At the annual Autodesk University Conference in Las Vegas last week, NVIDIA released the Quadro RTX 4000 graphics card. This is the first mid-range professional GPU released, powered by the Turing architecture and running the NVIDIA RTX platform.
NVIDIA’s graphics cards have been so popular in not only facilitating buffer-free game playing on home machines, like for Battlefield V (above), but for the creation of the visuals at large game developers like DICE, Blizzard and Microsoft.
There really was a seismic shift in computer graphics with the launch of Turing in August. The Turing technology features new RT Cores to accelerate ray tracing and next-gen Tensor Cores for AI inferencing. This makes real-time ray tracing possible. The manufacturing, architecture, engineering and media creation industries witnessed the greatest leap since the invention of the CUDA GPU in 2006. I spent many hours digging into CUDA with explanations in articles published on my old home CGSociety.org. I still have books and articles sitting in my study here, waiting to be re-read.
The Quadro RTX 4000 features a single-slot design that fits in a variety of workstation chassis. With 8GB of GDDR6 graphics memory, there is 40% more memory bandwidth than the Quadro P400, which is the previous release. There’s also 36 RT Cores designed to bring real-time ray tracing of characters, and environments. Additional to those uses for visual graphics, the RTX 4000 Quadro GPU card has 288 Turing Tensor Cores which can deliver 57 teraflops of neural network or deep learning performance. This is critical to AI-enhanced operations in many industries. There’s also hardware support for VirtualLink. This is the new open industry standard for the display and bandwidth demands for VR software and headset delivery.
I will be drilling down into the use of NVIDIA technology in both VFX and general science, over the next few months. Stand by and buckle up. This is going to be truly amazing.