Manual single precision floating point type

The algorithm of the eight digit index is a=127+（-3）=124

124 Binary system Said is 0111 1100

It can be understood that there are two levels of mapping , First from 8 Bit binary number , Mapping to 0~255 The decimal number of , This mapping relationship , Is the normal conversion of binary numbers and decimal numbers , Then the second layer is one of them 0 and 255 Corresponding to two special cases , and 1-254 Then corresponding in sequence -126-127