A 12-STEP SORTING NETWORK FOR 22 ELEMENTS

Sorting Networks are cost-effective multistage interconnection networks with sorting capabilities. Optimal sorting networks theoretically consume Θ(NlogN) comparisons. However, the fastest implementable sorting networks built so far consume Θ(Nlog 2 N) comparisons, and generally, use the Merge-sorting strategy to sort the input. This implies that faster networks can be developed — here we show a network that sorts 22 elements in only 12 steps, outperforming the Merge-sorting based solution for this problem which needs at least 13 steps.