2IC30·Computer Systems

Courses
Computer Systems
Exam Preparation
PP2 simulation exercises

2IC30 · assessment

PP2 simulation exercises

Thirteen editable PP2 assembly exercises imported from the original simulator bank.

1/13

Not started

Subroutine Sum receives n on the stack and returns 1+2+…+n in the caller-reserved result slot. The driver (given) loads n=5 and copies the returned value to memory[100] so it can be checked. Write the body of Sum . Expected: 15 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller / driver (given — do not change) ----
    LOAD R0 5         ; n = 5  (try other values)
    SUB  SP 1         ; reserve result slot
    STOR R0 [--SP]    ; push argument n
    BRS  Sum          ; call (return address pushed)
    LOAD R1 [SP+1]    ; read result (sits above return addr)
    STOR R1 [100]     ; publish to a fixed address for checking
    ADD  SP 2         ; pop arg + result slot
    BRA  end
; ---- subroutine Sum: stack = [SP]ret [SP+1]n [SP+2]result ----
Sum:
    ; TODO: read n from [SP+1], compute 1+2+...+n,
    ;       write the sum into the result slot [SP+2], then RTS.
    RTS
end:

Self-mark against the exam rubric

Calling convention stated (args on stack, result slot reserved by caller).
Callee reads argument n from the correct stack slot ([SP+1]).
Loop accumulates the sum and terminates (n reaches 0).
Result written to the caller-reserved slot, not just left in a register.
Stack balanced on return (SP restored; RTS pops the return address).

0/5 criteria met

Compute n! with a countdown loop. The driver loads n=5 and the result is published to memory[100]. Write the loop that leaves n! in R1. Expected: 120 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 5         ; n  (given)
    ; TODO: compute n! into R1 (remember 0! = 1! = 1)
    STOR R1 [100]     ; publish result (given)
end:

Self-mark against the exam rubric

Accumulator initialised to 1 (not 0).
Loop multiplies by each value from n down to 2.
Termination test correct (n ≤ 1 stops; handles n=0,1 → 1).
Final product stored to the result location.

0/4 criteria met

Determine whether n is prime by trial division. The driver loads n=7 (prime). Write code that sets R1 = 1 if n is prime, else 0. Result goes to memory[100]. Change LOAD R0 7 to 9 to test a composite. Expected: 1 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 7         ; n  (given): 7 = prime, 9 = not
    ; TODO: set R1 = 1 if n is prime, else R1 = 0.
    ;   hint: handle n<2 and n==2, then trial-divide d=2,3,... while d*d <= n;
    ;   compute n mod d by repeated subtraction.
    STOR R1 [100]     ; publish (given): 1 = prime, 0 = not
end:

Self-mark against the exam rubric

Handles small cases (n<2 not prime, n==2 prime).
Loops divisors only while d*d ≤ n (correct stopping condition).
Computes divisibility correctly (remainder via repeated subtraction).
Returns 0 on the first divisor found, 1 if none divides n.
Result placed at the agreed location; registers used are documented.

0/5 criteria met

Find the largest element of a 5-element array preloaded at memory[200..204] = {3,9,2,7,5}. The pointer (R0=200) and count (R1=5) are given. Walk the array and leave the maximum in R2. Result to memory[100]. Expected: 9 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 200       ; pointer to array (given)
    LOAD R1 5         ; element count (given)
    ; TODO: scan the 5 elements at [R0], [R0+1], ... and leave the max in R2.
    ;   load an element with  LOAD R3 [R0]  then  ADD R0 1  to advance.
    STOR R2 [100]     ; publish maximum (given)
end:

Self-mark against the exam rubric

Initialises max from the first element (not 0).
Iterates over exactly the remaining elements via pointer + counter.
Comparison/update keeps the larger value.
Result written to the result location after the loop.

0/4 criteria met

Iterative Fibonacci: fib(0)=0, fib(1)=1, fib(k)=fib(k-1)+fib(k-2). The driver loads n=10. Write the loop that leaves fib(n) in R1. Result to memory[100]. Expected: 55 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 10        ; n  (given)
    ; TODO: compute fib(n) into R1.
    ;   keep two running values a=fib(k-1), b=fib(k); update them n times.
    STOR R1 [100]     ; publish result (given)
end:

Self-mark against the exam rubric

Base cases handled (n=0 → 0, n=1 → 1).
Two running values updated in the right order each iteration.
Loop counter decremented and termination correct.
Final value stored to the result location.

0/4 criteria met

Greatest common divisor with Euclid: gcd(a,b) = gcd(b, a mod b) until b = 0. The driver loads a=48, b=36. Write the loop that leaves gcd in R0. Uses the MOD instruction. Result to memory[100]. Expected: 12 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 48        ; a  (given)
    LOAD R1 36        ; b  (given)
    ; TODO: while b != 0:  r = a mod b ; a = b ; b = r .  Leave gcd in R0.
    ;   hint:  MOD R2 R1  computes R2 = R2 mod R1.
    STOR R0 [100]     ; publish gcd (given)
end:

Self-mark against the exam rubric

Loop continues while b ≠ 0 and stops when b = 0.
Remainder computed with MOD (a mod b) each iteration.
a and b reassigned in the correct order (a←b, b←r).
gcd left in the agreed register and stored to the result location.

0/4 criteria met

Compute x n by repeated multiplication. The driver loads x=2, n=10. Write the loop that leaves the result in R2. Result to memory[100]. Expected: 1024 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 2          ; x  (given)
    LOAD R1 10         ; n  (given)
    ; TODO: set R2 = x^n  (remember x^0 = 1). Multiply R2 by x, n times.
    STOR R2 [100]      ; publish result (given)
end:

Self-mark against the exam rubric

Accumulator initialised to 1 (handles n=0 → 1).
Loop multiplies the accumulator by x exactly n times.
Counter decremented and termination test correct.
Result stored to the result location.

0/4 criteria met

Return the index of the first element equal to 7 in a 5-element array at memory[200..204] = {3,9,2,7,5}, or −1 if absent. Pointer R0=200, count R1=5, target R2=7 are given. Leave the index (or −1) in R3. Result to memory[100]. Expected: 3 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 200        ; array pointer (given)
    LOAD R1 5          ; count (given)
    LOAD R2 7          ; target (given)
    ; TODO: scan the array; set R3 = index of first match, else R3 = -1.
    ;   load an element with  LOAD R4 [R0]  then  ADD R0 1  to advance.
    STOR R3 [100]      ; publish index (given)
end:

Self-mark against the exam rubric

Iterates over exactly the n elements via pointer + counter.
Compares each element to the target and stops on first match.
Returns the correct 0-based index of the match.
Returns −1 when the value is not present (loop exhausted).

0/4 criteria met

Compute a × b using only addition — the classic 'no multiply instruction' exercise. The driver loads a=6, b=7. Write the loop that leaves the product in R2. Result to memory[100]. Expected: 42 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 6          ; a  (given)
    LOAD R1 7          ; b  (given)
    ; TODO: set R2 = a*b by adding a to R2 exactly b times. Do NOT use MUL.
    STOR R2 [100]      ; publish product (given)
end:

Self-mark against the exam rubric

Product accumulator initialised to 0.
Adds a to the accumulator exactly b times (counter loop).
No MUL instruction used (addition only).
Product stored to the result location.

0/4 criteria met

Reverse the 5-element array at memory[200..204] = {1,2,3,4,5} in place using two pointers that meet in the middle. Left pointer R0=200 and right pointer R1=204 are given. Swap from the ends inward. The check reads memory[200] (should be 5) into memory[100]. Expected: 5 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 200        ; left pointer (given)
    LOAD R1 204        ; right pointer (given)
    ; TODO: while left < right, swap [R0] and [R1], then R0++ and R1--.
    LOAD R4 [200]      ; (given) read new first element ...
    STOR R4 [100]      ; ... and publish it for checking
end:

Self-mark against the exam rubric

Two pointers start at the two ends of the array.
Each iteration swaps the elements at the pointers (uses temporaries).
Pointers move inward and the loop stops when they meet/cross.
Array is reversed in place (no second buffer).

0/4 criteria met

Recursive n! exercising the full calling convention: fac(n) = 1 if n≤1, else n·fac(n−1). The driver loads n=5, reserves a result slot, pushes n, calls with BRS , and publishes the returned value to memory[100]. Write the body of Fac — it must save n across its recursive call. Expected: 120 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller / driver (given — do not change) ----
    LOAD R0 5         ; n = 5
    SUB  SP 1         ; reserve result slot
    STOR R0 [--SP]    ; push argument n
    BRS  Fac          ; call
    LOAD R1 [SP+1]    ; read result (above return addr)
    STOR R1 [100]     ; publish for checking
    ADD  SP 2         ; pop arg + result slot
    BRA  end
; ---- Fac: stack on entry = [SP]ret [SP+1]n [SP+2]result ----
Fac:
    ; TODO: if n<=1 write 1 to [SP+2] and RTS.
    ;   else recurse on n-1 (reserve slot, push n-1, BRS Fac, read child result),
    ;   compute n * child, write to [SP+2], RTS.
    RTS
end:

Self-mark against the exam rubric

Calling convention stated (result slot reserved by caller, arg pushed, BRS/RTS).
Base case n ≤ 1 returns 1 into the caller's result slot.
Recursive case allocates a child frame and calls Fac on n−1 with BRS.
n is preserved across the recursive call (re-read after SP adjusts).
Stack fully balanced on every return (child frame popped; RTS pops return addr).

0/5 criteria met

Real 2018 exam, Question 5. Translate IsPrime(n) to a PP2 subroutine. The parameter and result share one stack slot: caller does PUSH R0 / BRS IsPrime / PULL R0 . The subroutine builds a boolean array a [0..n] on the stack ( SUB SP R0 allocates n+1 words; a [i] = RAM[SP+i]), marks all true, then sieves out multiples, and returns a [n]. n is at SP+1 on entry; the result goes back to SP+1. Driver loads n=7. Write the body of IsPrime . Expected: 1 (7 is prime).

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller (given) ----
    LOAD R0 7         ; n = 7  (try 9, 12, 13 …)
    PUSH R0           ; push n (same slot is reused for the result)
    BRS  IsPrime
    PULL R0           ; result: 1 = prime, 0 = not
    STOR R0 [100]     ; publish for checking
    BRA  progend
; ---- subroutine: n at [SP+1] on entry, result to [SP+1] on exit ----
IsPrime:
    ; TODO: allocate a[0..n] on the stack, set a[2..n]=true, sieve out
    ;   multiples (a[2i],a[3i],… = false), then return a[n].
    ;   register-indexed addressing [SP+Ri] reaches a[i]. Use BHI (unsigned i>n).
    RTS
progend:

Self-mark against the exam rubric

Calling convention: parameter and result share stack slot SP+1 (PUSH/BRS/PULL).
Array a[0..n] allocated on the stack (SUB SP n; SUB SP 1) and freed before RTS.
First loop sets a[2..n] = true using register-indexed addressing [SP+Ri].
Sieve marks multiples of each i (a[2i], a[3i], …) as false.
Returns a[n], stack restored, BHI used for the unsigned i>n bound.

0/5 criteria met

Real 2017 exam, Question 5. Translate the procedure addSquares(m,n,a) that computes c = Σ i<m Σ j<n a[i·n+j]² to a PP2 subroutine. Parameters are on the stack: m at SP+1, n at SP+2, base address a at SP+3; result to SP+4. Element a[i·n+j] is RAM[a+i·n+j]. Overflow ( BVS ) jumps to exception . The driver passes a 2×2 matrix {1,2,3,4} (m=n=2). Write the body of addSquares . Expected: 30 (1²+2²+3²+4²).

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller (given) ----
    SUB  SP 1         ; result slot
    LOAD R0 200       ; address of matrix a (preloaded at 200..203)
    PUSH R0           ; push a
    LOAD R0 2
    PUSH R0           ; push n
    LOAD R0 2
    PUSH R0           ; push m
    BRS  addSquares
    ADD  SP 3         ; release m, n, a
    PULL R0           ; result
    STOR R0 [100]
    BRA  progend
; ---- subroutine: m@SP+1, n@SP+2, a@SP+3, result@SP+4 ----
addSquares:
    ; TODO: nested loops i<m, j<n; k = RAM[a+i*n+j]; c += k*k (check BVS overflow).
    ;   store c at [SP+4], then RTS.
    RTS
exception:
    LOAD R0 -1
    STOR R0 [SP+4]
    RTS
progend:

Self-mark against the exam rubric

Calling convention: m@SP+1, n@SP+2, a@SP+3, result@SP+4; caller releases params.
Nested loops bounded by m (outer) and n (inner), counters start at 0.
Element address computed as a + i·n + j and dereferenced with LOAD R3 [R3].
k² accumulated into c; BVS overflow check after the multiply and the add.
Result stored at SP+4 and RTS returns; registers documented.

0/5 criteria met

Exam Preparation

Courses
Computer Systems
Exam Preparation
PP2 simulation exercises

2IC30 · assessment

PP2 simulation exercises

Thirteen editable PP2 assembly exercises imported from the original simulator bank.

1/13

Not started

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller / driver (given — do not change) ----
    LOAD R0 5         ; n = 5  (try other values)
    SUB  SP 1         ; reserve result slot
    STOR R0 [--SP]    ; push argument n
    BRS  Sum          ; call (return address pushed)
    LOAD R1 [SP+1]    ; read result (sits above return addr)
    STOR R1 [100]     ; publish to a fixed address for checking
    ADD  SP 2         ; pop arg + result slot
    BRA  end
; ---- subroutine Sum: stack = [SP]ret [SP+1]n [SP+2]result ----
Sum:
    ; TODO: read n from [SP+1], compute 1+2+...+n,
    ;       write the sum into the result slot [SP+2], then RTS.
    RTS
end:

Self-mark against the exam rubric

Calling convention stated (args on stack, result slot reserved by caller).
Callee reads argument n from the correct stack slot ([SP+1]).
Loop accumulates the sum and terminates (n reaches 0).
Result written to the caller-reserved slot, not just left in a register.
Stack balanced on return (SP restored; RTS pops the return address).

0/5 criteria met

Compute n! with a countdown loop. The driver loads n=5 and the result is published to memory[100]. Write the loop that leaves n! in R1. Expected: 120 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 5         ; n  (given)
    ; TODO: compute n! into R1 (remember 0! = 1! = 1)
    STOR R1 [100]     ; publish result (given)
end:

Self-mark against the exam rubric

Accumulator initialised to 1 (not 0).
Loop multiplies by each value from n down to 2.
Termination test correct (n ≤ 1 stops; handles n=0,1 → 1).
Final product stored to the result location.

0/4 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 7         ; n  (given): 7 = prime, 9 = not
    ; TODO: set R1 = 1 if n is prime, else R1 = 0.
    ;   hint: handle n<2 and n==2, then trial-divide d=2,3,... while d*d <= n;
    ;   compute n mod d by repeated subtraction.
    STOR R1 [100]     ; publish (given): 1 = prime, 0 = not
end:

Self-mark against the exam rubric

Handles small cases (n<2 not prime, n==2 prime).
Loops divisors only while d*d ≤ n (correct stopping condition).
Computes divisibility correctly (remainder via repeated subtraction).
Returns 0 on the first divisor found, 1 if none divides n.
Result placed at the agreed location; registers used are documented.

0/5 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 200       ; pointer to array (given)
    LOAD R1 5         ; element count (given)
    ; TODO: scan the 5 elements at [R0], [R0+1], ... and leave the max in R2.
    ;   load an element with  LOAD R3 [R0]  then  ADD R0 1  to advance.
    STOR R2 [100]     ; publish maximum (given)
end:

Self-mark against the exam rubric

Initialises max from the first element (not 0).
Iterates over exactly the remaining elements via pointer + counter.
Comparison/update keeps the larger value.
Result written to the result location after the loop.

0/4 criteria met

Iterative Fibonacci: fib(0)=0, fib(1)=1, fib(k)=fib(k-1)+fib(k-2). The driver loads n=10. Write the loop that leaves fib(n) in R1. Result to memory[100]. Expected: 55 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 10        ; n  (given)
    ; TODO: compute fib(n) into R1.
    ;   keep two running values a=fib(k-1), b=fib(k); update them n times.
    STOR R1 [100]     ; publish result (given)
end:

Self-mark against the exam rubric

Base cases handled (n=0 → 0, n=1 → 1).
Two running values updated in the right order each iteration.
Loop counter decremented and termination correct.
Final value stored to the result location.

0/4 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 48        ; a  (given)
    LOAD R1 36        ; b  (given)
    ; TODO: while b != 0:  r = a mod b ; a = b ; b = r .  Leave gcd in R0.
    ;   hint:  MOD R2 R1  computes R2 = R2 mod R1.
    STOR R0 [100]     ; publish gcd (given)
end:

Self-mark against the exam rubric

Loop continues while b ≠ 0 and stops when b = 0.
Remainder computed with MOD (a mod b) each iteration.
a and b reassigned in the correct order (a←b, b←r).
gcd left in the agreed register and stored to the result location.

0/4 criteria met

Compute x n by repeated multiplication. The driver loads x=2, n=10. Write the loop that leaves the result in R2. Result to memory[100]. Expected: 1024 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 2          ; x  (given)
    LOAD R1 10         ; n  (given)
    ; TODO: set R2 = x^n  (remember x^0 = 1). Multiply R2 by x, n times.
    STOR R2 [100]      ; publish result (given)
end:

Self-mark against the exam rubric

Accumulator initialised to 1 (handles n=0 → 1).
Loop multiplies the accumulator by x exactly n times.
Counter decremented and termination test correct.
Result stored to the result location.

0/4 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 200        ; array pointer (given)
    LOAD R1 5          ; count (given)
    LOAD R2 7          ; target (given)
    ; TODO: scan the array; set R3 = index of first match, else R3 = -1.
    ;   load an element with  LOAD R4 [R0]  then  ADD R0 1  to advance.
    STOR R3 [100]      ; publish index (given)
end:

Self-mark against the exam rubric

Iterates over exactly the n elements via pointer + counter.
Compares each element to the target and stops on first match.
Returns the correct 0-based index of the match.
Returns −1 when the value is not present (loop exhausted).

0/4 criteria met

Compute a × b using only addition — the classic 'no multiply instruction' exercise. The driver loads a=6, b=7. Write the loop that leaves the product in R2. Result to memory[100]. Expected: 42 .

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 6          ; a  (given)
    LOAD R1 7          ; b  (given)
    ; TODO: set R2 = a*b by adding a to R2 exactly b times. Do NOT use MUL.
    STOR R2 [100]      ; publish product (given)
end:

Self-mark against the exam rubric

Product accumulator initialised to 0.
Adds a to the accumulator exactly b times (counter loop).
No MUL instruction used (addition only).
Product stored to the result location.

0/4 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

    LOAD R0 200        ; left pointer (given)
    LOAD R1 204        ; right pointer (given)
    ; TODO: while left < right, swap [R0] and [R1], then R0++ and R1--.
    LOAD R4 [200]      ; (given) read new first element ...
    STOR R4 [100]      ; ... and publish it for checking
end:

Self-mark against the exam rubric

Two pointers start at the two ends of the array.
Each iteration swaps the elements at the pointers (uses temporaries).
Pointers move inward and the loop stops when they meet/cross.
Array is reversed in place (no second buffer).

0/4 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller / driver (given — do not change) ----
    LOAD R0 5         ; n = 5
    SUB  SP 1         ; reserve result slot
    STOR R0 [--SP]    ; push argument n
    BRS  Fac          ; call
    LOAD R1 [SP+1]    ; read result (above return addr)
    STOR R1 [100]     ; publish for checking
    ADD  SP 2         ; pop arg + result slot
    BRA  end
; ---- Fac: stack on entry = [SP]ret [SP+1]n [SP+2]result ----
Fac:
    ; TODO: if n<=1 write 1 to [SP+2] and RTS.
    ;   else recurse on n-1 (reserve slot, push n-1, BRS Fac, read child result),
    ;   compute n * child, write to [SP+2], RTS.
    RTS
end:

Self-mark against the exam rubric

Calling convention stated (result slot reserved by caller, arg pushed, BRS/RTS).
Base case n ≤ 1 returns 1 into the caller's result slot.
Recursive case allocates a child frame and calls Fac on n−1 with BRS.
n is preserved across the recursive call (re-read after SP adjusts).
Stack fully balanced on every return (child frame popped; RTS pops return addr).

0/5 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller (given) ----
    LOAD R0 7         ; n = 7  (try 9, 12, 13 …)
    PUSH R0           ; push n (same slot is reused for the result)
    BRS  IsPrime
    PULL R0           ; result: 1 = prime, 0 = not
    STOR R0 [100]     ; publish for checking
    BRA  progend
; ---- subroutine: n at [SP+1] on entry, result to [SP+1] on exit ----
IsPrime:
    ; TODO: allocate a[0..n] on the stack, set a[2..n]=true, sieve out
    ;   multiples (a[2i],a[3i],… = false), then return a[n].
    ;   register-indexed addressing [SP+Ri] reaches a[i]. Use BHI (unsigned i>n).
    RTS
progend:

Self-mark against the exam rubric

Calling convention: parameter and result share stack slot SP+1 (PUSH/BRS/PULL).
Array a[0..n] allocated on the stack (SUB SP n; SUB SP 1) and freed before RTS.
First loop sets a[2..n] = true using register-indexed addressing [SP+Ri].
Sieve marks multiples of each i (a[2i], a[3i], …) as false.
Returns a[n], stack restored, BHI used for the unsigned i>n bound.

0/5 criteria met

PP2 Assembly·simulate·pp2-browser

Source code for main.pp2

; ---- caller (given) ----
    SUB  SP 1         ; result slot
    LOAD R0 200       ; address of matrix a (preloaded at 200..203)
    PUSH R0           ; push a
    LOAD R0 2
    PUSH R0           ; push n
    LOAD R0 2
    PUSH R0           ; push m
    BRS  addSquares
    ADD  SP 3         ; release m, n, a
    PULL R0           ; result
    STOR R0 [100]
    BRA  progend
; ---- subroutine: m@SP+1, n@SP+2, a@SP+3, result@SP+4 ----
addSquares:
    ; TODO: nested loops i<m, j<n; k = RAM[a+i*n+j]; c += k*k (check BVS overflow).
    ;   store c at [SP+4], then RTS.
    RTS
exception:
    LOAD R0 -1
    STOR R0 [SP+4]
    RTS
progend:

Self-mark against the exam rubric

Calling convention: m@SP+1, n@SP+2, a@SP+3, result@SP+4; caller releases params.
Nested loops bounded by m (outer) and n (inner), counters start at 0.
Element address computed as a + i·n + j and dereferenced with LOAD R3 [R3].
k² accumulated into c; BVS overflow check after the multiply and the add.
Result stored at SP+4 and RTS returns; registers documented.

0/5 criteria met