After it was released, a bug in the floating-point math unit was discovered, requiring every operating-system maker to write special code to accommodate the problem.
This article proposes a quick sort method based on the machine code of floatingpoint numbers, which can sort the floating point numbers distributed at random.