Description

Computer Organization

Midterm Exam – Term 172

10:00 am – 12:00 noon

Computer Engineering Department

Student Name: ID:

Dr. Aiman El-Maleh COE 301 ICS 233
Dr. Marwan Abu-Amara COE 301 ICS 233
Dr. Muhamed Mudawar ICS 233

Q1 / 19 Q2 / 17
Q3 / 07 Q4 / 14
Q5 / 07 Q6 / 16
Total / 80

Q1. [19 points] Fill-in the Blanks
a) (1 point) In addition to being space and time efficient, programming in assembly language has the advantage of _____________________________________________.
b) (1 point) In addition to faster program development and maintenance, programming in high-level language has the advantage _______________________________________.
c) (1 point) The instruction set architecture of a processor provides an interface between _______________________________________.
d) (1 point) Assuming Array is defined as shown below:
Array: .word 10, 11, 12, 13, 14
The content of register $t1 (in hexadecimal) after executing the following sequence of instructions is ________________________. la $t0, Array lw $t1, 4($t0)
e) (2 points) Write a minimum sequence of MIPS basic instructions to implement the pseudo instruction: bgt $s1,$s2,Next

f) (2 points) Write a minimum sequence of MIPS basic instructions to implement the pseudo instruction: andi $t0,$t0,0x12345678

g) (1 point) Assuming that $a0 contains an Alphabetic character (uppercase or lowercase), write a MIPS instruction that will make the character stored in $a0 always lower case. Note that the ASCII code of character ‘A’ is 0x41 while that of character ‘a’ is 0x61.

h) (3 points) The following is a partial MIPS assembly language code:
Address Label Instruction
0x00400000

bgtz $a1, loop
. . .
0x00403000
loop:
addu $a0, $a1, $v0
. . .
0x00410000 bne $a0, $zero, loop
Calculate the 16-bit immediate value (in hexadecimal) for loop in the bgtz instruction.
Calculate the 16-bit immediate value (in hexadecimal) for loop in the bne instruction.

i) (2 points) Given that Array2 is defined as shown below:
Array2: .byte -1, 2, -3, -4, -5, 6
After executing the following sequence of instructions, the content of the two registers (in hexadecimal) is $t1=_________________ and $t2=____________________.
la $t0, Array2 lb $t1, 2($t0) lhu $t2, 2($t0)
j) (2 points) Given the following data definitions, and assuming that the first variable X starts at address 0x10010000, then the addresses for variables Y and Z will be ________________ and __________________.
.data
X: .byte 10, 11, 12
Y: .half 13, 14, 15 Z: .word 16, 17, 18
k) (3 points) Write a minimum sequence of MIPS basic instructions to multiply the signed integer value of register $t0 by 15.25 without using multiplication instructions. Put the final integer result in $t0. For example, if the initial value of $t0 is 5 then the final value will be 76. The additional fraction is truncated.

Q2. Floating-Point [17 points]
a) (3 points) Find the decimal value of the following single-precision float:
S Exponent Fraction
1 1000 1000 111 1100 1101 0000 0000 0000

b) (4 points) Find the normalized IEEE 754 single-precision representation of +59.625.

c) (2 points) Show the IEEE 754 representation of +Zero and -Infinity for single precision:

d) (2 points) Find the approximate decimal value of the largest positive denormalized float for single precision.

e) (6 points) Given that A and B are single-precision floats, compute the difference A–B. Use rounding to nearest even. Perform the operation using guard, round and sticky bits.
A = +1.111 0000 0000 1111 1100 0001 × 2-4
B = +1.000 1111 1111 0000 0000 1111 × 2+3

Q3. [7 points] Translate Nested IF statements
Using minimal number of instructions, translate the following nested IF statements into MIPS assembly code. Assume that variables a, b, c, and d are signed integers stored in registers $s0, $s1, $s2, and $s3, respectively. If needed, you can use pseudo-branch instructions only.
if ( ((a > b) || (c <= d)) && (a == c) ) { if (c != d) a = b + c;
else c = b – 2;
}

Q4. [14 points] Translate a Recursive Function
Translate the following high-level recursive function freq into MIPS assembly code. The freq function counts the number of times an integer i appears in an array A of n integers. The array A is already in memory. The function parameters are passed in registers $a0, $a1 and $a2, respectively. The freq function returns the result in register $v0. Use MIPS convention in saving and restoring only the necessary register(s) in the recursive function. Your MIPS implementation of the freq function must be recursive. Add comments to explain the utilization of registers.
int freq(int A[], int n, int i) { if (n == 0) return 0; int j = 0; if (A[n-1] == i) j = 1; return (j + freq(&A[0], n-1, i)); }

Q5. [7 points] Compute the Sum of Decimal Digits
Write a MIPS function sum_digits that computes and returns the sum of decimal digits in an unsigned integer. For example, the sum of decimal digits for 1536 is 1+5+3+6 = 15. The function sum_digits receives the unsigned integer argument in binary in register $a0. For example, 1536 = (0000 … 0110 0000 0000)binary. It should extract the decimal digits, compute, and return their sum, also in binary, in register $v0. Hint: divide the unsigned integer by 10 to extract the decimal digits.

Q6. [16 points] Write Loops to Traverse a Matrix

Given that M is a square matrix of N×N integers (N rows by N columns), which is already read and initialized in memory. The starting address of matrix M is stored in register $a0, and N is stored in register $a1. This is always the case for parts (a) and (b).

a) (7 points) Write MIPS code to compute the sum of all N elements of row i, where i < N. The value of i is stored in register $a2. The sum should be computed in $v0. Add comments to explain the utilization of registers. Do NOT use pseudo-instructions.