...
The Java language allows platforms to use available floating-point hardware that can provide extended floating-point support with exponents that contain more bits than the standard Java primitive type double (in the absence of the strictfp modifier). Consequently, these platforms can represent a superset of the values that can be represented by the standard floating-point types. Floating-point computations on such platforms can produce different results than would be obtained if the floating-point computations were restricted to the standard representations of float and double. According to the JLS, §15.4, "FP-strict Expressions" [JLS 2005]:
The net effect [of non-fp-strict evaluation], roughly speaking, is that a calculation might produce "the correct answer" in situations where exclusive use of the float value set or double value set might result in overflow or underflow.
...
FLP00-C. Understand the limitations of floating-point numbers | |
FLP00-CPP. Understand the limitations of floating-point numbers |
Bibliography
Ensuring the Accuracy of Floating-Point Numbers | |
[JLS 2005] | |
[JPL 2006] | 9.1.3, Strict and Non-Strict Floating-Point Arithmetic |
Making Deep Copies of Objects, Using strictfp, and Optimizing String Performance |
...
NUM05-J. Do not use denormalized numbers Rule 03: Numeric Types and Operations (NUM)