Intel’s 64-Chip Pohoiki Beach Neuromorphic System Models Human Brain To Boost CPU Performance
Just when you thought you might have finally been getting a grasp on all of Intel's lake-inspired codenames, such as Coffee Lake, Ice Lake, Rocket Lake, and the list goes on, here is another codename to add to your mental repository—Pohoiki Beach, which consists of an 8 million-neuron neuromorphic system comprising 64 brain-inspired Loihi research chips.
That is a lot to absorb. The takeaway is that this is a different kind of computing platform than a typical modern system. Loihi applies the principles found in biological brains to computer architectures, and according to Intel, it can process information up to 1,000 times faster and 10,000 times more efficiently than CPUs for specialized applications.
Now with the introduction of Pohoiki Beach, researchers can efficiently scale up novel neural-inspired algorithms, such as sparse coding, simultaneous localization and mapping (SLAM), and path planning, Intel says. Like the human brain, these algorithms can learn and adapt based on data inputs.
A Nahuku board, shown above, can house 8 to 32 Intel Loihi neuromprophic chips
"We are impressed with the early results demonstrated as we scale Loihi to create more powerful neuromorphic systems. Pohoiki Beach will now be available to more than 60 ecosystem partners, who will use this specialized system to solve complex, compute-intensive problems," says Rich Uhlig, managing director at Intel Labs.
Pohoiki Beach is not going to slip into a consumer roadmap, nor will brain-inspired processors, at least not anytime soon. However, Intel views neuromorphic computing as one way to keep Moore's Law relevant, albeit for specific applications.
"The Pohoiki Beach neuromorphic system demonstrates the benefits of a specialized architecture for emerging applications, including some of the computational problems hardest for the internet of things (IoT) and autonomous devices to support. By using this type of specialized system, as opposed to general-purpose computing technologies, we can expect to realize orders of magnitude gains in speed and efficiency for a range of real-world applications, from autonomous vehicles to smart homes to cybersecurity," Intel says.
A Nahuku board interfaced to an Intel Arria 10 FPGA development kit
While Pohoiki Beach may not be poised to infiltrate the home in a traditional PC sense, you may come across the tech at some point. For example, Intel says researchers are currently using Loihi systems for various projects, including adding adaptation capabilities to the AMPRO prosthetic leg, object racking using emerging event-based cameras, automating a foosball table with neuromorphic sensing and control, and so on.
This type of computing also provides benefits in power consumption and efficiency, which can be particularly useful in the IoT sector.
"With the Loihi chip we’ve been able to demonstrate 109 times lower power consumption running a real-time deep learning benchmark compared to a GPU, and 5 times lower power consumption compared to specialized IoT inference hardware," said Chris Eliasmith, co-CEO of Applied Brain Research and professor at University of Waterloo. "Even better, as we scale the network up by 50 times, Loihi maintains real-time performance results and uses only 30 percent more power, whereas the IoT hardware uses 500 percent more power and is no longer real-time."
Pohoiki Beach is an effort to accelerate neuromorphic computing and adoption. Sometime later this year, Intel will follow it up with an even larger Loihi system called Phoiki Springs.