- AMD is working on its own version of texture compression technology based on neural networks.
- AMD’s Neural Texture Block Compression focuses primarily on reducing the size of game packages.
- AMD emphasizes that its texture compression technology will be straightforward because it uses unchanged runtime execution.
AMD developers are working on their version of Texture Compression technology based on Neural Networks, similar to what Nvidia proposed last year.
Why it matters: As developers continue to create more detailed and lifelike games, file sizes have quickly gotten out of hand. Therefore, there is a huge need for competent compression.
AMD did not state the benefits of its texture compression technology compared to conventional techniques.
The launch of Neural Texture Block Compression only mentions the technique’s ability to shrink game packages in size, but it will be quite simple to implement. This is mainly because this technique uses unmodified runtime execution.
There is currently no information on how the new compression technology will be implemented in games or what hardware would be required to support it. However, AMD will provide more information during the EGSR conference 2024 on July 2.
Neural networks should become even more important for gaming in the future. Nvidia CEO Jensen Huang recently suggested that the company may someday concentrate on enhancing texture quality and employing AI to create objects and NPCs in games.
While Nvidia has a lot more planned for the future in utilizing AI’s benefits, AMD does not appear to be too far behind. The texture compression technology is just a small taste of everything that the company is currently working on.
Like Nvidia, fans can expect a new generation of AMD GPUs by next year. Dubbed RDNA 4, this generation could lead to a bigger focus on the mid-range market.
Thank you! Please share your positive feedback. 🔋
How could we improve this post? Please Help us. 😔