[Bug tree-optimization/65935] New: [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[Bug tree-optimization/65935] New: [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

            Bug ID: 65935
           Summary: [6 Regression] 433.milc in SPEC CPU 2006 is
                    miscompiled
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: hjl.tools at gmail dot com
                CC: rguenther at suse dot de
  Target Milestone: ---

On Linux/i686, r62283 miscompiled 433.milc in SPEC CPU 2006:

gcc -m32 -c -o com_vanilla.o -DSPEC_CPU -DNDEBUG -I. -DFN -DFAST
-DCONGRAD_TMP_VECTORS -DDSLASH_TMP_LINKS  -O3 -funroll-loops -msse2
-mfpmath=sse -ffast-math -fwhole-program -flto=jobserver -fuse-linker-plugin   
            com_vanilla.c
gcc -m32 -c -o io_nonansi.o -DSPEC_CPU -DNDEBUG -I. -DFN -DFAST
-DCONGRAD_TMP_VECTORS -DDSLASH_TMP_LINKS  -O3 -funroll-loops -msse2
-mfpmath=sse -ffast-math -fwhole-program -flto=jobserver -fuse-linker-plugin   
            io_nonansi.c
gcc -m32  -O3 -funroll-loops -msse2 -mfpmath=sse -ffast-math -fwhole-program
-flto=jobserver -fuse-linker-plugin          control.o f_meas.o gauge_info.o
setup.o update.o update_h.o update_u.o layout_hyper.o check_unitarity.o
d_plaq4.o gaugefix2.o io_helpers.o io_lat4.o make_lattice.o path_product.o
ploop3.o ranmom.o ranstuff.o reunitarize2.o gauge_stuff.o grsource_imp.o
mat_invert.o quark_stuff.o rephase.o cmplx.o addmat.o addvec.o clear_mat.o
clearvec.o m_amatvec.o m_mat_an.o m_mat_na.o m_mat_nn.o m_matvec.o make_ahmat.o
rand_ahmat.o realtr.o s_m_a_mat.o s_m_a_vec.o s_m_s_mat.o s_m_vec.o s_m_mat.o
su3_adjoint.o su3_dot.o su3_rdot.o su3_proj.o su3mat_copy.o submat.o subvec.o
trace_su3.o uncmp_ahmat.o msq_su3vec.o sub4vecs.o m_amv_4dir.o m_amv_4vec.o
m_mv_s_4dir.o m_su2_mat_vec_n.o l_su2_hit_n.o r_su2_hit_a.o m_su2_mat_vec_a.o
gaussrand.o byterevn.o m_mat_hwvec.o m_amat_hwvec.o dslash_fn2.o
d_congrad5_fn.o com_vanilla.o io_nonansi.o             -lm        -o milc

  Running 433.milc ref peak lto default

433.milc: copy 0 non-zero return code (exit code=1, signal=0)

WARMUPS COMPLETED
Unitarity problem on node 0, site 0, dir 0 tolerance=1.000000e-04
SU3 matrix:
0.991448 0.021247 -0.099836 -0.054012 0.051686 0.032000 
0.100302 -0.049616 0.989806 -0.028206 0.027581 -0.078778 
-0.059070 0.023841 -0.022247 -0.078163 0.994634 0.006470 
repeat in hex:
82a7e05f 9c8421c0 b159fbbd cc2c8301 eca1e2e4 4a88bfbf 
7c12f3d5 f7ba11aa 1939e0e5 a49c5e6c 0095ea4b a16c5754 
cfce59e0 209d3be4 702a8d28 2fdd37ea d75f9836 f0058a61 
...
Unitarity error count exceeded.
termination: Wed Apr 29 11:44:07 2015

Termination: status = 1

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-04-30
   Target Milestone|---                         |6.0
     Ever confirmed|0                           |1

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> ---
gcc -m32  -O3 -funroll-loops -msse2 -mfpmath=sse -ffast-math also miscompiled.

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

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

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

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
What exact revision do you mean?  r62283 is quite a bit old...

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
I suppose r222514.

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Richard Biener from comment #3)
> I suppose r222514.

Yes, it is.

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
Also reproduces with -Ofast -march=bdver2 -m32 so I suppose -O3 -ffast-math
-m32 -msse2 should also reproduce it.

I wont get to look into this before Monday, any help in bisecting
(-fno-tree-vectorize for which file(s) fix it?  for which function?)
appreciated
until then.

At least milc is C and not too big ;)

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Also fails with the test input.

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
Building rephase.c:rephase with -fno-tree-slp-vectorize makes it succeed.

The loop is basically

register int i,j,k,dir;
register site *s;

    for(i=0,s=lattice;i<sites_on_node;i++,s++){
 for(dir=0;dir<=3;dir++){
     for(j=0;j<3;j++)for(k=0;k<3;k++){
  s->link[dir].e[j][k].real *= s->phase[dir];
  s->link[dir].e[j][k].imag *= s->phase[dir];
     }
 }
    }

where the inner two loops are unrolled and SLP applies to them.  We need to
perform swapping on some of the mult operands and then somehow fail to use
the correct vectors to build up s->phase[dir] from parts.

  vect_cst_.35_44 = {_117, _121};
  vect_cst_.36_45 = {_110, _114};
  vect_cst_.37_46 = {_103, _107};
  vect_cst_.38_47 = {_92, _96};
  vect_cst_.39_48 = {_85, _89};
  vect_cst_.40_52 = {_78, _82};
  vect_cst_.41_53 = {_49, _71};
  vect_cst_.42_54 = {_31, _60};
  vect_cst_.43_55 = {_24, _27};

note how these should be all {_24, _24} but only the first one is correct.

I think this is a latent issue since we do the swapping tricks.

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Testcase that also fails with -m64.

/* { dg-do run } */
/* { dg-additional-options "-O3" } */
/* { dg-require-effective-target vect_double } */

#include "tree-vect.h"

extern void abort (void);
extern void *malloc (__SIZE_TYPE__);

struct site {
    struct {
        struct {
            double real;
            double imag;
        } e[3][3];
    } link[32];
    double phase[32];
} *lattice;
int sites_on_node;

void rephase (void)
{
  int i,j,k,dir;
  struct site *s;
  for(i=0,s=lattice;i<sites_on_node;i++,s++)
    for(dir=0;dir<32;dir++)
      for(j=0;j<3;j++)for(k=0;k<3;k++)
        {
          s->link[dir].e[j][k].real *= s->phase[dir];
          s->link[dir].e[j][k].imag *= s->phase[dir];
        }
}

int main()
{
  int i,j,k;
  check_vect ();
  sites_on_node = 1;
  lattice = malloc (sizeof (struct site) * sites_on_node);
  for (i = 0; i < 32; ++i)
    {
      lattice->phase[i] = i;
      for (j = 0; j < 3; ++j)
        for (k = 0; k < 3; ++k)
          {
            lattice->link[i].e[j][k].real = 1.0;
            lattice->link[i].e[j][k].imag = 1.0;
            __asm__ volatile ("" : : : "memory");
          }
    }
  rephase ();
  for (i = 0; i < 32; ++i)
    for (j = 0; j < 3; ++j)
      for (k = 0; k < 3; ++k)
        if (lattice->link[i].e[j][k].real != i
            || lattice->link[i].e[j][k].imag != i)
          abort ();
  return 0;
}

/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "slp1" } } */
/* { dg-final { cleanup-tree-dump "slp1" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #9 from Richard Biener <rguenth at gcc dot gnu.org> ---
Testing a fix.  The issue is latent with at least constant operands - but it's
probably impossible(?) to get bogus operand order on those.  In theory if the
loads from phase were just in a different BB the issue would reproduce as well.
OTOH then GCC 5 swaps in vect_get_and_check_slp_defs which already does swap
operands in the IL...

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

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

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

--- Comment #11 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.

[Bug tree-optimization/65935] [6 Regression] 433.milc in SPEC CPU 2006 is miscompiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65935

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
Author: rguenth
Date: Mon May  4 13:31:02 2015
New Revision: 222764

URL: https://gcc.gnu.org/viewcvs?rev=222764&root=gcc&view=rev
Log:
2015-05-04  Richard Biener  <rguenther@suse.de>

        PR tree-optimization/65935
        * tree-vect-slp.c (vect_build_slp_tree): If we swapped operands
        then make sure to apply that swapping to the IL.

        * gcc.dg/vect/bb-slp-pr65935.c: New testcase.

Added:
    trunk/gcc/testsuite/gcc.dg/vect/bb-slp-pr65935.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-vect-slp.c