Thank you! :) It's mostly thanks to nwagyu, who made an amazing port to the NumWorks calculator, which has (almost) the same CPU as the Playdate.
I tried really hard to push performance, but there's a hard bottleneck: the Playdate's fast memory is too small to hold the whole NES game, so it constantly has to read from slower memory, and that caps the speed :(
But I'll check out your homebrew game and see if any dedicated optimizations are possible. I'll let you know!