Bootstrapped & tested on x86_64-unknown-linux-gnu, applied.
Richard.
2019-11-07 Richard Biener <rguent...@suse.de>
PR tree-optimization/92405
* tree-vect-loop.c (vectorizable_reduction): Appropriately
restrict lane-reducing ops to single stmt chains.
Index: gcc/tree-vect-loop.c
===================================================================
--- gcc/tree-vect-loop.c (revision 277907)
+++ gcc/tree-vect-loop.c (working copy)
@@ -5865,6 +5865,18 @@ vectorizable_reduction (stmt_vec_info st
reduc_chain_length++;
}
+ /* For lane-reducing ops we're reducing the number of reduction PHIs
+ which means the only use of that may be in the lane-reducing operation.
*/
+ if (lane_reduc_code_p
+ && reduc_chain_length != 1
+ && !only_slp_reduc_chain)
+ {
+ if (dump_enabled_p ())
+ dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
+ "lane-reducing reduction with extra stmts.\n");
+ return false;
+ }
+
reduc_def = PHI_RESULT (reduc_def_phi);
for (i = 0; i < op_type; i++)
{
