Home / BBQ Opinions / Ashes of The Singularity DX12 Benchmark is anything but credible

Ashes of The Singularity DX12 Benchmark is anything but credible

We purchased three games for graphic card benchmarking- Ashes of The Singularity, DOOM and Hitman (2016).

I will not be using AOTS.

The game’S DX12 benchmark is ridiculously hyped but it doesn’t seem to have consistent quality with DX12. I ran the DX11 and DX12 variant with the GTX 1080 that I have for review (finally), but it hardly makes any noticeable difference between the presets. Though we don’t have any AMD GPUs and other Nvidia to compare it with and reach a conclusion, I suspect that the game has trouble with terrain shaders, which may explain why the gameplay quality looks ‘old’. That said, anyone bragging about their graphic card using Ashes of Singularity benchmark should be taken with a grain of salt. Just because this is the first DX12 that can be benchmarked, doesn’t mean it is doing a proper job.

Moreover, DX12 benchmark does not work but ironically the game can run in DX12. That sounds completely silly and Stardock doesn’t have a solid fix. -Observations were made using DX11 benchmark and the render quality in DX12 during playing. Hitman and ROTR in DX11 and DX12 show a decent amount of difference (and DOOM is OpenGL based) at least with a DX12 compatible card that I have. That said, we don’t get AMD cards now to make a proper side-by-side comparison.  AOS’s definition of the highest setting (crazy) profile is completely silly because none of the details are preset to a maximum setting, including anti-aliasing and texture detail rendering. Sure, we can use the custom setting and set everything at max, but it just doesn’t feel right to use this game for benchmarking. Why call the preset as crazy when details are set to high and not ultra??

I’ve played RTS like Caesar III, AOE, AOE II and AOM- and those games were a lot more entertaining than AOS. Alas, we got aboard the hype train (happens to the best of us) and found this game’s graphics is a bunch of hooey.  The AOTS that we used was V1.12, and judging by the AOTS related content and benchmark, was well before its release date- 26th May. AMD also was under fire for the rendering quality video of the game when talking about RX 480 which was then clarified on Reddit that the crazy preset did not include the highest detailing  in ‘Crazy’ preset. Unfortunately, there is no option to roll back (something that we should have by now) the patch updates.

The game workload in our benchmark will be as follows:

  • Hitman 2016 (DX11 and DX12)
  • ROTR (DX11 and DX12)
  • DOOM (OpenGL)

These games are chosen simply because of the amount of fanbase it has, and that it is multi API. DOOM will be having Vulkan API support at some point as well. Other workloads will stay the same and 3DMark API overhead test under Futuremark benchmarks will be used from now onwards. If AOTS is fixed in the near future, we’ll include it. That would seem unlikely since people would find it difficult to relate with the game.

Leave a Reply

Your email address will not be published. Required fields are marked *

*