From: Richard Biener <rguenther@suse.de>
Date: Wed, 1 Oct 2025 09:26:45 +0000 (+0200)
Subject: tree-optimization/122110 - do not reject all bit-precision reductions
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=c6865e7e15bc9a1337df00d2ca03604e1712a2dd;p=thirdparty%2Fgcc.git

tree-optimization/122110 - do not reject all bit-precision reductions

We can handle bitwise-operation reductions and reductions on mask
vectors just fine.

	PR tree-optimization/122110
	* tree-vect-loop.cc (vectorizable_reduction): Relax restriction
	to mode-precision operations.
---

diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
index 18360375e29..1d549e4a03e 100644
--- a/gcc/tree-vect-loop.cc
+++ b/gcc/tree-vect-loop.cc
@@ -7187,7 +7187,11 @@ vectorizable_reduction (loop_vec_info loop_vinfo,
     return false;
 
   /* Do not try to vectorize bit-precision reductions.  */
-  if (!type_has_mode_precision_p (op.type))
+  if (!VECTOR_BOOLEAN_TYPE_P (vectype_out)
+      && !type_has_mode_precision_p (op.type)
+      && op.code != BIT_AND_EXPR
+      && op.code != BIT_IOR_EXPR
+      && op.code != BIT_XOR_EXPR)
     return false;
 
   /* Lane-reducing ops also never can be used in a SLP reduction group