AMD Radeon RX Vega 56 benchmarks: the best Vega GPU

The AMD Radeon RX Vega 56 is an upper mid-range graphics card, intended to beat Nvidia’s GeForce GTX 1070 at a similar price point – a job it performs admirably as the benchmarks on this page demonstrate. Its performance level was enough to push Nvidia into creating an entirely new SKU, the GTX 1070 Ti, which slots into place between 1070 and 1080 and in turn outperforms the Vega 56.

The Vega 56 is also the more popular of AMD’s two Vega models thus far, as it’s able to provide almost all of the high-end tech that makes the platform desirable with only a few clever cutbacks to bring its price to a competitive level. It also has a slim advantage against its bigger brother when it comes to power usage and thermal performance, although Nvidia’s Pascal lineup outperforms it in both respects.

The Vega 56 generally provides good performance at 1080p and 1440p, but struggles when stretching to 4K resolution unless paired with a FreeSync display, where variations in frame-rate can be smoothed out somewhat. Provided you have a fast processor (we’d recommend forgoing AMD and picking up a seventh or eighth generation Core i7), the Vega 56 can also work well in concert with high refresh rate 1080p monitors – though owing to high driver overhead under DX11, we’d probably recommend a GTX 1070 or its Ti counterpart for that task over AMD’s offering. Regardless, thanks to the Vega 56’s strong performance offering and reasonable choice, it is still the best AMD graphics card on the market right now.

In order to give you a good idea of what frame-rates you can expect in different games, we’re going to show you benchmarks from nine different games at 1440p and 4K with very High or ultra graphical settings enabled. We’re including five cards in total in our comparison: the Vega 56 and Vega 64, plus an Nvidia trio: the GTX 1070, GTX 1070 Ti and GTX 1080. We’ll also share a cross-generational benchmark, letting you see how the Vega 56 compares to past AMD cards.

Rather than using the traditional benchmark graphs or videos with permanently “burnt in” performance data, we’ve taken a different approach. If you’re viewing this page on desktop rather than mobile, you’ll see that each benchmark is comprised of an embedded YouTube video with live data shown below. You can jump around the video, and the data will remain synchronised. More importantly, you can also use the controls to the right of the video to add or remove different graphics cards at different resolutions. This makes it easy to see just the data you’re interested in.

There’a also a handy graph below the live data, which shows average and best or worst case results from the entire run. You can mouse over the graph to see different metrics, and you can click on the barchart to toggle between real frame-rates and percentage-based values. For more information on the benchmarking system, check out our guide to how the Digital Foundry benchmarking system works right here.

We begin with the 2014 release in Ubisoft’s long-running Assassin’s Creed series, AC Unity. The game is set on the soon-to-be blood-slick streets of Paris on the cusp of revolution, providing opportunities to render fancy dresses and shiny guillotines in equal measure. At 1440p, the Vega 56’s performance is 10 per cent behind the Vega 64 and 13 per cent behind the GTX 1070 it ought to be competing against. The Vega 56 does a little better at 4K though, claiming a tenuous lead over the Nvidia card. However, the actual frame-rate of 26fps is hardly playable. AMD cards suffer here as performance drops significantly whenever the depth of field effect kicks in.

Next up we have Ashes of the Singularity, a 2016 title tested under DirectX 12 and replete with shiny lasers and low-flying spaceships (or high-flying ground-ships?). The Vega 56 heartily outguns the GTX 1070 here, with a sizeable 10 per cent lead at 1440p. Meanwhile, the next better Nvidia card, the GTX 1070 Ti, has a 12 per cent lead over the Vega 56, meaning the AMD card slots in pretty evenly between the two cards in terms of performance. At 4K, the gap between the Vega 56 and the GTX 1070 Ti closes to just five per cent, with the Vega 56 turning in a quite playable 57 frames per second.

2016 release Battlefield 1 and the killing fields of the Great War are where we’re headed next, as we follow a British tank making the journey through no man’s land. The Vega 56 does well here, coming close to the GTX 1070 Ti with 97fps to 102fps, a deficit of just four per cent. At 4K there’s a big gap between the two cards, with the Vega 56 at 56 frames per second and the GTX 1070 Ti sitting 12 per cent behind.

Remarkably, Crysis 3 somehow remains a challenge for modern graphics cards more than five years after its debut in 2013. The Vega 56 is on a level pegging with the GTX 1070 here, sitting just a frame per second behind at both the resolutions we tested. The 1440p result of 67 frames per second is outstanding, while the 4K result of 32 frames per second is perhaps a little more ‘cinematic’ but still quite playable.

The Division’s snowy metropolis tends to render faster on AMD hardware, and we see that the Vega 56 is able to eclipse the the GTX 1070 by 13 per cent at 1440p and 14 per cent at 4K. The 4K frame-rate of 38fps is not the fastest we’ve seen, but it should be broadly playable if the settings are turned down from the Ultra preset, FreeSync is turned on, or ideally both.

Far Cry Primal, the 2016 stepping stone between Far Cry 4 and Far Cry 5, is our next benchmark. Despite this title’s slight preference for Nvidia hardware, the Vega 56 comes close to the GTX 1070 Ti in the benchmark. The AMD card provides an effortless 70 frames per second at 1440p and a less comfy 39 frames per second at 4K. These results are eight per cent and twelve per cent faster than the GTX 1070 at 1440p and 4K, respectively.

Our most recent and most demanding title is Ghost Recon Wildlands, released in 2017. On the challenging ultra preset, the Vega 56 manages to just score 44 frames per second, between the GTX 1070 at 40 and the GTX 1070 Ti at 45. Meanwhile, the Vega 64 sits three frames per second ahead. At 4K, the Vega 56 can only muster 27 frames per second. To be honest, the high or very high presets are recommended for essentially all cards targeting 4K.

2016 saw the return of born survivor Lara Croft in Rise of the Tomb Raider. The three-part benchmark is tested under the DX12 render path. The Vega 56 boasts a 10 per cent lead over the GTX 1070 regardless, and sits close to the GTX 1070 Ti once again with a deficit of just four per cent. The game is also playable at 4K based on these results, with a frame-rate of 42fps, ten per cent better than the GTX 1070 can manage. What we will say is that this benchmark flatters to deceive – in actual gameplay, careful settings management is required to get decent performance.

2015 classic The Witcher 3 is another challenging title at its ultra preset, particularly at 4K. The Vega 56 easily exceeds the 60 frames per second we’d want to see at 1440p, but only manages to deliver 42 frames at 4K. However, that 4K result is still a comfortable 10 per cent ahead of its GTX 1070 competition. It even passes the GTX 1070 Ti at the same test by a small but solid five per cent.

We’ll conclude with a look at how the Vega 56 and Vega 56 compare to their predecessors in Assassin’s Creed Unity. Going up against the Vega cards in this unfair fight are the cheaper RX 570 and RX 580 and AMD’s last high-end card, the R9 Fury X. You can see that the R9 Fury X is 12 per cent behind the Vega 56, while the RX 580 is considerably slower with a frame-rate that’s 23 per cent behind the Vega 56.

With that, our look at the benchmarks of the Vega 56 are finally complete. We encourage you to check out our full review of the Vega 56.

Now that you’ve seen the benchmarks for one card, why not check out see which GPUs we recommend? Click through to see Digital Foundry’s updated selections for the best graphics cards and for the best gaming monitors of 2018.

Leave a Reply

Your email address will not be published. Required fields are marked *