#include <cruntime.h>
float __uint16_to_float32(uint16_t arg);
__uint16_to_float32 converts the 16-bit unsigned integer arg to a 32-bit floating value and returns the floating value as the result. As all 16-bit unsigned integers can be represented exactly in 32-bit floating point format, rounding is never necessary.
__uint16_to_float32 is a provided on all implementations of CrossWorks.