NVIDIA Takes Computer Graphics to the Next Level

NVIDIA unveils its Turing Architecture, finally allowing for real-time ray tracing.

Computer graphics have come a long way since Ben Laposky created the first machine-generated graphic images, an Oscilloscope, in the 1950s. While innovation has been awe-inspiring since, there was always one seemingly unattainable level: real-time ray tracing.

That is about to change. NVIDIA recently unveiled a fundamental game changer for computer graphics with its Turing Architecture. It fuses ray tracing, artificial intelligence, simulation and rasterization, making real-time ray tracing a reality.

“Turing is NVIDIA’s most important innovation in computer graphics in more than a decade,” said Jensen Huang, founder and CEO of NVIDIA, at the annual SIGGRAPH conference. “Hybrid rendering will change the industry, opening up amazing possibilities that enhance our lives with more beautiful designs, richer entertainment and more interactive experiences. The arrival of real-time ray tracing is the Holy Grail of our industry.”

Turing Architecture chips have three types of cores: CUDA, ray tracing and Tensor AI.

Turing Architecture chips have three types of cores: CUDA, ray tracing and Tensor AI.

Simply put, the ray-tracing technique renders images by tracing pixel light paths and creating the effects as they encounter virtual objects. The results are photorealistic renderings used in a variety of platforms, from video games to movies. For film, this method could take up to hours to complete.

As technology has become more interactive, and our expectations have grown, developing real-time imaging became even more important. Turing features RT Cores that drastically speed up light and sound computations. This causes the speed of ray tracing to increase by up to 25 times that of current methods and the speed of final-frame rendering to increase to more than 30 times that of CPU nodes.

The addition of artificial intelligence (AI) features—such as deep learning anti-aliasing, denoising, resolution scaling and video retiming—pave the way for new capabilities and quicker computations. It also supports augmented reality (AR) and virtual reality (VR), making it possible for immediate rendering.

The potential uses for this new technology go beyond traditional graphic uses. For AR and VR purposes, the ability to change parameters could allow a car buyer to choose the desired options and features of a vehicle and then take it for a virtual spin with photorealistic views of the vehicle’s interior and exterior, as well as the road they are “traveling” on.

Architectural, interior and landscaping designers also could benefit from real-time ray-tracing possibilities. Turing’s ability to match reality allows for realistic lighting, which could allow for more realistic views.