Nvidia GeForce RTX 3080 review: the super-performers

Let's begin with three new entrants to our 2020 graphics card test suite - Doom Eternal, Control and Borderlands 3 - and a returning favourite, Shadow of the Tomb Raider. All four games show significant improvements running on the RTX 3080 compared to previous-gen cards, whether you're looking at the RTX 2080, RTX 2080 Ti or even the venerable GTX 1080.

These 'super-performers' are perhaps unsurprisingly ones that Nvidia opted to focus on in our first reveal of RTX 3080's horsepower and represent something of a best-case scenario for the graphics card. Each title is relatively recent - Shadow of the Tomb Raider is the oldest at around two years old - and all four use a modern graphics API, such as DirectX 12 and Vulkan. In each case, the CPU is well-utilised to ensure that we're not bottlenecked, allowing the graphics card to run at full tilt even at lower resolutions.

Depending on how you've accessed this review, our performance metrics are presented in one of two ways. If you're reading this on a mobile device, you'll get the basics: a table with average frame-rate and lowest one per cent measurements. However, if you're on a desktop or laptop, you get the full-fat DF experience. You can play the YouTube videos to see our recorded frame-rate and frame-time metrics from each card throughout our test scene, even if you skip around or adjust the playback speed. Beneath that you'll see our barcharts, which are generated from the frame-time metrics. Remember that you can mouse over for various stats and click to swap between the raw averages and percentage differences.

It's a system designed to give you as much in-depth data as you want - or indeed as little, for at-a-glance viewing - all derived from video captures of each respective GPU to ensure accuracy. Let's get into the results; I'm sure you're as keen to see them as we are to share them!

Doom Eternal

Following on from the triumphant Doom (2016), Doom Eternal is another cleverly crafted game that uses the forward-looking id Tech engine. Now at version seven, this engine drops support for the DirectX family of APIs entirely, instead relying on the Vulkan API to great effect. This allows for efficient use of high core count processors, like the Core i9 10900K on our test bed, so we see a big uptick in performance from the 2080 to the 3080 even at 1080p as we're not as CPU-limited as we'd be in many other titles.

The percentage improvement from 2080 to 3080 is 53 per cent at 1080p, growing to 71 per cent at 1440p and 87 per cent at 4K. If you're lucky enough to instead be rocking the RTX 2080 Ti, the fastest available consumer card available until the launch of the 3080, you'll still see a 24 per cent boost to frame-rates at 1080p, rising to around 35 per cent at the higher resolutions we tested. We're talking about results in the hundreds of fps here, so it's not as if performance was bad on Nvidia's previous flagship, but at 4K the RTX 3080 is able to comfortably max out a 4K 144Hz monitor, something the RTX 2080 Ti couldn't achieve.

If you're instead upgrading from a 10-series GTX card, you'll see even greater improvements. Using the GTX 1080 we tested as a baseline, performance increases by around 150 per cent at 1080p, rising to about 175 per cent at 1440p and just about 200 per cent at 4K. Even accounting for the two generations' difference between these two cards, that is a mighty result.

You may be wondering why we have included ultra textures benchmark variants for the GTX 1080 and RTX 2080 at 4K resolution. The truth is that the huge boosts RTX 3080 delivers here come from two sources: the chip itself and the fact it has more VRAM capable of handling the so-called 'ultra nightmare' textures. Here, by including both ultra nightmare and ultra nightmare with ultra textures benches, you can see how much of the perf boost comes from the architecture and how much comes from the new chip, plus the extra RAM.

Doom Eternal: Vulkan, Ultra Nightmare, 8x TSSAA

Borderlands 3

While Borderlands 3 used the DirectX 11 API by default on the game's launch in 2019, its DX12 implementation has since left beta status and allows for much better utilisation of modern multi-threaded CPUs. That allows the raw FP32 compute power of the RTX 3080 to achieve incredible results compared to previous-gen cards.

With settings at the punishing 'Bad Ass' preset, we see a noticeable performance improvement even at 1080p, the lowest resolution we tested. The RTX 2080 and 2080 Ti turn in an average of 87 and 114fps, respectively, but the 3080 delivers a crispy 143fps, a lively boost of 64 per cent over the RTX 2080 (and 28 per cent over the RTX 2080 Ti). As we ramp up the resolution, the RTX 3080 is able to increase its lead over the 2080 - we're looking at a 74 per cent gen-on-gen increase at 1440p, rising to 81 per cent at 2160p. That's pretty nuts, but if you're instead coming from the perfectly respectable GTX 1080, you're looking at a 150 per cent frame-rate advantage at 4K, as we transition from a cinematic 24fps to a comfortable 60fps average.

Just a quick note here about the missing 1080p and 1440p numbers for GTX 1080 - we hope to add those soon but in the here and now there's a bug in how Nvidia's driver handles v-sync off in some DX12 games and by extension the way we use video capture to get our benchmarks. This curious issue applies to Borderlands 3 and a couple of other titles in our line-up.

Borderlands 3: Bad Ass, DX12, TAA

Control

Remedy's Control has become the poster child for the advantages of ray tracing, and it's also home to Alex Battaglia's Corridor of Doom - a jaunt in the maintenance section that can bring even powerful graphics cards to their knees. This area sadly isn't easily repeatable, so we're opting for a slightly easier sequence from the start of the game. In this test, we're leaving the RT and DLSS features disabled to give the GTX series cards a fair fight, so this is purely a test of traditional rasterised performance. (If you're more interested in RT performance, don't worry - we'll return to Control with all the bells and whistles enabled later on this very review!)

Like Borderlands 3 and Doom Eternal, there's a clear division between each of the cards we tested: GTX 1080, RTX 2080, RTX 2080 Ti and RTX 3080. Looking at 1080p initially, the RTX 3080 leads the RTX 2080 by a healthy 65 per cent and the GTX 1080 by more than 175 per cent. If we instead consider the RTX 2080 Ti, there's just under a 30 per cent advantage for Nvidia's latest over their previous flagship.

As we raise the resolution, the 3080 is allowed to stretch its legs a little, with a 32 per cent lead over the 2080 Ti and 71 per cent advantage over the vanilla 2080. It's a similar story at 4K, with the 3080 holding a 34 per cent advantage over the 2080 and a 75 per cent lead against the 2080 Ti. Just for fun, let's look at the GTX 1080 - it manages a lowly 18fps at 4K, compared to 58fps for the RTX 3080, a lead of 217 per cent. That's without even accounting for DLSS and RTX, which when enabled together would result in a more accurate image and boost frame-rates even higher. Remember, we'll see those results towards the end of this review!

Control: High, DX12, TAA

Shadow of the Tomb Raider

Shadow of the Tomb Raider concludes our quartet of super-performers. You may well be familiar with the game if you've checked out our reviews before, but if you haven't then it's important to keep in mind that this long integrated benchmark is made up of three scenes of varying difficulty - and therefore, varying degrees of separation between different graphics cards. Therefore, if you have the time and you're on the desktop version of the site, do press play on the embedded YouTube video to see how each scene plays out.

As the oldest game we've seen so far, it's perhaps not surprising that even our 5.0GHz Core i9 10900K CPU bottlenecks the system at 1080p. Here, there's only a 31 per cent advantage for the 3080 over the 2080, with the RTX 2080 Ti coming within 11 per cent of the new flagship. Things change rapidly at higher resolutions though, with 1440p seeing the RTX 3080 move out to a 58 per cent lead over the 2080 and 4K letting the 3080 achieve an average result that's 71 per cent higher than its last-gen counterpart. For a $700 card, that's pretty outstanding.

Shadow of the Tomb Raider: Highest, DX12, TAA

The results so far are outstanding for the RTX 3080, but not all games are able to take such good advantage of the new hardware. Let's take a look at some perhaps more typical performers in our next group of games, which includes two new additions to the Digital Foundry GPU test suite.

Nvidia GeForce RTX 3080 Analysis

• Introduction, Hardware, Power
• Doom Eternal, Control, Borderlands 3, Shadow of the Tomb Raider - Game Benchmarks Part 1 [This Page]
• Gears 5, Far Cry 5, Hitman 2 - Game Benchmarks Part 2
• Metro Exodus, Dirt Rally 2, Assassin's Creed Unity - Game Benchmarks Part 3
• Death Stranding, Ghost Recon Wildlands, Assassin's Creed Odyssey - Game Benchmarks Part 4
• Control, Metro Exodus, Battlefield 5, Quake 2 RTX - RTX Game Benchmarks
• Nvidia GeForce RTX 3080 - The Digital Foundry Verdict

Will you support the Digital Foundry team?

Digital Foundry specialises in technical analysis of gaming hardware and software, using state-of-the-art capture systems and bespoke software to show you how well games and hardware run, visualising precisely what they're capable of. In order to show you what 4K gaming actually looks like we needed to build our own platform to supply high quality 4K video for offline viewing. So we did.

Our videos are multi-gigabyte files and we've chosen a high quality provider to ensure fast downloads. However, that bandwidth isn't free and so we charge a small monthly subscription fee of £4.50. We think it's a small price to pay for unlimited access to top-tier quality encodes of our content. Thank you.

Find out more about the benefits of our Patreon

ncG1vNJzZmivp6x7psHRqJ6apZWne6%2Bx02ibop%2BZqa6tss6upZ2qqWJ%2FcX6PZqWvoZSerm6zxJ%2Bmq5uVYr%2B1xIxsZ3FoXaeyt7XEsHapmZeainM%3D