AMD has just announced a new version of its Radeon Software Adrenalin Edition, 21.4.1, and with it a heap of new features, such as AMD Link for PC (opens in new tab) and Ryzen CPU monitoring (opens in new tab). However, one notable omission from today's release is FidelityFX Super Resolution, and according to AMD there's not much more it wishes to say on the subject just yet.
With Nvidia DLSS proving a mighty performance boon for all GeForce cards in supported games, it's no surprise that Radeon fans are itching for a red team alternative. That will come in FidelityFX Super Resolution, so says AMD, a new, open source feature coming to a game near you… sometime in 2021.
The last we heard of the feature (opens in new tab) was Scott Herkelman's chat to PC World (opens in new tab) last month, where the AMD exec confirmed it would not require machine learning to work. He does, however, say that it still needs more time in the oven.
The Radeon software team doesn't have much more to add to that at this time. When asked about the feature's arrival during a pre-briefing on the new Radeon Software package last week, AMD senior product manager, Adrian Castelo, told us "when the time's right, and the time's appropriate" then we'll hear more about the feature and its release.
Furthermore, Castelo explains that development is still ongoing for an upscaling feature, and that it targets a broad range of games.
At least it's comforting to know that FidelityFX Super Resolution hasn't been canned. But how could it? Nvidia is sweeping performance in DLSS compatible games, and in games as demanding as say, Cyberpunk 2077, it provides the extra oomph necessary for when the GPU silicon can't muster up enough.
An important feature for AMD to get down just right, no doubt. It may be on to something with its open approach at the very least. A unilateral standard could help win AMD some initial favour where Nvidia's proprietary feature falters.
For the time being, Radeon GPU owners can utilise FidelityFX Contrast Adaptive Sharpening (CAS), which offers a way to clean up an image rendered at lower than native resolution.