The use of lists instead of trees results in a large reduction of RAM usage (Figure 1). In addition, the parallelization of the code allows for drastic reduction of the computation times. This is shown on the speed-up curves (Figure 2) where the speed-up is the ratio between the computation time with p processors and the computation time with 1 processor. Ideally, the speed-up curve must be close to S_p=p, which is the case here.

Figure 1: Comparison of RAM usage between lists and trees.

Figure 2: Speed-up curves for 3 test situations.