Sometime later this year, Intel plans to release what it’s calling “the industry’s first silicon for neural network processing.” Designed specifically for machine learning and artificial intelligence-focused computing, the Intel Nervana Neural Network Processor (NNP) will feature a new type of architecture compared to standard silicon chips. It’s being built on the technology Intel acquired last year through its $408 million purchase of Nervana, a California-based startup.
Intel said the purpose-built NNP will support the development of “entirely new classes of applications” that can draw intelligent insights from large volumes of data. Those capabilities could enable better healthcare diagnoses, improve personalization on social media, speed up advances in autonomous vehicles, and provide more accurate weather and hurricane forecasts, according to the company.
Changing the Silicon Portfolio
Being developed in collaboration with Facebook, Intel’s Neural Network Processor was unveiled yesterday at the Wall Street Journal’s D.Live global technology conference in Laguna Beach, Calif. Intel CEO Brian Krzanich, who spoke at the conference, wrote afterward in a commentary that the chip will help drive progress in cognitive and AI technologies that will be transformative to the industry and the world.
“We have multiple generations of Intel Nervana NNP products in the pipeline that will deliver higher performance and enable new levels of scalability for AI models,” Krzanich noted. “This puts us on track to exceed the goal we set last year of achieving 100 times greater AI performance by 2020.”
In a separate announcement yesterday, former Nervana CEO and co-founder Naveen Rao added that the new NNP has been in development for more than three years, and features a design aimed at supporting more advanced machine learning and deep learning workloads.
“We’ve been listening to our customers and applying changes to Intel’s silicon portfolio to deliver superior machine learning performance,” said Rao, who is…