On Wed, May 15, 2013 at 2:07 PM, Marek Polacek <pola...@redhat.com> wrote:
> This is an attempt to add more dumping to scev_cprop. I don't know what
> exactly
> would you want to see in there, so now it prints e.g.:
>
> (replace_stmt
> (virtual_p_1 with 0)
> )
>
> and
>
> (final_value_replacement
> o__lsm.3_2 = PHI <_6(4)>
> with
> o__lsm.3_2 = pretmp_10 + 20;
> )
>
> Comments? Maybe that first hunk isn't really neccessary.
>
> 2013-05-15 Marek Polacek <pola...@redhat.com>
>
> * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
>
> --- gcc/tree-scalar-evolution.c.mp 2013-05-15 13:43:31.020698846 +0200
> +++ gcc/tree-scalar-evolution.c 2013-05-15 13:43:58.080788354 +0200
> @@ -3295,7 +3295,17 @@ scev_const_prop (void)
>
> /* Replace the uses of the name. */
> if (name != ev)
> - replace_uses_by (name, ev);
> + {
> + replace_uses_by (name, ev);
> + if (dump_file && (dump_flags & TDF_SCEV))
should be without dump_flags checking
> + {
> + fprintf (dump_file, "(replace_stmt \n (");
> + print_generic_expr (dump_file, name, 0);
> + fprintf (dump_file, " with ");
> + print_generic_expr (dump_file, ev, 0);
> + fprintf (dump_file, ")\n) \n");
and no need to do it the LISP-y way ;)
> + }
> + }
>
> if (!ssa_names_to_remove)
> ssa_names_to_remove = BITMAP_ALLOC (NULL);
> @@ -3391,6 +3401,12 @@ scev_const_prop (void)
>
> /* Eliminate the PHI node and replace it by a computation outside
> the loop. */
> + if (dump_file && (dump_flags & TDF_SCEV))
> + {
Likewise.
> + fprintf (dump_file, "(final_value_replacement \n ");
> + print_gimple_stmt (dump_file, phi, 0, 0);
> + fprintf (dump_file, " with \n ");
> + }
> def = unshare_expr (def);
> remove_phi_node (&psi, false);
>
> @@ -3398,6 +3414,11 @@ scev_const_prop (void)
> true, GSI_SAME_STMT);
> ass = gimple_build_assign (rslt, def);
> gsi_insert_before (&bsi, ass, GSI_SAME_STMT);
> + if (dump_file && (dump_flags & TDF_SCEV))
Likewise.
> + {
> + print_gimple_stmt (dump_file, ass, 0, 0);
> + fprintf (dump_file, ") \n");
> + }
> }
> }
> return 0;
I would have liked to see failed attempts as well, then with TDF_DETAILS.
Failed attempts for the "real" final value replacement stuff (I'm not sure
the constant propagation part is worth keeping ... how often does it trigger?)
Thanks,
Richard.
> Marek
