Unlike Nvidia, AMD Wants ‘AI Acceleration’ For NPC Behavior, Movement And More

Expert Verified By

AMD Is Coming Up With New Approach To AI Accelerators!

Head of AMD Radeon Technology and Senior Executive David Wang recently interviewed with 4Gamer. During the interview, David Wang discussed many things, including the future of AMD and the strategy for the next generation of Radeon cards based on RDNA4.

David Wang started by discussing Mesh Shaders; the current Radeon RDNA3 architecture features MDIA, also known as Multi Draw Indirect Accelerator. RDNA3 can also process MDI at a hardware level, which is why AMD succeeded in achieving 2.3x performance compared to RDNA2.

David also states that the company sees MDIA as a direct replacement for the current LoD Level of Detail system and plans to implement better and more advanced shaders and techniques in the future for RDNA4.

David Wang also has an entirely different approach to how AI acceleration should be used on AMD Radeon cards. According to David, the decision to include an AI accelerator on RDNA3 was more of a business decision based on what users want or don’t want.

AMD is also not planning to limit the use of AI accelerator on the RDNA3 architecture to just upscaling methods like DLSS, Xess, or FSR. It’s looking to use AI acceleration to improve other things within graphics cards, like better NPC Behavior and Movement or image processing.

Developers can use the RDNA3 AI accelerator to provide users with a much better gaming experience. David also states,

“The reason NVIDIA is actively trying to use AI technology even for applications that can be done without AI technology is that NVIDIA has installed a large-scale inference accelerator in the GPU.

To make effective use of it, it seems they are working on a theme that needs to mobilize many inference accelerators.

That’s their GPU strategy, which is excellent, but I don’t think we should have the same strategy.

We are focused on including the specs that users want and need to give them enjoyment in consumer GPUs. Otherwise, users are paying for features they never use. We believe that inference accelerators should be implemented in gamers’ GPUs to make games more advanced and fun”.    

As a closing note, the AMD representatives promised us a more robust architecture next year that will perform much better than what RDNA3 has to offer. What are your thoughts on what AMD RDNA4 provides, and what do you expect?

You Might Like: Recent PlayStation PC Ports Have Been Underwhelming On Steam

Was our article helpful? 👨‍💻

Thank you! Please share your positive feedback. 🔋

How could we improve this post? Please Help us. 😔

Gear Up For Latest News

Get exclusive gaming & tech news before it drops. Sign up today!

Join Our Community

Still having issues? Join the Tech4Gamers Forum for expert help and community support!

Latest News

Join Our Community

104,000FansLike
32,122FollowersFollow

Trending

Framework RAM Prices For Computers Jump Again by 32.5% After Recent 50% Increase

Framework increases the price of the computer RAM once again by 32.5% and says that it stilll charges less than apple.

AI Set to Consume 20% of Global DRAM Supply as HBM and GDDR7 Drive the Next Phase

AI will consume 20% of the Global DRAM supply in 2026, and the next phase will focus on memory capacity and inference costs.

Arthur Morgan Was Originally Designed To Be Less Attractive In Red Dead Redemption 2

Arthur Morgan's final design is more attractive and...

Industry Should Focus More on Fun Games Instead of “Monetization and Customer Values”, Says Former PS Boss

Former PS boss says that the industry is now focused more on monetization; instead, they should make games more creative and fun.

Generative AI Will Reduce AAA Development Cycle To Just 2 Years, Level-5 CEO Claims

Akihiro Hino stated that generative AI has the potential to substantially decrease the AAA development period to just two years.