AMD plans to showcase its new neural texture block compression technology in London, competing with Nvidia's texture compression research.

AMD's GPUOpen team announced that it will present its new paper "Neural Texture Block Compression" at the European Graphics Rendering Symposium on July 2. This research is similar to Nvidia's previously announced neural texture compression technology, both focusing on using neural networks for image compression of video game textures to save storage space with almost no loss in visual quality. AMD's method uses small neural networks to compress textures, allowing games to integrate it easily without changing runtime execution. Although the actual performance of AMD's neural texture compression technology has not been seen yet, Nvidia's neural texture compression research has already demonstrated its potential.$AMD(AMD.US) $NVIDIA(NVDA.US)