> scheduling every single opcode perfectly based on the deepest microarchitectural

Is that even possible knowing OOE+ branch execution and many other CPU tricks?