Jacob Roach / Digital Trends

AMD is set to reveal a enquiry paper about its technique for neural texture block compression at the Eurographics Symposium on Rendering ( EGSR ) next week . It sounds like some technobabble , but the idea behind neural compression is passably mere . AMD says it ’s using a nervous net to compact the monolithic textures in games , which cuts down on both the download size of a game and its need on yourgraphics poster .

We ’ve heard about similar tech before . Nvidia introduced a paper onNeural Texture Compressionlast twelvemonth , and Intel comply up with a paper of its own thatproposed an AI - drive level of detail(LoD ) technique that could make models look more naturalistic from farther away . Nvidia ’s claims about Neural Texture Compression are specially telling , with the paper maintain that the technique can store 16 times the datum in the same amount of blank space as traditional block - based compression .

AMD has n’t uncover its research yet , so there are n’t a net ton of particular about how its method would play . The key with Nvidia ’s approach is that it leverage the GPU to depressurise textures in real prison term . This has been an issue in several games released in the past span of twelvemonth , fromHalo InfinitetoThe Last of Us Part ItoRedfall . In all of these game , you ’ll notice low - quality textures if you track down out of VRAM , which is peculiarly noticeable on 8 GB nontextual matter posting like theRTX 4060andRX 7600 .

One detail AMD did reveal is that its method should be easier to desegregate . The tweet announcing the newspaper reads , “ unchanged runtime capital punishment allows easy biz integrating . ” Nvidia has n’t say if its technique is particularly hard to integrate , nor if it will ask specific hardware to work ( though the latter is belike a safe bet ) . AMD has n’t made mention of any special hardware requirements , either .

We'll present & quot;Neural Texture Block Compression" ; @#EGSR2024 in London .

Nobody like downloading huge game packages . Our method acting compact the grain using a neural internet , reducing data point sizing .

unaltered runtime execution of instrument allows easy game integration.https://t.co/gvj1D8bfBfpic.twitter.com/XglpPkdI8D

& mdash ; AMD GPUOpen ( @GPUOpen)June 25 , 2024

At this tip , neuronic compaction for textures is n’t a feature usable in any game . These are just research papers , and it ’s severe to say if they ’ll ever turn into features on the level of something likeNvidia ’s DLSSorAMD ’s FSR . However , the fact that we ’re seeing AI - force compaction from Nvidia , Intel , and now AMD suggests that this is a new trend in the macrocosm of personal computer gaming .

It produce good sense , too . feature like DLSS have become acornerstone of forward-looking nontextual matter card game , serve as an umbrella for a large swath of performance - boost features . Nvidia ’s chief operating officer has saidthe company is look into more way to leverage AI in games , from generating objective to enhancing textures . As characteristic like DLSS and FSR continue become more large , it makes sentience that AMD , Nvidia , and Intel would take care to expand their capabilities .

If we do see nervous grain densification as marketable features , they ’ll likely show up with the next generation of graphics identity card . Nvidia is expect to unveil itsRTX 50 - series GPUsin the second one-half of the twelvemonth , AMD could showcase its next - genRDNA 4 GPUsin a like time frame , andIntel ’s Battlemagearchitecture is arriving in laptops in a matter of calendar month through Lunar Lake CPUs .