The industry-standard 3DMark benchmark has been updated with a new test dedicated to the AMD vs. Nvidia ray tracing showdown.

With AMD only a few days away from entering the ray-tracing game with its RX 6000 series RDNA 2-based graphics cards, UL, the benchmark provocateur, is now able to effectively test the pure ray-tracing guts of both AMD and Nvidia's GPUs. released this all-new feature test that can be used to effectively test the pure ray tracing guts of both AMD and Nvidia GPUs.

I already feel that this benchmark will be a great direct matchup once AMD's RX 6800 XT is released on November 18. This Radeon GPU is touted to be a true competitor to the RTX 3080 in rasterized rendering workloads (workloads used primarily in modern games), but where AMD's RDNA 2 architecture falls on the performance spectrum for cutting edge ray-traced effects It is not yet known where it will be positioned in the performance spectrum of state-of-the-art ray trace effects.

Nvidia claims a significant improvement in ray tracing performance with its Ampere architecture, and the RTX 3080 offers a fairly healthy improvement over the ray tracing-enabled RTX 2080 Ti.

The two companies have different approaches to how to accelerate ray tracing; we already know that Nvidia combines RT Core and Tensor Core to both compute bounding volume hierarchy and denoise ray-traced scenes but AMD's RDNA 2 approach is shrouded in thick pre-release fog.

Once both architectures were available for testing and purchase (in theory), 3DMark released an entirely new benchmark built around Microsoft's DirectX Raytracing API in order to get a bare-bones feel for the land of ray tracing.

The aptly named "3DMark DirectX Raytracing feature test" is "designed to make ray tracing performance a limiting factor" and is drawn in a single ray tracing pass. This allows us to see exactly how these cards align with little regard for conventional performance.

UL has already provided one ray tracing benchmark. That is Port Royal. The difference here is that Port Royal is built to use a game-like mix of traditional DX12 rendering and ray tracing, whereas the new DirectX Raytracing benchmark uses only ray tracing to render the scene It is.

"In this feature test, camera rays are traced across the field of view at small random offsets to simulate depth-of-field effects. The frame rate is determined by the time it takes to trace and shade a set number of samples of each pixel, combine the result with previous samples, and display the output on the screen.

The benchmark itself is run at 1440p to see how changing the number of samples affects performance.

There is also an interactive mode that allows users to play in the sci-fi 3DMark world created for UL's benchmark suite.

Currently, both 3DMark ray trace benchmarks can be accessed.

In this benchmark, the RTX 2080 (with Threadripper 2970WX) was run quick and dirty with 12 samples and averaged 20.48 fps, while the RTX 3080 (with Intel 10700K) recorded 46.27 fps. [On the other hand, the RTX 3080 (with Intel 10700K) struggled to 46.27 fps.