This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: BCT optimization
- To: David Edelsohn <dje at watson dot ibm dot com>
- Subject: Re: BCT optimization
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Wed, 30 Sep 1998 20:16:38 -0600
- cc: Michael Hayes <m dot hayes at elec dot canterbury dot ac dot nz>, egcs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <9810010200.AA48040@marc.watson.ibm.com>you write:
> A problem with assigning a pseudo and allowing reload to choose
> the register is that this optimization is specifically scheduling the use
> of this resource based on additional criteria about when it is
> beneficial. I am a little concerned about not stuffing in the count or
> loop register early if reload thinks it is smarter and wants to override
> the cost-benefit decision about that resource.
Note, we still use the same criteria for creating these patterns in loop.
For the ppc, we'll never have more than one pseudo which prefers the count
register live at the same time. Every pseudo that is used in this manner
should be allocated to the count register by global register allocation. If
global or reload ever spills the count reg or uses it for some other purpose,
then we should figure out why since that shouldn't be happening if we do things
correctly.
jeff