1. Design the datapath of FPU, which includes mantissa part and exponent part.

2. Of course, both exponent and mantissa are encoded using a binary system rather than decimal system.

3. If the exponent of a floating point number is zero, the number is stored in integer format in the mantissa.

4. This is a standard floating point representation that the first bit of mantissa must be 1.

5. An IEEE floating point number dedicates 1 bit to the sign of the number, 8 bits to the exponent, and 23 bits to the mantissa, or fractional part.

6. Furthermore, various elegant properties of Benford's law, especially the mantissa distribution of data sets, are discussed.

7. The decimal part of a logarithm. In the logarithm 2.95424, the mantissa is 0.95424.

8. The decimal part of a logarithm. In logarithm 2.95424, the mantissa is 0.95424.