AMD’s Radeon RX Vega has been the talk of the town this week following the official reveal of the Vega 64 and Vega 56 on Monday morning. AMD has been extremely cautious to not let slip any relevant performance benchmarks to the public, and they have intentionally made their slides as vague as possible, either to avoid tipping off NVIDIA, or to avoid being taken to task by their fans should they fail. Early performance numbers always manage to find their way onto the internet though, and Tweaktown managed to get their hands on some of these from a close source. The results are quite interesting, so hit the jump to see what’s up!
|Battlefield 1||Civilization VI||Call of Duty Infinite Warfare||DOOM|
|NVIDIA GeForce GTX 1080 8GB||99.5 fps||75.3 fps||104.9 fps||99.1 fps|
|AMD Radeon RX Vega 56 8GB||95.4 fps||85.1 fps||99.9 fps **||101.2 fps|
|AMD Radeon R9 Fury X 4GB||84.0 fps||57.9 fps||85.2 fps||89.5 fps|
|NVIDIA GeForce GTX 1070 8GB||81.3 fps||74.5 fps||86.3 fps||78.4 fps|
|NVIDIA GeForce GTX 980 Ti 6GB||71.5 fps||70.0 fps||75.9 fps||73.3 fps|
The table above includes two sources of data. The Radeon RX Vega 56 numbers come from Tweaktown’s “industry source” who provided them with the benchmark numbers as well as the tested settings. The card was tested on a machine with an Intel Core i7-7700K, 16GB of G.Skill RAM, and Windows 10 64-bit. However, I decided to not include the source’s comparative numbers for the GTX 1070 they also claimed to have tested, because there is an error in the data that shows identical performance in two games on the same card. I instead took recent and relevant results from TechpowerUp’s review of the Zotac GeForce GTX 1080 Mini and included them in here. All the results from the other cards come from that review.
The results do look quite good, considering that AMD has said nothing about the RX Vega 56’s performance compared to the GTX 1070, let alone the more expensive GTX 1080. If they are really preparing a GTX 1080 killer at the $399 price point, then they’ve kept things extremely quiet on the best value card in the RX Vega family. Pulling in wins in Civilization VI, a game which has traditionally shown bad performance with AMD hardware, and DOOM, is nothing to sniff at. Even the Battlefield 1 numbers are quite impressive. If this is an accurate representation of what AMD’s going to be delivering to its fans this month, then I think there’s no need to worry about the success of Vega or Radeon Technologies group.
The outlier in the results is Call of Duty: Infinite Warfare. Tweaktown noted that their source told them that the tests for Infinite Warfare were run at the High preset, instead of the Ultra preset that most other review sites roll with. The only website I know that runs tests at the high preset is GameGPU, but I can also recall that AMD’s prescribed settings for Infinite Warfare require that you use the High preset instead of Ultra.
For the moment, no-one has come forth with any similar performance claims, and it may all be made up and stuck on the internet to see how much people fuss over it. It wouldn’t be the first time that someone released fake benchmark numbers just to see if websites will pick up on an obviously fake story without checking their data first. Who knows.
But if the numbers are correct, and if it games as well as Tweaktown’s source claims, well… we’ll have a bloody price war on our hands coming very soon! Competition is always good for everyone involved.