Intel Corporation (NASDAQ:INTC) in May conducted its inaugural Artificial Intelligence (AI) Developers Conference that saw it showcase its technologies and leadership. The acquisition of Nervana for the training workloads, Altera for reprogrammable FPGA acceleration hardware and Movidius and MobileEye for real-time processing have significantly strengthened the company’s portfolio of AI technologies.

Understanding the Neural Networks and the training

It is possible for one to use the Neural Networks without understanding how the training works. The training process has been automated by most of the modern machine learning libraries. However, it is important to understand the process since that is the way to gain valuable insights on the working of the neural nets, the application and the reconfiguration.

The update on the Nervana Neural Network Processor (NNP) roadmap was the biggest announcement in the event. Before its purchase by Intel, projections had been made that Nervana was going to provide a fabric-enabled NNP accelerator that would beat an unspecified GPU by a purported 10X.

Recent developments

The first-generation chip is currently being sampled by the company to the major AI customers. There are plans underway to incorporate their enhancement and input requests in the first-generation NNP part and that might take place towards the end f 2019.


Start your workday the right way with the news that matters most.

Your information is 100% secure with us and will never be shared Disclaimer & Privacy Policy

Reports indicate that Chip X has been overstating its performance. Chip X is fundamentally the NVIDIA Corporation (NASDAQ:NVDA) Volta GPU.

It was sometime back that Facebook, Inc (NASDAQ:FB) forwarded quite interesting data pointing out to the fact that it had been using (Xeon) CPUs for all inference work and select training jobs as well. Aside from that, it had been using Recurrent Neural Networks (RNN) for speech and language translation whereas for the training Convolutional Neural Networks (CNNs) it used the GPUs.

They are such tasks that compel Intel to employ Nervana. The main reason as to why Facebook employs CPUs is basically because it already has got quite a significant number of them.

Intel’s spokesperson has outlined that a large number if enterprises especially at night have surplus CPU. He outlined that it was a great move that they had eased the software burden by resorting to the use of the various existing resources for most if the learning workloads.