New DX12 Benchmark: 1300% Better Than DirectX 11

This is some serious performance gains.

Posted By | On 27th, Mar. 2015 Under News | Follow This Author @@Martintoney2012


DirectX 12. Get used to seeing that splashed everywhere in the coming months, because if the reports are to be believed, it’s going to have a massive impact for PC games development. 3DMark, the extremely popular cross platform benchmark, released  DX12 Vs DX11 Vs Mantle benchmarking software. PC World then picked it up and ran with it. The final results were far better than what many had predicted.

During their tests, PC World found that the AMD R9 290X saw a speed of 1300% in 3DMark’s API test, with Mantle scoring 1200% over the DX11 API test score. The DX12 API score however, takes the lead by a considerable margin. PC World also went ahead and swapped out the R9 290X for Nvidia’s beefy GTX Titan X and found that the two ran almost identically, despite Nvidia’s GTX Titan X costing nearly three times as much as the R9.

Intel’s Iris Pro Integrated GPU saw a boost in performance of X4 and the test revealed that Intel’s Quad Core CPUs (with Hyper Threading on) gave the fastest performance, with a noticeable difference in the DX12 Multi Threaded tests. Obviously, this is all fantastic stuff. But this is a benchmark, real world results are likely to vary from machine to machine. But all said, this is good news.


Awesome Stuff that you might be interested in

  • Psionicinversion

    Thats basically a draw calls test so basically DX12 allows for “unlimited” drawcalls atm 😀

  • Oxybelis
  • James

    Draw Calls do not determine real-time performance. More draw calls actually create overhead, so the goal wouldn’t be to create as many as possible, but to use them as efficiently as possible. Synthetic benchmarks are good at measuring single precision/double precision compute performance, but they’re not so good at representing an actual video game(s) performing.

    • Brian Blair

      Shh! You are not supposed to make logical comments! You are supposed to just pretend and go along with it! And then worship them! How dare you have independant thought! You must be a conspiracy theorist! Or a terroist!

  • Graeme Willy

    3DMark only implemented the CPU overhead test, since that’s what DX12 and Mantle are all about. Even in DX11, there’s very little in the way of overhead on the GPU side. Which is why they claim that DX12 may, or may not make a difference in the GPU department. In the overhead test, the GPU is only used to reference/ demonstrate how many draw calls the CPU was able to issue to the GPU. I don’t think there’s an actual DX12 GPU-centric benchmark, yet.


Copyright © 2009-2017 All Rights Reserved.