New TNT routines for parallel computing with MPI.

Mol Phylogenet Evol

Unidad Ejecutora Lillo (Fundación Miguel Lillo - Consejo Nacional de Investigaciones Científicas y Técnicas), Miguel Lillo 251, S. M. de Tucumán 4000, Argentina; American Museum of Natural History, Central Park West at 79th Street, New York, United States. Electronic address:

Published: January 2023

Phylogenetic inference, which involves time-consuming calculations, is a field where parallelization can speed up the resolution of many problems. TNT (a widely used program for phylogenetic analysis under parsimony) allows parallelization under the PVM system (Parallel Virtual Machine). However, as the basic aspects of the implementation remain unpublished, few studies have taken advantage of the parallelization routines of TNT. In addition, the PVM system is deprecated by many system administrators. One of the most common standards for high performance computing is now MPI (Message Passing Interface). To facilitate the use of the parallel analyses offered by TNT, this paper describes the basic aspects of the implementation, as well as a port of the parallelization interface of TNT into MPI. The use of the new routines is illustrated by reanalysis of seven significant datasets, either recent phylogenomic datasets with many characters (up to 2,509,064 characters) or datasets with large numbers of taxa (up to 13,921 taxa). Versions of TNT including the MPI functionality are available at: http://www.lillo.org.ar/phylogeny/tnt/.

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ympev.2022.107643DOI Listing

Publication Analysis

Top Keywords

computing mpi
8
pvm system
8
basic aspects
8
aspects implementation
8
tnt
6
tnt routines
4
routines parallel
4
parallel computing
4
mpi
4
mpi phylogenetic
4

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!