Next: Successive Over Relaxation Example
Up: Scalar Variables
Previous: Scalar Variables
Given the following assignments:
A(2:99) = (A(:98)+A(3:))/2 ! neighbour calculations
B(22:56)= 4.0*ATAN(1.0) ! section of B calculated
C(:) = SUM(D,DIM=1) ! Sum down a column
Assuming the `owner-computes' rule,
the following distributions would be examples of good HPF programming,
!HPF$ DISTRIBUTE (BLOCK) ONTO P :: A
!HPF$ DISTRIBUTE (CYCLIC) ONTO P :: B
!HPF$ DISTRIBUTE (BLOCK) ONTO P :: C ! or (CYCLIC)
!HPF$ DISTRIBUTE (*,BLOCK) ONTO P :: D ! or (*,CYCLIC)
The following facts were used in deciding which distribution methods were
to be used:
- A is involved in neighbourhood calculations. A(3) is
combined
with A(1) and assigned to A(2)
- B an `odd shaped' inner section of B is used -
in order to balance the load a CYCLIC distribution should be used.
- the calculation of an element of C involves a whole column
of D. This means that every element of C should be aligned
with the corresponding whole column of D.
Now try this question
Return to corresponding overview page
Next: Successive Over Relaxation Example
Up: Scalar Variables
Previous: Scalar Variables
©University of Liverpool, 1997
Wed May 28 20:20:27 BST 1997Not for commercial use.