Published in News

An owl’s brain powers Chipzilla's new computer

by on18 April 2024

Who needs quantum cats when neuromorphic computing takes flight

Intel has heralded a significant leap in neuromorphic with the unveiling of Hala Point, billed as the world's most substantial computer.

Quantum computing is sucking up all the headlines, but its promise remains on the distant horizon due to scaling issues. Meanwhile, neuromorphic computing emerges as the immediate game-changer, drawing design cues from the human brain to revolutionise chip architecture.

Intel's Loihi 2 processor powers Intel’s Hala Point. It mimics the cognitive capacity of an owl's brain, performing up to 20 quadrillion operations per second with staggering energy efficiency.

Hala Point eclipses traditional CPU and GPU systems in real-time AI tasks, offering a fiftyfold speed increase and a hundredfold energy reduction. Intel invites enterprises to explore this technology's potential through the Intel Neuromorphic Research Community, a consortium of over 200 members, including industry giants and academic institutions.

Intel's commitment to traditional manufacturing techniques and digital circuits has enabled the practical application of neuromorphic principles, such as the unification of memory and processing. The Loihi 2 system can learn in real-time, adapting its neural communications dynamically, a stark contrast to conventional AI systems' static training.

This adaptive capability renders neuromorphic computing ideal for edge computing scenarios, including streaming media analysis and wireless signal processing, and holds promise for data centres and high-performance computing applications.

Despite the strides made by Intel and others like IBM, BrainChip, and Prophesee, neuromorphic computing faces significant adoption barriers. The shift to event-based spike processing demands a radical overhaul of programming languages and a departure from established AI models designed for traditional computing infrastructures.

The technology's nascent state, coupled with a limited developer ecosystem and a shortage of tools, underscores the challenges ahead. Yet, the potential for neuromorphic computing to redefine the landscape of AI and optimisation remains undiminished.

Last modified on 18 April 2024
Rate this item
(1 Vote)