PyNN and Nest

If you want to build HBP brain models, the preferred way is to write your neural topology in PyNN (a Python package for simulator-independent specification of neuronal network models).
Once you've done this, you run the model through a simulator like Nest, Brian or Neuron.
They're working on making Nest distributed so you can run it on High Performance Computer Platforms.
There are standard neural and synaptic models that are supposed to run the same on different platforms.
There are problems with IO for agents.
Instead of a simulator, you can download your code onto neuromorphic chips, and run it there.
This really is better than our FLIF model written in Java because other people can use it more easily. (It's worse because I don't control the under lying code.)