Calculate the arithmetic mean of a single-precision floating-point strided array, ignoring NaN values, using a two-pass error correction algorithm with extended accumulation, and returning an extended precision result. - View it on GitHub