Nvidia Announces Multi Frame Generation with DLSS 4 For GeForce RTX 50-Series GPUs

by

At CES 2025, Nvidia not only showcased a fresh array of graphics cards but also introduced innovative software technologies that leverage the AI capabilities built into these new GPUs. The spotlight technology among the new features is the latest version of Nvidia’s well-known DLSS (Deep Learning Super Sampling).

Referred to as DLSS 4, this technology leverages the Tensor cores found in RTX graphics cards, and in addition to its existing capabilities of producing images at higher resolutions, it now includes Multi-Frame Generation. This is an advancement from Frame Generation, a feature that can now produce three extra frames for every single frame produced by the system.

Nvidia asserts that Multi Frame Generation allows them to potentially boost a game’s frame rate up to eightfold, surpassing conventional rendering methods.

The updated DLSS 4 comes with enhancements to its collection of features, such as Ray Tracing Reconstruction, Super Resolution, and Deep Learning Anti-Aliasing (DLAA). These improvements will now utilize transformer architecture, similar to that used in AI models like ChatGPT, Flux, and Gemini. This marks the first instance of real-time application of this technology for graphics rendering.

Additionally, Nvidia introduced Reflex 2 alongside DLSS 4. According to Nvidia, this technology syncs CPU and GPU tasks to potentially lower in-game latency by as much as 75%.

Or more informally:
Nvidia unveiled Reflex 2 together with DLSS 4. They say it’s all about making the CPU and GPU work together smoothly, which could slash in-game lag by up to 75%!

It’s important to mention that for these latest additions to function properly, they will necessitate the use of an Nvidia GeForce RTX 50-series graphics card.


Read More

2025-01-07 16:41