AMD Radeon RX 5700/ RX 5700 XT review: head-to-head with Nvidia Super

AMD’s long-awaited Navi graphics cards are finally here, combining the firm’s new RDNA architecture with the advanced 7nm process that spawned the Radeon 7. The resultant $349 Radeon 5700 and $399 Radeon 5700 XT are capable of superior frame-rates than last year’s RTX competitors at a similar price point, while also consuming less power and generating less heat than their Vega predecessors.

It looks like an open goal for Team Red, but Nvidia put two new players into the game: the RTX 2060 Super and 2070 Super. These cards are more expensive than their Radeon counterparts – thanks to a price drop from AMD two days before launch – but Nvidia’s new cards are powerful performers in their own right. So, which mid-range card offers the best value for money? That’s what this review intends to answer, as we take a look at how the Radeon RX 5700 and RX 5700 XT measure up against both Nvidia’s original RTX 2060/2070 duo and the firm’s last-minute Super substitutes.

Considered outside the scope of Nvidia’s new releases, the Radeon 5700 cards are impressive. The 7nm process and RDNA architecture means that AMD’s new GPUs produce considerably more graphical horsepower than the 14nm Vega 64 – 1.25x the performance per clock, according to AMD, despite possessing fewer compute units and less memory bandwidth – so gen-on-gen teraflop comparisons mean nothing here, except to emphasise what an impressive job AMD’s engineers have done.

The new Navi cards look impressive too, with a aluminium alloy shroud, a single cooling fan in a ‘blower’ configuration and a seven-phase VRM for overclocking. In terms of power connectors, both cards require one eight-pin and one six-pin input. I/O is quite standard, with three DisplayPort 1.4 connectors and one HDMI 2.0 port; you don’t get the DVI-D input of the RTX 2060 Super or the USB-C VirtualLink port found on both RTX Super cards.

It’s also worth considering the new features AMD is introducing with the RX 5700 series. There’s no hardware-accelerated ray tracing, the hallmark of Nvidia’s RTX series, but AMD introducing a more widely applicable alternative to another RTX feature: DLSS upscaling. Rather than add support for this AI-powered upscaling on a per-game basis, which is Nvidia’s approach, AMD instead relies on a more broadly applicable sharpening effect called Radeon Image Sharpening (RIS). The idea is that you can run games below the native resolution of your monitor to increase frame-rates, then turn on this sharpening feature to get a cleaner upscaled image. RIS is contrast-sensitive, which AMD claims will help it avoid the halo artifacts that crop up in more naive sharpening methods. You can also use RIS at your monitor’s native resolution to counteract the softening effect of anti-aliasing methods like TAA.