With FSR 2.0, AMD wants to offer an alternative to Nvidia’s DLSS that also works on GPUs without machine learning units.
Enlarge
FSR 2.0 should be implemented quickly, especially in DLSS games.
© amd.com
AMD’s Fidelity FX Super Resolution (FSR) should offer an alternative to Nvidia’s image enhancer DLSS. In contrast to Nvidia technology, the temporal upscaler does not require hardware units for machine learning on the graphics card. In this way, AMD technology can also be used on older graphics cards, even on those of the competition. AMD has now announced that FSR 2.0 can be implemented in games that are already prepared for DLSS in just three days.
Nvidia graphics cards are also supported
Compared to FSR 1.0, FSR 2.0 will place higher demands on the graphics card. But even on older GPUs, the new technology should be able to achieve a higher frame rate. And FSR 2.0 should even work on the graphics cards of the competition, specifically AMD promises support down to the Pascal generation from Nvidia.
Little work for developers
FSR 2.0 offers Quality, Balanced, Performance, and Ultra-Performance presets. Depending on the setting, resolutions between 33 and 67 percent of the native resolution are used as a basis. The final version of FSR 2.0 will be available on an open source basis for all interested game makers in the next quarter. While the implementation in games with DLSS support should be particularly quick, games without decoupled render resolutions require several weeks of work. So players can hope for a quick implementation of FSR 2.0, at least for games with DLSS support. The interest of the players in a performance improvement apart from DLSS seems to be great.
AMD FSR and Nvidia DLSS in direct performance and quality comparison
