so we can be sure there are no float instructions. Nothing in here uses floating point, and this saves space by avoiding the FPU_EMULATOR in the kernel.