fild{l|ll}
Convert the integer operand (signed) into extended-real and load it onto the floating-point stack.
fild 2(%eax)