This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Proposed patch for loop.c
- To: Joern Rennecke <amylaar at cygnus dot co dot uk>
- Subject: Re: Proposed patch for loop.c
- From: Bernd Schmidt <bernds at cygnus dot co dot uk>
- Date: Fri, 1 Sep 2000 18:42:15 +0100 (BST)
- cc: gcc-patches at gcc dot gnu dot org
On Fri, 1 Sep 2000, Joern Rennecke wrote:
> In article <Pine.SOL.3.90.980528110457.28555H-100000.cygnus.egcs.patches@matlock.informatik.rwth-aachen.de> you wrote:
> ...
> : I found two problems. First, general_induction_var returns the benefit of
> : a giv, but it always returns 1 even if it calculated a benefit of zero.
> : In combine_givs, the benefits of multiple givs can be added up in some cases,
> : possibly increasing the error quite a bit. Thus I think it's better to
> : have general_induction_var return zero or nonzero depending on whether X
> : is a giv or not, and accept a pointer to an int in which it can store the
> : benefit.
>
> Why not return a negative number for 'no giv'?
Eh? The problem has been fixed for about two years now, so we're not exactly
looking for a solution anymore...
Bernd