dbl_fsqrt
int dbl_fsqrt(dbl_floating_point *, dbl_floating_point *, unsigned int *);
return(dbl_fsqrt(&fpregs[r1],