Intel Unveils Loihi 2: Its Second-Generation Neuromorphic Artificial Intelligence Research Chip

In 2018, Intel launched its first-generation Loihi chip, which was created using a proven 14nm process node and integrated 128k Neurons at the same time. Now, Intel unveils the second version of its Loihi neuromorphic chip, “Loihi 2,” an artificial intelligence processor that more accurately reflects human brain processes than other AI technology. The new chip has a grain size of 31 mm² and can package up to 2.3 billion transistors; it’s an incredible advancement for future technology. This new chip also uses pre-production Intel 4 process and grows to 1 million neurons.

The Intel Loihi 2 processor is better because it incorporates some of the most complex and advanced chip processes that rely heavily on Extreme Ultraviolet (EUV) lithography technology. The new tech further integrates up to 1 million neurons for faster processing speeds 10 times those found in first-generation Loihi units.

The Loihi 2 is a chip that can be programmed to have up to 4,096 states and organize them into 128 types of neural networks connected by a network-on-chip(NoC) Neuromorphic Cores. The processor has 192KB SRAM with 25MB total memory for further flexibility in programming it all together, which will let you use less power doing your tasks. The Loihi 2 chip is 15 times more powerful than the first-generation version of it and has made significant leaps forward in terms of processing power.

Loihi 2 is the next-generation chip for a revolutionary new system that will change everything we know about computing. It has better compression rates and faster data exchanges, with an off-chip parallel interface to extend neural networks into physical systems. Loihi 2 chips will support Ethernet interfaces, glueless integration with a broader range of event-based vision sensors, and large meshed networks.


Other Source:

The post Intel Unveils Loihi 2: Its Second-Generation Neuromorphic Artificial Intelligence Research Chip appeared first on MarkTechPost.

This post was originally published on this site