Doom Vulkan Benchmarks: Up to 40% Performance Boost for Radeon Cards, GTX 1070 Mixed
The PC version of Doom 2016 has finally received the Vulkan support. The update offers significant boost in performance, especially for AMD owners who are getting some game-changing improvements. Initial Doom Vulkan vs OpenGL benchmarks suggest that AMD Radeon cards see roughly 30 to 40 per cent increase in gaming performance but this could be higher.
For those who don’t know, Vulkan is a low-overhead API derived from AMD’s Mantle, offering many of the same advantages as the Microsoft’s DirectX 12. The async compute in particular sees big gains for Radeon hardware, and the feature is used extensively in Doom.
id Software itself is pretty vocal of the potential of async compute. Lead programmer Billy Khan said in a statement:
The time is now, really. Doom is already a clear example where async compute, when used properly, can make drastic enhancements to the performance and look of a game. Going forward, compute and async compute will be even more extensively used for idTech6. It is almost certain that more developers will take advantage of compute and async compute as they discover how to effectively use it in their games.
Doom Vulkan vs OpenGL benchmarks: Fury X vs GTX 1080 / GTX 1070 / GTX 980 Ti And More!
Digital Foundry has just released a comparison video to show Doom performance using four high-end GPU: GTX 1080, GTX 1070, GTX 980 Ti and R9 Fury X.
We can see in the video that AMD’s Fury X underperforms on OpenGL against its nearest rivals, the GTX 1070 and GTX 980 Ti however, the performance is significantly improved when Vulkan is introduced. In fact, the Fury X moves ahead of both GeForce cards.
For the Polaris based Radeon RX 480, the results remain the same. The RX 480 performs seven per cent slower than the GeForce GTX 970, and 24 per cent than the GTX 980. As soon as Vulkan is engaged, the RX 480 beats the GTX 970, delivering performance on par with the GTX 980.
What are your thoughts on the advantages of Vulkan and async compute? Let us know in the comments below.