• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

NVIDIA's Neural Texture Compression Cuts VRAM Use From 6.5 GB to 970 MB

erek

erek

Fully [H]
2FA
Joined
Dec 19, 2005
Messages
17,421
Re: “Goodbye rasterization, hello NVIDIA neural rendering
“You've got neural shaders to "unlock new compression and approximation techniques for...”

—

“Neural Texture Compression works by training a neural network to understand what a Texel, the base pixel of a texture map, looks like on a specific material. NVIDIA has trained these neural networks for nearly every game material, making them ready for real-world game deployment. This process is so refined that the output can either provide a more realistic version on top of the base texture layer that the game uses or maintain the same texture appearance for significant VRAM savings. For a full technical explanation and a look at what Neural Texture Compression does behind the scenes, you can watch the in-depth YouTube video about the technology.“


View: https://youtu.be/-H0TZUCX8JI

Source: https://www.techpowerup.com/348003/...mpression-cuts-vram-use-from-6-5-gb-to-970-mb
 
DukenukemX said:
Yea but is this usable on AMD and Intel cards?
Click to expand...
I presume NV is pushing hard for proprietary in some way. About three years since it first came up and there's still nothing from it, wonder if it has consistency problems or other artifacts.
 
Prince Valiant said:
I presume NV is pushing hard for proprietary in some way. About three years since it first came up and there's still nothing from it, wonder if it has consistency problems or other artifacts.
Click to expand...
It is not proprietary, the core functionality for this was part of the D3D12 Cooperative Vector update in early 2025.
It works well, but until recently there wasn’t enough supported hardware in the wild for developers to bother with it.
It makes extensive use of Matrix Multiplication which AMD doesn’t effectively support in the bulk of the RDNA architecture with 3.5 being the first iteration to actually accelerate it in a usable way.
Intel ARC and newer supports it, but it’s not like Intel has a sizeable enough footprint in the gaming space for them to currently be a deciding factor…
 
Marees said:
if the PS6 implements this tech then it would be usable on (future) AMD cards. need to wait n watch
Click to expand...
Project Amethyst (the AMD-Sony name for the PS6’s GPU), will support “Neural Arrays” and have “Radiance Cores”, which will support the functionality.
 
bigdogchris said:
Sweet, Nvidia's plan to ship 2/4/8 GB graphics cards in the 6000 generation is coming to fruition.

4GB GPU will be overkill shortly thanks to this and AI up scaling.
Click to expand...
also HyperMemory-style features
 
You must log in or register to reply here.
Back
Top