NVIDIA Launches Quadro RTX GPUs for Mobile Workstations26 May, 2019 By: Cadalyst Staff
New Turing-based graphics processing units are intended to boost mobile machines’ performance with real-time ray tracing, virtual reality, and other demanding visual tasks.
NVIDIA has announced a new family of Quadro graphics processing units (GPUs) for mobile workstations, including three Turing architecture–based Quadro RTX models. Turing, which was unveiled last year with the launch of Quadro RTX GPUs for desktop workstations, is now being brought to the mobile form factor to “address some of those limitations and constraints that have plagued our mobile systems in the past,” said Allen Bourgoyne, senior product marketing manager at NVIDIA, during a media briefing on the launch.
With the expansion to mobile, NVIDIA is “freeing professionals to do full-featured, full-functionality workflows wherever they may be,” Bourgoyne continued. Those workflows include a variety of demanding applications that are heavy on the visuals, including 3D modeling, 8K video editing, and more. According to NVIDIA, Quadro RTX–powered mobile workstations are capable of handling “the latest advancements” in real-time photorealistic rendering, artificial intelligence (AI) acceleration, and virtual reality. Workstation vendors offering the new GPUs in their mobile systems include Dell, HP, Lenovo, and MSI.
NVIDIA Quadro RTX GPUs are available in 15- and 17-inch mobile workstations. Image courtesy of NVIDIA.
At the top end of the new GPU family are the Quadro RTX 5000, 4000, and 3000, which feature “very fast” GDDR6 memory, Bourgoyne said, and are designed to enable desktop-level performance and capabilities in a mobile form factor. The Quadro T2000, T1000, P620, and P520 round out the new lineup with smaller amounts of memory and fewer cores. (Every member of the Quadro family has some number of CUDA processing cores, but only the Quadro RTX 5000, 4000, and 3000 have NVIDIA RT cores [to accelerate ray tracing] and Tensor cores [for AI inferencing] as well.) Together, these seven models replace the previous generation of Quadro mobile GPUs, which comprised the P5200, P4200, P3200, P2000, P1000, P600, and P500.
RTX technology enables photorealistic virtual reality (VR), the application of AI to graphics problems such as denoising, and real-time ray tracing. “[With RTX,] all the reflections are accurate, all the shadows are accurate, the lighting,” explained Bourgoyne. These capabilities are useful in fields ranging from automotive design to architecture — anywhere that interactive visualization and realistic shadows, reflections, and lighting support design and customer communication processes. “[In] segments like architecture, customers can make very … rapid design decisions” on the spot, Bourgoyne explained, without time-consuming exchanges of e-mailed renderings and the like. “The goal is to minimize any rework that has to happen.”
Cobus Bothma, director of applied research at the Kohn Pedersen Fox architecture firm, confirmed that idea in an NVIDIA release: “The real-time ray-tracing and AI capabilities inherent in Quadro RTX GPUs has transformed the way we as architects visualize and process the data for our designs,” he said. “The ability to interact with physically accurate models in real time means we’re making smarter design decisions faster. Now, with the latest RTX mobile workstations, we are excited about the possibilities out in the field, further freeing us to focus on creating our best work while saving clients time and money.”
Bourgoyne also pointed out the value of RTX in simulation workflows: With ANSYS Discovery Live, for example, users typically hand off jobs to a server group, then wait for them to be completed. By using GPUs to accelerate the rendering process instead, “it really allows designers to take advantage of simulation as part of their product development process.” Users can verify that whatever they’re creating can survive stresses “so when it comes time to simulate the final product, you have a high degree of quality,” Bourgoyne continued. And because RTX allows them to work on mobile machines, engineers can run simulations locally, wherever they are.
RTX-accelerated software applications include Unreal Engine, Unity, and Autodesk Arnold; the list will soon encompass SOLIDWORKS, V-Ray, Siemens NX, and more. NVIDIA will continue to work with “all the major CAD design vendors,” said Bourgoyne.