This came in before the exam (obviously), but my email access in Turkey was not terrific.  Since the
 questions continue to be relevant, I am answering them now. 

     Dr. Patt,

     I have a question about the first part on old Exam 2, 
     posted on the 
     306 site. I'm not sure you 
     can answer them (I'm guessing you're not giving 
     solutions for a reason), but maybe you can at least give me a push in the 
     right direction. On the first part of 
     the first question, you say that a load 
     instruction specifies the location xF401 and you ask 
     how we know whether to 
     load from KBDR or the location xF401. 
     I thought the memory location xF401 IS 
     the KBDR... so I am totally not understanding what you're getting at here.

I would say you are totally on top of it.  Memory address xF401 IS the address of the keyboard data
 register.  There is NO location in the memory with that address.  After all, it there were, the computer
 would not know which xF401 the program had in mind.

        My second question deals with part 2 of the same question (a similar 
	question also is on PS6, so I'm guessing it's a 
	pretty important idea). You 
	ask how many accesses to memory different instructions 
	need to make. I just 
	want to make sure I understand what this means. Just for starters, every 
	instruction has at least 1 access - during 
	the fetch phase, when it loads the 
	MAR with the PC and loads the value 
	at that location into MDR to decode it? 

Absolutelty.  Right on target.

	      So RET would only have 1 access? 

Yup.

	If an instruction has to load a value into a 
	location and it loads the MAR with that 
	location and the MDR with the value, 
	that would be 1 more access, correct?

Yup.

	       My final question is about the final part 
	       of this question. You ask us 
	       what value will be in R1 after the 
	       instruction: ASCII   LD  R1, ASCII. You 
	       give us the location of ASCII (you 
	       say "the symbol table entry... is x4F08"), 
	       but nothing else... how is it remotely 
	       possible to know what R1 would contain 
	       with this information? Don't you need the value in the location labelled 
	       ASCII to be able to say this?

You sure do.  And what is that value?  How about: 0010001100001000, which just happens to be 
the assembled version of LD R1,ASCII.  How about that!

        I think I got most of the rest of the test so far, 
	thankfully, but I 
	had the **** scared out of me 
	when I couldn't even get any part of the first question.

	Thank you,
	<<name withheld to protect the user of ****>>

Good luck on the final.

Yale Patt