FSR 2.0 will be temporary scaling, departing from the spatial methods used in the first iteration FSR. It is said that this technology provides better image quality than native permission. It will also include smoothing, which will significantly improve the quality compared to FSR 1.0.

Unlike NVIDIA DLSS and Intel Xess, AMD FSR 2.0 will not rely on machine learning algorithms, which means that the tensor kernel is not required or the XMX acceleration. Technology will work with some competitors video cards

AMD provided parallel screenshots from the game called Deathloop. It shows performance and quality modes compared to Native 4K and FSR 1.0.:

** First view on AMD FidelityFx Super Resolution 2.0 in Deathlooop

The company confirmed that FSR 2.0 will follow FSR 1.0 from the point of view of the following open source solutions. The technology code will be available on the Gpuopen website so that developers can integrate it into existing and new games.

AMD will share more detailed information about FSR 2.0 while performing at GDC 2022 March 23. At sessions called "Scaling the image of the next generation for games", most likely, equipment requirements will be described in detail.