SpiNNaker
Massively parallel neuromorphic supercomputer for brain simulation
SpiNNaker is a manycore supercomputer built by the University of Manchester to simulate spiking neural networks. Released in 2019, it uses over one million ARM cores to model human brain activity in real time.
- Contains 1,036,800 ARM968 cores
- Simulates up to a billion neurons in real time
- Uses 7 TB of mobile DDR SDRAM
- Requires 100 kW of power
- Successor SpiNNcloud launched in 2025
What it is
SpiNNaker is a specialised supercomputer architecture designed for simulating spiking neural networks. Developed by the Advanced Processor Technologies Research Group at the University of Manchester, it functions as a massively parallel computing platform. The system aims to replicate the behaviour of large aggregates of neurons in real time, making it a key tool for neuromorphic research and brain modelling projects.
How it came about
The design emerged from the Department of Computer Science at the University of Manchester. It was created to support the Human Brain Project by providing hardware capable of handling complex neural simulations. On 14 October 2018, the project announced it had reached the milestone of one million active cores. The full system was completed and announced in 2019, marking a significant step in large-scale brain simulation capabilities.
Inside the chip
The architecture relies on manycore processing rather than traditional high-speed single cores. It consists of 57,600 processing nodes, each containing 18 ARM968 processors running at 200 MHz. Every core is paired with 128 MB of mobile DDR SDRAM and emulates 1,000 neurons. This structure allows the system to manage over one million cores and 7 TB of memory, facilitating the real-time simulation of up to a billion neurons through spiking neural network models.
How it was made
The complete machine is housed in ten 19-inch racks, with each rack containing more than 100,000 cores. The chips are mounted on cards within five blade enclosures. The system requires a 240 V power supply and draws approximately 100 kW of energy. It operates in an air-conditioned environment to manage the heat generated by the dense arrangement of processing nodes and memory modules.
Where it went
SpiNNaker serves as a primary component of the neuromorphic computing platform for the Human Brain Project. Researchers use it to simulate the behaviour of biological neural networks. By modelling up to a billion neurons in real time, the system helps scientists understand brain function and test theories of neural connectivity without the limitations of conventional serial computing architectures.
Versions and family
The initial system achieved its million-core target in 2018. A second generation, known as SpiNNcloud, was funded by an 8 million euro grant awarded to TU Dresden in September 2019. This successor machine entered operation in 2025, expanding on the original architecture to provide even greater simulation capacity for neuromorphic research applications.
Why it matters
The project demonstrated that low-power, manycore architectures could effectively simulate large-scale neural systems. By reaching the one-million-core milestone, it proved the viability of using modified ARM processors for neuromorphic tasks. This work has influenced the development of future brain-inspired computing platforms and contributed significantly to the goals of the European Human Brain Project.
Questions people ask about SpiNNaker
What is the main purpose of SpiNNaker?
It is designed to simulate spiking neural networks in real time. The system models the behaviour of up to a billion neurons, helping researchers study brain function within the Human Brain Project.
How many processors does the system contain?
The machine comprises 57,600 nodes, each with 18 ARM968 processors. This totals 1,036,800 cores, all running at 200 MHz to handle massive parallel neural simulations.
Who developed SpiNNaker?
The Advanced Processor Technologies Research Group at the University of Manchester designed the architecture. Steve Furber is credited as the developer, and it is part of the Manchester computers family.
What came after the original SpiNNaker?
A second generation called SpiNNcloud was funded in 2019. Built by TU Dresden, this successor system went into operation in 2025 to continue neuromorphic research efforts.
Good reading and sources
- Official websiteOfficial site
- Efficient Reward-Based Structural Plasticity on a SpiNNaker 2 PrototypearXiv, Article
The facts on this page come from the Wikipedia article "SpiNNaker", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryChip rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.
It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.