first time to ask questions in stack overflow. please share your valuable experiences, i am trying to model FP16 num operations, like fp16 a,b a+b=?, i need set to different rounding mode and make sure the result complies/align with ieee754 standard. i searched around, no one seems do the same thing before, i am trying to use gmpy2 lib, it has a context, but the context create function ieee() only accept 32,64..., 16 is a eligal number, I know we can set context.emax, context.emin, context.precision, but which value should i assign them? seems context.emax=15, context.emin=-14, context.precison=10 seems not work.

looking forward to comments? thanks ahead guys.

i want to know gmpy2.context.emax/emin/precision value, when FP16 context used. and why we use that value.