[Bug tree-optimization/40254] New: [4.5 Regression] SPEC2006 403.gcc miscompares

[Bug tree-optimization/40254] New: [4.5 Regression] SPEC2006 403.gcc miscompares

[rguenth at gcc dot gnu dot org]

rev. 147829 (SLP vectorization) causes 403.gcc to miscompare when built with
-O3 -funroll-loops -fpeel-loops -ffast-math -march=native on x86_64 AMD K8
and K10.


-- 
           Summary: [4.5 Regression] SPEC2006 403.gcc miscompares
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: rguenth at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[rguenth at gcc dot gnu dot org]

------- Comment #1 from rguenth at gcc dot gnu dot org  2009-05-26 13:36 -------
*** Bug 40245 has been marked as a duplicate of this bug. ***


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hjl dot tools at gmail dot
                   |                            |com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2009-05-26 14:34 -------
Does -fno-strict-aliasing help?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[rguenther at suse dot de]

------- Comment #3 from rguenther at suse dot de  2009-05-26 14:36 -------
Subject: Re:  [4.5 Regression] SPEC2006 403.gcc
 miscompares

On Tue, 26 May 2009, pinskia at gcc dot gnu dot org wrote:

> ------- Comment #2 from pinskia at gcc dot gnu dot org  2009-05-26 14:34 -------
> Does -fno-strict-aliasing help?

That is of course already enabled.

Richard.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[irar at il dot ibm dot com]

------- Comment #4 from irar at il dot ibm dot com  2009-05-27 08:43 -------
The bug is in data-refs analysis for basic blocks: two accesses that are not
adjacent (reload.c:1370) are considered as adjacent, and, therefore, get
vectorized together, causing the wrong code generation.


-- 

irar at il dot ibm dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |irar at il dot ibm dot com
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2009-05-27 08:43:46
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[irar at il dot ibm dot com]

------- Comment #5 from irar at il dot ibm dot com  2009-05-27 09:59 -------
I'll test this patch tomorrow:

Index: tree-data-ref.c
===================================================================
--- tree-data-ref.c     (revision 147903)
+++ tree-data-ref.c     (working copy)
@@ -718,17 +725,26 @@ dr_analyze_innermost (struct data_refere
       base_iv.no_overflow = true;
     }

-  if (!poffset || !in_loop)
+  if (!poffset)
     {
       offset_iv.base = ssize_int (0);
       offset_iv.step = ssize_int (0);
     }
-  else if (!simple_iv (loop, loop_containing_stmt (stmt),
-                      poffset, &offset_iv, false))
+  else
     {
-      if (dump_file && (dump_flags & TDF_DETAILS))
-       fprintf (dump_file, "failed: evolution of offset is not affine.\n");
-      return false;
+      if (!in_loop)
+        {
+          offset_iv.base = poffset;
+          offset_iv.step = ssize_int (0);
+        }
+      else if (!simple_iv (loop, loop_containing_stmt (stmt),
+                          poffset, &offset_iv, false))
+        {
+          if (dump_file && (dump_flags & TDF_DETAILS))
+            fprintf (dump_file, "failed: evolution of offset is not"
+                                " affine.\n");
+          return false;
+        }
     }

   init = ssize_int (pbitpos / BITS_PER_UNIT);


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[irar at gcc dot gnu dot org]

------- Comment #6 from irar at gcc dot gnu dot org  2009-05-28 09:03 -------
Subject: Bug 40254

Author: irar
Date: Thu May 28 09:02:53 2009
New Revision: 147945

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=147945
Log:

        PR tree-optimization/40254
        * tree-data-ref.c (dr_analyze_innermost): Take POFFSET into account
        in analysis of basic blocks.


Added:
    trunk/gcc/testsuite/gcc.dg/vect/pr40254.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-data-ref.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[rguenth at gcc dot gnu dot org]

------- Comment #7 from rguenth at gcc dot gnu dot org  2009-05-28 09:12 -------
Fixed.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254

[Bug tree-optimization/40254] [4.5 Regression] SPEC2006 403.gcc miscompares

[hjl at gcc dot gnu dot org]

------- Comment #8 from hjl at gcc dot gnu dot org  2009-05-30 13:50 -------
Subject: Bug 40254

Author: hjl
Date: Sat May 30 13:49:33 2009
New Revision: 148004

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=148004
Log:
2009-05-30  H.J. Lu  <hongjiu.lu@intel.com>

        Backport from mainline:

        2009-05-28  Dodji Seketeli  <dodji@redhat.com>

        PR c++/39754
        * g++.dg/template/canon-type-1.C: New test.
        * g++.dg/template/canon-type-2.C: Likewise.
        * g++.dg/template/canon-type-3.C: Likewise.
        * g++.dg/template/canon-type-4.C: Likewise.
        * g++.dg/template/canon-type-5.C: Likewise.
        * g++.dg/template/canon-type-6.C: Likewise.
        * g++.dg/template/canon-type-7.C: Likewise.

        2009-05-28  Ira Rosen  <irar@il.ibm.com>

        PR tree-optimization/40254
        * gcc.dg/vect/pr40254.c: New test.

        2009-05-26  Richard Guenther  <rguenther@suse.de>

        PR middle-end/40252
        * gcc.c-torture/compile/pr40252.c: New testcase.

        2009-05-26  Dodji Seketeli  <dodji@redhat.com>

        PR c++/40007
        * g++.dg/template/typedef18.C: New test.
        * g++.dg/template/typedef19.C: Likewise.
        * g++.dg/template/typedef20.C: Likewise.

        2009-05-25  Ira Rosen  <irar@il.ibm.com>

        PR tree-optimization/40238
        * gcc.dg/vect/pr40238.c: New test.

        2009-05-24  Richard Guenther  <rguenther@suse.de>

        PR middle-end/40233
        * gcc.c-torture/compile/pr40233.c: New testcase.

Added:
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-1.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-1.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-2.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-2.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-3.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-3.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-4.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-4.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-5.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-5.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-6.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-6.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/canon-type-7.C
      - copied unchanged from r148002,
trunk/gcc/testsuite/g++.dg/template/canon-type-7.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef18.C
      - copied unchanged from r148003,
trunk/gcc/testsuite/g++.dg/template/typedef18.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef19.C
      - copied unchanged from r148003,
trunk/gcc/testsuite/g++.dg/template/typedef19.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef20.C
      - copied unchanged from r148003,
trunk/gcc/testsuite/g++.dg/template/typedef20.C
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr40233.c
      - copied unchanged from r148003,
trunk/gcc/testsuite/gcc.c-torture/compile/pr40233.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr40252.c
      - copied unchanged from r148003,
trunk/gcc/testsuite/gcc.c-torture/compile/pr40252.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.dg/vect/pr40238.c
      - copied unchanged from r148003,
trunk/gcc/testsuite/gcc.dg/vect/pr40238.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.dg/vect/pr40254.c
      - copied unchanged from r148003,
trunk/gcc/testsuite/gcc.dg/vect/pr40254.c
Modified:
    branches/gcc-4_4-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40254