[Bug tree-optimization/63185] Improve DSE with branches

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63185

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
Created attachment 44146
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=44146&action=edit
testresults

Related I checked

Index: gcc/passes.def
===================================================================
--- gcc/passes.def      (revision 260317)
+++ gcc/passes.def      (working copy)
@@ -209,8 +209,6 @@ along with GCC; see the file COPYING3.
       NEXT_PASS (pass_return_slot);
       NEXT_PASS (pass_fre);
       NEXT_PASS (pass_merge_phi);
-      NEXT_PASS (pass_thread_jumps);
-      NEXT_PASS (pass_vrp, true /* warn_array_bounds_p */);
       NEXT_PASS (pass_chkp_opt);
       NEXT_PASS (pass_dce);
       NEXT_PASS (pass_stdarg);
@@ -224,6 +222,7 @@ along with GCC; see the file COPYING3.
       NEXT_PASS (pass_ch);
       NEXT_PASS (pass_lower_complex);
       NEXT_PASS (pass_sra);
+      NEXT_PASS (pass_vrp, true /* warn_array_bounds_p */);
       /* The dom pass will also resolve all __builtin_constant_p calls
          that are still there to 0.  This has to be done after some
         propagations have already run, but before some more dead code

which as quite some regressions (attached).  Ignore any auto-fdo fails.
Will check w/o the removal of pass_thread_jumps.