Multiple Sequence Alignment is an essential tool in the analysis and comparison of biological sequences. Unfortunately, the complexity of this problem is exponential. Currently feasible methods are, therefore, only approximations. The progressive multiple sequence alignment algorithms are the most widespread among these approximations. Still, the computation speed of typical problems is often not satisfactory. Hence, the well known progressive alignment scheme of ClustalW has been subject to parallelization to further accelerate the computation. In the course of this action a unique scheme to parallelize sequence alignment in particular and dynamic programming in general was discovered, which yields an average of n/2 parallel calculations for problem size n. The scalability of O(n) tasks for problem size n can be even maintained for slower networks.