HDNNSim is a closed-source project. The HDNNApp format, however, is disclosed under a permissive license (MIT Expat).

The project is privately funded, with a little help from the Netherlands Enterprise Agency (


HDNNSim is written in C, using OpenMP, and has a minimal set of dependencies, predominantly for I/O related functions:

  • SQLite
  • JSMN, a lightweight json parser

The available tooling includes a drop-in replacement for PyNest (python) and an adaptor for TensorFlow's SavedModel.

Contact us 

by email: j.koppen -at-