AMD Shares Future GPU Strategy

Recently, David Wang, Senior Vice President of Engineering R&D of AMD Radeon Technology Division, and Rick Bergman, Executive Vice President of AMD Computing and Graphics Division, accepted an interview with 4Gamer and shared the future GPU strategy and talked about the RDNA 4 architecture.
David Wang said that the RDNA 3 architecture GPU has MDIA, which is a multi-draw indirect accelerator that can handle MDI at the hardware level. AMD expects more advanced shaders and techniques to emerge for RDNA 4 architecture GPUs as part of its new GPU programming model standard specification. At the same time, AMD’s recent commitment to the ROCm open software platform is comparable to Nvidia’s CUDA platform.

David Wang praised Nvidia’s AI strategy but didn’t think it would work with AMD’s GPUs. According to AMD, some other uses of AI accelerators may be game NPCs, applying learning models to achieve better and smoother effects of standard programming models, so that players have a better gaming experience.

AMD’s GPU has steadily developed from RDNA to RDNA 3. Rick Bergman promised that the performance of the RDNA 4 architecture GPU will be stronger than that of the RDNA 3 architecture, and will not disappoint players’ expectations. AMD’s GPU IP has entered the smartphone field through Samsung’s Exynos 2200 and is also supplied to Tesla’s in-vehicle systems, providing GPU technology for various platforms.

AMD has various RDNA 3 architecture-based GPU release plans in 2023. According to its GPU roadmap, RDNA 4 architecture GPUs may arrive in 2024, corresponding to the Radeon RX 8000 series graphics cards.