After the arrival of the RTX 4090, it is now time for the GeForce RTX 4080. Nvidia is once again bringing a luxurious and expensive video card to the market, but this time it is a slightly lower price point than the flagship that was released in mid-October. The GeForce manufacturer originally planned to release two different cards called ‘RTX 4080’, but later backed off. In this review, we look at the model that is allowed to go through life as an RTX 4080 and compare its performance with a range of other video cards from both Nvidia and AMD.
Where the x80 series of GeForce video cards was positioned just below the real top model in the past and offered a more or less acceptable price-performance ratio for the enthusiast, the RTX 4080 should not only cost 1469 euros, but originally also had to sharing its name with an inferior card based on a lower quality chip with significantly lower specifications. Nvidia’s goal with this seemed clear: lure potential buyers to the performance of ‘the’ (faster) RTX 4080 and entice them with the price of the slower model. RTX 4080 performance for an RTX 4080 price, except one doesn’t equal the other.
|RTX 4080 16GB
|“RTX 4080” 12GB
|November 15, 2022
That Nvidia finally shot down the slower ’12GB version’ of the RTX 4080 will no doubt have at least as much to do with competitor AMD’s announced line-up as with all the feedback from gamers and hardware enthusiasts about the confusing, if not misleading, positioning of the two RTX 4080 cards. In the past, the manufacturer has often raised a finger not the thumb at the consumer, and the introduction of the RTX 4080 is further confirmation that the company is, again and again, pushing the boundaries of what the customer just accepts.
The green light never changes to red
So what remains is the ‘real’ RTX 4080, with 16GB memory and the larger GPU chip. Unchanged is the suggested retail price, we will mention it again, of 1469 euros. Due to inflation, the exchange rate of the euro against the dollar and a series of decisions at Nvidia headquarters, the MSRP of the RTX 4080 is twice as high as the RTX 3080 at launch in September 2020 . It must now be said that Nvidia’s suggested retail price of 719 euros for the RTX 3080 was completely unrealistic at the time, and only the most fortunate for that amount were able to get hold of this video card shortly after it came out.
Still, the suggested retail price of the RTX 4080 in euros illustrates well what kind of world the PC gamer has unfortunately ended up in. In addition, Nvidia has been paying close attention over the past two years to what its customers are apparently willing to pay for GeForce video cards. Nvidia’s strategy of an imaginative ultimate flagship builds on its reputation for premium quality and features, with matching premium suggested retail prices across its portfolio. Despite sharply increased competition from the Radeon camp in recent years, Nvidia’s market share hardly seems to suffer from this. The fact that the tech community breathed a sigh of relief when the bogus RTX 4080 was revealed to have been “canceled” speaks volumes about the self-reinforcing effect of Nvidia’s closed technologies and ” walled gardens’ on its market share. Hooray, the one and only RTX 4080 is here!
Odd man out: AD103
The Nvidia GeForce RTX 4080 uses the AD103 GPU, which is currently exclusive. Namely, the canceled 12GB version of the RTX 4080 , which appears to be rebranded as an RTX 4070 Ti , uses the AD104 chip, which is even smaller than the AD103. In the naming, the AD103 GPU also stands out because Nvidia rarely uses odd numbers, and specifically the ‘103’ has never been used before by the manufacturer for desktop cards. Like the RTX 4090, the RTX 4080 is also based on the Ada Lovelace architecture. In our review of the GeForce RTX 4090 you can read more about the new features of the RTX 40 series, including DLSS3 and AV1 encoding.
Apart from the name, the AD103 cannot be called groundbreaking compared to both the RTX 4090 and previous generations. This is first apparent from the die-size, which at 379mm² is rather on the small side for a GeForce x80 card when compared to previous RTX cards. Thanks to the 4nm process, the number of transistors on this ‘small’ GPU is of course considerably higher than on older Nvidia models.
On the RTX 4080, 9728 CUDA cores are available, while the AD103 GPU contains a total of 10752. That means that almost 10 percent of the computing cores are disabled. For the tensor and RT cores we encounter the same proportions, as do the texture units and rops . With the available memory controllers, the memory bus totals to 256bit. That is narrower than the 320bit on the RTX 3080, but thanks to the higher memory speed, this is largely compensated on the RTX 4080. In addition, a significantly larger L2 cache ensures less dependence on the vram on the RTX 40 series. Nvidia has set the total graphics power at 320W, identical to the RTX 3080.
|RTX 3080 (10GB)
|Suggested retail price FE
Our starting point in GPU tests is always that we want the rest of the system to form the smallest possible bottleneck. We have therefore upgraded our test system in recent weeks, after having used an overclocked AMD Ryzen 9 5950X for a long time. Our new gpu test system consists of a water-cooled Intel Core i9-13900K processor, overclocked to 5.5GHz allcore. Because the e-cores of this processor cause performance degradation in some of our tests, we have completely disabled the e-cores on our platform. This means that only all 8 performance cores with 16 threads remain active. As working memory, we use 32GB of fast DDR5-7200 with relatively tight timings, to ensure that both the CPU and the RAM are the limiting factor for measuring GPU performance as little as possible.
The exact specifications of the test system can be found in the table below.
|Test system GPUs
|Intel Core i9-13900K (p-cores @5.5GHz, e-cores disabled)
|Gigabyte Aorus Z790 Master
|Random access memory
|G.Skill Trident Z RGB F5 32GB (2x 16GB) DDR5-7200 CL34-45-45-115
|Silicon Power XS70 4TB
|FSP Hydro PTM Pro ATX3.0 1200W
|Alphacool Eisblock XPX, Alphacool XT45 480mm radiator, Alphacool D5 water pump, be quiet Pure Wings 2 fans
Tested video cards
For our GPU tests, we always use reference cards or Founders Editions, unless the GPU in question has not been released as such or we do not have one available. In that case we use a custom model that is as close as possible to the reference specification in terms of clock speeds.
For this review we used the following models:
- Nvidia GeForce RTX 4090 Founders Edition
- Nvidia GeForce RTX 4080 Founders Edition
- Gigabyte GeForce RTX 3090 Ti Gaming OC
- Nvidia GeForce RTX 3090 Founders Edition
- AMD Radeon RX 6950 XT reference card
- Nvidia GeForce RTX 3080 Founders Edition
- AMD Radeon RX 6800 XT reference card
- Gigabyte GeForce RTX 3070 Ti Gaming OC
- Nvidia GeForce RTX 2080 Ti Founders Edition
- Nvidia GeForce RTX 2080 Founders Edition
- Nvidia GeForce GTX 1080Ti Founder’s Edition
Drivers and measurement method
We tested all video cards for this review with the latest driver available when we started. For the AMD Radeon cards, we used Radeon Software 22.10.3; for Nvidia GeForce cards it was GeForce 526.72.
Using PresentMon, we measure the performance in each tested game, from which we calculate the average frame rates, fps, as well as the frame times of the 99th and 99.9th percentiles and report the latter two in milliseconds .
The time it takes to render images within a 3D game and therefore within our benchmark varies from frame to frame. With our frametime measurement, the rendering times of all individual frames are stored. Then we discard the 1 percent slowest frames. The highest render time of the remaining 99 percent of the frames, or the slowest frame, is the 99th percentile frametime.
At the request of some readers, we’ve also included the 99.9th percentile values. So for this we only disregard the 0.1 percent slowest frames. In theory this is even more precise, but in practice incidental causes and measurement errors sometimes throw a spanner in the works. For now, we’ve listed them in the review, so keep that in mind when looking at these results.
We regularly review the range of games and make these choices taking into account each game’s API, engine, genre, AMD/Nvidia ratio, age and technical benchmark details to arrive at the most representative suite possible. to come out.
|Assassin’s Creed: Valhalla
|Anvil Next 2.0
|Call of Duty: Modern Warfare II
|red engine 4
|Id tech 7
|EGO Engine 4.0
|Forza Horizon 5
|Guardians of the Galaxy
|Metro Exodus (+Enhanced)
|Red Dead Redemption 2
|Total War: Warhammer III
|TW Engine 3
In addition to performance, we measure the power consumption of video cards. We perform the current measurement with a riser card from the manufacturer Adex, which we place between the PCIe slot and the video card. That way we can not only measure the current that runs through the loose power cables, but also the power that the video card draws directly from the PCIe slot. For the measurement we use several copies of the Tinkerforge Voltage/Current Bricklet 2.0 . For the final measurement, the riser card is equipped with such a bricklet. To measure the current of the PEG cable, the bricklet was placed together with the necessary connectors on a printed circuit board specially designed for our test. Depending on the number of power cables that a video card requires, we obviously use various meters.
3DMark and Unigine Superposition
The RTX 4080 is over 60 percent faster than the RTX 3080 in Time Spy and the RTX 4090 is in turn almost 30 percent faster than the RTX 4080. In Fire Strike Ultra we measure the same ratios between the RTX 4080 and its predecessor, but here the RTX 4090 even ends up more than 43 percent above the RTX 4080.
3DMark Port Royal
Port Royal is 3DMark’s benchmark in which raytracing performance is measured. It is a synthetic test that produces both a score and an average frame rate. The RTX 4080 outperforms the RTX 3080 by 55 percent and the RTX 3090 Ti by 22 percent in this test. The RTX 4090 is now more than 42 percent faster than the RTX 4080.
Superposition is a benchmark of Unigine and can be run in both OpenGL and DirectX 11. For our tests we use DX11 and two of the available graphics presets. We also see the RTX 4080 perform significantly better than its predecessor in Superposition.