I've reviewed your code, and unfortunately, it's based on an old version of my build. We've made significant progress since then, particularly in performance optimization. As it stands, your build will struggle quite a bit on weak CPUs.
Your NVIDIA integration looks good, though. I'll incorporate it into my release this weekend
Hey Croque, glad you took a look at it and that you're planning to implement some things in your build!!!
I actually started working on this the very day you first shared your code, trying to get it running on my hardware and experimenting with other implementations. I've also made quite a few general updates to the code since then, but I'll check out the latest changes in your fork to see if there are any useful optimizations I might have missed. Thanks for the heads-up!
So far, it's been running pretty well on the hardware I've tested, although I haven't tested it on that many different setups, and none of them have been particularly low-end since I don't really have any modest hardware at home. I guess that could explain why I haven't encountered the issues you mentioned, which would make sense.
The lowest-end hardware it's been tested on so far is a borrowed laptop with an Intel i5-12450H. I also want to test it on a couple of laptops I have at home, but neither is exactly representative of lower-end hardware (one has a Ryzen 9 5980HX with an RX 6800M, and the other has an i7-12650H with an RTX 3080 Ti Mobile).
So if anyone else gives it a try and notices any performance issues, please let me know!