CoreNEURON is a compute engine for the NEURON simulator optimised for both memory usage and computational speed on modern CPU/GPU architectures. Some of its key goals are to:

  • Efficiently simulate large network models
  • Reduce memory usage
  • Support execution on specialised hardware, GPUs
  • Support optimisations such as vectorisation and cache-efficient memory layout (e.g. Structure-Of-Arrays)

CoreNEURON is tightly integrated with NEURON, and can transparently simulate a large number of NEURON models. Certain conditions that must be met are covered in the CoreNEURON Compatibility section.