Related Experiment Video
Updated: May 17, 2026
09:13
A Fully Automated and Highly Versatile System for Testing Multi-cognitive Functions and Recording Neuronal Activities in Rodents
Published on: May 3, 2012
Supercomputers ready for use as discovery machines for neuroscience
Moritz Helias1, Susanne Kunkel, Gen Masumoto
1Institute of Neuroscience and Medicine (INM-6), Computational and Systems Neuroscience, Jülich Research Centre Jülich, Germany ; RIKEN Brain Science Institute Wako, Japan.
Frontiers in Neuroinformatics
|November 7, 2012
Summary
The NEST simulator now leverages the K supercomputer for large-scale neural network simulations. This advancement makes complex brain modeling practical and accessible for computational neuroscience research.
Area of Science:
- Computational Neuroscience
- High-Performance Computing
- Neural Network Simulation
Background:
- The NEST simulator is a standard tool for modeling spiking neural networks.
- Previous limitations hindered the simulation of large-scale brain networks on supercomputers.
Purpose of the Study:
- To enhance the NEST simulator for efficient utilization of supercomputing resources.
- To enable large-scale neural network simulations previously unattainable.
Main Methods:
- Developed multi-threaded components for network wiring and simulation, utilizing 8 cores per MPI process.
- Applied a mathematical model of memory consumption to guide optimization.
- Exploited the hierarchical structure of the K supercomputer for managing large network communication.
Main Results:
- Achieved excellent computational scaling on the K supercomputer.
- Enabled simulation of networks up to 10^8 neurons and 10^12 synapses.
- Reduced simulation turn-around times to minutes, facilitating interactive research.
Conclusions:
- The optimized NEST simulator effectively harnesses supercomputing power for neuroscience.
- Large-scale neural network simulations are now a practical tool for computational neuroscience.
- Usability of supercomputers for network simulations approaches that of a single PC.
