Software floating point number.
Constructs Fp from hardaware floating point number.
Uses approximate division for now TODO: use full precision division for void when Fp division is ready
coefficient size in bits
See Implementation
Software floating point number.