Nvidia Unveils New Chips for Artificial Intelligence and Virtual Reality
Nvidia, the company known for manufacturing state-of-the-art graphics cards, is going into artificial intelligence and virtual reality. According to a report in Fortune, the American technology company announced that it is working on chips for artificial intelligence (i.e. learning machines) and virtual reality at the recently concluded GPU Technology Conference. The company's CEO Jen-Hsun Huang presented during the conference that they are making "new products" for companies who are at the forefront of these emerging technologies.
With the graphics chip company focusing on the graphics processing, it hopes to help virtual technology companies with Iray VR, which is a new video rendering technology that "will improve the way 360 degree video displays on virtual reality headsets." This new technology will be available in June, according to Huang. With this technology, more photo-realistic designs can be easily done which can be very useful in building interiors and car design, the report cited.
In terms of artificial intelligence, Huang explained why they have decided to go into "deep learning." He said in his speech, as reported in Forbes, "Finally computers powered by deep learning have the opportunity to do things that we once only dreamed about doing. It's no longer just a field or an app. Deep learning is bigger than that. That's why our company has gone all in on deep learning. I've been talking about it for five years. Every year, we do more and more in this area. This utterly changes computation."
Get Our Latest News for FREE
Nvidia introduced a new chip called the Tesla P100 GPU which performs "deep learning neural network tasks" 12 times faster than what the company previously made. According to The Verge, this new chip is a product of $2 billion worth in research and development, equipped with 150 billion transistors on a single chip. While the chip can also do other high-performance computing tasks, Huang emphasized that "it's really good at machine learning."
