Nvidia RTX 3080 Review

Nvidia’s next generation of video cards is here, and our introduction to the new RTX 30-series comes to us in the form of the RTX 3080. It’s been about two years since the RTX 20-series launched, and along with it came the ability for ray traced lighting, shadows, and reflections in games, and AI-driven image quality improvements. Those features were in their infancy, but things have evolved, and the RTX 30-series is a way to embrace these innovations in PC gaming graphics.

We have the RTX 3080 and put it to the test to see exactly how it stacks against its previous-generation counterparts. With Nvidia claiming very high performance jumps compared to the RTX 2080, and the RTX 2080 Ti being the most powerful consumer-grade card of that generation, it makes sense to see how the RTX 3080 stacks against them in particular. When looking at the following results, keep in mind that the RTX 3080 launches with a $700 USD suggested retail price, similar to the RTX 2080, while the RTX 2080 Ti goes for upwards of $1000.

For this review, we have Founder’s Editions of each card which were provided by Nvidia ahead of the RTX 3080’s September 17 launch date.

Technical Details

The new RTX 30-series cards are powered by the Ampere architecture, which is an evolution of the 20-series Turing architecture. They share similar approaches with three processing cores to take care of different tasks: the main Streaming Multiprocessor (SM), the Tensor Core for AI-based tasks and deep learning technologies, and the RT Core that’s dedicated to real-time ray tracing. Ampere packs more of all that, but enhanced for better performance and efficiency this generation, and also moves to faster GDDR6X for video memory.

A few other things have changed with the RTX 3080 as well. The power pins are tucked into the side of the card and require you to use the packaged dongle that connects to the card’s miniature 12-pin slot and lets you hook in the two necessary 8-pin power cables. For the RTX 3080, it’s recommended to have a power supply that provides at least 750 watts; the card can consume up to 320 watts under load (without any manual adjustments), which is quite the jump from its predecessors.

You may also notice a new type of cooling solution. The RTX 3080 features two independent push-and-pull fans, one on each side of the card–the topside fan pushes air through the card upwards while the underside fan brings air into the card to be blown out the rear exhaust vents. As for ports, the card has three DisplayPort 1.4a outputs and one HDMI 2.1 output. It’s worth noting that Nvidia hasn’t used the same metrics for specs between the two generations, making it a bit complicated to directly compare cards, but for the more granular technical specs, check the table below.

Methodology And Test Bench

Now, let’s establish our methodology. With our situation working from home, we had access to limited resources, but managed to put together a well-suited test bench. This includes the following specs:

• MSI Mag Z490 Tomahawk motherboard
• Intel Core i7-10700 (non-K) 8-core/16-thread CPU clocked at 4.7GHz
  • Corsair H100i AIO Liquid Cooler
• 16GB (2x8GB dual-channel) Corsair Vengeance DDR4-3000 RAM
• Samsung 960 Pro 512GB NVMe SSD
• Seasonic Prime 1000W (80 Plus Platinum) PSU

Our tests included three video cards: the RTX 3080, of course, the RTX 2080 Ti, and the base RTX 2080, all Founder’s Editions from Nvidia. The driver version supplied for testing was the GeForce Game Ready 456.16 driver.

As for benchmarks, we used 10 different PC games to compare performance between our three video cards. These included Metro Exodus, Control, Shadow of the Tomb Raider, Wolfenstein: Youngblood, Final Fantasy XV, Final Fantasy XIV: Shadowbringers, Red Dead Redemption 2, Assassin’s Creed Odyssey, Horizon Zero Dawn, and Forza Horizon 4. All of these are modern games, some more graphically demanding than others, but they represent a good variety to test things like RTX ray tracing and DLSS supersampling, as well as raw performance metrics.

In this review, we focus on performance with max settings and Nvidia graphics features turned on when available at both 4K and 1440p resolutions. And we measure each card in average FPS calculated by the game’s benchmark tools (except for Control, which was done using FRAPS). Keep in mind that you always have options to adjust graphics features to improve with FPS output in your own situations–here, we’re trying to see how the cards compare when pushed as far as games let them.

Benchmark Results

With the important technical details out of the way, let’s get into the numbers–lots of numbers.

Metro Exodus

Metro Exodus is one of the most demanding games out there, and it includes full ray tracing options as well as DLSS. And the built-in benchmark is a good stress test.

At 4K with every setting maxed, average framerate came out to 51.6 FPS for the RTX 3080, 41.4 FPS for the RTX 2080 Ti, and 31.5 FPS for the RTX 2080. This comes out to a 24.6% increase over the 2080 Ti and a 63.8% improvement over the 2080. At 1440p, average framerate came out to 67.2 FPS for the 3080, 55.8 FPS for the 2080 Ti, and 43.4 FPS for the 2080–a 20.4% and 54.8% increase over the 2080 Ti and 2080 respectively.

Control

Control is visually enticing, demanding, and full of graphics options. We turned every ray tracing option on, maxed out settings, and used DLSS. Since Control doesn’t have a built-in benchmark, we made our own by running through an early section of the Old House that was easily repeatable and used FRAPS to record the average FPS.

At 4K, average framerate came out to 64 FPS on the RTX 3080, 47.4 FPS on the 2080 Ti, and 31.5 FPS on the 2080. This makes for a 35% jump from the 2080 Ti and a whopping 103.2% increase over the 2080. At 1440p, average FPS counts came out to 104.6 FPS for the 3080, 82.3 FPS for the 2080 Ti, and 65.5 for the 2080, which marks a 27.1% and 59.7% increase over the respective cards.

Shadow of the Tomb Raider

Shadow for the Tomb Raider remains a graphical showcase with lush, detailed environments, and features ray tracing and DLSS.

Running at 4K with everything maxed, the RTX 3080 got a 70 FPS average, while the 2080 Ti got 57 FPS and the 2080 got 43 FPS. These are 22.8% and 62.8% increase over the 2080 Ti and 2080 respectively. Running at 1440p, the 3080 hit an average of 100 FPS, while the 2080 Ti got 86 FPS and the 2080 got 67 FPS, which comes out to 16.3% and 49.5% increases over the respective cards.

Wolfenstein Youngblood

Wolfenstein Youngblood isn’t the most graphically demanding game but here we can measure performance in the idTech 6 engine with ray tracing and DLSS. We used the Riverside benchmark run with everything at max settings.

At 4K, we got averages of 117 FPS for the 3080, 89 FPS for the 2080 Ti, and 69 for the 2080. These results make for a jump of 31.5% and 69.6% with the 3080. At 1440p, averages came out to 171 FPS for the 3080, 144 FPS for the 2080 Ti, and 118 for the 2080, meaning smaller gains of 18.8% and 45% with the 3080.

Final Fantasy XV

Final Fantasy XV is a graphical powerhouse with a robust benchmark test. While it doesn’t have ray tracing, it has other Nvidia features including DLSS. For 1440p, the FFXV bench does not allow DLSS, so here we just used the “High Quality” settings preset, which should be judged a different scale in comparison to our 4K results. Note that this benchmark does not give average FPS, rather produces a total score.

At 4K max settings, the 3080 scored 7281, the 2080 Ti scored 6061, and the 2080 scored 4720. This comes out to 20.1% and 54.3% performance increases over the 2080 Ti and 2080. Running 1440p, the 3080 scored 11381, while the 2080 Ti scored 9963, and the 2080 scored 8341, which comes out to 14.2% and 36.4% performance increases.

Final Fantasy XIV: Shadowbringers

Well, what can I say. Shadowbringers is the best expansion for an all-time favorite game in Final Fantasy XIV. You know I had to do it to ’em. There’s a free benchmark tool out there, too. This one also produces a score rather than an average FPS.

At 4K max, the RTX 3080 scored 14138, the 2080 Ti got 11578, and the 2080 got 8950. These scores indicate performance jumps of 22.1% and 58% for the 3080. At 1440p, we have the 3080 at 17626, the 2080 Ti at 16914, and the 2080 at 15667. Here, we see a lower spec game like FFXIV reach its upper limit of GPU reliance and become more CPU bound with minimal increases between cards.

Red Dead Redemption 2

Red Dead Redemption 2 is certainly graphically demanding with a slew of options to help tweak your experience, and an intense benchmark tool to get real scenario performance. We turned everything up to the max, but just went with TAA for antialiasing.

At 4K, average framerate came out to 43.1 FPS for the 3080, 37.1 FPS for the 2080 Ti, and 28.3 FPS for the 2080. This marks increases of 16.2% and 52.3% for the RTX 3080. At 1440p, average framerate came out to 62.9 FPS with the 3080, 52 FPS with the 2080 Ti, and 39 FPS with the 2080. These numbers indicate increases of 21% and 61.3% for the RTX 3080.

Assassin’s Creed Odyssey

Assassin’s Creed Odyssey remains a visually stunning and demanding game across a vast and detailed open-world version of Ancient Greece. Though it tends to be CPU-heavy, it’s still important to see how high-end video cards fare here.

At 4K, the 3080 got an average of 62 FPS, while the 2080 Ti got 52 FPS and the 2080 got 41 FPS. This comes out to respective gains of 19.2% and 51.2% in performance. Using 1440p, the 3080 hit an average of 78 FPS, while the 2080 Ti got 69 FPS and the 2080 got 62 FPS. Like FFXIV, Assassin’s Creed Odyssey at 1440p starts to show where the workload leans heavier into the CPU, limiting the effectiveness of a beefy GPU.

Horizon Zero Dawn

Horizon Zero Dawn just got ported to the PC with mixed results, though it still can look incredible. It’s had some improvements since its launch and has a decent benchmark tool to stress our video cards.

At 4K, average FPS counts came out to 75 FPS for the 3080, 57 FPS for the 2080 Ti, and 45 FPS for the 2080. These results show 31.6% and 66.7% increases for the 3080. Using 1440p, average FPS numbers were 109 FPS with the 3080, 96 FPS with the 2080 Ti, and 78 FPS with the 2080, coming out to 13.5% and 39.7% jumps for the 3080 over the respective cards.

Forza Horizon 4

And finally our last test, Forza Horizon 4, a beautiful open world racing game with a wildly detailed benchmark tool.

At 4K everything maxed, including MSAA, the 3080 hit a 96 average FPS while the 2080 Ti hit 75 FPS and the 2080 hit 61 FPS. These are respective jumps of 28% and 50.8% for the 3080. Going to 1440p, average FPS counts came out to 129 FPS for the 3080, 112 for the 2080 Ti, 95 FPS for the 2080, which comes out to 15.2% and 35.8% with the RTX 3080.

Heat Check

In terms of heat, the RTX 3080 stays relatively cool even under intense workloads. The card’s idle temperature sat around 35 C and peak load temperature never exceeded 75 C. Even with the high power draw, the RTX 3080 remains around the temps of the RTX 2080 Ti and RTX 2080, which speaks to the design of the new card’s cooling system. For reference, the RTX 2080 Ti’s idle temps were around 34 C and peak load temp didn’t exceed 78 C, while the RTX 2080 would idle at about 33 C and hit peak load temp at 74 C.

The Impact of DLSS And Ray Tracing

Of course, ray tracing is one biggest factors with the RTX cards, and while the Turing and Ampere architecture make this possible with the RT Core, it still impacts framerate. To get an idea of how much it affects performance, we tested both Metro Exodus and Control with ray tracing options maxed and without any ray tracing options enabled.

However, the real MVP of the RTX cards is the Tensor Core that makes DLSS possible. Deep learning supersampling allows you to render your game at a slightly lower resolution, then applies supersampling to project the target resolution using the separate AI-driven processor on the card. The result is image quality that’s either on par or better than simply using native resolution while taking the entire antialiasing/supersampling workload off the video card’s main processor and asking it render a lower resolution. It’s a win-win because it provides a major boost in performance, too. Just look at the comparison in Metro Exodus and Control using ray tracing with DLSS on and off.

Essentially, DLSS makes it possible to yield high, playable framerates while running 4K with ray tracing enabled. DLSS alone is a strong selling point for RTX cards now that more games have adopted the feature.

Verdict

So after all those numbers, benchmarks, and graphics settings, the takeaway is this: The RTX 3080 is a powerful video card with big generational jumps, especially when considering its $700 MSRP. It shows decent gains over the RTX 2080 Ti which came out of the gate much more expensive and shows significant improvements over the base RTX 2080. Nvidia’s claim of “up to twice as fast as the 2080” is a bit of a stretch, but tempered expectations by saying that we’d see closer to 60% to 70% gains in games, and that much is true based on our test results.

Now, whether or not you should get an RTX 3080 depends on how you want to run your games. It’s your best option for 4K, capable of 60 FPS at max settings, with little to no compromises. However, the improvements tend to taper off at 1440p. Regardless, it’s a better value than the 2080 Ti and 2080 at this point, even if the gains aren’t as apparent at 1440p.

But if you don’t care about 4K or aren’t trying your hardest to get the most out of a high refresh rate 1440p monitor, I’d consider the RTX 3080 overkill, for now. If those aspects aren’t your jam, you probably want to keep your eye on the upcoming RTX 3070. Either way, if you’re still on the GTX 10 series, this might be a good generation to hop with great features like DLSS for simultaneous image quality and performance improvements, and ray tracing which is increasingly relevant in games.

The RTX 3080 is an incredible card, but just make sure it’s something that you’ll actually get the most out of.