Best solution: add or addu

Question

What is the best solution for the translation of the following C code into MIPS assembly?

x = A[i];

Note that x ⇨ $t0, A[] ⇨ $s0, i ⇨ $s1.

With add:

sll $t0, $s1, 2
add $t0, $t0, $s0

or with addu:

sll $t0, $s1, 2
addu $t0, $t0, $s0

I know the difference between add and addu, but I not understand when to use one or the other.

If you understand the difference then you should understand when to use each. Do you want a trap on overflow or not? — Konrad Lindenbach, Jul 07 '14 at 18:38
I don't know. On what basis I must choose if I want a trap or not? — ᴜsᴇʀ, Jul 07 '14 at 18:43
If you need to ask if you need an overflow trap, you do not need an overflow trap. — markgz, Jul 07 '14 at 19:28

score 3 · Answer 1 · answered Jul 07 '14 at 20:05

3

C code won't throw an exception on integer overflow. So assembly shouldn't, either. Use addu.

answered Jul 07 '14 at 20:05

Seva Alekseyev

59,826
25
160
281

score 0 · Answer 2 · answered Jul 08 '14 at 11:06

add in MIPS throws an exception on signed overflow.
Let's assume you want to index into a large array, and the array crosses between (signed)INTPTR_MAX and (signed)INTPTR_MIN. If you use add, this will throw an exception, although you may have been well within the array bounds.
What you are more likely to need to care for in array indexing is unsigned overflow, crossing from (unsigned)UINTPTR_MAX to zero, but MIPS doesn't help you there, AFAIK.

Best solution: add or addu

2 Answers2