What Does NVIDIA’s Ray Tracing News Mean for the CAD Market?

Herrera on Hardware: At NVIDIA’s GPU Technology Conference in March, the graphics processing unit (GPU) developer announced “real-time ray tracing” — but is it? And what does this development portend?

Real-time ray tracing of 3D graphics for professional visualization markets has finally arrived … sort of. Long coveted for its high-fidelity images — particularly by professionals in media, as well as architecture, product design, and styling — ray tracing has come a long way. It has evolved from an intriguing but thoroughly impractical way to visualize 3D, to one that could be accomplished reasonably well on high-performance servers, to one that we can now start considering for real-time use on our CAD workstations.

But “start” is the operative word there. Because while NVIDIA managed to impress at its 2018 GPU Technology Conference (GTC) last month with a product promising real-time ray tracing for your workstation, the tried-and-true raster-based graphics technology that’s long supported interactive CAD workflows won’t be going away anytime soon.

Markets Other Than Graphics Are Now Sharing NVIDIA’s Attention

Indispensable in every CAD professional’s IT toolbox, graphics processing units (GPUs) have long been measured strictly on their ability to render 3D graphics with both high fidelity and real-time, interactive performance. It’s the main — or only — reason any of us running AutoCAD, SOLIDWORKS, or similar modeling and simulation apps might concern ourselves with which GPU is in our next workstation. But while we may have always relied on GPUs for the same thing, GPU vendors such as NVIDIA and AMD have found new reasons for independent software vendors (ISVs) and users alike to covet high-performance GPUs in their applications.

The emergence of workstations in the 1980s and PC graphics in the ‘90s spurred countless original equipment manufacturers (OEMs) and startups alike to bet their fortunes on 3D graphics. NVIDIA was one of the few to survive, and arguably the only one to thrive. During the ‘00s, GPUs underwent a fundamental change in architecture, moving from fixed hardware implementing 3D graphics-specific pipelines, to programmable and far more general-purpose architectures. In retrospect, that shift was truly a seminal moment in computing history, as it opened the door to a wealth of new applications for the GPU’s superior highly parallel, floating point performance, well beyond traditional raster-based 3D graphics. That shift proved an inflection point for what NVIDIA has become: a company developing GPUs for a wide range of otherwise performance-throttled applications.

While the graphics market never took a back seat in regards to volume and revenue — not even close — most would argue it has when it comes to the company’s marketing and messaging. Over time, NVIDIA’s outward attentions shifted noticeably away from graphics to focus on today’s priorities: artificial intelligence (AI) and machine learning. That focus has gotten so pronounced, with the past few years’ GTC shows putting so much emphasis on “deep learning,” that conference attendees could easily forget “graphics” was what the G in GPU stood for. But the message and vibe at GTC ‘18 was a little different, bringing the company back full circle, leveraging all NVIDIA has learned on the journey to those new markets to deliver on the ultimate achievement in computer graphics: real-time, ray-traced rendering.

Traditional Markets Benefit from NVIDIA’s Journey into HPC and AI

AI and machine learning are gradually finding compelling applications in virtually every corner of the computing landscape, including the market those “graphics” processing units have always served: CAD. Consider the AI algorithms used to generate the internal composition of objects to be 3D printed, taking into account the printing materials to create the optimal structure to balance weight, materials, and strength. And deeper in CAD workflows, AI is already taking on some of the burden of designing the form and function of the object itself — something Autodesk is pursuing with its Dreamcatcher initiative. Soon, we’ll likely all see AI not as some stand-alone application, but as a tool naturally integrated into our traditional workflow, a right-hand assistant helping us accomplish much more than we’d do on our own.


NVIDIA CEO Jensen Huang introduces the Quadro GV100 GPU and RTX technology at the company’s GPU Technology Conference (GTC) in March 2018.

1 2 3