Martin Piper (he/him) 💙💛🌻💉 on Nostr: npub17lg0m…awl3g I would probably try to find a good implementation of IEEE ...

npub17lg0mxsz7jwfuu42m4ua7eum0lqjk66snd5j6skhf6n5hvcavrrsdawl3g (npub17lg…wl3g) I would probably try to find a good implementation of IEEE floating point in plain old C/C++ using 32 bit integers. Then compile that to 6502 while hoping the compiler optimisations are good enough.

Then if really needed hand optimise that compiler output.

The full IEEE implementation is quite large and complex. It's a trade-off between optimisation and maintainability of the 6502 code.