4 Bold Statements From Nvidia CEO Jensen Huang On The Future Of Computing

Connecting The Dots

While Nvidia is perhaps best known for the GPU technology that makes video game graphics so intense, that technology is also key to adding performance to enterprise workloads and data centers, as well as a major component in some of the world's fastest computers.

During the Nvidia GPU Technology Conference, Nvidia CEO Jensen Huang spotlighted those connections when he tied ray tracing to high-performance computing and to data center and enterprise workloads.

While Nvidia’s yet-to-close acquisition of Mellanox was not a part of the conversation, the tie was evident in his focus on the need for increased performance in the data center. Here’s a look at how Huang sees the world.

Data Center Workloads For GPUs Will Expand

“We announced that we have server architecture with graphics for the data center, whether it’s graphics rendering or graphics workstations. We've created this thing called Omniverse [3-D collaboration platform] that makes it possible for us to achieve ray tracing on a data center scale for collaboration among multiple teams, multiple tools, coming together. We can create a wonderful render for them.”

High-Performance Computing To Move To Enterprise Intelligent Workloads

“In terms of high-performance computing, you know that supercomputing has always been an area that we are very good at. This year, of the high-performance computing market top 500, more than 50 percent of the new computational capability came from Nvidia's P100. And so supercomputing continues to be an important market for us. However, high-performance computing is going to grow out beyond supercomputing data centers, beyond cloud service providers, and into enterprise intelligent workloads.”

The Rise Of Data Science

“[Data science provides] data-driven methods for problems that are too gigantic and too complicated. … We finally have a convergence of three things that make it all possible. A large amount of data has been made available because of digital technologies, and since there's more and more people clicking on the internet and websites and services, we're collecting a lot of information. The second thing is state-of-the-art machine- learning approaches. And then the third thing is more is more computation [including our GPUs]. These three things coming together have ignited what is now known as data science. It is the most popular course in school, the most over-subscribed course in school. It is likely to be a required class for every single field from mathematics to computer science to oceanography to biology and chemistry. … We all have so much data. And for the very first time, we have a methodology to predict the future with the data that we have today. Data science is the new high-performance computing workload.”

Ray Tracing Is Here

“When a movie is made and you see the images, it's rendered in software. It's very, very slow. The point of [our RTX ray tracing technology] is not to make ray tracing possible. The point of RTX is to make ray tracing fast. And so, when every game engine adopts ray tracing, and we have ray tracing everywhere, I think it’s fantastic news. I'm super excited about that. And the more people using ray tracing, the more opportunity we have to make it go faster. DXR [DirectX ray tracing] is about ray tracing. The next version of Unreal Engine 4.22 is about ray tracing. Unity has just announced they have ray tracing. So, ray tracing, ray tracing, ray tracing. It's all about ray tracing, all the time. The next generation is thinking about ray tracing. Ray tracing is here. And if you're not completely clear about it, ray tracing is pretty close to my heart. I really love ray tracing.”