These rules permit the compiler to generate efficient machine code while maintaining the accuracy of floating-point computations.

• 这些规则使编译器可以在保持浮点计算精确性的前提下生成高效的机器码。