Inner-loop optimization regression from 3.3 to 3.4

[Jan Hubicka]

> On Sun, Oct 19, 2003 at 11:55:51AM +0200, Falk Hueffner wrote:
> > Maybe it's easier to do that in the machine description, that way it
> > gets also optimized if the user explicitely specified x >> 63.
> 
> Eh.  I'd really rather not...  There are an uncountable multitude
> of such.  I'd hope that the user example would be handled by 
> combine.

OK, I will use the branch costs directly and lets see how far I get.

Honza
> 
> 
> r~