IEEE 754 — precision and rounding
Not every decimal number has a finite binary representation. When more than 23 fraction bits would be required, the value is rounded to the nearest representable float.
Why 0.1 is not exact
In binary, 0.1 produces a repeating fraction. In single precision the stored pattern is:
0 01111011 10011001100110011001101
The stored float, shown with more precision, is 0.10000000149011612. The error relative to 0.1 is 1.4901161138336505E-09.
Adjacent numbers
Every float has adjacent representable values. For 0.1:
previous: 0.09999999403953552
stored: 0.10000000149011612
next: 0.10000000894069672
Practice
Progress: 0/5 — correct: 0.
Is the number 1.5 exactly representable in IEEE 754 single precision?
This fraction has a finite binary representation because, in lowest terms, its denominator is a power of 2.