AMD has revealed its full range of RX 6000 graphics cards. Dubbed “Big Navi,” these new GPUs are directly competing with Nvidia’s latest RTX 30- series, with the RTX 3070 launching soon. AMD’s latest range starts with the new RX 6800XT, a massive card that features 16GB of GDDR6 memory, 72 compute units, a base clock of 2015MHz, and a boost clock of 2250MHz. It’s also less power hungry than Nvidia’s RTX 3080, requiring only 300W (a difference of 20W). The RX 6800 keeps the same 16GB of GDDR6 memory, but drops compute units down to 60 while also reducing base and boost clocks to 1815MHz and 2105MHz respectively.
The news comes hot on the heels of its Ryzen 5000 series CPU announcement earlier in October.
AMD’s own benchmarks positions the RX 6800XT as a direct competitor to the RTX 3080, while the RX 6800 is compared to the older RTX 2080Ti. In both sets of results below there are instances where Nvidia’s cards are ahead, AMD is confident that its cards either match or better its competition in the majority of cases. AMD omitted any comparisons to the RTX 3070, but given its delayed launch date that isn’t too surprising.
Last and rounding out the 6000 series is the RX 6900XT, AMD’s answer to the massive RTX 3090. It also keeps the same 16GB of GDDR6 memory as well as the base and boost clock speeds of the RX 6800XT, but increases total compute units to 80. This gives the RX 6900XT more focused gaming performance when compared to the RTX 3090, despite AMD’s own benchmarks showing mixed results in a direct comparison.
The entire range of cards is out soon, too. The RX 6800XT launches on November 18 for $649, while the RX 6800 will retail for $579 and launch on the same day. The flagship, the RX 6900XT, launches on December 5 for $999.
AMD is also launching software features for the 6800XT that allow for one-click overclocking and more optimized memory delegation when using the new GPU with AMD’s latest Ryzen 5000 CPUs (similar to DirectX 12’s DirectStorage feature that will start being supported next year). These two features allow performance gains of up to 13% in some titles, according to AMD.
AMD has chosen to keep the same memory across its entire range of cards but also not go with the faster GDDR6X memory that Nvidia decided on for its RTX 30-series line. AMD says that with its own Infinity Cache technology, it’s able to boost memory bandwidth performance and address the bottlenecks created by increasing resolutions to 4K without having to resort to more expensive memory. This means that all of AMD’s cards feature more memory than most of Nvidia’s line, with the RTX 3070 featuring 8GB, the RTX 3080 with 10GB, and only the RTX 3090 beating the lot with 24GB.
The biggest omission from the event was AMD’s answer to Nvidia’s Deep-Learning Super Sampling technology, or DLSS. This allows Nvidia GPU to render games at lower internal resolutions and upscale them using AI, producing nearly indistinguishable difference and increasing performance significantly. Speaking to The Verge, AMD says that it is currently working on a super sampling solution that it will make open and cross-platform when ready, while also ensuring that the feature will have widespread support when it launches. That won’t be this year, but AMD hopes it will be soon and also support both the Xbox Series X and PS5.
AMD is set to have a busy November, with both the RX 6800XT and RX 6800 launching on November 18 for $649 and $549 respectively. The flagship, the RX 6900XT, launches on December 5 for $999. AMD is also launching its Ryzen 5000 series CPUs starting on November 5.
Now Playing: Nvidia RTX 3080 Review