This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/61980] ICE: in compute_affine_dependence, at tree-data-ref.c:4233 with -fcheck-data-deps
- From: "sabrinadfs at gmail dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Fri, 01 Aug 2014 11:29:55 +0000
- Subject: [Bug tree-optimization/61980] ICE: in compute_affine_dependence, at tree-data-ref.c:4233 with -fcheck-data-deps
- Auto-submitted: auto-generated
- References: <bug-61980-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61980
--- Comment #6 from Sabrina Souto <sabrinadfs at gmail dot com> ---
OK, Thank you.
(In reply to Richard Biener from comment #5)
> (In reply to Sabrina Souto from comment #4)
> > I don't have much experience with GCC, so I'm a bit confused with your 2
> > answers, please help me understanding what is happening: Based both on the
> > documentation and the results, I understood that the result from the
> > dependency analysis (Banerjee dist vectors: 0 1 , Omega dist vectors: 0) is
> > wrong or not expected, but this is due to the "unreliably" work of
> > -fcheck-data-deps or, since -fcheck-data-deps is used for debug, it is
> > working well and this result must be investigated?
>
> Clearly either Omega or Banerjee is wrong. Banerjee is used by GCC for
> code-generation and Omega is used to produce a comparison result only.
>
> If Omega is wrong we don't have to worry (it only makes -fcheck-data-deps
> less useful), if Banerjee is wrong then we have to worry.
>
> I can't tell nor didn't investigate who is wrong (but from past experience
> it's usually Omega).