Coincidentally, one of Tom 7's SIGBOVIK 2018 papers [1] actually describes the emulation of uint8_t with floating point operations and nothing else. A working l

Coincidentally, one of Tom 7's SIGBOVIK 2018 papers [1] actually describes the emulation of uint8_t with floating point operations and nothing else. A working library is available [2]. [1] http://tom7.org/papers/fluint.pdf http://tom7.org/papers/fluint.pdf [2] https://github.com/ixchow/fluint8 https://github.com/ixchow/fluint8