Parallelise the cluster (regular mixture model) algorithms more efficiently

dsteinberg / libcluster

An extensible C++ library of Hierarchical Bayesian clustering algorithms, such as Bayesian Gaussian mixture models, variational Dirichlet processes, Gaussian latent Dirichlet allocation and more.

GNU Lesser General Public License v3.0

140 stars 22 forks source link

Parallelise the cluster (regular mixture model) algorithms more efficiently #5

Open dsteinberg opened 8 years ago

dsteinberg commented 8 years ago

At the moment only the 'grouped' mixtures and more hierarchical algorithms are parallelised in an efficient manner.

dsteinberg commented 8 years ago

This code is really slow compared to just dividing the data into groups and running e.g. the parallel G-LDA algorithm...