Nvidia reveals DLSS 2.0 – More detail and no downsides

The promise of Nvidia’s DLSS (Deep Learning Super Sampling) technology was to improve it with time through the use of deep learning algorithms; and with DLSS 2.0, Nvidia has seeming mitigated the major downsides of its AI upscaling technology. 

When Nvidia first announced DLSS, we said that it had the potential to be more revolutionary than raytracing, at least in the short term, but lower than expected levels of visual quality sent that dream crashing down to earth. To judge the quality of Nvidia’s initial DLSS implementation, you need only look at our PC performance analysis for Metro Exodus.   

DLSS 2.0 – A fresh start

With DLSS 2.0, Nvidia promises significant performance increases and “beautiful, crisp game images”. delivering higher levels of graphical fidelity than before and increased hardware efficiency across Nvidia’s entire RTX graphics card lineup. 

Another improvement is that DLSS 2.0 uses the same network for all games, which will allow game developers to integrate DLSS more easily, and allow Nvidia to update its DLSS network without breaking game support. This should help DLSS’ quality improve more over time. 

Nvidia’s latest AI network can also execute much faster than before, allowing the algorithm to update over 2x faster than before. This has allowed Nvidia to remove the resolution limitations of older DLSS implementations, which is great news for all Geforce users. 

DLSS 2.0 Games

DLSS 2.0 is already available in Wolfenstein: Youngblood and Deliver Us The Moon, though the feature will come to Mechwarrior 5: Mercenaries and Control later this week. 

More games will gain DLSS 2.0 support over time.