No DX12 Software is Suitable for Benchmarking *spawn*

What GPU?

 

I’d guess if you’re not VRAM limited you won’t see much improvement except maybe from some bandwidth saving if any of the test is bandwidth limited.

 

Redundant feature ?

 

This test has nothing to do with VRAM, as it's using the Texture Space Shading part of Sampler Feedback, not the more interesting streaming technique!

Performance varies greatly though depending how the base performance of the benchmark is. I got nearly a 20% more performance while my 2060 laptop was on battery.

 

I'd imagine that the more shading power a GPU has the lower the gains from using texture space shading will be.

 

Sampler feedback - just like the original tiled resources - have always been in a weird space. Even the theoretical benefit is pretty marginal... it's basically cases where you have an anisotropic kernel walk across the edge of a tile boundary and you're unwilling to do borders. Obviously borders are usually completely reasonable for texture data.

For texture-space shading you may want to expend a bit of effort to avoid shading border pixels but it's still kinda in the noise IMO. Without more details on what the "software solution" is and why they need a software *sampling* solution in the second pass (presumably because they are doing indirections per tap instead of borders?), I'm not sure if there's a meaningful way to interpret the results of the test to be honest.

 

Nv's 496.13 driver seem to add about 10% of performance for Ampere in AC Valhalla.

 

https://developer.nvidia.com/blog/tag/advanced-api-performance/

 

Despite all of these years, DX12 remains slower than DX11, even in a recent game like Kena, @4K, the 3080 is 15% slower in DX12 than DX11, 6800XT is 7% slower.
https://www.computerbase.de/2021-10.../#diagramm-directx-11-vs-directx-12-3840-2160

 

They do mention, however, that the game has frame pacing issues and that they're worse in DX11

 

UE4 isn't great. Most developers who use their own engine achieve better performance on AMD GPUs.

 

UE4 is open sourced to AMD (and other IHVs), they can and do submit into it anything which they need to get better performance.
It is in fact a much more fair engine for assessing cross-vendor performance than whatever "most developers" do for AMD based consoles.

 

Indeed, but that is an optimization issue, not an API issue, the developer should go ahead and optimize the DX11 path more and the issue should go away, many console games suffered this exact issue at launch, and the developers fixed it later with patches.

 

No, the issue is of reduced DX12 fps plagued many engines even on AMD hardware, have a tour in this thread and see for yourself.

 

It happens yes, but more than 50% of the time(most) DX12 is faster for AMD.

 

This is simply not true, out of my memory, most of the games that featured both APIs worked slower on both AMD and NVIDIA in DX12.

 

When top CPUs were used.

 

6 years ago, DX12 came with the promise of reducing CPU load, while also increasing scene complexity, we achieved little of that sadly. There are still games that suffer fps losses due to DX12 to this day. Even when the game has DXR, the developers can't get the DX12 to perform better than DX11 in many cases, I recently tried the game Deliver Us The Moon without DXR, and DX11 gave me 72fps in one scene, but DX12 gave me just 64fps on my 2080Ti, that's a 12% difference in fps I could have used elsewhere. Ghostrunner behaves the same, Control too, Resident Evil and several others.

This is really bad, on the PC space, we sometimes lose fps due to Hyper Threading, CPU security patches, and now DirectX 12 as well. Those fps can really come in handy if we enable Ray Tracing, where every little bit helps.

There are other games that have DX12 equal to DX11, or slightly better: like Metro Exodus, Division 2, Hitman 3 or Shadow of the Tomb Raider, and this needs to become the norm.