Panjo is a parallel implementation of the neighbor-joining algorithm commonly used in Phylogenetics. It is implemented using distributed memory and the total memory usage per-processor should be on the order of (N^2)/P where P is the number of processors and N is the number of things to put in a tree.

There is not much documentation available yet, but you can read a report here which is more about the implementation and some benchmark numbers which might be useful. report

download platform release date status maintainer release notes
panjo_0.0.1.tar.gz (mac osx, unix) unstable bullard & stat,berkeley,edu ...