Question

Translate the following C code into MIPS assembly code. Assume a in $s0, i in $s1, and base address of b[] in $s2, respectively. Note that the element’s data type in array b is integer (word). Please do not change the C code structure and please comment your MIPS code.

for (i = 100; i>a; i--) {         

b[i] = 16*i;

b[i+1] = 16*(i+1);

b[i+2] = 2*i + 16;

}

Answer 1

loop:
   slt $t1, $s0, $s1       # if a<i set $t1 = 1 else $t1 = 0
   beq $t1, 0, end         # if $t1 is 0, end loop
   sll $t2, $s1, 2           # calculate offset address by multiplying i by 4
   add $t2, $t2, $s2       # adding base address to offset address
   sll $t3, $s1, 4           # calculating 16*i by shifting 4 bits
   sw $t3, ($t2)           # store word to b[i]
   addi $t3, $t3, 16       # add 16 to previously calculated (16*i) , 16 * (i+1) = (16 * i) + 16
   addi $t2, $t2, 4       # increasing address by adding 4
   sw $t3, ($t2)           # store word to b[i+1]
   add $t3, $s1, $s1       # adding i+i = 2*i and storing to $t3
   addi $t3, $t3, 16       # adding 16, so $t3 = ((2*i)+16)
   addi $t2, $t2, 4       # increasing address by adding 4
   sw $t3, ($t2)           # store word to b[i+1]
   addi $s1, $s1, 1       # incrementing i
   j loop                   # back to loop