Are you a gamer? If you are, I bet you know about the game-changing technology called AMD FidelityFX Super Resolution (FSR). It’s meant to boost your graphics performance, and it works on almost any GPU, even your old one! With FSR, you can enjoy higher frame rates and smoother gameplay, regardless of whether you have a Radeon or a GeForce card. However, although FSR really is for everyone, does it offer the same performance boost on an NVIDIA graphics card as it does on an AMD GPU? I wanted to learn the answer to this question on my own, so I decided to test FSR on two similar GPUs, one made by AMD and the other by NVIDIA. Would you like to see the results and find out which is better at FSR? Then, by all means, read on:
NOTE: In making this article, we teamed up with TechniQualities and created this YouTube video which summarizes our findings. If you prefer our full in-depth work, scroll down past this video, and continue reading.
Both AMD and NVIDIA have their own upscaling technologies. AMD’s technology is called FidelityFX Super Resolution (FSR), while NVIDIA’s is named Deep Learning Super Sampling (DLSS). One of the major differences between DLSS and FSR is compatible hardware. NVIDIA’s DLSS is locked down to its GeForce RTX GPUs, but AMD’s FSR is open-source and GPU-agnostic, thus available on far more hardware.
In other words, AMD FSR can work with any graphics card, from any manufacturer. You can use FSR with any AMD card, as well as with any NVIDIA or Intel GPU, as long as they support DirectX 11 or DirectX 12. For instance, if your favorite game supports AMD FSR (and there’s a good chance it does), you can turn AMD FSR on no matter if your graphics card is, let’s say, an AMD Radeon RX 6800 or an NVIDIA GeForce RTX 3070.
AMD FSR, just like NVIDIA’s DLSS technology, aims to boost performance and image quality in games by using upscaling. This technique renders a game at a lower resolution than your monitor’s native resolution and then uses an algorithm to scale it up to fit your screen. For example, a game might not run very well at native 4K resolution, but with FSR, you can render it at 1080p or 1440p and then upscale it to 4K. The end result will be higher performance and sharper detail than native 1080p. You’ll get a higher FPS and smoother gameplay without sacrificing too much visual fidelity.
FSR has five different modes: Ultra Quality, Quality, Balanced, Performance, and Ultra Performance. Depending on your preference and hardware, each preset offers a different trade-off between performance and image quality.
I’ve tried AMD FSR in more than just a few games, and I have to say I’m impressed by how well it works. With its help, I can play demanding games like Cyberpunk 2077 or The Callisto Protocol in 4K resolution, at high settings, with excellent framerates. The image quality is really good, and even on FSR’s Balanced mode, I can barely notice any difference from the native 4K resolution.
One question I had and couldn’t get out of my head lately was: which is better at FSR? If I ran a game using FSR on an AMD graphics card, would I get better performance than with a similar NVIDIA card? Does AMD somehow favor its own GPUs when using FSR? Or is this upscaling technology truly GPU-agnostic, as AMD says? To get an answer, I decided to run a little experiment:
Test AMD FSR on AMD and NVIDIA… sure, but how? Well, I ran a couple of games on the same desktop computer, with and without AMD FSR, using first the AMD Radeon card and then the NVIDIA card. I chose two GPUs that are direct competitors and have similar performances and prices:
The other hardware (and software) components of the test PC were these:
- Motherboard: ASUS Prime X670E-Pro WiFi
- CPU cooler: Cooler Master MasterLiquid ML360R RGB
- Memory: Kingston Fury Beast RGB DDR5-6000 32GB
- Storage: Kingston KC3000 NVMe PCIe 4.0 SSD (2GB)
- Monitor: ASUS ROG Strix XG32UQ (4K, 160Hz)
- Power Supply Unit: ASUS ROG Thor 850W Platinum
- Operating System: Windows 11 Pro version 22H2
Then, once I’d chosen the hardware, the next step was to decide what resolutions and what games to test. As 1080p seems too low for today’s expectations, the resolutions I ended up with were 1440p and 4K. After all, this is where you can easily notice the benefits of upscaling technologies like AMD FSR.
Next, the games. Unfortunately, time didn’t allow me to include tens of AAA titles in this experiment. Although just one or two games would have been too few, I believe the four that I chose are enough to get a relatively clear picture of what to expect from AMD FSR. They have steep system requirements and are quite demanding in terms of graphics performance. The games are Cyberpunk 2077, The Callisto Protocol, Assassin’s Creed Valhalla, and F1 22. I used the highest graphics settings (with ray-tracing effects enabled) for each game and ran them with FSR off and on, set on the Balanced preset. The charts show the results of each GPU/resolution/FSR scenario. Here are my findings:
I began with Cyberpunk 2077, one of the most demanding games today, especially if you want to enjoy ray-tracing effects. In 1440p, turning on FSR resulted in a framerate increase of 97% on the AMD Radeon RX 7900 XT and almost 98% more fps on the GeForce RTX 4070 Ti. In 4K, FSR offered an even higher framerate increase with the Radeon RX 7900 XT: 122%. With NVIDIA’s GeForce RTX 4070 Ti, the increase was smaller but still significant: 52% more fps.
The Callisto Protocol is a survival horror game that can use ray tracing to create realistic effects on shadows and reflections. In 1440p resolution, FSR delivered a noticeable framerate increase of almost 30% with the Radeon RX 7900 XT and 21% more frames per second with the GeForce RTX 4070 Ti. In 4K, FSR’s benefits are even more visible: on the Radeon RX 7900 XT, the fps increase was 85%, while on the GeForce RTX 4070 Ti, it delivered 69% more frames per second.
Assassin’s Creed Valhalla is a beautiful action role-playing game, but it can require a lot of performance from your GPU to run at its best. When testing the game in 1440p resolution, turning on AMD FSR, the increase in the number of frames per second was over 37% with the Radeon RX 7900 XT and 39% with NVIDIA’s GeForce RTX 4070 Ti. In 4K, FSR led to a framerate boost of 72% with the Radeon RX 7900 XT, and 69% with the GeForce RTX 4070 Ti.
The fourth and last game I tested was F1 22, a car racing game that also needs a powerful computer if you want to play it in high resolutions and at top visual quality. When running it in 1440p and with AMD FSR, I witnessed a 46% framerate step up with the Radeon RX 7900 XT and 60% with the GeForce RTX 4070 Ti. In 4K, enabling FSR meant an even more significant increase in framerate. With the Radeon RX 7900 XT, I measured an fps boost of almost 82%, while with the GeForce RTX 4070 Ti, the increase was over 96%.
As you can see, AMD FSR lives up to the hype and offers impressive performance gains across different games, and it does so without compromising too much on image quality. Unfortunately, looking at the framerates in each game and resolution, with and without FSR, still doesn’t give a clear answer to which one’s better: AMD or NVIDIA. For that, I had to compile these results and see what the statistics show. Let’s get to it:
While the benefits of AMD FSR may vary depending on your hardware configuration and personal preference, the general statistics should hold true. To create them, I calculated the geometric mean for each game and for each resolution tested, both for the scenario where AMD FSR is enabled and when it’s not.
Are you wondering what a geometric mean is and why I chose to compile the results this way? Well, what you should know is that this is a method of calculating the average of a group of numbers by multiplying them all together and then getting the nth root. While it may sound complicated, it actually isn’t (see this Wikipedia article for more insight). Furthermore, the geometric mean is quite helpful when you need to compare the growth rates of various things. For instance, if you’d like to know how much money interest rates will get you yearly in a bank account, you should use the geometric mean. The same applies to this AMD FSR experiment: I want to get a better idea on the average framerate over the entire pack of games, all put together, so that I can then calculate the percentage increase when using AMD FSR. This is what I learned:
As you can see in the table, the average increase across all games is impressive regardless of what graphics card I used. The statistics tell us that, in 1440p, enabling FSR for the GeForce RTX 4070 Ti leads to an average fps increase of 53%, while for the AMD Radeon RX 7900 XT, the increase is 50%. These are pretty similar improvements, but you could say that NVIDIA benefits more than AMD from using FSR in 1440p resolution (3%).
In 4K resolution, using AMD FSR delivers even more impressive results, both for AMD and NVIDIA. However, this time, it looks like the AMD GPU makes the most out of FSR (89% fps increase), although NVIDIA also puts this upscaling tech to good use (71% framerate increase). It’s clear that both graphics cards offer remarkable framerate gains in 4K, more than in 1440p. However, the numbers also say that the AMD GPU is better than the NVIDIA one when using FSR in 4K (18%).
So, in the end, the answer to the question that gave birth to this experiment is a mixed one: both AMD and NVIDIA cash in when using AMD’s FSR upscaling technology. In lower resolutions, NVIDIA seems to be the winner, with AMD very close behind, while in 4K, AMD wins by a landslide.
That’s all for today’s gaming experiment. I hope you enjoyed it, and if you did, please leave a comment below, letting me know what you think about AMD FSR. Do you use it? In what games do you turn it on? What improvements did you notice on your system? Tell me your story, and until next time, happy gaming!