For x86 div operation, can someone please help understand why the EAX register is not used when the operand/dividend size is a double word, instead of splitting them into DX
