This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Recent changes to cse.c
- To: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- Subject: Re: Recent changes to cse.c
- From: law at redhat dot com
- Date: Wed, 25 Jul 2001 07:54:35 -0600
- cc: Richard dot Earnshaw at arm dot com, gcc at gcc dot gnu dot org
- Reply-To: law at redhat dot com
In message <10107251021.AA04107@vlsi1.ultra.nyu.edu>you write:
> Why on earth would we want to split a compare and branch if we can't
> schedule the insns separately?
>
> It's not that you can *schedule* them separately, but that splitting it
> gives more accurate information as to when registers are used and will help
> better schedule other insns that interact with those insns.
Remember, scheduling happens before dbranch, so the compare and branch
*must* be consecutive on a HAVE_cc0 target, no exceptions.
>From a register usage/scheduling standpoint there is nothing to be gained by
splitting the compare/branch.
jeff