In the book of E. Tannebaum “Computer Architecture” There is a section “Appendix B. Floating Point Numbers”. Quote from there:
I do not understand how the number is calculated in the first case, but specifically how the mantissa is calculated.
The author calculates the Mantissa: 1 × 16 -3 sup>+ b × 16 -4 sup>. But after all, the categories -3 and -4 are filled with zeros: 0000 and 0000.
In my opinion, Mantissa is 0 × 16 -1 sup>+ b × 16 -2 sup>+ 0 × 16 -3 sup>+ 0 × 16 -4 sup>= b × 16 -2 sup>= 0,04296875 10 . Consequently, the whole number is + (16 5 sup>× 0,04296875 10 ) = +45056 10 . Isn’t there?
And it is not yet clear why for normalization we shifted the Mantissa on two hexadecimal discharge left. After all, the second category (16 -2 sup>) – nonzero.
Answer 1, Authority 100%
I was always hit by the babies of publishers and translators, but this is some kind of new level.