Add benchmark comparing igraph and motifcensus

[david-beauchesne]

I think that the advantage on motifcensus over igraph is not only that it considers species position in the census, but also that it is much more efficient. It could be interesting to include a simple benchmark in the package description.