Returns the floating-point, single-precision value from the lower half of the 64-bit MMX value in the source operand.
float _m_to_float( __m64 m );
There is no error return.
The compiler correctly ensures that an implict FEMMS is issued before any attempt to use the result of _m_to_float( ) operation.
END Microsoft Specific