[计算机组成原理] 浮点数表示

定点数

假设位数32

浮点数表示

根据的IEEE标准不同，阶数和尾数的偏置情况不同，可能为biased exponent，也可能biased significant

偏置

Why we bother biasing the exponent in this way?
Because

1. want to support negative values, negative exponents. So by having our biased exponent the result of adding 127 to the actual exponent. We can support negative numbers.
   BUT we already have a negative number represetation two's complement why not just use that?
   BECAUSE with a bias exponent, it very easy to compare exponents and quickly find out which one is larger

conclusion

how to represent ZERO

reference

演示如何将科学记数法中的二进制数存储为称为Binary32的32位浮点格式
浮点数的表示