Aligning auxiliary arrays
Want to minimize communication in largest parallel loop:
FORALL (I = K+1:N, J = K+1:N)
& A(I, J) = A(I, J) – COL(I) * ROW(J)
Suggests replicated alignments for COL and ROW:
!HPF$ ALIGN COL(I) WITH T(I, *)
!HPF$ ALIGN COL(J) WITH T(*, J)