Home

Welcome to the libpll-2 wiki!

libpll-2 is the new official fork of libpll. It implements the new site repeats optimization to speedup computations.

Please start reading libpll's wiki for more information about libpll. Then learn more about site repeats usage.

Projects that are already using libpll-2 and site repeats, and reported speedups compared with the tip pattern optimization:

• RAxML-NG: speedup ranges between 1.2 and 1.5
• ModelTest-NG: speedup around 2
• EPA-ng: no speedup (likelihood computation is not the main bottleneck) but memory footprint reduced by 30%.