Figure 12.1 Two orderings for summing four values. Shown in (a) is the ordering typically used by parallel algorithms. Shown in (b) is the natural "do loop" ordering.