Skip to main content


Table 3 Running times on large data sets

From: Towards a practical O(n logn) phylogeny algorithm

  # of sequences
  20,000 40,000 78,132
weighted majority, 5 quartets 6m 41s 15m 52s 34m
FastTree (NJ phase only) 13m 52s 41m 15s 116m
  1. Running times of the random walk algorithm compared to FastTree. We used the huge.1 alignment from the original FastTree paper[9]. Smaller data sets were created by choosing a random subset of sequences from the large alignment. Our algorithm runs 2.1 to 3.4 times faster than FastTree on these very large data sets.