Hello,

my question is about exercise 3 b) in the exam of WS 17. I tried to convert the last example -41.1 to ReSyFloat, but I don't understand why the correct solution is NaN.

We have B=2, e=3, m=4 so beta should be 3. I used the algorithm on slides 23-25 in the chapter Floating Point Numbers. In the first step I get E=floor(log_2(41.1))=5, but then we have E+beta = 5+3=8>2^3-1, so we can't represent the exponent using 3 bits and therefore my solution would be overflow and not NaN.

Why is the solution NaN?