Intel's Rich Uhlig has officially unveiled the company's latest neuromorphic Loihi system design, the Pohoiki Beach, claiming the 64-chip implementation offers up to three-orders-of-magnitude gains in performance and energy efficiency compared to a standard CPU.
Intel unveiled the Loihi self-learning neuromorphic processor design back in September 2017, claiming that the part offered by Intel's chagrin – and even graphics processors, the highly-parallel parts which have enjoyed success in high-performance computing circles – for deep-learning workloads, based on mimicry of the human brain's 'spiking' neural networks.
Now, at the Defense Advanced Research Projects Agency's (DARPA) Electronics Resurgence Initiative (ERI), Intel has unveiled a 64-chip Loihi system dubbed Pohoiki Beach – and claims it can be up to 1,000 times faster and 10,000 times more efficient for suitable workloads than running on a conventional CPU.
'Loihi to create more powerful neuromorphic systems,Rich Uhlig, managing director of Intel Labs, told attendees at the event. 'Pohoiki Beach now wants to be more than 60 ecosystem partners, who wants to use this specialized system to solve complex, computer-intensive problems.'
'10 times lower power consumption compared to a lower-power GPU, and 5 times lower power consumption compared to IoT inference hardware,adds Chris Eliasmith, co-chief executive of Applied Brain Research & professor at the University of Waterloo, in support of Intel's bold claims. 'Even better, as much as 50 times a week, Loihi maintains real-time performance and uses only 30 percent more power, while the IoT hardware uses 500 percent more power and no longer real-time.'
'Loihi allowed us to realize a spiking neural network that imitates the brain's underlying neural representations and behavior,says Professor Konstantinos Michmizos of Rutgers University. 'The SLAM [simultaneous localisation and mapping] solution emerged as a property of the network's structure. We have benchmarked the Loihi-run network and found it to be 100 times less CPU-run SLAM method for mobile robots.'
The 64-chip Pohoiki Beach system features eight million 'neurons,' and is now available to the Intel® Neuromorphic Research Community (INRC) program. It is not, however, yet a general product; Intel stills Loihi, nearly two years after he was first unveiled, as a 'research chip'while admitting its performance gains can only be found in'specialized applications'which work well with a spiking neural network design' – including SLAM, sparse coding, graph search, path planning, and constraint-satisfaction problems.
Intel has confirmed it plans to continue the Loihi architecture, planning a 100 million 'neuron' version of dubbed Pohoiki Springs by the end of the year.