Analyze the Time Complexity of Nested Loops

Work out how the running time of the C-style pseudocode below grows, and express it in Big-O / Big-Theta notation:

for (int i = 0; i < n; i++) {
    for (int j = 0; j < i * i; j++) {
        for (int k = i; ~k; k -= 1) {
            temp++;
        }
    }
}

The expression ~k denotes bitwise NOT. Integers are stored in two's-complement form, and you may assume $n \ge 1$.