The American multinational corporation and technology company Intel Corporation known for manufacturing semiconductor chip revealed its first-ever Artificial Intelligence (AI) chip that is meant for machine learning workloads which they named as “Spring Hill”.
Spring Hill or the Nervana Neural Network Processor for Inference (NNP-I 1000) was developed at Intel’s lab in Haifa. The chip was a 10 nm chip that allows it to handle high workloads. According to Intel, Spring hill is the first AI product after the company invested in Israel AI startup’s Habana Labs and NeuroBlade.
Naveen Rao, General Manager of Intel’s Artificial intelligence products group said, “In order to reach a future situation of ‘AI everywhere’, we have to deal with huge amounts of data generated and make sure organizations are equipped with what they need to make effective use of the data and process them where they are collected.”
Spring hill is small enough that it can be set up in data centers via an M.2 storage device. “Most inference workloads are completed on the CPU even though accelerators like Nvidia’s T-Series offer higher performance,” said Patrick Moorhead, founder of Moor Insights and Strategy.
“When latency doesn’t matter as much and raw performance matters more, then accelerators are preferred. Intel’s Nervana NNP-I is intended to compete with discrete accelerators from Nvidia and even Xilinx FPGAs.” Moorhead added.
Intel said that the chip’s performance can be expanded at a wide power range and can achieve a high degree of programmability without compromising performance or power efficiency.
“Intel is pulling expertise in power and storage and looking for synergies in its processor suite,” said Mueller. “Since Springhill is deployed via an M2 device and port, which is something that was called a coprocessor a few decades ago, this effectively offloads the Xeon processor. But we’ll have to wait and see how well Springhill can compete with more specialized, typically GPU-based processor architectures,” he added.
The company has two more generations of the NNP-I to release yet. According to Intel, Facebook is already started using the said chip.