This is the mail archive of the gcc@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]

Re: Haney's real matrix test regression


Richard Kenner wrote:

> This is because G++ uses alias set zero for all aggregate types, thus
> disabling most alias analysis.
>
> I have no idea what's involved in improving that or whether it will be
> done in time for GCC 3.1

I'm very confused.
For sure the "O-O C++" version performed better than the "hard coded C"
version of the test case until a week ago, and this is the case indeed
the case for gcc3.0.2.

You mean that your recent checkins, while improving the optimization of C
code have a negative impact for C++ code until some radical modifications
are implemented to improve C++ alias analysis?

Cheers,
Paolo Carlini.



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