Consider a general recursive mechanism $R$, whose input is an array A() initially with $n$ elements. The pseudo code of $R$ is listed below. It takes a parameter $k$, use $\sqrt{n}$ ($n$ is the size of A when applied, $n$ may change) steps to partition A, and may call another function $Q$, whose complexity may vary and depends on the size of its input.

Procedure R (array A with size n, int k)
1. if n<k exit;
2. partition A into k equal size parts
   A1, A2, ..., Ak (This takes sqrt(n) steps)
3. for (i=1 to k)
   . call R(Ak, k); // may change Ak
4. merge A1, A2, ..., Ak into new A with size n;
   (This takes k steps)
5. call Q(A);
6. return()

There are 3 derived algorithms X, Y, Z, based on $R$. X sets $k$ as 2, and its Q's complexity is $\theta(m)$, $m$ is the size of Q's input. Y sets $k$ as 4, and its Q's complexity is $\theta(\sqrt{m})$, $m$ is input size. Z sets $k$ as 9, and its Q's complexity is $\theta(m)$, $m$ is input size.

What following comparisons are true?