Should gcc-4.0 apply mathematical associative rules for addition or multiplication

[Jeffrey A Law]

On Mon, 2004-10-04 at 17:07, Andrew Pinski wrote:
> On Oct 4, 2004, at 6:56 PM, Fariborz Jahanian wrote:
> > /* test */
> > static const double C = 0x1.0p1023;
> >
> > double foo(double x)
> > {
> >         return ( ( (x * C) * C ) * C );
> > }
> >
> > This is done in routine: simplify_rhs_and_lookup_avail_expr(...) which 
> > is preceded with this comment:
> >
> >  /* If we have z = (x OP C1), see if we earlier had x = y OP C2.
> >      If OP is associative, create and fold (y OP C2) OP C1 which
> >      should result in (y OP C3), use that as the RHS for the
> >      assignment.  Add minus to this, as we handle it specially below.  
> > */
> 
> This patch should work:
> Index: tree-ssa-dom.c
> ===================================================================
> RCS file: /cvs/gcc/gcc/gcc/tree-ssa-dom.c,v
> retrieving revision 2.58
> diff -u -p -r2.58 tree-ssa-dom.c
> --- tree-ssa-dom.c	4 Oct 2004 13:19:20 -0000	2.58
> +++ tree-ssa-dom.c	4 Oct 2004 23:05:28 -0000
> @@ -1598,7 +1598,9 @@ simplify_rhs_and_lookup_avail_expr (stru
>        assignment.  Add minus to this, as we handle it specially below.  
> */
>     if ((associative_tree_code (rhs_code) || rhs_code == MINUS_EXPR)
>         && TREE_CODE (TREE_OPERAND (rhs, 0)) == SSA_NAME
> -      && is_gimple_min_invariant (TREE_OPERAND (rhs, 1)))
> +      && is_gimple_min_invariant (TREE_OPERAND (rhs, 1))
> +      && (!FLOAT_TYPE_P (type)
> +	  || flag_unsafe_math_optimizations))
>       {
>         tree rhs_def_stmt = SSA_NAME_DEF_STMT (TREE_OPERAND (rhs, 0));
I believe we should also avoid these transformations for pointer
types as well.

Long term this stuff is really something for a tree combiner.

jef