Nvidia GeForce RTX 5080 FE Review: More Performance, Same Price

Cyberpunk 2077

The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

While the RTX 4090 and RTX 5090 can be talked about as at least semi-professional cards due to their price and memory capacity, the cheaper RTX 5080 is already a model aimed primarily at demanding gamers. It mainly targets 4K gaming, but with 16 GB of memory it’s also quite usable for application use and working with AI models. Let’s start by recapping the RTX 5080’s specs.

 
A detailed description of the new technologies that the new graphics cards and architecture bring is provided by Jan Olšan in the article Blackwell: GeForce RTX 5000 architecture and innovations. It presents the individual chips that Nvidia uses in the new graphics cards, as well as the new DDR7 memory, changes to the cache, compute units in the GPU, changes to the shaders and Tensor Cores, and improved compute units for ray tracing. Support for the new standards DisplayPort 2.1b and PCIe 5.0 has been added to the cards as well.

With the new generation of cards, the capabilities of DLSS are also advancing, which can now use Multi Frame Generation to generate up to three intermediate frames, and new is also the second generation of Nvidia Reflex with Frame Warp technology, which can adjust the movement of objects and the camera in rendered frames based on the most recent information from peripherals, which it receives only during or after the frame is rendered. Everything is described in more detail in the mentioned article.

The unboxing of the RTX 5080 Founders Edition is already quite different from what we’re used to. The box is in natural colors. In order to get inside, you have to tear off the tape around the perimeter. When you open the lid and flip up the sides, a smaller sturdy stylized card-shaped package also made of corrugated cardboard peeks out at you. Underneath is an even smaller box with a power adapter.

Inside the package is the card itself. In the small box you will find a brief installation guide and a power octopus with three tentacles of eight pins.

While the RTX 4080 Founders Edition was a massive three-slot brick, the RTX 5080 FE looks like an ordinary card again. It’s wider, it’s longer, but otherwise it has the compact design that was previously common with reference cards and fits easily into regular sized cases.

The design is no different from the higher performance RTX 5090. The only thing different is the colour shade of the frame part, it is brighter on the RTX 5080.

In terms of diagnostic applications, we still won’t see much more than when the RTX 5090 was launched.

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

From the bottom, it looks like any other card. Two axial fans draw air from under the card and push it into the fins of the heatsink. Only the rear bracket lets you know that there is something unusual about the RTX 5080 FE. It’s perforated on virtually all higher-performance cards nowadays and air goes through it. If the cooler pushes too much air through the perforation, it can be a significant source of noise. This was the main reason why cards with radial fans were extremely noisy, and why they were eventually replaced by coolers with axial fans.

Where there used to be a full PCB and later a backplate, you can now see finning. We’re used to it being behind the PCB these days. The design of the card directly from Nvidia, however, surprises by the fact that aslo the half of the card near the rear bracket is passable by air. The PCB just is in the middle of the card.

You’ll notice that you can see through the card not only at the back, but also at the front when you look at it from above. The heatsink is passable under both fans, the only part that is full is in the middle where the PCB is.

 

The air blows diagonally upwards over the card from the vents on the sides of the frame as well.

The card is about 30.6 cm long, only 3 cm wider than the bracket and takes up only two slots in height.

The configuration of outputs is standard – there are three DisplayPorts and one HDMI. With the last generation, Nvidia was behind AMD in implementing support for the latest DisplayPort interface standards, with the RTX 50 they remedy the situation and the DisplayPorts support the DP 2.1b standard with UHBR20.

There is a pair of screws in the rear wall of the frame that serve as plugs for the metric threaded mounting holes. These also allow the back of the card to be secured to the case to avoid putting too much strain on the PCI Express slot. By the way, there are magnets around the perimeter of the Founders Edition frame. Also on the side. Apparently they make the cooler assembly easier, but they can also hold the screws while you are tinkering with your PC.

The power supply is provided by a 16-pin 12V-2×6 connector, which we already know from the RTX 40 Super series cards.

Nvidia solved the problematic cable bending at the connector by adding an extremely soft adapter.

As with the RTX 5090, I’ll point out that when comparing it to the cables that come with most PSUs, I’d recommend reaching for the included adapter. The wires are like rubber, and compared to most comparatively stiffer cables, you’re less likely to pull out the connector by pushing on the cable, worsening the contact, raising the transient resistance, and causing the connector to overheat and melt.

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

Test build

I don’t test graphics cards openly on a desk like most reviewers, I use a classic gaming PC for measurements. The components are mounted in a Fractal Design Meshify 2 case. To give you an idea, for now I’ll borrow a photo of another build in the same case, which you may have come across in articles dedicated to system cooling. It also features a quartet of 140mm fans in all available positions. All system fans have their speeds reduced to about 780 rpm. The build is not completely silent, but it can cool even high performance cards with high power consumption.

Inside is the newer Aorus X870 Elite WiFi7 motherboard, which is equipped with an AMD Ryzen 9 9800X3D processor. The cooling of the processor is taken care of by the new Noctua NH-D15 G2 in the offset position. The processor is complemented by 64GB of DDR5-6000 memory in two modules from Kingston, and the system and data are stored on a 4TB SSD from the same manufacturer. The whole setup is powered by the top-of-the-line 1300W Be quiet! Dark Power Pro 13 PSU.

Below is a picture of the configuration and current system settings and the version of Windows used for testing. The memories are set to an EXPO profile with DDR5-6000 32-38-38-38-1,350V with XMP/EXPO High Bandwidth Support set to Enabled, Infinity Fabric Frequency and Dividers manually set to 2000 MHz and the UCLK DIV1 MODE divider set to UCLK=MEMCLK.

The SVM Enable setting is deactivated in the advanced processor settings for higher performance and the “Core Isolation” setting in Windows Security is also deactivated.

Furthermore, the Radeon graphics adapter integrated in the processor and the integrated SATA controller are manually deactivated.

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

Black Myth: Wukong

Metro Exodus Enhanced Edition

Satisfactory

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

Cyberpunk 2077

Global setting RT Ultra

Performance with DLAA, DLSS Performance and DLSS Frame Gen

1920 × 1080

2560 × 1440

3840 × 2160

Global setting RT Overdrive

Performance with DLAA, DLSS Performance and DLSS Frame Gen

2560 × 1440

3840 × 2160

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

F1 24

Performance with DLAA, DLSS Performance and DLSS Frame Gen

 

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

3DMark Port Royal

 

UL Procyon, Stable Diffusion 1.5 (FP16) Overall score

Blender

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

I measured the operating characteristics in the F1 24 benchmark at a resolution of 3840 × 2160 pixels with the Ultra setting and DLAA anti-aliasing. With this setting, the GPU is fully loaded. This is a benchmark with five laps on the Singapore map in wet conditions.

The frame rate hovers around 60 frames per second.

The card’s clock speed holds steady around 2730 MHz.

With this setting, the power consumption of the card is around 333 W according to the monitoring, which is below the power limit.

The power consumption of the whole computer measured by a UT71E multimeter was then around 456 W.

The temperatures of the GPU are significantly lower than those of the RTX 5090.

The speed of both fans is the same and stays around 1620 rpm after warming up, at which the measured noise level reaches around 42.3 dBA. And it’s a shame, the curve could have been shifted lower, this way the card is on the worse side in comparison to the noise level of common non-reference mid-range models.

⠀
The GeForce RTX 5080 series gaming graphics cards go on sale today, and simultaneously, we can now publish performance test results. The biggest improvement over the last generation is an interesting performance boost in the most demanding titles at 4K and support for new standards and technologies. The edge is especially gained in demanding ray tracing scenarios, and vice versa.

Closing summary

Although we’re not testing many games yet, it’s clear that next to the RTX 4080, the RTX 5080 is better suited to scenarios with higher graphics demands. With lower resolutions and less demanding visuals, the gap between the two models narrows. The RTX 5080 handles demanding ray tracing significantly better. But games are also moving more in this direction, and in titles that don’t use ray tracing yet, graphics cards usually don’t have a performance problem.

If we had stuck to the suggested prices at launch for the evaluation, we would have had reason to rejoice. The last-gen RTX 4080 was launched with an MSRP of 1199 USD. The RTX 5080 has an MSRP of 999 USD, down 200 USD. But after that came the RTX 4080 Super for the same 999 USD. So the newcomer matches its price and offers more performance for it. But the increase in raw performance is not very high, Nvidia is betting more on an increase in performance thanks to Multi Frame Generation. To some extent, this is a repeat of the last generation scenario.

From a quick look at the stores, it doesn’t look like any big RTX 4080 discounts are happening. Given that the RTX 5080, which is about a fifth more powerful in 4K in the most optimistic scenarios, should start selling tomorrow, I’m already worried about what the prices of non-reference cards will look like. One can suspect that manufacturers will set prices for new models so that RTX 4080 Super stocks will sell out without big discounts.

DLSS 4: Two extra frames with similar responsiveness

I’ll point out again that the idea that the card only renders every fourth frame in 4× multi frame generation mode tempts people to think that the response time will be longer than in single frame generation. This is not the case, the response time in 3x and 4x mode is not significantly worse compared to what we know with frame generation with a single intermediate frame from the last generation. In both single frame generation and triple frame generation, only one rendered frame is still being deferred. The only difference is that the card stacks more frames generated at a higher speed in front of the rendered frame.

However, the basic principle that the card postpones rendering the last rendered frame remains. So the same applies to Multi Frame Generation as it does to frame generation in RTX 40 – it’s not suitable for situations where you’re starting from too low a frame rate and interpolating to, say, a smooth 80. In that case, while the image is not jerky, it has a noticeable lag behind mouse movement. It’s similar to playing at 25 frames per second. When rendering at a low 25 frames per second, the long response time is natural and will not be improved by adding three generated frames before the rendered frame.

Therefore, it is a good idea to adjust the detail settings to achieve higher base frame rates and then interpolate from there to higher frame rates for fast monitors. If the graphics card doesn’t have enough horsepower to run games with the given settings, this means either lowering the detail or using classic DLSS in one of the Quality, Balanced or Performance modes along with the frame rate generation. Once you get to a base 40–50 frames per second, it still won’t be for pro gaming or frenetic shooters, but everything else can be played comfortably.

Higher performance for higher consumption

What was true of the RTX 5090 in relation to power consumption can also be said of the RTX 5080. The manufacturing process node has not improved significantly, and so the RTX 5080 also delivers higher performance, but at a higher power consumption. In the case of the Founders Edition, it’s clear that a similarly designed cooler handles the lower-power card better. Still, the fan control could have been a bit quieter at the cost of higher chip temperatures.

When it comes to performance, in some scenarios the difference is big, but especially at lower resolutions the edge is not so significant. But the trend in game development is clear, demanding ray tracing is becoming the standard for big titles and developers are equally taking advantage of the opportunities that upscaling gives them to push the demands on the graphics card much further.

The performance difference isn’t enough to give RTX 4080 owners a reason to upgrade. It’s more a question of whether to choose the RTX 4080 Super or the RTX 5080 when buying a new card. At similar prices to the RTX 4080 Super, or at a small premium, the new card is clearly more prospective, in which case buying the RTX 5080 can be clearly recommended.

It will ultimately depend primarily on the market prices of non-reference models and the actual availability of the cards. Unless the RTX 5080 will sell for significantly more, there is no reason to prefer RTX 4080 Super.

Pros and cons

+ High performance for 4K gaming
+ Compact design while maintaining good performance
+ Decent performance gains in the most demanding scenarios
+ Compared to RTX 40, the bonus of DLSS 4 with Multi Frame Generation
+ Support for new standards
+ New features for professional use (AI, video processing)

– With less demanding ray tracing and at lower resolutions, the performance increase is smaller
– Fans could run at lower speeds
– Increase in power consumption, performance/watt ratio similar to last generation

• Contents
• RTX 5080 Founders Edition: introduction and parameters
• Card design, connectors
• Test build
• Black Myth: Wukong, Metro Exodus EE, Satisfactory
• Cyberpunk 2077
• F1 24
• UL 3Dmark, Procyon AI, Blender benchmark
• Operating characteristics – F1 24
• Closing summary