On the eve of the Consumer Electronics Show in Las Vegas this week, Intel staged a launch event that covered a new version of its Nervana AI processor and a demonstration of the next-generation Xeon 10nm chip.
The Nervana Neural Network Processor for Inference – NNP-I – is, according to Intel, a new class of chip for accelerating AI inference for companies with high workload demands and is expected to go into production in the second half of this year. Intel said Facebook is a NNP-I development partner.
“This is a really big deal for us,” said Navin Shenoy, Intel EVP of the Data Center Group, “it expands our position in AI above and beyond what we’ve done with Xeon into a new domain.”
Additionally, Intel said it expects to have a neural network processor for training, code-named “Spring Crest,” available later this year, following the schedule announced in August.
Shenoy also provided the first public demonstration of Intel’s 10nm Ice Lake server silicon, running a machine learning simulation. “Compatible with the upcoming 14nm Cooper Lake, Ice Lake processors targeting server are expected to deliver performance improvements, new hardware-enhanced security features and more, with shipments targeted for 2020,” noted the company.
Lastly, Intel announced that at the tail end of 2018 it started shipping its next-generation Intel Xeon processors, code-named “Cascade Lake,” the 14-nm platform that introduces support for Intel Optane DC persistent memory and Intel DL Boost, designed to accelerate AI deep learning inference. Cascade Lake is set for broad availability in the first half of this year.