PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

compiler bug for array pointer assignments

 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling
View previous topic :: View next topic  
Author Message
tesfortran



Joined: 04 Dec 2012
Posts: 3
PostPosted: Fri Dec 07, 2012 10:23 am    Post subject: compiler bug for array pointer assignments Reply with quote
I have come across what appears to be a compiler bug.
I have a very simple program that demonstrates this. I want to point a 3D array at an equivalent sized 1D array

here is the code

Code:

      MODULE array_1D
       integer, allocatable, target, save :: array(:)
      end Module array_1D
     
C-----------------------------------------------------

      PROGRAM main
       call alloc
       call write1
       call write2
       call sub
       call write1
       call write2
      end
     
C-----------------------------------------------------

      Subroutine alloc   
       use array_1D
       l=2
       n=3
       m=4
       allocate(array(l*n*m))
       array =0
      return
      end
       
C-----------------------------------------------------

      Subroutine sub
       use array_1D
       l=2
       n=3
       m=4
       do i=1,l*n*m
        array(i)=i
       end do
      return
      end
       
C-----------------------------------------------------

      Subroutine write1
       use array_1D
       l=2
       n=3
       m=4
       write(*,'(24I4)')(array(i),i=1,l*n*m)
      return
      end
       
C-----------------------------------------------------

      Subroutine write2
       use array_1D
       integer, pointer :: array3d(:,:,:)
       l=2
       n=3
       m=4
       array3d(1:l,1:n,1:m)=>array
       write(*,'(24I4)')(((array3d(i,j,k),i=1,l),j=1,n),k=1,m)
      return
      end
       
       
             


when I compile this I don't get any compile time errors; however when I run it I get the following results.

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
1 2 3 4 5 6 6 7 8 9 10 11 11 12 13 14 15 16 16 17 18 19 20 21

As you can see it appears that the compiler is miscalculating the offset in the memory for the third dimension. I get similar results with different ranges of l,n,m. I got the same results using PGF90 or PGFORTRAN

I compiled the exact same source code using another brand of compiler and got the expected results:
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Is there something wrong with the compiler or am I missing something here?
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 4995
Location: The Portland Group Inc.
PostPosted: Fri Dec 07, 2012 11:14 am    Post subject: Reply with quote
Hi tesfortran,

Thanks for the report. I was able to recreate the issue here and have sent it to our compiler engineers for further investigation (TPR#19037).

- Mat
Back to top
View user's profile
jtull



Joined: 30 Jun 2004
Posts: 233
PostPosted: Tue May 21, 2013 6:02 pm    Post subject: TPR 19037 is fixed as of 13.1 Reply with quote
19037 - UF: Array remapping gives incorrect results

was fixed in the 13.1 release.

thanks,
dave
Back to top
View user's profile
Display posts from previous:   
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling All times are GMT - 7 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2002 phpBB Group