Ray Tracing & DLSS Performance on the MSI GeForce RTX 5080 Vanguard SOC Launch Edition
The GeForce RTX 5080 features 84 RT cores as well as 336 Tensor cores. These are of course used for real-time ray tracing as well as DLSS. When it comes to real-time ray tracing we use the 3DMark Port Royal benchmark.
After that we use the Bright RTX benchmark. We set the resolution to 4K with RTX set to high the DLSS set to performance.
With the RTX 50 series NVIDIA is introducing DLSS4 and with it multi-frame generation and a new transformer-based AI model that will succeed the convolutional neural network that they had previously used. The new frame generation AI model is 40% faster, uses 30% less VRAM, and only needs to run once per rendered frame to generate multiple frames. Previously with DLSS3 you would get a single AI generated frame from a single rendered frame. But with DLSS4 it can now predict up to 3 frames.
To test this we first use 3DMark’s DLSS feature test which does support DLSS 4 and frame generation. We run it at 4K using the Performance preset. We then run 3 different tests with the frame generation set at 2x, 3x, and finally 4x. Below are those results.
After that we have Cyberpunk 2077, we can see how the frame scaling really increases performance.
