[Bug tree-optimization/33246] New: Missed opportunities for vectorization due to data ref analysis

[Bug tree-optimization/33246] New: Missed opportunities for vectorization due to data ref analysis

[spop at gcc dot gnu dot org]

Here is a reduced testcase for Polyhedron:fatigue.f90:928.  The non
vectorized loops that compute stress_tensor show up on the top user
times in the profile of that benchmark:

module perdida_m

integer, parameter, private :: LONGreal = selected_real_kind(15,90)

contains

      subroutine perdida (lambda, mu, stress_tensor, strain_tensor,
plastic_strain_tensor)
      real (kind = LONGreal), intent(in) :: lambda, mu
      real (kind = LONGreal), dimension(:,:), intent(in) :: strain_tensor
      real (kind = LONGreal), dimension(:,:), intent(inout) ::
plastic_strain_tensor
      real (kind = LONGreal), dimension(:,:), intent(out) :: stress_tensor
      stress_tensor(:,:) = (1.0_LONGreal - damage) *                           
            &
                                      generalized_hookes_law
(strain_tensor(:,:) -          &
                                                     
plastic_strain_tensor(:,:), lambda, mu)
      end subroutine perdida


      function generalized_hookes_law (strain_tensor, lambda, mu) result
(stress_tensor)
      real (kind = LONGreal), dimension(:,:), intent(in) :: strain_tensor
      real (kind = LONGreal), intent(in) :: lambda, mu
      real (kind = LONGreal), dimension(3,3) :: stress_tensor
      end function generalized_hookes_law

end module perdida_m


This is due to data ref analysis problems:

./fatigue.f90:14: note: not vectorized: data ref analysis failed
(*stress_tensor.0_16)[D.1508_168] = D.1513_173
./fatigue.f90:14: note: bad data references.

and

./fatigue.f90:14: note: not vectorized: data ref analysis failed D.1489_133 =
(*strain_tensor.0_41)[D.1488_132]
./fatigue.f90:14: note: bad data references.


-- 
           Summary: Missed opportunities for vectorization due to data ref
                    analysis
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: spop at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33246

[Bug tree-optimization/33246] Missed opportunities for vectorization due to data ref analysis

[spop at gcc dot gnu dot org]

------- Comment #1 from spop at gcc dot gnu dot org  2007-08-30 03:07 -------
Created an attachment (id=14136)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14136&action=view)
testcase

I just realized that all these inlined testcases were scrambled by bugzilla.
I'm submitting the testcases as attachments.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33246

[Bug tree-optimization/33246] Missed opportunities for vectorization due to data ref analysis

[steven at gcc dot gnu dot org]

------- Comment #2 from steven at gcc dot gnu dot org  2007-08-30 08:07 -------
Having this case vectorized might help us look better compared to other
compilers.  gfortran produces a 65% slower "fatigue" binary than the faster
compiler (Sun f95, see http://www.eleves.ens.fr/home/coudert/benchmark.pdf)


-- 

steven at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2007-08-30 08:07:00
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33246

[Bug tree-optimization/33246] Missed opportunities for vectorization due to data ref analysis

[dorit at gcc dot gnu dot org]

------- Comment #3 from dorit at gcc dot gnu dot org  2007-08-31 13:57 -------
...
> This is due to data ref analysis problems:
> ./fatigue.f90:14: note: not vectorized: data ref analysis failed
> (*stress_tensor.0_16)[D.1508_168] = D.1513_173
> ./fatigue.f90:14: note: bad data references.
> and
> ./fatigue.f90:14: note: not vectorized: data ref analysis failed D.1489_133 =
> (*strain_tensor.0_41)[D.1488_132]
> ./fatigue.f90:14: note: bad data references.

The data-ref analyzer reports:
   failed: evolution of offset is not affine.

As a result, the DR fields that represent the access relative to the inner-most
loop are almost all empty:

        base_address:
        offset from base address:
        constant offset from base address:
        step:
        aligned to:
        base_object: (*(real8[0:D.1433] *) D.1437_15)[0]
        symbol tag: SMT.79

However note that the DR fields relative to the outer-loop are computable:

        outer base_address: &A.23
        outer offset from base address: 0
        outer constant offset from base address: 0
        outer step: 24
        outer aligned to: 128

If the data-ref analyzer can return the expression for the evolution in the
inner-loop, instead of failing, we would at least have a chance to do
outer-loop vectorization. 

This is a duplicate of PR33113.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33246

[Bug tree-optimization/33246] Missed opportunities for vectorization due to data ref analysis

[spop at gcc dot gnu dot org]

------- Comment #4 from spop at gcc dot gnu dot org  2007-08-31 14:19 -------
Duplicate.

*** This bug has been marked as a duplicate of 33113 ***


-- 

spop at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33246