Hello,
I have just one question to ask. How difficult would it be to implement FPU into miniming core? How come such feature were not implemented already?
Hello,
I have just one question to ask. How difficult would it be to implement FPU into miniming core? How come such feature were not implemented already?
Implementing a sort of FPU wouldn’t be so hard, but implementing an feature complete 040 FPU or better yet a full featured 68882 is another story. The main problem (once functions are implemented) is precision. For some software, like probably most games (quake) lower precision is good enough, but for other software that expenct and depend on full 80 bit precision, you end up with rounding errors which in worst case can cause crashes. Only rather recently the apollo core 68080 on V2 cards got good enough precision to that 5 * 1.3 is 6.5 and not 6.4999999… for example.