Future Technology

Future of Deep Learning Accelerates With 15-Billion Transistor Chip

By on
Future of Deep Learning Accelerates With 15-Billion Transistor Chip
PHOTOGRAPH: GDJ/Pixabay |

A chip with a whopping 15 billion transistors is unveiled by Nvidia. Designed towards deep learning artificial intelligence technology, exploring this subfield of machine learning is significantly accelerated with the latest tech.

During the GPUTech conference held in San Jose, California, Nvidia’s CEO Jen-Hsun Huang unveiled the Tesla P100, a 15-billion transistors chip. Huang also reveals they have decided to go “all-in on AI.”

Focal Point on Deep Machine Learning

Nvidia has been known as a company that develops chips to be used in gaming rigs and workstations. But they proved to have evolved into a company that brings chips for amplifying human intelligence for deep structured learning which has become their current fastest growing business.

While it was risky, the company went all out in their transition with Tesla P100 which includes establishments for new architectures and processes. And the road to the creation of chips for deep machine learning is only starting.

It is worth noting that Nvidia previously developed two chips called the Tesla M4 and Tesla M40. The two chips immediately got sold out. They were then prompted for a volume production of the Tesla P100.

Huang says it is the  “largest FinFET chip” ever created. FinFET is known as the Fin Field Effect Transistor which is a type of non-planar transistor used in modern processors. Because of its design, FinFET allows more computing power packed into small spaces.

Huang also reveals their strategy which is to “accelerate deep learning everywhere.” And with increasingly being known as “the AI computing company,” their goal would no longer be that far off from reality.

Future of AI

Hierarchical learning is undoubtedly quickly advancing. Even Jurgen Schmidhuber, known as the father of deep structured learning, predicts in few years’ time, artificial intelligence would be capable of functioning independently.

Since it is a utilization of large neural networks, deep learning has been highly important in the development of robots and AI. And with many revolutionary technological advancements coming our way, augmenting human and artificial intelligence would easily be done in various ways soon enough.

Comments

About the author

To Top