Intel has announced the development of Hala Point, which is being touted as the world's largest neuromorphic system. It is aimed at the development of sustainable artificial intelligence, Intel said.
Hala Point's main goal is to support research in the field of brain-inspired artificial intelligence and to solve the problems of efficiency and sustainability common in modern AI.
Hala Point is a significant advance over Intel's first-generation Pohoiki Springs large-scale research system. It has improvements that provide more than 10 times more neuron capacity and up to 12 times higher performance.
The cost of computing modern artificial intelligence models is growing at an unacceptable rate. The industry needs fundamentally new approaches capable of scaling. For this reason, we developed Hala Point, which combines the power of deep learning with the new capabilities of brain-inspired learning and optimization. Mike Davis, Director of Neuromorphic Computing, Intel Labs
The Hala Point system can support up to 20 quadrillion operations per second, and its efficiency exceeds 15 trillion 8-bit operations per second.
Intel promises to enable real-time continuous learning in AI applications across a variety of areas, including solving scientific and engineering problems, logistics, urban infrastructure management and large language models (LLM).