Question

I’d love your help with the following problem.

I was asked to implement the function is_cube, which receives n as an argument and checks if it is a cube, in MIPS assembly. For example 8 (2^3) and 1000 (10^3) are cubes.

I wrote the following code:

# UNTITLED PROGRAM

.data   

str: .asciiz "Please enter your number >"
str1: .asciiz "The number is a cube"
str2: .asciiz "The number is not a cube"


.text

 main:  

   li $v0 4
   la $a0 str
   syscall

   li $v0 5
   syscall


   move $t0 , $v0
   li $t1, 0
   blt $t0, $zero, negative


  negative:
    sub $t5, $zero, 1
mul $t0, $t0, $t5



 is_cube: 

  addi $t1, $t1, 1
  sgt $t2, $t1, $t0
  bne $t2, $zero, There_is_not 
  mul $t3, $t1, $t1
  mul $t4 ,$t3, $t1
  beq $t4, $t0, There_is

  jal is_cube



 There_is:

   li $v0 4
   la $a0 str1
   syscall

   jal end

 There_is_not:

   li $v0 4
   la $a0 str2
   syscall

   jal end


 end: 

and it works, without saving $s0 in a stack by $sp and all of this process. My question is: Is it Ok not saving it? and if so, When should I use it?

Thanks a lot.

Answer 1

You haven’t implemented a function. You have implemented a loop, which happens to use jal instead of a normal branch instruction. Execution starts in main and then falls through from negative to a loop starting at is_cube, and then when the loop terminates at There_is or There_is_not you use jal end instead of another branch instruction.

This code happens to work on MIPS because the jal instruction does not push the current value of the program counter to the stack, unlike the x86.

If this is homework, you should go back to your notes on writing functions in assembler before submitting this code.