fmaxfmina

fmaxfmina is a hypothetical function described in some numerical libraries as a single-call operation that yields both the maximum and the minimum of two single-precision floating-point numbers. It is not part of the standard C math library and its existence depends on a particular library’s API design.

In typical designs, fmaxfmina would take two inputs, a and b, and return a pair of results: max and min. One common approach is to return a small structure containing two floats, such as a pair { max, min }. Another approach is to return a two-element tuple or to provide outputs via reference parameters. The exact interface is library-specific, and there is no universally accepted signature.

NaN handling is a key consideration for any such function, given the IEEE 754 behavior of NaN values. A conservative design would mirror the semantics of fmaxf and fminf: for the max component, if one argument is NaN and the other is not, the non-NaN value is chosen; if both are NaN, the max output would be NaN. Similarly for the min component. Some implementations may instead define NaN propagation differently, so users should consult the specific library’s documentation.

Rationale and use cases: combining max and min in a single call can reduce function-call overhead or simplify code paths in algorithms that repeatedly need both extremes. However, because fmaxfmina is not standardized, its availability, behavior, and performance are library-dependent. As an alternative, many programmers simply compute fmaxf and fminf separately in sequence.

See also: fmaxf, fminf, IEEE 754 NaN handling, numerical library design.