Re: -fno-align-functions still adds padding on i386 on gcc3.1

[Richard Henderson <rth at redhat dot com>]

On Thu, May 16, 2002 at 11:06:48PM -0400, John Baldwin wrote:
> Thanks.  Someone else discovered that the additional function padding
> wasn't actually making up that much of a difference but that using
> -fno-guess-branch-probability cut down on the generated code size
> considerably.  Perhaps -Os should not enable this optimization?

Hum.  Better would be to modify -freorder-blocks so that
it had a mode in which it tried to minimize the number
(and size) of branches.

For those cases in which the number of branches is identical,
we should still honor the branch probability.


r~