AMD aims to catch up to Nvidia's DLSS with "neural supersampling"

AMD develops neural supersampling to compete with Nvidia's DLSS.

: AMD is working on a neural supersampling technique to match Nvidia's DLSS in real-time path tracing. This innovation combines upscaling and denoising into a single AI model. It is aimed at improving image quality on RDNA GPUs but lacks a release date and compatibility details.

AMD is developing a new neural supersampling technique to rival Nvidia’s DLSS technology. This approach seeks to combine upscaling and denoising processes into a single AI model capable of enhancing real-time path tracing performance on RDNA GPU architectures.

The technique aims to generate high-quality, denoised, and supersampled images at resolutions higher than the rendered resolution. By focusing on Monte Carlo denoising, AMD hopes to achieve improvements in handling demanding ray-traced scenes.

Details about the release or compatibility across different RDNA generations remain unclear. However, this technology may be introduced as part of a future FSR 4 release, potentially enhancing AMD's competitive standing against Nvidia.

Neural SupersamplingAI modelgaming