E. Xor-Grid Problem
time limit per test
5 seconds
memory limit per test
256 megabytes
input
standard input
output
standard output

Given a matrix $$$a$$$ of size $$$n \times m$$$, each cell of which contains a non-negative integer. The integer lying at the intersection of the $$$i$$$-th row and the $$$j$$$-th column of the matrix is called $$$a_{i,j}$$$.

Let's define $$$f(i)$$$ and $$$g(j)$$$ as the XOR of all integers in the $$$i$$$-th row and the $$$j$$$-th column, respectively. In one operation, you can either:

  • Select any row $$$i$$$, then assign $$$a_{i,j} := g(j)$$$ for each $$$1 \le j \le m$$$; or
  • Select any column $$$j$$$, then assign $$$a_{i,j} := f(i)$$$ for each $$$1 \le i \le n$$$.
An example of applying an operation on column $$$2$$$ of the matrix.

In this example, as we apply an operation on column $$$2$$$, all elements in this column are changed:

  • $$$a_{1,2} := f(1) = a_{1,1} \oplus a_{1,2} \oplus a_{1,3} \oplus a_{1,4} = 1 \oplus 1 \oplus 1 \oplus 1 = 0$$$
  • $$$a_{2,2} := f(2) = a_{2,1} \oplus a_{2,2} \oplus a_{2,3} \oplus a_{2,4} = 2 \oplus 3 \oplus 5 \oplus 7 = 3$$$
  • $$$a_{3,2} := f(3) = a_{3,1} \oplus a_{3,2} \oplus a_{3,3} \oplus a_{3,4} = 2 \oplus 0 \oplus 3 \oplus 0 = 1$$$
  • $$$a_{4,2} := f(4) = a_{4,1} \oplus a_{4,2} \oplus a_{4,3} \oplus a_{4,4} = 10 \oplus 11 \oplus 12 \oplus 16 = 29$$$

You can apply the operations any number of times. Then, we calculate the $$$\textit{beauty}$$$ of the final matrix by summing the absolute differences between all pairs of its adjacent cells.

More formally, $$$\textit{beauty}(a) = \sum|a_{x,y} - a_{r,c}|$$$ for all cells $$$(x, y)$$$ and $$$(r, c)$$$ if they are adjacent. Two cells are considered adjacent if they share a side.

Find the minimum $$$\textit{beauty}$$$ among all obtainable matrices.

Input

The first line contains a single integer $$$t$$$ ($$$1 \le t \le 250$$$) — the number of test cases.

The first line of each test case contains two integers $$$n$$$ and $$$m$$$ ($$$1 \le n, m \le 15$$$) — the number of rows and columns of $$$a$$$, respectively.

The next $$$n$$$ lines, each containing $$$m$$$ integers $$$a_{i,1}, a_{i,2}, \ldots, a_{i,m}$$$ ($$$0 \le a_{i,j} < 2^{20}$$$) — description of the matrix $$$a$$$.

It is guaranteed that the sum of $$$(n^2 + m^2)$$$ over all test cases does not exceed $$$500$$$.

Output

For each test case, print a single integer $$$b$$$ — the smallest possible $$$\textit{beauty}$$$ of the matrix.

Example
Input
4
1 2
1 3
2 3
0 1 0
5 4 4
2 3
0 2 4
4 5 1
3 3
1 2 3
4 5 6
7 8 9
Output
1
3
13
24
Note

Let's denote $$$r(i)$$$ as the first type operation applied on the $$$i$$$-th row, and $$$c(j)$$$ as the second type operation applied on the $$$j$$$-th column.

In the first test case, you can apply an operation $$$c(1)$$$, which assigns $$$a_{1,1} := 1 \oplus 3 = 2$$$. Then, we'll receive this matrix:

23

In the second test case, you can apply an operation $$$r(1)$$$, which assigns:

  • $$$a_{1,1} := g(1) = 0 \oplus 5 = 5$$$
  • $$$a_{1,2} := g(2) = 1 \oplus 4 = 5$$$
  • $$$a_{1,3} := g(3) = 0 \oplus 4 = 4$$$

The resulting matrix after performing the operation is:

554
544

In the third test case, the best way to achieve minimum $$$\textit{beauty}$$$ is applying three operations: $$$c(3)$$$, $$$r(2)$$$, and $$$c(2)$$$. The resulting matrix is:

046
456