Sandia Fires Up a Brain-Like Supercomputer That Can Simulate 180 Million Neurons
1 min read
Summary
Sandia National Laboratories has switched on a neuromorphic supercomputer called SpiNNaker2, which is capable of simulating between 150 and 180 million neurons and is 18 times more energy efficient than existing graphics processing units (GPUs).
The computer operates using bursts of electricity to mimic the way in which the human brain communicates, only using neurons when they have important information to transmit, therefore using less power.
The computing industry is betting that brain-inspired computer hardware could provide a more energy efficient way of running machines, as the energy requirements for mining AI continue to grow.
Some estimates suggest AI accounts for up to 20% of global data centre power demand.
