Lab 9 - Arrays

Arrays are common in programs. They store a series of values instead of a single (scalar) value.

The code below defines an array in the data section, then the code prints each value out.

    extern  printf      ; We will use this external function

    section .data       ; Data section, initialized variables

mystr: db "%d", 10, 0   ; String format to use (decimal), followed by NL

myarr: dd 10, 20, 30, 40, 50, 60
temp:  dd 0

    section .text
    global main
main:

   xor eax, eax                ; A = 0
   mov ecx, myarr              ; C points to myarr

myloop:
   mov   ebx, DWORD [ecx+4*eax]  ; Get the value  B = myarr[A]
   mov   [temp], ebx
   push  rax
   push  rcx
                               ; Now print the result out
   mov   rdi, mystr            ; Format of the string to print
   mov   rsi, [temp]           ; Value to print
   mov   rax, 0
   call  printf

   pop   rcx
   pop   rax
   add   eax, 1                ; A++
   cmp   eax, 6                ; Does A == 6?
   jl    myloop                ; if less, jump to myloop

   mov  rax, 0
   ret

• The code uses "xor eax, eax". What is the equivalent "mov" instruction? What advantage does "xor" have over the "mov" instruction?
• Before the "mov ebx, DWORD [ecx+4*eax]" command, which registers much be set to a value? If they are not, what will happen? (You can simply explain this.)
• The "4*eax" is done intentionally. Why is the "4*" there? What happens if the command is "mov ebx, DWORD [ecx+eax]" instead? (Show this as a program, and explain its output.)
• Notice that the loop has the length of 6, and that the "cmp" instruction compares A to 6. However, "hard-coding" the array length is a bad idea. How could we use the dollar sign and "equ" to code the array length in a better way? (Refer to chapter 3 in your book.)
• Notice how the code performs "push rax" then "push rcx", then pops them in reverse order: "pop rcx" and "pop rax". Is this a mistake? Why or why not?

The objective for this lab is to add to the code, to determine the minimum value, the maximum value, and the average value. You will need to define storage for each of these in the .data section (or the .bss section).

Keep the part that prints each value in the array, only alter it to print the values on a single line. Also, have the program print something before the values, like "The array has the following values:".

In the program, the first thing to do is to initialize the minimum and maximum values. Some programmers try to use values that are clearly wrong for initialization, such as 100,000 for the minimum and 0 for the maximum. Then, looking at each value, replace the minimum and maximum with anything that is smaller or larger, respectively.

Questions:

• Suppose that the array has values 90, 120, 10. Describe (in a step-by-step manner) how a program would examine each one, and update the minimum and/or maximum. When finished, would we have the correct minimum and maximum?
• Why are the values 0 and 100,000 not approproiate? Show an example list of numbers that would make these initial values wrong for the minimum, then another example list for the maximum.

The idea of looking at each value in turn and updating the minimum and/or maximum is good. However, the initialization should be with a value in the array. To do this, we can copy the first array value to the minimum, and the maximum.

Have the program print some text, such as "The minimum is:" then the minimum value. Likewise, it should print text followed by the maximum value. And it should print some appropriate text followed by the average.

Questions:

• Suppose that the array has values 90, 120, 10. Initialize the minimum and maximum to the first value. Then describe (in a step-by-step manner) how a program would examine each one, and update the minimum and/or maximum. When finished, would we have the correct minimum and maximum?

Now code this algorithm into the assembly language program. It should initialize the minimum and maximum with the first value of the array, then examine each value in the array (starting with the second value), and update the minimum and maximum as needed.

Questions:

• Suppose that the array has only the value 90. How should you alter the program to make it work with only one value?
• Suppose that the array is empty, that there are no values in it. How should you alter the program to make it still work?

To find the average, we can make a sum of the values. Then divide the sum by the number of values in the array. Store this average in another memory location.

Questions:

• The average of 10, 11 is reported as 5. Why is it 5, and not 5.5?
• Would the average still work if some of the values are negative? Why or why not?
• Would the average still work if all of the values are negative? Why or why not?
• If the array is empty, what is a reasonable output?
• Suppose that the array is very long, such as thousands of values. Would the minimum, maximum, and average code still work? Why or why not?
• Given code that prints an array, it sounds trivial to add code to also find and print the minimum, maximum, and average. How does the original array printing code compare to the final code? Is this what you expected, and why or why not?

Include comments, and meaningful identifiers. Make sure to show the final version of the .asm program, and the assembly, link, and test run(s).