- Nvidia has detailed a clever new RTX Neural Shaders technology
- This includes neural texture compression to include more textures in VRAM.
- Helps understand why Nvidia chose more efficient video RAM loads with some RTX 5000 GPUs
Nvidia just unveiled its new RTX 5000 GPUs at CES 2025, but along with that reveal came a more low-key, but still crucial, announcement about new technology to help speed up the gaming performance of these graphics cards.
Namely, RTX Neural Shaders, which Nvidia describes as “small neural networks” (AI) that are being introduced into programmable shaders on its next-generation GPUs. (A shader simply refers to a small GPU program that runs on a graphics card, and is not limited to facilitating shading or lighting effects in a game, but that is the original derivation of the name.)
Nvidia explains in a blog post: “RTX Neural Shaders SDK allows developers to train their game data and shader code on an RTX AI PC and accelerate their neural representations and model weights with Nvidia Tensor Cores at runtime.” “.
RTX Neural Shaders actually power a trio of new technologies, but the secret weapon for the new RTX 5000 GPUs when it comes to VRAM (and there’s been a lot of concern about the amount of onboard video memory being quite unstable) is RTX . Neural texture compression.
To make the most of the available VRAM, RTX Neural Texture Compression uses compression (obviously) to fit more game textures into the graphics card’s RAM.
Here’s Nvidia’s key boast: “Neurally compressed textures save up to 7 times more VRAM or system memory than traditional block-compressed textures with the same visual quality.”
In addition to this texture compression magic, Nvidia also incorporates RTX Neural Materials. Team Green notes that this uses AI to “compress complex shading code that is typically reserved for offline materials and built with multiple layers, such as porcelain and silk.” The claim is 5x faster material processing, making it “possible to render cinema-quality assets at gaming-ready frame rates.”
Finally, the third piece of the puzzle here is RTX Neural Radiance Cache, which is a method of using AI to improve ray tracing performance (specifically, giving gamers “better performance and indirect illumination with path tracing”).
Look
Analysis: Hair today, RTX Hair tomorrow (and much more)
As always, take any press announcements with a grain of salt, but the performance boost mentioned with the texture compression trick for RTX 5000 graphics cards is potentially huge.
When the 12GB video RAM load for the RTX 5070 and the 16GB configuration for the RTX 5080 were passed around in the rumor mill some time ago, as well as the stinging sense of disappointment that Nvidia was “scrimping once more in VRAM”, there was an undercurrent of speculation that the RTX 5000 series could come with a special sauce to better adapt the textures to the memory available on the cards next generation graphics.
And that’s exactly what happened, meaning that while 12GB seems like a meager allocation for the RTX 5070, it will go much further thanks to RTX Neural Texture Compression. The problem, of course, is that game developers have to use the technology, and 12GB still seems a little shaky in terms of future-proofing, at least to me.
We’ll also need to test the performance of this AI-powered texture compression and compare the results to Nvidia’s lofty claims, but it sounds very promising. And of course, it also helps make sense of Nvidia’s VRAM decisions.
All of this comes alongside a bunch of other new tricks for the RTX 5000 generation, including RTX Neural Faces and RTX Character Rendering SDK for improved face, hair, and skin graphics. Indiana Jones and the Great Circle is getting an update featuring RTX Hair (yes, that’s really its name) that uses “linear sweep sphere rendering.” [to run] faster, with improved image quality and using less VRAM” compared to previous hair rendering methods. (See the results in the image above.)
That game will also be one of many titles to benefit from DLSS 4, which is, of course, a key piece of software for RTX 5000 GPUs. It includes DLSS Multi Frame Generation (MFG) to further increase frame rates via inserting more artificial frames compared to DLSS 3 Frame Generation. (Also note that Nvidia is incorporating Reflex 2 for further input lag reduction, up to 75% now, to accompany MFG and mitigate any additional latency.)
Nvidia also told us that there will be 75 games and apps that support DLSS 4 MFG capability when its Blackwell graphics cards launch in late January (the RTX 5090 and 5080 to start), a big head start for buyers of the new ones. GPU. (RTX 5070 models will follow in February.)
It’s worth noting that RTX 4000 owners also get something here in the form of improved DLSS frame generation with higher performance and lower memory usage, but this is not MFG, which is exclusive to the new Blackwell GPUs.
Via Neowin
TechRadar will cover this year’s edition extensively CESand will bring you all the important announcements as they happen. Go to our CES 2025 News page for the latest stories and our practical verdicts on everything from 8K From foldable TVs and screens to new phones, laptops, smart home devices and the latest in artificial intelligence.
And don’t forget follow us on tiktok and WhatsApp For the latest from the CES fair!
