[Bug fortran/30374] New: aliasing amoung dummy arguments: possibly missing optimization

[Bug fortran/30374] New: aliasing amoung dummy arguments: possibly missing optimization

[burnus at gcc dot gnu dot org]

The following program is (more or less) well-defined for the human reader, but
non-standard Fortran (which cannot be easily detected at compile time, cf. PR
30373).

gfortran outputs:
            1           1           1
while ifort, NAG f95 and sunf95 output the expected:
            4           5           6

program test
  implicit none
  integer, dimension(3,3) :: n
  n(:,1)=(/1,2,3/)
  n(:,2)=(/1,2,3/)
  n(:,3)=(/1,2,3/)
  call sub(n(1,:),n(1,:))
  write(6,*) n(1,:)
contains
  subroutine sub(a,b)
    implicit none
    integer, dimension(3), intent(inout) :: a
    integer, dimension(3), intent(in) :: b
    a=(/4,5,6/)
  end subroutine
end program test

Expected:

- Give same result as the other compilers ?

- Check, whether one could do some code-generation improvements; Richard Main
wrote:

"in this particular case, there is are simple optimizations that would avoid
problems. Some compilers might do variants of those optimizations. In
particular, passing a non-contiguous array slice (such as a row) here is likely
to cause copy-in/copy-out. Such copy-in/copy-out is one of the things that
the standard is explicitly written to allow; it is also sometimes the source of
behavior that can be surprising. (Or conversely, in other situations, the lack
of copy-in/copy-out can result in behavior that might seem surprising to some.)

"I suspect that some compilers avoid the copy out here, possibly because of the
intent(in), or possibly because of b being unused (in which case, even the copy
in could be avoided). It seems an obvious enough optimization provided that the
compiler has the right information at the right time to do it. However, the
standard does *NOT* require this optimization. Failure to do the optimization
does not constitute a compiler bug in terms of implementing the standard
(though it might indicate that the compiler has room to improve its
optimization)."

http://groups.google.com/group/comp.lang.fortran/browse_thread/thread/2b4df0a44e704b7f/70c44bc5ea8cafb0


-- 
           Summary: aliasing amoung dummy arguments: possibly missing
                    optimization
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: burnus at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374

[Bug fortran/30374] aliasing amoung dummy arguments: possibly missing optimization

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2007-01-04 22:29 -------
What GCC is doing is calling _gfortran_internal_pack/_gfortran_internal_unpack
in a different order than most other compilers.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374

[Bug fortran/30374] aliasing amoung dummy arguments: possibly missing optimization

[pault at gcc dot gnu dot org]

------- Comment #2 from pault at gcc dot gnu dot org  2007-01-05 09:30 -------
(In reply to comment #1)
> What GCC is doing is calling _gfortran_internal_pack/_gfortran_internal_unpack
> in a different order than most other compilers.

It should not call _gfortran_internal_unpack at all for the intent in variable
- this is the missed optimisation to which Richard was referring.  I am
regetesting a patch with the patches for PR23232, PR27996 and PR27998.

Paul


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374

[Bug fortran/30374] aliasing amoung dummy arguments: possibly missing optimization

[jv244 at cam dot ac dot uk]

------- Comment #3 from jv244 at cam dot ac dot uk  2007-10-26 07:25 -------
paul, the patch you have been regtesting (9 months ;-) seems not yet in. I've
marked this as a missed-optimization.


-- 

jv244 at cam dot ac dot uk changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu dot org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |missed-optimization
   Last reconfirmed|0000-00-00 00:00:00         |2007-10-26 07:25:36
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374

[Bug fortran/30374] unpacking intent(IN) arguments

[jv244 at cam dot ac dot uk]

------- Comment #4 from jv244 at cam dot ac dot uk  2009-05-21 08:34 -------
Paul, 2 years ago you had a (one liner?) patch for this one, but gfortran still
seem to unpack for intent(IN) data:

MODULE M1
CONTAINS
  SUBROUTINE S1(I)
   INTEGER, DIMENSION(3), INTENT(IN) :: I
  END SUBROUTINE
END MODULE

USE M1
INTEGER, TARGET, DIMENSION(9) :: I
INTEGER, DIMENSION(:), POINTER :: IP
IP=>I(1:9:3)
CALL S1(IP)
END


-- 

jv244 at cam dot ac dot uk changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |4.5.0
   Last reconfirmed|2007-10-26 07:25:36         |2009-05-21 08:34:53
               date|                            |
            Summary|aliasing amoung dummy       |unpacking intent(IN)
                   |arguments: possibly missing |arguments
                   |optimization                |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374

[Bug fortran/30374] unpacking intent(IN) arguments

[jv244 at cam dot ac dot uk]

------- Comment #5 from jv244 at cam dot ac dot uk  2009-05-21 08:40 -------
(In reply to comment #4)
> Paul, 2 years ago you had a (one liner?) patch for this one, but gfortran still
> seem to unpack for intent(IN) data:

ugh... no. I tested with 4.3 instead of 4.5. Instead we can close this, as this
seems solved.


-- 

jv244 at cam dot ac dot uk changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|4.5.0                       |
      Known to work|                            |4.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374

[Bug fortran/30374] unpacking intent(IN) arguments

[jv244 at cam dot ac dot uk]

------- Comment #6 from jv244 at cam dot ac dot uk  2009-05-21 08:45 -------
fixed


-- 

jv244 at cam dot ac dot uk changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30374