This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [lno] fix division by zero





Richard Henderson wrote:
> > At the end of tree-vectorizer.h you can set
> > #define DBG_VECT(A)
> > to
> > #define DBG_VECT(A) A
> > to get information on why vectorization did not occur.
> > Setting
> > #define DBG_VECT2(A) A
> > will spit even more details.
> >
> > It should be changed to be cotrolled by a flag.
>
> It should *always* be emitted to the debug file
> and *never* to stderr.

Yes. This should be cleaned up. Adopting Daniel's interpretation of the
flags (http://gcc.gnu.org/ml/gcc-patches/2004-01/msg00196.html), the above
macros can be mapped roughly as follows -

printouts under DBG_VEC --> TDF_stats ("how much of something we did").

printouts under DBG_VECT --> TDF_details ("*why* things happened. Like why
we didn't vectorize something").

printouts under DBG_VECT2 --> TDF_debug ("Debugging information").

The code may not follow this convention strictly, but at least that was the
general intention.

dorit



                                                                       
                      Richard Henderson                                
                      <rth@redhat.com>         To:       Dorit Naishlos/Haifa/IBM@IBMIL
                                               cc:       Daniel Berlin <dberlin@dberlin.org>, Toon Moene
                      05/01/2004 02:52          <toon@moene.indiv.nluug.nl>, Pop Sébastian <pop@gauvain.u-strasbg.fr>,
                                                gcc-patches@gcc.gnu.org
                                               Subject:  Re: [lno] fix division by zero
                                                                       




On Sun, Jan 04, 2004 at 10:22:27AM +0200, Dorit Naishlos wrote:
> At the end of tree-vectorizer.h you can set
> #define DBG_VECT(A)
> to
> #define DBG_VECT(A) A
> to get information on why vectorization did not occur.
> Setting
> #define DBG_VECT2(A) A
> will spit even more details.
>
> It should be changed to be cotrolled by a flag.

It should *always* be emitted to the debug file
and *never* to stderr.


r~



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]