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: Fw: ping [PATCH] A few cleanups in tree-vect-*


Zdenek Dvorak <rakdver@atrey.karlin.mff.cuni.cz> wrote on 15/01/2007
12:14:09:

> Hello,
>
> > Dorit Nuzman/Haifa/IBM wrote on 15/01/2007 00:48:54:
> >
> > > http://gcc.gnu.org/ml/gcc-patches/2007-01/msg00004.html
>
> > Bootstrapped on powerpc-linux.
> > Bootstrapped with vectorization enabled on i386-linux.
> > Tested on the vectorizer testcases on these two platforms.
>
> You should run whole testsuite (it is required by testing standards,
> and also I think there are some testcases that enable vectorization
> outside of the vect directories).

yes, I know, I meant testing on all the *.exp that enable -ftree-vectorize
(which, I admit, doesn't cover a few testcases in gcc.dg - I'll cover that
too).

For this particular case of simple obvious cleanups inside the vectorizer I
felt that bootstrap-with-vectorization-enabled along with the above level
of testing is safe enough. Usually, if I touch anything outside the
vectorizer, or even if I don't, but there's a functionality change in the
patch (not just a cleanup) - I do a full make-check. But in cases like
this...??

Having said that, I'll do a full make check before committing this patch.

thanks,

dorit

>
> Otherwise, this is patch is OK.
>
> Zdenek


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