CS 271 Computer Architecture and Assembly Language

Final Exam

 Calculator and 8.5 x 11 note page permitted.

 Multiple-Choice problems, 2 pts each. Others as marked.
 For all MASM questions, WriteDec is a procedure that displays the contents of the eax register
as an integer. CrLf is a procedure that moves the cursor to the beginning of the next line.

1. (5 pts) What is the correct order of steps in the instruction execution cycle?
_____ Update the instruction pointer to point to the next instruction
_____ Fetch the instruction operand(s) if necessary
_____ Execute the instruction
_____ Fetch the instruction at the instruction pointer into the instruction register
_____ Decode the instruction in the instruction register
_____ Go back to the beginning of the instruction execution cycle

;code fragment R
mov eax,1
2. (3 pts) In code fragment R, suppose that variables y and z are mov ebx,3
declared as DWORD, and z contains a non-negative value. Assuming that mov ecx,z
the value of z is small enough so that no overflow occurs, what is cmp ecx,0
calculated and stored in memory location y? (Your answer should be an je store
top:
algebraic expression, not a literal value.) mul ebx
loop top
store:
___________________ mov y,eax
Use the following declarations and address information for problems # 3 - 4
MAXSIZE = 100
.data
list DWORD MAXSIZE DUP (?)

Recall that DWORD specifies 4 bytes (32 bits). After the "program" is loaded and relocated, the label list is
equivalent to absolute address 1000h. I.E., during program execution, the list array starts at memory address
1000h. All addresses are given in hexadecimal; other values are in decimal.

3. (2 pts) How many bytes of memory are reserved by the declaration of list? ____________

4. _____ After the statement mov edi,OFFSET list is executed, which of the following
correctly assigns the value 42 to the 10th element of the list array?
A. mov list[edi+36],42
B. mov list[9],42
C. mov [edi+36],42
D. all of the above
E. none of the above
;code fragment S
mov edi,OFFSET list
5. (3 pts) After code fragment S is executed, what are the mov eax,MAXSIZE
contents of the first 6 elements of the list array. top:
cmp eax,0
jle quit
mov [edi],eax
dec eax
add edi,TYPE list
jmp top
quit:
;code fragment T
6. (3 pts) After code fragment T is executed, what are the mov edi,0
contents of the first 6 elements of the list array. mov eax,1
mov ecx,MAXSIZE
top:
mov list[edi],eax
add eax,2
add edi,4
loop top

7. (3 pts) After code fragment U is executed, what are the ;code fragment U
mov edi,OFFSET list
contents of the first 6 elements of the list array.
mov ebx,0
mov eax,1
mov ecx,MAXSIZE
top:
mov [edi+ebx],eax
inc eax
add ebx,TYPE list
loop top
8. _____ Pipelining is a form of instruction-level parallelism that is implemented by running multiple
instruction execution cycles so that
A. sequential instructions can be in different stages of the cycle at the same time
B. sequential instructions can be executed simultaneously
C. each instruction can be processed by multiple CPUs to verify accuracy
T
D. an n-stage pipeline can guarantee that a program will run in n time (where T is the time to run the
program without pipelining).
E. none of the above

9. _____ Synchronization of parallel processes may be required when

A. a single program is running on a multi-processor or a multi-computer system
B. two or more programs are running simultaneously on a single processor
C. a program implements an algorithm that uses threads
D. all of the above
E. none of the above

10. _____ The speed of processor-level parallelism as implemented in hardware depends on

I. the CPU speed of individual processors
II. the I/O speed of individual processors
III. the speed of the interconnection network
IV. scalability of the parallel architecture
A. I, II, III B. II, III, IV C. I, III, IV D. I, II, III, IV
E. none of A - D is correct

11. _____ The speed of processor-level parallelism as implemented in software depends on the
I. parallelizability of algorithms
II. features of parallel application programming languages
III. parallel operating system software
IV. efficiency of parallel system libraries
A. I, II, III B. II, III, IV C. I, III, IV D. I, II, III, IV
E. none of A - D is correct

In problems # 12 - 16, A, B, and C are Boolean expressions. Select True if the statement is a Boolean identity,
False otherwise.

12. _____ 1 A  A
A. True B. False

13. _____ A B  A B
A. True B. False

14. _____ A  B  A  B
A. True B. False

15. _____ A  B  A B
A. True B. False

A  ( B  C )  ( A  B )( A  C )
16. _____
A. True B. False
Use this information to answer questions # 17 - 18.
An algorithm takes 20 seconds to execute on a single 2.4G processor. 40% of the algorithm is sequential.
Assuming zero latency and perfect parallelism in the remaining code …

17. (3 pts) Approximately how long should the algorithm take on a parallel machine made of 8 2.4G
processors?

____________

18. (3 pts) Suppose that we can add (with perfect scalability) any number of 2.4 G processors to the system.
What is the fastest time that can be achieved?

____________

Use the circuit diagram at the right to answer question # 19.

19. _____ What is the name of this circuit?

A. half adder B. full adder
C. 1-bit ALU D. multiplexer
E. none of the above
 ;code fragment V
Use this information and the code at the right to answer mov edx,OFFSET string
questions # 20 - 21. mov ecx,MAXSIZE
Given the following declarations for an IA-32 processor: dec ecx
MAXSIZE = 10 call ReadString
mov ecx,eax ;number of
.data
; digits entered
string BYTE MAXSIZE DUP (?) mov val,0 ;initialize val
val DWORD ? mov esi,OFFSET string
The ReadString procedure accepts the address of the cld
memory destination in edx, and the maximum number of top:
characters to read in ecx. ReadString stores the user’s input mov eax,0
characters in the memory destination, and the actual number lodsb
of characters in eax. The ASCII code for character ‘0’ is 48. sub eax,48
mov ebx,eax
Suppose that the user enters the string “5738” (without the
mov eax,val
quotes) when the ReadString procedure is called. mov edx,10
mul edx
20. (3 pts) What is displayed? add eax,ebx
mov val,eax
_____________________________________ loop top
done:
mov eax,val
21. _____ What is stored in memory at val? call WriteDec
A. 0x5 0x7 0x3 0x8
B. 0x8 0x3 0x7 0x5
C. 0x6A 0x16 0x0 0x0
D. 0x0 0x0 0x16 0x6A
E. none of the above
Use this information and the code at the right to answer questions # 22 - 23. ;code fragment W
Space is a macro that displays a specified number of blank spaces. mov eax,1
mov ebx,0
top:
22. (3 pts) What is the output of MASM code fragment W? call WriteDec
Space 2
_____________________________________________ cmp eax,13
jg quit
mov ecx,ebx
mov ebx,eax
23. _____ WriteChar displays the ASCII character in the AL register. Which add eax,ecx
of the following correctly implements the Space macro? jmp top
quit:
A. C.
Space MACRO x Space MACRO x
push EAX LOCAL again
push ECX push EAX
mov AL, ’ ’ push ECX
mov ECX,x mov AL, ’ ’
again: mov ECX,x
call WriteChar again:
loop again call WriteChar
loop again
pop ECX
pop EAX pop ECX
ENDM pop EAX
ENDM

B. D. B and C
Space MACRO x
LOCAL again E. none of these
push ECX
push EAX
mov AL, ’ ’
mov ECX,x
again:
call WriteChar
loop again
pop EAX
pop ECX
ENDM

Use this information to answer questions # 24 - 25.

Suppose that a program's data and executable code require 3200 bytes of memory. A new section of code must be
added; it will be used with various values 20 times during the execution of a program. When implemented as a
macro, the macro code requires 60 bytes of memory. When implemented as a procedure, the procedure code
requires 192 bytes (including parameter-passing, etc.), and each procedure call requires 5 bytes.

24. (3 pts) How many bytes of memory will the entire program require if the new code is added as a macro?
____________

25. (3 pts) How many bytes of memory will the entire program require if the new code is added as a
procedure?

____________
Use this information to answer questions # 26 - 31.
Given the following MASM data segment declarations:
.data
list DWORD 50 DUP(?)
matrix DWORD 20 DUP(5 DUP(?))
value DWORD 10597059 ; decimal
string BYTE ”Computer Architecture”,0
The base address of list is 1000H

26. (2 pts) What is the hexadecimal base address of matrix?

___________________

27. (2 pts) What is the hexadecimal address of list[5*TYPE list] ?

___________________

28. (2 pts) In a high-level language, the 27th element of list is referenced as list[26]. What is the hexadecimal
address of list[26] ?
___________________

29. (2 pts) In a high-level language, the 4th column of the 6th row of matrix is referenced as matrix[5][3].
What is the hexadecimal address of matrix[5][3] ?
___________________

30. (2 pts) What is the character in BYTE PTR string+6 ?

___________________

31. (2 pts) What is the decimal value in BYTE PTR value+2 ?

___________________

32. (3 pts) Show the postfix (RPN) form of the infix expression a - (b + c) / (d * e)

_____________________________________________

33. (3 pts) Find the integer value of the postfix (RPN) expression
2 7 + 3 2 - + 5 / 2 *

_____________________________________________

34. (3 pts) Show the infix form of the postfix (RPN) expression a b c d + - e / *

_____________________________________________
Assume that a, b, c, d, e, and z have all been declared as REAL10. Use the IA-32 FPU instructions in this table:

Instruction Meaning
finit Initialize the FPU
fld var Push value of var onto the register stack
fstp var Pop value from register stack into var
fadd Pop two values from register stack, add, and push result onto register stack
fsub Pop two values from register stack, subtract first popped from second popped,
and push result onto register stack
fmul Pop two values from register stack, multiply, and push result onto register stack

35. _____ (3 pts) Which of the following correctly implements the assignment statement:

z = a * (b + c) + (d - e)

NOTE: Don’t forget operator precedence !!

A. C.
finit finit
fld a fld a
fmul fld b
fld b fld c
fadd fadd
fld c fmul
fadd fld d
fld d fld e
fsub fsub
fld e fadd
fstp z fstp z

D.
B. finit
finit fld a
fld b fld b
fld c fld c
fadd fadd
fld a fld d
fmul fld e
fld d fsub
fld e fadd
fsub fmul
fadd fstp z
fstp z
E. none of these
36. (5 pts) The code below uses the Space macro defined previously. What output is generated by the MASM
"program"?

main PROC
push 1
push 1
push 5
call rfinal
exit
main ENDP

rfinal PROC
push ebp
mov ebp,esp
mov eax,[ebp+16]
mov ebx,[ebp+12]
mov ecx,[ebp+8]
mul ebx
mov [ebp+16],eax
cmp ebx,ecx
jge unwind

inc ebx
push eax
push ebx
push ecx
call rfinal
unwind:
mov eax,[ebp+16]
call WriteDec
Space 2
pop ebp
ret 12
rfinal ENDP

_____________________________________________