If you compare the graphics cards of AMD and NVIDIA, in addition to the performance, light pursuit, energy efficiency, price and other specifications, another place that can not be ignored is that AI,NVIDIA has added a physical AI unit from the RTX 20 series graphics card, which together with the RT light tracking unit has become the symbol of the N card.
The AI capability of graphics cards was not popular a few years ago, but now a lot of software already supports AI.Just whether it is necessary to do a special AI unit in the end, the concept of AMD in this respect is different from that of NVIDIA.
The senior executives of the AMD GPU department were interviewed by the Japanese media at the end of last year, which was only recently announced. In addition to talking about the RDNA4 architecture to continue to improve the performance, David Wang, senior vice president of the AMD graphics card department, also talked about the AI issue of the graphics card.
David Wang is still very objective, first praised NVIDIA's approach, saying that they actively try to use AI technology on game cards, GPU uses a large AI reasoning accelerator (refers to dedicated AI unit), may work where accelerators are needed, this is NVIDIA's strategy, which is great.
But David Wang says AMD has a different strategy and is currently focusing on using GPU where users need and care about in order to enjoy the game card experienceOtherwise, users may have to pay for GPU features they will never use.AMD believes that AI in game cards can and should be used to make video games more advanced and more interesting.
To sum up their point of view, AMD does not completely disagree that game cards have the ability to accelerate AI.It's just that they think that NVIDIA has allocated too much GPU resources to AI in this regard.And this part of the cost is the player has to pay, AMD hopes that AI can bring benefits to the game.
Recommended use China IT News APP
Download flyfish app to read news