This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: scheduler speedup
> On Fri, Oct 26, 2001 at 02:43:35AM +0200, Jan Hubicka wrote:
> > In case you do have two overlapping regions (to my understanding this can
> > happen)
>
> No, it cannot. Every block is a member of exactly one region;
> in the degenerate case, the region is itself.
Aha, that is the bit I was missing.
(I was quite surprised that the sanity checking didn't abort, but I suspected
that the large regions are used only pre-reload scheduling as comments
in the top of file suggests and the code did handle case where some regions
are large I suspected to avoid quadratic behaviour. Now I am quite puzzled
why it is done so at all).
But would be the patch OK with the suggested change if I bootstrap it at
some machine that do use pre-reload scheduling?
Honza
>
>
> r~