Nvidia GeForce RTX 3050 review: revisiting the super-performers

We begin with games that we call 'super-performers' because they use relatively modern graphic APIs like DX12 or Vulkan - so they're better equipped to make full use of high-end CPUs like the Core i9 10900K in our test system. That lets our GPU run at full speed, hitting higher frame-rates than we'd see in more CPU-limited titles that rely on a single thread for the majority of their game processing, and making the gaps between different levels of graphics performance more apparent. Nvidia's upper-tier RTX 30-series cards have been the most impressive here, but I wouldn't rule out the RTX 3050 also outperforming to a small degree.

Note that our benchmark results are more than simple bar charts. On desktop, browsers, you'll get embedded YouTube videos of each test scene. Press play on the video to replay our benchmark run, and see how each card handles the scene with live frame-time and frame-rate metrics. You can even choose exactly what GPUs at what resolutions you're interested in by ticking and unticking the boxes to the right side of the benchmark widget. Below the real-time stuff is a bar chart, which you can mouse over to see different measurements and click to switch between frame-rates and percentage differences. All the data here is derived from video captured directly from each GPU, ensuring an pinpoint accurate replay of real performance. On mobile browsers, you'll get a simplified view - just a table of the lowest one percent and average frame-rates for each card at each resolution; we recommend returning with a desktop browser to get the full picture!

With that crash course out of the way, let's see how the RTX 3050 performs - we're hoping for something akin to the RTX 2060, but performance may be closer to the GTX 1660 Super - let's see.

Doom Eternal

Doom Eternal isn't a difficult game to run at high frame-rates, but its Ultra Nightmare preset does enact a harsh toll on sub-standard video memory allocations. The RTX 3050's 8GB holds it in good stead, despite rather limited memory bandwidth, with an average of 156fps at 1080p and 109fps at 1440p - only 4K requires turning down settings as we only reach an average of 53fps in our test scene. Moving back to 1080p, the focus of this card, the RTX 3060 outperforms the 3050 by 44 percent, which is a pretty sizeable margin for neighbouring cards in Nvidia's stack. The RX 6600 is also significantly faster than the 3050 here, by 26 percent, but the lower-tier RX 6500 XT can't run the game at all due to its 4GB of VRAM.

Borderlands 3

Good old Borderlands 3 comes next, and here we have our first look at performance from the RTX 2060 and GTX 1660 Super - cards that couldn't be retested in Doom Eternal thanks to their lower VRAM allocations. The 3050 is 10 percent faster than the GTX 1660 Super, but about the same amount slower than the RTX 2060. The game is still more or less playable at 1440p, but at 38fps you'd be best served going for significantly lower graphical settings. However, that's still far better than the cheaper RX 6500 XT, which manages 12.7fps average at the same 1440p resolution.

Control

Control's best known for its excellent ray tracing implementation, but its traditional rasterised gameplay is still remarkably challenging at its high preset. Let's take a look at performance in its non-ray-traced mode then, at 1080p. The RTX 2060 is 22 percent faster than the 3050, a pretty significant difference as the cheaper card dips below the 60fps mark more frequently. The RX 6600 costs more than the 3050, but also delivers better performance - 27 percent faster, to be exact. It'll be interesting to see how that differential shifts when RT is enabled - any predictions? Turn to page five to find out...

Shadow of the Tomb Raider

Shadow's integrated benchmark includes a party scene, a jungle scene and a village scene, each with different performance characteristics. It's rare to see most outlets test more than one scene in a given game - including us - due to the significant amount of time each game takes to test in every graphics card, so seeing how cards can swap places in the different phases of this test underlines the importance of seeking out multiple sources to get a fuller picture. After all, reviewers could get legitimately opposing results when comparing two cards, just by choosing different scenes, so looking at a plurality of reviews should provide a more balanced outlook.

The RTX 3050 scores an average of 90fps here at 1080p, a pretty comfortable experience, and is dead-level on a 60fps average at 1440p. The RTX 2060 is 14 percent faster, while the 2060 Super extends the margin to 31 percent. The difference between the 3050 and the next card up is also pretty noticeable - 39 percent at 1440p.

Compared to the RX 6500 XT we reviewed earlier this week, the 3050 is much better - 55 percent faster at 1080p, 65 percent faster at 1440p and 120 percent faster (!) at 4K. Given that the Nvidia card is ostensibly 25 percent more expensive, that's a pretty good deal - especially as you get DLSS in the bargain.

So far, the RTX 3050 looks to be a capable card - but let's take a look at some games that aren't so kind of Nvidia's Ampere architecture.

Nvidia GeForce RTX 3050 analysis

• Introduction and hardware analysis
• Doom Eternal, Control, Borderlands 3, Shadow of the Tomb Raider - Game Benchmarks Part 1 [This Page]
• Death Stranding, Far Cry 5, Hitman 2, Assassin's Creed Odyssey - Game Benchmarks Part 2
• Metro Exodus, Dirt Rally 2, Assassin's Creed Unity - Game Benchmarks Part 3
• Control, Metro Exodus, Battlefield 5 - RT game benchmarks
• Nvidia GeForce RTX 3050 - the Digital Foundry verdict