public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
From: Richard Biener <rguenth@gcc.gnu.org>
To: gcc-cvs@gcc.gnu.org
Subject: [gcc r12-2347] tree-optimization/101462 - fix signedness of reused reduction vector
Date: Fri, 16 Jul 2021 07:41:49 +0000 (GMT)	[thread overview]
Message-ID: <20210716074149.816F9389902D@sourceware.org> (raw)

https://gcc.gnu.org/g:d6aa28bb93c6fc9042ee87ff7addac60647dbddb

commit r12-2347-gd6aa28bb93c6fc9042ee87ff7addac60647dbddb
Author: Richard Biener <rguenther@suse.de>
Date:   Fri Jul 16 08:50:40 2021 +0200

    tree-optimization/101462 - fix signedness of reused reduction vector
    
    This fixes the partial reduction of the reused reduction vector to
    carried out in the correct sign and the correctly signed vector
    recorded for the skip edge use.
    
    2021-07-16  Richard Biener  <rguenther@suse.de>
    
            * tree-vect-loop.c (vect_transform_cycle_phi): Correct sign
            conversion issues with the partial reduction of the reused
            vector accumulator.

Diff:
---
 gcc/tree-vect-loop.c | 36 +++++++++++++++++++++++++-----------
 1 file changed, 25 insertions(+), 11 deletions(-)

diff --git a/gcc/tree-vect-loop.c b/gcc/tree-vect-loop.c
index fc3dab0d143..00a57b2ba62 100644
--- a/gcc/tree-vect-loop.c
+++ b/gcc/tree-vect-loop.c
@@ -7706,21 +7706,35 @@ vect_transform_cycle_phi (loop_vec_info loop_vinfo,
   if (auto *accumulator = reduc_info->reused_accumulator)
     {
       tree def = accumulator->reduc_input;
-      unsigned int nreduc;
-      bool res = constant_multiple_p (TYPE_VECTOR_SUBPARTS (TREE_TYPE (def)),
-				      TYPE_VECTOR_SUBPARTS (vectype_out),
-				      &nreduc);
-      gcc_assert (res);
-      if (nreduc != 1)
-	{
-	  /* Reduce the single vector to a smaller one.  */
+      if (!useless_type_conversion_p (vectype_out, TREE_TYPE (def)))
+	{
+	  unsigned int nreduc;
+	  bool res = constant_multiple_p (TYPE_VECTOR_SUBPARTS
+					    (TREE_TYPE (def)),
+					  TYPE_VECTOR_SUBPARTS (vectype_out),
+					  &nreduc);
+	  gcc_assert (res);
 	  gimple_seq stmts = NULL;
-	  def = vect_create_partial_epilog (def, vectype_out,
-					    STMT_VINFO_REDUC_CODE (reduc_info),
-					    &stmts);
+	  /* Reduce the single vector to a smaller one.  */
+	  if (nreduc != 1)
+	    {
+	      /* Perform the reduction in the appropriate type.  */
+	      tree rvectype = vectype_out;
+	      if (!useless_type_conversion_p (TREE_TYPE (vectype_out),
+					      TREE_TYPE (TREE_TYPE (def))))
+		rvectype = build_vector_type (TREE_TYPE (TREE_TYPE (def)),
+					      TYPE_VECTOR_SUBPARTS
+						(vectype_out));
+	      def = vect_create_partial_epilog (def, rvectype,
+						STMT_VINFO_REDUC_CODE
+						  (reduc_info),
+						&stmts);
+	    }
 	  /* Adjust the input so we pick up the partially reduced value
 	     for the skip edge in vect_create_epilog_for_reduction.  */
 	  accumulator->reduc_input = def;
+	  if (!useless_type_conversion_p (vectype_out, TREE_TYPE (def)))
+	    def = gimple_convert (&stmts, vectype_out, def);
 	  if (loop_vinfo->main_loop_edge)
 	    {
 	      /* While we'd like to insert on the edge this will split


                 reply	other threads:[~2021-07-16  7:41 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210716074149.816F9389902D@sourceware.org \
    --to=rguenth@gcc.gnu.org \
    --cc=gcc-cvs@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).