tree-vect-loop.c (vect_analyze_loop_operations): Not relevant PHIs are ok.

2017-06-02  Richard Biener  <rguenther@suse.de>

	* tree-vect-loop.c (vect_analyze_loop_operations): Not relevant
	PHIs are ok.
	* tree-vect-stmts.c (process_use): Do not mark backedge defs
	for inductions as relevant.

From-SVN: r248820
This commit is contained in:
Richard Biener 2017-06-02 10:26:10 +00:00 committed by Richard Biener
parent e7754973ed
commit 643a9684fa
3 changed files with 22 additions and 2 deletions

View file

@ -1,3 +1,10 @@
2017-06-02 Richard Biener <rguenther@suse.de>
* tree-vect-loop.c (vect_analyze_loop_operations): Not relevant
PHIs are ok.
* tree-vect-stmts.c (process_use): Do not mark backedge defs
for inductions as relevant.
2017-06-02 Richard Biener <rguenther@suse.de>
* tree-vect-loop.c (get_initial_def_for_induction): Inline into ...

View file

@ -1708,8 +1708,7 @@ vect_analyze_loop_operations (loop_vec_info loop_vinfo)
are not used in the outerloop (unless it is double reduction,
i.e., this phi is vect_reduction_def), cause this case
requires to actually do something here. */
if ((!STMT_VINFO_RELEVANT_P (stmt_info)
|| STMT_VINFO_LIVE_P (stmt_info))
if (STMT_VINFO_LIVE_P (stmt_info)
&& STMT_VINFO_DEF_TYPE (stmt_info)
!= vect_double_reduction_def)
{

View file

@ -579,6 +579,20 @@ process_use (gimple *stmt, tree use, loop_vec_info loop_vinfo,
gcc_unreachable ();
}
}
/* We are also not interested in uses on loop PHI backedges that are
inductions. Otherwise we'll needlessly vectorize the IV increment
and cause hybrid SLP for SLP inductions. */
else if (gimple_code (stmt) == GIMPLE_PHI
&& STMT_VINFO_DEF_TYPE (stmt_vinfo) == vect_induction_def
&& (PHI_ARG_DEF_FROM_EDGE (stmt, loop_latch_edge (bb->loop_father))
== use))
{
if (dump_enabled_p ())
dump_printf_loc (MSG_NOTE, vect_location,
"induction value on backedge.\n");
return true;
}
vect_mark_relevant (worklist, def_stmt, relevant, false);
return true;