efficient way of using large array subsections

[Kholikov__Shukur]

Hi all,

I need some help to speed up a small portion of my fortran code.

Following code is inside several loops.

    ix1 and iy1 variables are assigned in outer loops

      ij = 0
      do i = 1,4
         ix1pi = ix1+i
         do j = 1,4
            ij = ij+1
            q(ij,:) = dcube(:,ix1pi,iy1+j)
            uxuy(ij,:) = ux(i)*uy(j)
         enddo
      enddo
      sig2 = sum(q*uxuy,dim=1)

 "dcube" is allocatable array. 

My question is how to avoid copying columns or any other way of speeding this portion of code.

Thanks,

Shukur

[Steve_Lionel]

For one, I'd suggest reversing the bounds of q and uxuy (if you can), as the array sections you are assigning to are discontiguous and this is going to require an extra loop.

[Kholikov__Shukur]

Thanks Steve!!!  2.3 faster.  I think this is the case because this portion of the code is within 3 outer loops.

I am HAPPY.

Shukur