Understanding Non-Numerical Floating-Point Values: 1.#INF00, -1.#IND00, and -1.#IND

IEEE 754 floating-point representation defines several non-numerical values to indicate exceptional conditions during floating-point operations. These values include positive and negative infinity, Not-a-Number (NaN), and indeterminate (IND).

Positive Infinity

• 1.#INF00 on Windows
• inf on Linux

This value represents a positive number that is too large to be represented as a finite floating-point number. It is typically encountered when performing operations like dividing a very large number by a very small number.

Negative Infinity

• -1.#IND00 on Windows
• -inf on Linux

This value represents a negative number that is too large (absolute value) to be represented as a finite floating-point number. It can arise from operations like dividing a negative number by a very small number.

NaN (Not-a-Number)

• -1.#IND on Windows
• nan on Linux

NaN indicates that the result of an operation is undefined or invalid. This can occur when attempting to perform operations like taking the square root of a negative number or dividing by zero.

Indeterminate (IND)

• 1.$NaN

This value is not explicitly defined in IEEE 754 but may appear in some implementations. It typically indicates an indeterminate result, such as 0/0 or ∞/∞.

These non-numerical values can be useful for debugging floating-point code. For instance:

• Positive/negative infinity can help identify overflow or division by zero issues.
• NaN can point to invalid operations, such as taking the square root of a negative number.

To determine if a floating-point value is valid or not, you can use functions like isfinite() and isnan() in C .