this isnt for the end user, and im mainly doing this as a performance scaling test for my game across different device performance tiers. i have a 3080, so im really just looking to limit performance downwards. obviously, this isnt ideal compared to getting gpu directly, but im not keen on spending thousands of dollars on gpus just for testing.
limiting cuda cores and clocks would be ideal, but if there’s some other way of doing it then that also works. ideally it wouldn’t be tied to unreal engine, as i want to compare performance in games like gta to get an idea of where my fake gpu sits in terms of performance tier.