“The game was reconstructing the whole lighting of a scene in each frame even when nothing was changing, which was obviously a huge strain on CPU,” LTSharp explained.
LTSharp detailed the challenging process, emphasizing that it wasn’t simply a matter of inputting a command into Claude and receiving an immediate fix. Instead, it involved “two weeks of intense work” and extensive trial and error. The key improvement came from optimizing CPU usage by distributing tasks across more cores, increasing CPU utilization from 35-40% to 50-60% on a rig equipped with a 5800X processor.
Despite having no formal background in software development, LTSharp’s achievement stands as a testament to the potential of AI in gaming. They shared their work on GitHub, suggesting that more experienced developers could build on it to further enhance performance.
“After upgrading from 2070 Super to 9070 XT, the game continued to run poorly, with low fps and stuttering,” LTSharp lamented, motivating them to find a solution.
This fan-led project contrasts with some controversial uses of AI in the gaming industry. For instance, Epic Games delayed a Star Wars shooter to adopt an AI-first approach, prohibiting hand-written code. An Xbox Business Analyst also criticized the industry’s reluctance to embrace AI, claiming that developers ignoring AI’s potential risk falling behind.
While AI’s role in gaming remains contentious, LTSharp’s work demonstrates its capacity to empower individuals to enhance gaming experiences. This grassroots solution could inspire developers to explore AI’s potential in optimizing game performance, potentially leading to broader industry adoption.
Article Source: www.ign.com
