Down to one digit
digsum(n) adds up the decimal digits of n: digsum(1234) is 1 + 2 + 3 + 4 = 10. Apply it again and again and the answer shrinks to one digit, the number's digital root: 1234 gives 10, and 10 gives 1. The starter's main reads numbers until the input runs out and prints both results for each one, calling digsum once for the digit sum and once more for every step toward the root. Write digsum.
Split off the lowest digit with a division by 10: the remainder (udiv, then msub) is the digit, and the quotient is what is left. Stop when nothing is left, so digsum(0) is 0. The input is never negative.
digsum calls nothing, which makes it a leaf: it needs no frame of its own as long as it works only in x0 and x9 to x15. main keeps its values in x19, x20 and x21 across the calls, so a digsum that borrows those breaks main.
The starter's input prints:
1234: digit sum 10, digital root 1
99999: digit sum 45, digital root 9
What is checked
- the two lines above
- the program exits with status 0
- the same checks on other inputs the checker keeps hidden, including 0, the largest 64-bit number, and no input at all
specification
We run your program on the input above and on the hidden ones, and compare what it does. Nothing is matched against a stored solution.