Comparison of thinning algorithms on distributed memory machines