Nvidia: Native resolution gaming is obsolete, the future belongs to DLSS

Recently, Digital Foundry released a roundtable discussion video on YouTube featuring luminaries like Jakub Knapik from CD Projekt Red and Jacob Freeman and Bryan Catanzaro from Nvidia. Together, they delved into the significance of Nvidia’s latest DLSS 3.5 technology in “Cyberpunk 2077” and the advantages of AI-generated frame rate enhancements and rendering in gaming.

During the discourse, when queried about DLSS and whether Nvidia plans to prioritize native resolution performance in their GPUs, Bryan Catanzaro, Vice President of Deep Learning Research at Nvidia, posited that native resolution is no longer the optimal solution for achieving maximal graphical fidelity in games. The gaming graphics industry is ostensibly gravitating towards a more AI-centric image reconstruction and AI-based graphical rendering paradigm.

This perspective is notably contentious. Elaborating further, Catanzaro contended that since “Moore’s Law” has met its terminus, brute-forcing graphical fidelity enhancements is no longer the ideal. There’s an exigency for more sophisticated techniques, such as DLSS, to augment visual fidelity and circumvent the present-day challenges of marginal graphics hardware performance increments.

Citing “Cyberpunk 2077” as a case in point, both Valadas and Jakub Knapik of CD Projekt Red attested that without DLSS, especially the upscaling and frame generation technologies, implementing full path tracing in the game would have been insurmountable. They further highlighted the astuteness with which DLSS deftly augments modern GPU’s performance bandwidth, enabling “Cyberpunk 2077” to run real-time, hyper-realistic light simulations. Compared to traditional gaming rendering methods, DLSS-rendered images are markedly more realistic and detailed. When probed about the prospects of machine learning in gaming graphics, Catanzaro went so far as to assert that DLSS/AI would eventually supersede conventional rendering techniques altogether.

It’s evident that Nvidia harbors immense confidence in its DLSS/AI technology. Yet, Catanzaro’s perspective is bound to ignite debates among the gaming community, given that DLSS isn’t universally embraced by all gamers. However, with the so-called “demise” of Moore’s Law, AI might very well be the solitary force propelling 3D graphics in the foreseeable future. Viewing from another angle, Catanzaro’s statements seemingly corroborate that Nvidia will prioritize GPU’s AI capabilities over traditional raster performance in the future.