AMD FSR 2.0 performs worse than AMD FSR 1.0

The debut of AMD's FSR 2.0 technology sparked a lot of excitement, and there were lots of reasons for it. The Sunnyvale behemoth stated that this latest version of its well-known rescaling algorithm addresses a critical drawback of FSR 1.0, the inability to employ temporary pieces.

With FSR 1.0, the rescaling was created using just spatial elements, which reduced the burden that said technique required and allowed for a significant performance gain even in less capable graphic systems but had a significant influence on image quality in the end.

The performance mode results in such a significant loss in sharpness and image quality that it is simply incomprehensible.

The FSR 2.0 technology, as expected, works with temporal and spatial aspects, which means it uses information from prior frames to provide more exact and higher quality rescaling.

This translates to a much higher result than what we could obtain with FSR 1.0 technology; however, it is clear that it has not been a "free" improvement, and in the end, it represents an additional workload, which makes FSR 2.0 offer less performance improvement than FSR 1.0, regardless of the graphics card used.

That is why there is such a performance reduction compared to FSR 1.0, even on current-generation graphics cards such as the powerful Radeon RX 6800 XT and GeForce RTX 3080. We don't want to get into technical details since We want all of our readers to be able to understand it without difficulty. Consider what happens when you use a rescaling algorithm like FSR 2.0:

• Reduced GPU workload as a result of rendering resolution clipping. The total amount of pixels might range from 67 percent of the original resolution to only 33 percent depending on the setting selected.

• Although the effort is lessened in that sense, the method introduces a new task that consumes rendering time on each frame. If you simply have to deal with spatial aspects, the procedure will be completed faster than if you have to deal with temporal elements as well.

• Because it does not use specialized hardware to speed the workload that the algorithm represents, the impact on performance is greater than, for example, with second-generation DLSS, though the difference is minimal, at least in Deathloop.

AMD has done well and made some progress, but NVIDIA's second-generation DLSS technology remains the main challenger, both in terms of image quality and performance, at least until we see what Intel XeSS is capable of. The whole HardwareUnboxed comparison may be found in the linked video.