This repository has been archived by the owner. It is now read-only.
Deploy the Neural Simulation Tool (NEST) over large Beowulf MPI-clusters on DigitalOcean. Ready in 15 minutes.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

NEST-simulator Beowulf cluster

The deloy system requires at least python 2.7 and DigitalOcean python-bindings (pip install -U python-digitalocean). It was tested up to 1/2 million neurons (Brunnel balanced network) on 400vcpus with 1.28To RAM for ~19$ hourly. This script is provided as is and comes with no warranty nor liability.


This was part of a bachelor semester project carried out at the Laboratory of Computational Neuroscience from Swiss Institute of Technology in Lausanne. The command was used to evaluate different approaches of implementing spike-timing dependent plasticity on various size and networks. You can find the resulting work here.


Digital Ocean token is stored in .key file at command root. Be sure to never share it with anyone.

  • ./ key [token]: setup Digital Ocean token (empty for removing it)
  • ./ list: list current cluster
  • ./ create [2|3|...] [1gb|2gb|...]: create n-cluster of given size and setup ssh everywhere
  • ./ install [script]: execute given bash script on each node
  • ./ run [1..] [program]: run given python (PyNEST) program on the cluster using mpi and n processes
  • ./ delete: delete: remove current cluster

Common scenario

Global installation takes 15 minutes on average.

./ key digitalOceanToken
./ create 4 8gb
./ list
./ install install_env.bash
./ install install_nest.bash
./ install install_stdpmodule.bash
./ run 16
./ run 8
./ run 4
./ delete