures.
except for
* sage.rings.polynomial.polynomial_element.Polynomial.inverse_mod: likely
numerical noise
Of course all these failures might be GCC 4.7.2 non-related.
>
>
> -leif
>
> --
> () The ASCII Ribbon Campaign
> /\ Help Cure HTML E-Mail
>
>
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
