This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Low overhead looping support
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: Low overhead looping support
- From: Jeffrey A Law <law at hurl dot cygnus dot com>
- Date: Mon, 22 Feb 1999 01:32:55 -0700
- cc: Michael Hayes <m dot hayes at elec dot canterbury dot ac dot nz>, dje at watson dot ibm dot com, egcs-patches at cygnus dot com
- Reply-To: law at cygnus dot com
In message <19990219202441.B17904@cygnus.com>you write:
> On Fri, Feb 19, 1999 at 01:58:21AM -0700, Jeffrey A Law wrote:
> > of enclosed loops? Did I misread something. Note it doesn't look like
> > you called it pattern in your patch.
>
> Huh? He's just passing in the loop nesting depth to d_a_b_o_c.
Parse error on my part. Somehow I read that as a new argument, named
pattern.
> > Any specific reason not to pass the iteration count (either as a constant
> > rtx or an varying rtx if it isn't constant) to a branch_on_count
> > expander? I'd like to see some discussion about the pros/cons for each
> > approach.
>
> How would you envision this working? The init goes at the beginning
> of the loop and the branch_on_count goes at the end.
I guess that makes sense. Presumably we add the insn before the jump to the
loop test for loops where we have not removed that jump. Right Michael?
jeff