public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "froydnj at codesourcery dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/46259] [4.6 Regression] Revision 166102 failed 416.gamess in SPEC CPU 2006
Date: Mon, 01 Nov 2010 14:57:00 -0000	[thread overview]
Message-ID: <20101101145700.dWMXMIy2SUL9VY9-tj5nzKC1STznMu6i2aWciZ2M87A@z> (raw)
In-Reply-To: <bug-46259-4@http.gcc.gnu.org/bugzilla/>

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46259

--- Comment #5 from froydnj at codesourcery dot com <froydnj at codesourcery dot com> 2010-11-01 14:57:32 UTC ---
If I had to guess, one of these two hunks is probably responsible:

--- trunk/gcc/tree-vect-loop.c    2010/10/31 01:58:12    166101
+++ trunk/gcc/tree-vect-loop.c    2010/10/31 02:20:22    166102
@@ -3021,14 +3017,7 @@
             break;
           }

-        for (i = nunits - 1; i >= 0; --i)
-          t = tree_cons (NULL_TREE, init_value, t);
-
-        if (TREE_CONSTANT (init_val))
-          init_def = build_vector (vectype, t);
-        else
-          init_def = build_constructor_from_list (vectype, t);
-
+    init_def = build_vector_from_val (vectype, init_value);
         break;

       default:

--- trunk/gcc/tree-vect-stmts.c    2010/10/31 01:58:12    166101
+++ trunk/gcc/tree-vect-stmts.c    2010/10/31 02:20:22    166102
@@ -987,9 +987,7 @@
   loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
   tree vec_inv;
   tree vec_cst;
-  tree t = NULL_TREE;
   tree def;
-  int i;
   enum vect_def_type dt;
   bool is_simple_use;
   tree vector_type;
@@ -1055,13 +1049,7 @@
         if (vect_print_dump_info (REPORT_DETAILS))
           fprintf (vect_dump, "Create vector_inv.");

-        for (i = nunits - 1; i >= 0; --i)
-          {
-            t = tree_cons (NULL_TREE, def, t);
-          }
-
-    /* FIXME: use build_constructor directly.  */
-        vec_inv = build_constructor_from_list (vector_type, t);
+        vec_inv = build_vector_from_val (vector_type, def);
         return vect_init_vector (stmt, vec_inv, vector_type, NULL);
       }

Can you try reverting each one individually and see if those fix the
problem?  Thanks.


  parent reply	other threads:[~2010-11-01 14:57 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-11-01 14:37 [Bug tree-optimization/46259] New: " hjl.tools at gmail dot com
2010-11-01 14:37 ` [Bug tree-optimization/46259] " hjl.tools at gmail dot com
2010-11-01 14:41 ` froydnj at codesourcery dot com
2010-11-01 14:48 ` hjl.tools at gmail dot com
2010-11-01 14:49 ` hjl.tools at gmail dot com
2010-11-01 14:50 ` hjl.tools at gmail dot com
2010-11-01 14:57 ` froydnj at codesourcery dot com [this message]
2010-11-01 16:21 ` hjl.tools at gmail dot com
2010-11-01 16:33 ` hjl.tools at gmail dot com
2010-11-01 18:58 ` froydnj at gcc dot gnu.org
2010-11-01 20:04 ` jakub at gcc dot gnu.org

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=20101101145700.dWMXMIy2SUL9VY9-tj5nzKC1STznMu6i2aWciZ2M87A@z \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@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).