The battlefield where AMD RDNA 2 will charge at Nvidia Ampere has been selected.
The industry standard 3DMark benchmark has just been updated with a new test aimed specifically at distinguishing between AMD and Nvidia on the ray tra ray detection battlefield.
And is now only a short time away from being able to participate in the ray trace ray tracking game with its RX 6000-series RDNA 2-based graphics cards, UL, which has released a test of this all-new feature that can effectively test pure ray detection against both AMD and Nvidia’s GPUs.
I would have said that this benchmark will create a great confrontation when AMD’s RX 6800 XT is available from November 18. The Radeon GPU is said to be the true competitor of the RTX 3080 in its rapid display workload (those are mostly used in modern games), but we still don’t know where AMD’s RDNA 2 architecture will be on the performance spectrum for advanced ray tracking effects.
Nvidia is touting a significant improvement in ray detection performance with ampere architecture and RTX 3080 offering pretty good improvements over the RTX 2080 Ti with ray detection enabled.
The approach of both companies on how they accelerate tracing will vary. We already know nvidia uses a combination of RT Core and Tensor Core to calculate the Limited Volume Hierarchy and eliminate the ray tracking scene, but AMD’s RDNA 2 approach is still somewhat covered in dense fog before release.
To figure out where the ray tracking lands are located, when both architectures are available for testing or purchasing (well, theoretically purchased), 3DMark has announced a brand new benchmark built around Microsoft’s DirectX Raytracing API, both types of GPUs built for PC use.
The 3DMark DirectX Raytracing feature test is appropriately named “designed to make ray detection performance a limiting factor” and is performed in a single ray pass. So we’ll look at exactly how these tags link together without caring about their traditional performance.
UL has provided a ray tracking benchmark. It’s Port Royal. The difference here is that while Port Royal is built to use a game-like combination of traditional DX12 and ray tracking, the new DirectX Raytracing benchmark uses proprietary ray tracking to display scenes.
“In this feature test,” a post on the 3DMark blog said, “the camera beams are tracked through the field of view with small random numbers to simulate the depth-of-field effect. Frame rate is determined by the time taken to track and shade a set of samples per pixel, combining results with previous templates, and presenting outputs on the screen. “
The benchmark itself is running 1440p and you can change the number of templates to see how that affects performance.
There is also an interactive mode included for good measure, allowing you to play in the science fiction 3DMark world created exclusively for UL’s benchmark set.
You can access both ray tracking benchmarks in 3DMark right now, although we’ll have to wait until November 18 to really get anything interesting to compare.
For reference when that time came, the fast and fast running of the RTX 2080 (with Threadripper 2970WX) in the benchmark, with the sample number of 12, gave us an average of 20.48 frames per second, while the RTX 3080 (with Intel 10700K) was too fast at 46.27 frames per second.
Does anyone like to bet on where the RX 6800 XT will land?