Let . Note that 0.1 has infinite base 2 representation (where the last overlined binary digits repeat). Compute the following:

a. floating point number in (this is the normalized single precision system). means floating point.

b. Compute exactly the relative error at x=0.1

- 0.1 =E2 + 2 -1) i=1 - 0.1 =E2 + 2 -1) i=1