Write Great Code

Chapter 4. Floating-Point Representation

Floating-point arithmetic is an approximation of real arithmetic that solves the major problem with integer data types — the inability to represernt fractional values. Although floating-point arithmetic is often slower than integer arithmetic, modern CPUs incorporate well-designed floating-point units, thus reducing the performance difference between integer and floating-point arithmetic.

For this reason, the stigma surrounding floating-point arithmetic has diminished since the days when floating-point results were computed using software (rather than hardware). One unfortunate aspect of floating-point’s increasing ...