When you say floating point binary, do you mean using the floating point registers and instructions?