Keywords: Intel, Xeon chips

The world of artificial intelligence (AI) boom, faced with huge potential application chip business opportunities, chip makers into the taking, drives the increasingly fierce competition, the NVIDIA indicates that, by their GP - GPU computing initiative had great progress in the data center market, the current using the GPU chip processing deep learning (DL) neural network this kind of AI application has become a mainstream,

Venture Beat website reported that Intel (Intel) data center solutions team and corporate vice President, general manager of Jason Waxman post back, says Xeon chip is AI server markets around the world the absolute leadership of the product.

Waxman pointed out that the Xeon chips AI server market dominant in the whole world, and more than 97% of global support AI application server chip market share, but less than 3% of the AI server equipment is using the GPU, even if has always been the outside world has a lot of talking about the GPU to the value of the machine learning (ML). Equal to the claims of NVIDIA.

Waxman, said the industry needs to break through the scalable and open operation ability, in order to promote a wide range of innovative developer community. Intel in the developer BBS (IDF) on, also explain how to launch the AI industry leading platforms.

Intel has revealed that the symbol "Knights Mill" for the next generation of Xeon Phi processors will be published in 2017, and lock in the field of AI. Intel said Xeon Phi processor operation ability will be better than the GPU chip on the market at present, said Xeon Phi training speed is 2.3 times quicker GPU, NVIDIA have written against Intel, let both sides of the barge intensified fire from a distance.

After Intel to more than $350 million acquisition of Nervana Systems, so as to strengthen their own AI technology. Visible as the chip market increasingly competitive, operators are used to strengthen its own technical advantage routes will more and more.