[Bug tree-optimization/49957] New: Fails to SLP in 410.bwaves

[Bug tree-optimization/49957] New: Fails to SLP in 410.bwaves

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

           Summary: Fails to SLP in 410.bwaves
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: rguenth@gcc.gnu.org
                CC: spop@gcc.gnu.org, irar@gcc.gnu.org


While the loop

      subroutine shell(nx,ny,nz)
      implicit none
      integer i,j,k,l,nx,ny,nz
      real*8 q(21,nx,ny,nz),dq(21,nx,ny,nz)
      do k=1,nz
         do j=1,ny
            do i=1,nx
               do l=1,21
                  q(l,i,j,k)=q(l,i,j,k)+dq(l,i,j,k)
               enddo
            enddo
         enddo
      enddo
      return
      end

is vectorized using loop vectorization the following variant is not
(as appearing in 410.bwaves):

      subroutine shell(nx,ny,nz)
      implicit none
      integer i,j,k,l,nx,ny,nz
      real*8 q(5,nx,ny,nz),dq(5,nx,ny,nz)
      do k=1,nz
         do j=1,ny
            do i=1,nx
               do l=1,5
                  q(l,i,j,k)=q(l,i,j,k)+dq(l,i,j,k)
               enddo
            enddo
         enddo
      enddo
      return
      end

first of all dependence checking on the innermost unrolled loop fails:

(compute_affine_dependence
  (stmt_a =
D.1639_140 = *q_54[D.1638_139];
)
  (stmt_b =
*q_54[D.1638_157] = D.1647_160;
)
(subscript_dependence_tester
(analyze_overlapping_iterations
  (chrec_a = {((pretmp.33_209 + 6) + pretmp.33_213) + offset.5_32, +, 5}_3)
  (chrec_b = {((pretmp.33_209 + 7) + pretmp.33_213) + offset.5_32, +, 5}_3)
(analyze_siv_subscript
siv test failed: unimplemented.
)

as pretmp.33 is signed and we thus do not associate the constant offset
(well, I think this might be the problem at least).

That shouldn't prevent vectorization (and it doesn't).  But then
(probably due to the same reason) we get

t.f:7: note: === vect_analyze_data_ref_accesses ===
t.f:7: note: not consecutive access D.1639_140 = *q_54[D.1638_139];
t.f:7: note: not vectorized: complicated access pattern.
t.f:7: note: bad data access.
t.f:1: note: vectorized 0 loops in function.

so in the end it _does_ seem to be the underlying issue.

I will see what can be done here.

[Bug tree-optimization/49957] Fails to SLP in 410.bwaves

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.08.03 11:10:57
     Ever Confirmed|0                           |1

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-08-03 11:10:57 UTC ---
When associating the index expression "properly" it works.

[Bug tree-optimization/49957] Fails to SLP in 410.bwaves

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

Tobias Burnus <burnus at gcc dot gnu.org> changed:

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

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-08-03 14:05:13 UTC ---
Cf. RFC Fortran patch at http://gcc.gnu.org/ml/fortran/2011-08/msg00012.html to
change the index order into the canonical form.

[Bug tree-optimization/49957] Fails to SLP in 410.bwaves

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-08-03 14:41:49 UTC ---
The patch now makes us vectorize

shell_lam.f:303: note: LOOP VECTORIZED.
shell_lam.f:262: note: LOOP VECTORIZED.
shell_lam.f:205: note: LOOP VECTORIZED.

compared to just

shell_lam.f:262: note: LOOP VECTORIZED.

without.

[Bug tree-optimization/49957] Fails to SLP in 410.bwaves

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

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

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

--- Comment #4 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-08-03 15:17:42 UTC ---
Mine.

[Bug tree-optimization/49957] Fails to SLP in 410.bwaves

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.7.0

--- Comment #6 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-08-04 12:23:10 UTC ---
Fixed.

[Bug tree-optimization/49957] Fails to SLP in 410.bwaves

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49957

--- Comment #5 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-08-04 12:22:47 UTC ---
Author: rguenth
Date: Thu Aug  4 12:22:42 2011
New Revision: 177368

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=177368
Log:
2011-08-04  Richard Guenther  <rguenther@suse.de>

    PR fortran/49957
    * trans-array.c (add_to_offset): New function.
    (gfc_conv_array_ref): Build the array index expression in optimally
    associated order.
    (gfc_walk_variable_expr): Adjust for the backward walk.

    * gfortran.dg/vect/O3-pr49957.f: New testcase.

Added:
    trunk/gcc/testsuite/gfortran.dg/vect/O3-pr49957.f
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-array.c
    trunk/gcc/testsuite/ChangeLog