Since the Brandes Algorithm only increments one piece of data for each starting node, you can do the calculations for each starting node independently on different threads. So that's what we did. This doesn't make the computational complexity lower, but does make the overall runtime of the algorithm lower. (From about 50 minutes on my machine down to about 7 minutes).
Since the Brandes Algorithm only increments one piece of data for each starting node, you can do the calculations for each starting node independently on different threads. So that's what we did. This doesn't make the computational complexity lower, but does make the overall runtime of the algorithm lower. (From about 50 minutes on my machine down to about 7 minutes).