Right now nPhase is deterministic, meaning it doesn't use any useful time-saving heuristics. It also makes no use of parallelization. Generally it isn't optimized. It would be good to work on these points to make it more appropriate for use on larger/more heterozygous genomes.