From: Richard Biener <rguenther@suse.de>
Date: Fri, 4 Feb 2022 08:26:57 +0000 (+0100)
Subject: tree-optimization/103641 - improve vect_synth_mult_by_constant
X-Git-Tag: basepoints/gcc-13~1266
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=876e70d4681332a600492173af0c7259e5a438c6;p=thirdparty%2Fgcc.git

tree-optimization/103641 - improve vect_synth_mult_by_constant

The following happens to improve compile-time of the PR103641
testcase on aarch64 significantly.  I did not investigate the
effect on the generated code but at least in theory
choose_mult_variant should do a better job when we tell it
the actual mode we are going to use for the operations it
synthesizes.

2022-02-04  Richard Biener  <rguenther@suse.de>

	PR tree-optimization/103641
	* tree-vect-patterns.cc (vect_synth_mult_by_constant):
	Pass the vector mode to choose_mult_variant.
---

diff --git a/gcc/tree-vect-patterns.cc b/gcc/tree-vect-patterns.cc
index bea049921603..2baf974627e1 100644
--- a/gcc/tree-vect-patterns.cc
+++ b/gcc/tree-vect-patterns.cc
@@ -3041,6 +3041,9 @@ vect_synth_mult_by_constant (vec_info *vinfo, tree op, tree val,
   bool cast_to_unsigned_p = !TYPE_OVERFLOW_WRAPS (itype);
 
   tree multtype = cast_to_unsigned_p ? unsigned_type_for (itype) : itype;
+  tree vectype = get_vectype_for_scalar_type (vinfo, multtype);
+  if (!vectype)
+    return NULL;
 
   /* Targets that don't support vector shifts but support vector additions
      can synthesize shifts that way.  */
@@ -3050,16 +3053,13 @@ vect_synth_mult_by_constant (vec_info *vinfo, tree op, tree val,
   /* Use MAX_COST here as we don't want to limit the sequence on rtx costs.
      The vectorizer's benefit analysis will decide whether it's beneficial
      to do this.  */
-  bool possible = choose_mult_variant (mode, hwval, &alg,
-					&variant, MAX_COST);
+  bool possible = choose_mult_variant (VECTOR_MODE_P (TYPE_MODE (vectype))
+				       ? TYPE_MODE (vectype) : mode,
+				       hwval, &alg, &variant, MAX_COST);
   if (!possible)
     return NULL;
 
-  tree vectype = get_vectype_for_scalar_type (vinfo, multtype);
-
-  if (!vectype
-      || !target_supports_mult_synth_alg (&alg, variant,
-					   vectype, synth_shift_p))
+  if (!target_supports_mult_synth_alg (&alg, variant, vectype, synth_shift_p))
     return NULL;
 
   tree accumulator;