Five more games are getting AMD FSR 2.0 support

[Tudor Cibean]

The big picture: AMD's FidelityFX Super Resolution 2.0 upscaling tech improves image quality and game performance compared to native rendering. Hopefully, the list of supported titles keeps growing as the feature benefits all gamers without DLSS-enabled GPUs, not just AMD graphics card owners.

Five additional games have been added to AMD's FSR 2.0 temporal upscaling lineup, bringing the total count to 19. The five games that have joined the upcoming FSR 2.0 titles are Abyss World, Hitman 3, Rescue Party: Live!, Super People, and The Callisto Protocol. As Hitman 3 got released over a year ago, it will receive FSR 2.0 support through an update, while other games will have the tech at launch.

FidelityFX Super Resolution 2.0 is an open-source alternative to Nvidia's DLSS, with the main advantage being that it doesn't require specialized Tensor cores to function. While DLSS exclusively runs on relatively-new RTX GPUs, FSR 2.0 supports far older graphics cards and iGPUs, including ones from Nvidia and Intel.

Deathloop, God of War, and Farming Simulator 22 are currently the only available games that implement FSR 2.0. Slides provided by AMD show the performance difference between the three quality presets in the latter two titles.

When we recently compared FSR 2.0, FSR 1.0, DLSS, and native rendering in Deathloop, we found that FSR 2.0 produces far better results than its predecessor, getting very close to DLSS in both image quality and performance.

You can also check out our coverage on how FSR 2.0 works and what differs from the previous version.

Permalink to story.

https://www.techspot.com/news/94911-five-additional-games-getting-amd-fsr-20-support.html

 

[darkzelda]

This is something every new game should have.

 

[Peter Farkas]

I hoped Elite Dangerous: Odyssey might get it in its latest release as a surprise.

 

[yRaz]

Can't wait for nVidia to begrudgingly adopt FSR in the same way that G-sync is now just freesync with a badge on the front that monitor companies pay for.

"we can code specifically for 1 type of GPU or we can just copy and paste this code that supports all GPUs"

 

[punctualturn]

Can't wait for nVidia to begrudgingly adopt FSR in the same way that G-sync is now just freesync with a badge on the front that monitor companies pay for.

"we can code specifically for 1 type of GPU or we can just copy and paste this code that supports all GPUs"

The only way that will ever happens is if developers start using only FSR and sadly, we know that those devs really love the sponsorship moneis they get from nvidia .

But lets say nvidia does start using FSR instead of dlss, you can bet good money they will poison the name as you mentioned what they did with gsync.

 

[envirovore]

The Digital Foundry video on FSR2 in GoW shows that while FSR2 does indeed do a way better job than FSR1, it's an absolute hot mess when it comes to resolving fast motion in the foreground, thin objects on high contrast background, and cannot do particle effects at all.

So, to those who said after Deathloops reveal of it that we don't need fancy AI hardware on our GPUs to get upscaling to work cleanly, yes, we do.

 

[Puiu]

The Digital Foundry video on FSR2 in GoW shows that while FSR2 does indeed do a way better job than FSR1, it's an absolute hot mess when it comes to resolving fast motion in the foreground, thin objects on high contrast background, and cannot do particle effects at all.

So, to those who said after Deathloops reveal of it that we don't need fancy AI hardware on our GPUs to get upscaling to work cleanly, yes, we do.

It depends a lot on your expectations. These things can be improved over time (from both devs and AMD). As a starting point, it's a lot better than anyone thought that AMD could do with the first iteration of FSR 2.0 with an algorithm that doesn't require specialized hardware.

It actually proved that you don't need AI for it to be good. They just need to improve the ghosting (something DLSS also struggled until recently) and the fizzle.

 

[Nobina]

Hope FH5 gets it since it has FSR 1.0. It's not as good as DLSS but it's getting there and should be more widely adopted. It also beats not having any upscaling at all.

 

[envirovore]

It depends a lot on your expectations. These things can be improved over time (from both devs and AMD). As a starting point, it's a lot better than anyone thought that AMD could do with the first iteration of FSR 2.0 with an algorithm that doesn't require specialized hardware.

It actually proved that you don't need AI for it to be good. They just need to improve the ghosting (something DLSS also struggled until recently) and the fizzle.

Agreed, I'm happy to see AMD throw an open software attempt at a solution to this, and overall it's a very impressive showing. The more options for this tech, the better, especially if one of them can be a viable solution *without* the need for dedicated hardware.
I'd like to see some sort of update to address that fizzling and severe ghosting, get that (well and the particle effects issues) sorted out and it looks like they've got a winner in their hand.

Likewise, it's still pretty new, especially compared to DLSS which also wasn't all that great until version 2.+, and even now still has issues with ghosting and shimmer. No doubt FSR will improve with time.

I opted for Ampere, but will not scoff at giving FSR a go if needed and the game doesn't support DLSS, same with XeSS when that's released. If that ends up doing better than dlss, I'll opt for that if available then.

 

[Red34jfp]

Can't wait for nVidia to begrudgingly adopt FSR in the same way that G-sync is now just freesync with a badge on the front that monitor companies pay for.

"we can code specifically for 1 type of GPU or we can just copy and paste this code that supports all GPUs"

Back when it was released AMD said that implementing FSR 2.0 is easier, just a matter of days, for games that already supports DLSS. If both algorithms require a similar amount of work it's more likely that both technologies will coexist, especially as long as DLSS can provide an edge over FSR 2.0 (which appears to change from game to game, God of Wars for instance favours DLSS more when compared to Deathloop, apparently).

 

[yRaz]

Back when it was released AMD said that implementing FSR 2.0 is easier, just a matter of days, for games that already supports DLSS. If both algorithms require a similar amount of work it's more likely that both technologies will coexist, especially as long as DLSS can provide an edge over FSR 2.0 (which appears to change from game to game, God of Wars for instance favours DLSS more when compared to Deathloop, apparently).

Don't get me wrong, DLSS is hands down the superior upscaling tech. The problem is that you need nVidias blessing to code it into a game engine while(this is a gross over simplification) you can copy and past FSR open source code into a game engine. For triple A developers where cost is no object, DLSS and FSR co-existing is a no-brainer. But, for smaller devs or devs that are reaching close to their budget, pasting in FSR to their code is an amazing option. It benefits such a wide audience for no money that it just makes sense.

to add to that, both PlayStation and Xbox now use AMD's ecosystem. AMD has manipulated the market such that DLSS doesn't make sense for developers making console games.

I will never argue that FSR is better than DLSS, just that it is easier and makes more sense for developers to use

 

[poltevo]

FSR 2.0 is not copy and paste, that's why AMD said it would take less time for engines which *already* support DLSS. Integrating either FSR 2 or DLSS is not easy, but once a developer has done one, it is easier to do the other. The much more limited FSR 1, on the other hand, was much easier to integrate.