This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: C++ PATCH: multiple inheritance optimization
law@cygnus.com (Jeffrey A Law) writes:
> >
> > Right -- that's what Richard indicated earlier. He wasn't sure if the
> > reason GCSE doesn't handle hard registers was just that y'all weren't
> > sure it would work -- or if there was a known reason why it would not
> > work.
> I'm not aware of a fundamental problem why it wouldn't work once the code
> was enhanced to deal with hard registers.
>
> You do have to come up with a reasonable scheme to decide when it is safe to
> move a clobber expression like that -- noting of course that you don't have
> register life information at this point in the compilation.
... and probably add code to throttle gcse back when it would cause too
much register pressure. Often x86 functions get a lot slower when you
add a new variable and spilling is starting all over. I hope gcse won't
add to that.
-Andi
--
This is like TV. I don't like TV.