something to look forward to: Microsoft, NVIDIA, and now Intel are working to convert gaming assistants into a new sales point for generative AI. While Nvidia argues that its GPU is the best tool for local AI workload, Windows and Intel have tried to display the benefits of using NPU to unload the tasks from graphics chips during the gameplay. However, none of the proposed assistants have been tested wide real -world.
Intel recently showed how generic AI programs are running on their new NPUs, which can benefit gamers. Although many companies have promoted NPU’s gaming-related capacity, these benefits are largely unproven.
In Intel’s performance, a general AI model-a general third-party auxiliary-black myth provided the gameplay advice during a session of the myth: Wukong, a unwarded arrow lake lake core Ultra 200HX processor, possibly in a gaming laptop. Assistant responded to a spoken question using real -time voice integration and recovered an online video as to how to beat a boss.
It is not clear whether the performance has shown the actual gameplay or simply concept footage. Microsoft first used concept reels to introduce its upcoming Copilot Gaming AI coach, aimed at providing real -time tips in sports such as Minecraft and Overwatch while walking on the Qualcomm Snapdragon X chip. No actual footage exists yet, but the inner formulas of the Xbox can start testing the feature on mobile devices next month.
Also read: NPU vs. GPU: What is the difference?
Intel’s presentation also highlighted the possible performance benefit from running the AI model on NPU. The activist activated Wukong’s Framaret from 103 to 108 FPS, while running a single model on the GPU led to a decline of 94 FPS.
NVIDIA gaming assistant, which runs on GPU with at least 12GB Vram, is available in the latest version of the company’s desktop app. However, instead of offering gameplay guidance, this performance focuses on adaptation.
Arrow Lake is an interesting option for the performance of Intel, given that core Ultra 200HX and desktop 200k processors AI achieve only 13 tops (operating trillions per second).
If gaming AI performs according to auxiliary promise, they should be theoretically run on new notebooks and AMD’s upcoming Lunar Lake Core Ultra 200V or AMD’s Strikes Point Rezen AI 300 on new notebooks and handheld platforms, which can reach 50 tops. Benchmarking an assistant on a laptop or new handheld gaming PC can be equipped with chips that can give interesting results.
