AMD executive talks AI acceleration in new interview

In context: Workforce Pink revealed its AI acceleration {hardware} alongside the Radeon RX 7000 graphics playing cards in December however did not say a lot about what it plans to do with them – not less than not in English. Japanese outlet 4Gamer spoke with AMD about how its AI ambitions differ from Nvidia’s.

In a recently-published interview, senior figures from AMD expressed curiosity in AI purposes in graphics playing cards that transcend picture reconstruction, which Nvidia made well-liked. The machine-translated feedback mark the primary time the corporate has talked at size on the topic since unveiling its first GPUs incorporating AI accelerators late final 12 months.

Nvidia continuously boasts about its DLSS expertise which depends on the RTX graphics card’s tensor cores to enhance picture high quality and efficiency by machine studying. Workforce Pink’s counterpart, FSR, strives towards the identical aim with out hardware-based acceleration, making it out there on extra GPUs however doubtlessly much less efficient relying on the scenario.

When AMD revealed and launched the RX 7900 XT and 7900 XTX with AI accelerators in December, onlookers assumed they had been the corporate’s reply to Workforce Inexperienced’s tensor cores. That could be the case, however AMD does not need to cease there.

David Wang, senior vp of engineering at Radeon Applied sciences, urged utilizing AI acceleration to enhance the AI and motion of NPCs in video games. He additionally stated AI might speed up the graphics pipeline by neural methods, mentioning Secure Diffusion amongst different applied sciences.

Wang stated AMD desires to make sure the AI acceleration in its gaming GPUs primarily focuses on bettering the gaming expertise. His feedback distinction Workforce Pink’s imaginative and prescient with Workforce Inexperienced’s numerous non-gaming purposes that use tensor cores.

Nvidia has unveiled RTX-exclusive AI capabilities like noise canceling, simulating eye contact on video, and upscaling internet movies. Wang indicated that a few of these purposes do not essentially require AI and that Nvidia makes use of them to advertise its proprietary expertise. Wang does not criticize Nvidia’s stance however suggests prospects could be paying for options they by no means use.

Wang’s interview additionally covers topics just like the transition from primitive to mesh shaders. The corporate can also be researching expertise that would change the connection between the CPU and GPU when drawing graphics.

Source link