[Bug c/45971] New: ice in vect_update_ivs_after_vectorizer

[Bug c/45971] New: ice in vect_update_ivs_after_vectorizer

[dcb314 at hotmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

           Summary: ice in vect_update_ivs_after_vectorizer
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: dcb314@hotmail.com


Created attachment 22017
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22017
gzipped C source code

I just tried to compile package homebank-4.3 with the latest 4.6
snapshot 20101009 on a Fedora Linux box. The compiler said

preferences.c: In function 'homebank_pref_load':
preferences.c:474:10: internal compiler error: in
vect_update_ivs_after_vectorizer, at tree-vect-loop-manip.c:1779
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Source code attached. Flag -O3 required.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0
            Summary|ice in                      |[4.6 Regression] ice in
                   |vect_update_ivs_after_vecto |vect_update_ivs_after_vecto
                   |rizer                       |rizer

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2010.10.11 23:13:10
                 CC|                            |ktietz at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2010-10-11 23:13:10 UTC ---
It is caused by revision 159912:

http://gcc.gnu.org/ml/gcc-cvs/2010-05/msg00969.html

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> 2010-10-11 23:17:20 UTC ---
(In reply to comment #1)
> It is caused by revision 159912:
> 
> http://gcc.gnu.org/ml/gcc-cvs/2010-05/msg00969.html

This does not make sense as this is fixing an ICE which would have happened
earlier.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

H.J. Lu <hjl.tools at gmail dot com> changed:

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

--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> 2010-10-12 00:20:58 UTC ---
The bug was introduced between 159879 and 159912.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|hjl.tools at gmail dot com, |spop at gcc dot gnu.org
                   |ktietz at gcc dot gnu.org   |

--- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> 2010-10-12 01:52:49 UTC ---
It is caused by revision 159883:

http://gcc.gnu.org/ml/gcc-cvs/2010-05/msg00939.html

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

--- Comment #5 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-04 12:34:43 UTC ---
Reduced testcase:

typedef unsigned long gsize;
typedef int gint;
enum { NUM_LST_DSPOPE = 12 };
struct Preferences {
    gint lst_ope_columns[NUM_LST_DSPOPE+1];
};
extern struct Preferences *PREFS;
void
homebank_pref_load(gint *bsrc, gsize length)
{
  gint i; 
  for(i=0;  i<NUM_LST_DSPOPE-1 ;  i++)  
    { 
      PREFS->lst_ope_columns[i] = (bsrc[i] == (!(0))) ? i+1 : -(i+1);  
    }
}

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |matt at use dot net

--- Comment #6 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-04 13:24:38 UTC ---
*** Bug 46077 has been marked as a duplicate of this bug. ***

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-08 08:43:51 UTC ---
Slightly more reduced testcase:

void
foo (int *x, int *y)
{
  int i;
  for (i = 0; i < 11; i++)
    y[i] = (x[i] == 1) ? i + 1 : -(i + 1);
}

The following:

void
bar (int *x, int *y)
{
  int i;
  for (i = 0; i < 64; i++)
    y[i] = (x[i] == 1) ? i + 1 : -(i + 1);
}

ICEs with
internal compiler error: in get_initial_def_for_induction, at
tree-vect-loop.c:2659
instead.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[irar at il dot ibm.com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Ira Rosen <irar at il dot ibm.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |irar at il dot ibm.com

--- Comment #8 from Ira Rosen <irar at il dot ibm.com> 2010-11-09 09:12:43 UTC ---
Looking at this loop:

void
foo (int *x, int *y)
{
  int i;
  for (i = 0; i < 11; i++)
    y[i] = (x[i] == 1) ? i + 1 : -(i + 1);
}


I see that first time that we try to analyze phi i_19 = PHI <prephitmp.8_25(4),
0(2)> (to detect induction), scev analysis succeeds:

(analyze_scalar_evolution
  (loop_nb = 1)
  (scalar = i_19)
(get_scalar_evolution
  (scalar = i_19)
  (scalar_evolution = {0, +, 1}_1))
(set_scalar_evolution
  instantiated_below = 2
  (scalar = i_19)
  (scalar_evolution = {0, +, 1}_1))
)

pr45971.c:5: note: Access function of PHI: {0, +, 1}_1
pr45971.c:5: note: step: 1,  init: 0


But when we try to check the same phi later (in transformation), evolution
analysis fails:

pr45971.c:5: note: vect_update_ivs_after_vectorizer: phi: i_19 = PHI
<prephitmp.8_25(4), 0(8)>
(analyze_scalar_evolution
  (loop_nb = 1)
  (scalar = i_19)
(get_scalar_evolution
  (scalar = i_19)
  (scalar_evolution = ))
(analyze_initial_condition
  (loop_phi_node =
i_19 = PHI <prephitmp.8_25(4), 0(8)>
)
  (init_cond = 0))
(analyze_evolution_in_loop
  (loop_phi_node = i_19 = PHI <prephitmp.8_25(4), 0(8)>
)
  (evolution_function = scev_not_known))
(set_scalar_evolution
  instantiated_below = 8
  (scalar = i_19)
  (scalar_evolution = i_19))
)

pr45971.c:5: note: Access function of PHI: i_19

causing ICE.

Here is the code that if-cvt generates:

  # i_19 = PHI <prephitmp.8_25(4), 0(2)>
  # ivtmp.9_15 = PHI <ivtmp.9_23(4), 11(2)>
  D.2688_4 = (long unsigned int) i_19;
  D.2689_5 = D.2688_4 * 4;
  D.2690_7 = y_6(D) + D.2689_5;
  D.2692_11 = x_10(D) + D.2689_5;
  D.2693_12 = *D.2692_11;
  iftmp.0_13 = i_19 + 1;
  D.2707_22 = D.2693_12 == 1;
  iftmp.0_14 = ~i_19;
  pretmp.7_24 = i_19 + 1;
  D.2708_21 = D.2693_12 != 1;
  iftmp.0_2 = [cond_expr] D.2693_12 == 1 ? iftmp.0_13 : iftmp.0_14;
  prephitmp.8_25 = [cond_expr] D.2693_12 == 1 ? iftmp.0_13 : pretmp.7_24;
  *D.2690_7 = iftmp.0_2;
  ivtmp.9_23 = ivtmp.9_15 - 1;
  if (ivtmp.9_23 != 0)
    goto <bb 4>;
  else
    goto <bb 5>;

prephitmp.8_25 is actually always i_19 + 1, so the second cond_expr is
redundant.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Sebastian Pop <spop at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |spop at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #9 from Sebastian Pop <spop at gcc dot gnu.org> 2010-11-09 20:03:22 UTC ---
Mine.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

--- Comment #10 from Sebastian Pop <spop at gcc dot gnu.org> 2010-11-09 20:05:13 UTC ---
Created attachment 22353
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22353
patch

The attached patch solves the problem.  Not yet tested other than this PR's
testcase.

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

--- Comment #11 from Sebastian Pop <spop at gcc dot gnu.org> 2010-11-10 16:06:51 UTC ---
Author: spop
Date: Wed Nov 10 16:06:47 2010
New Revision: 166544

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166544
Log:
Fix PR45971: do not predicate condition phi nodes that are scev analyzable.

2010-11-10  Sebastian Pop  <sebastian.pop@amd.com>

    PR tree-optimization/45971
    * tree-if-conv.c (predicate_scalar_phi): Do not generate a COND_EXPR
    for phi nodes analyzable by scev.

    * gcc.dg/vect/O3-pr45971.c: New.

Added:
    trunk/gcc/testsuite/gcc.dg/vect/O3-pr45971.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-if-conv.c

[Bug tree-optimization/45971] [4.6 Regression] ice in vect_update_ivs_after_vectorizer

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45971

Sebastian Pop <spop at gcc dot gnu.org> changed:

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

--- Comment #12 from Sebastian Pop <spop at gcc dot gnu.org> 2010-11-10 16:08:23 UTC ---
Fixed.