Intel has unveiled a new neuromorphic system with over 1.15 billion neurons using the Loihi 2 processors. This system, known as Hala Point, aims to revolutionize the approach to Artificial Intelligence (AI) by offering a more efficient and modular solution.
The company believes that the industry needs new scalable approaches, and Hala Point is expected to push technology forward in terms of efficiency and adaptability at scale. The system consists of 1.152 Loihi 2 processors on an Intel 4 process node, supporting up to 1.15 billion neurons and 128 billion synapses across 140,544 neuromorphic processing cores.
Hala Point integrates processing, memory, and communication channels in a parallelized fabric, providing significant memory and communication bandwidth. With capabilities to perform 20 quadrillion operations per second, the system is designed for real-time continuous learning in various AI applications such as scientific problem solving, logistics, smart city infrastructure management, and AI agents.
Compared to its predecessor, Pohoiki Springs, Hala Point offers significant improvements in performance and efficiency for real-time processing tasks. The system is capable of executing its full capacity 20 times faster than the human brain and up to 200 times faster at lower capacity, especially in bioinspired models of spiking neural networks.
Researchers at Sandia National Laboratories will utilize Hala Point to solve scientific computing problems in various fields like device physics, computer architecture, and computer science. Intel plans to deploy this neuromorphic research prototype to its research collaborators as part of a new family of large-scale neuromorphic systems.
Overall, Hala Point represents a major breakthrough in the field of AI and could pave the way for even more advanced solutions in the future.