Nvidia may have the clear advantages over AMD thanks to it’s DLSS, but it seems that the company is holding back as it has been stated that they are launching their FidelityFX Super Resolution feature when it’s ready. This includes cross-support for the latest next-gen consoles, the PS5 and Xbox Series.
- Related Reading: AMD Radeon RX 6700 XT Release Date Set for March 18 With a MSRP $479
Speaking to the Linus Sebastian of Linus Tech Tips, Linus confirmed the reasoning as to why AMD’s FidelityFX Super Resolution has yet to release.
Apparently rather than rushing it [FidelityFX Super Resolution] out the door on only one new top-end card, they want it to be cross-platform in every sense of the word. So they actually want it running on all of their GPUs, including the ones inside consoles, before they pull the trigger.
It would’ve been nice to have it ready by the time the cards launched, but as we saw with Nvidia’s DLSS 1.0 versus DLSS 2.0, it could be for the best.
An example of DLSS
Now for all you console players out there who don’t know what DLSS and FidelityFX Super Resolution, they’re essentially a type of AI upscaling technique that uses machine learning to upscale lower resolution to higher res, all while maintaining the original performance. This means that picture quality can look a whole lot better without the need to sacrifice frame-rate. There are tons of other benefits as detailed on Nvidia’s website for their own DLSS.
The results of DLSS vary a bit, because each game has different characteristics based on the game engine, complexity of content, and the time spent on training. Our supercomputer never sleeps, and we continue to train and improve our deep learning neural network even after a game’s launch. When we have improvements to performance or image quality ready, we provide them to you via NVIDIA software updates.
DLSS is designed to boost frame rates at high GPU workloads (i.e. when your framerate is low and your GPU is working to its full capacity without bottlenecks or other limitations). If your game is already running at high frame rates, your GPU’s frame rendering time may be shorter than the DLSS execution time. In this case, DLSS is not available because it would not improve your framerate. However, if your game is heavily utilizing the GPU (e.g. FPS is below ~60), DLSS provides an optimal performance boost. You can crank up your settings to maximize your gains. (Note: 60 FPS is an approximation — the exact number varies by game and what graphics settings are enabled)
To put it a bit more technically, DLSS requires a fixed amount of GPU time per frame to run the deep neural network. Thus, games that run at lower frame rates (proportionally less fixed workload) or higher resolutions (greater pixel shading savings), benefit more from DLSS. For games running at high frame rates or low resolutions, DLSS may not boost performance. When your GPU’s frame rendering time is shorter than what it takes to execute the DLSS model, we don’t enable DLSS. We only enable DLSS for cases where you will receive a performance gain. DLSS availability is game-specific, and depends on your GPU and selected display resolution.
Of course, the description above is only for Nvidia’s version, though we do expect AMD to work in a similar fashion. It’s a pretty exciting technique, one that has worked wonders on many PC titles running on the RTX 3000-series line of cards.
A certain well known Xbox fanboy called Colt claimed only Xbox Series X was getting this upgrade, wonder how he’ll damage control this news lol.
Colts a clueless Xtard shill
It’s ironic that DLSS only works on cards that don’t actually need it! Glad to see AMD’s solution is thinking about us peasants playing on a potato! Not that that will help me, as I play on an old Nvidia card…
I do wonder what this means for the PS5 and XSX/S though, as only the latter has customised hardware for machine learning, (albeit, not nearly as optimised for AI as the Nvidia line, but still). I assume it will take advantage of that hardware, if not, then I assume Microsoft will have their own solution. I suspect Microsoft and AMD are working on a bespoke Xbox solution though, but we’ll see!
All RDNA 2 hardware has mixed integer(including PS5) Microshaft never added any hardware, it all standard from AMD
Wrong, Microsoft and AMD made some modifications to the Series hardware to better support AI, something that is definitely not in the PS5. (Maybe I’ll look it up if you don’t believe me).
Besides, the PS5 is not full RDNA2, I thought that was common knowledge at this point! Microsoft waited longer for it. Sony chose to go into production earlier.
Thats nonsense. Trashbox isn’t FULL RDNA 2 either. Trashbox has no high clocks, infinty cache or smart ram.
So there are some differences, it’s a console vs PC situation there, some things are better, some things worse. But the main features are present, unlike with the competition. If you want to get pedantic, then perhaps it’s :
Series X/S = RDNA 1.95
PS5 = RDNA 1.5
It may turn out that Sony were smart with that choice, or perhaps it will really hurt them in the long term vs short term gains.
Kinda like the SSD situation, Sony’s is more ambitious and expensive but hasn’t paid off at all so far, all of this hardware and software needs time to mature. I don’t think anyone really knows the true gap in power and won’t for several years.
Saying that, we’re starting to see the Series X pull ahead in resolutions now and it’s only a few months into this gen, I’m curious to see how much of a gap there will be after a year or two, when developers really get to grips with it.
Microbrain GDK has been embarrassing so far, hopefully things will improve a lot there as well. Sony went with a simpler architecture and stuck with last gens GDK for short term gains and easier development.
Microborg went for a more complicated architecture (a lot more cores for one) and an all new GDK, which is built around all of their consoles and PC, for more streamlined development, short term pain for potential long term gains.
Both consoles should do well though and I’m excited to see what Sony comes up with regarding the PSVR2!
I’m a PC gamer BTW, I’m not even a Microwave fan!
FidelityFx won’t be AI based. AMD cards have no machine-learning capabilities like what Tensor Cores do for Nvidia cards. As a result it will generally be at least a little worse than DLSS on average when it comes to image quality.
Dont need Tensor cores because the normal CUs can handle INT8 data sets.
Mixed integer(int4/in8 & dot4/dot8 is standard across RDNA/RDNA2 incuding both PS5/Trashbox
It looks like XBOX is more “ready” for this then PS5 but I would guess they would both need a firmware update to exploit the tech. Maybe the lack of availability of the consoles might be a win for those of us who have not bought one yet as it may take something like a revision 1 motherboard to support DLSS.
Doubt its gonna be a firmware feature for consoles, more a dev tool feature