The mathematical behavior of infinity in the floating-point system is defined by IEEE-754.

In general infinity is used to represent any positive number larger than the finite range of double. A finite number divided by infinity is the appropriately signed zero value. Most other operations involving infinity result either in another infinity or nan.

Note that negative infinity also exists, and can be easily specified as -infinity.

Dividing a non-zero, finite floating-point number by zero results in an infinity of the appropriate sign.