I have an $(kn \times kn)$ dimensional square positive semi-definite symmetric matrix of the form:
$$A=\begin{bmatrix}A_{11} & A_{21} & ... & A_{n1} \\ A_{21}' & A_{22} & ... & A_{n2} \\ \vdots & \vdots & \ddots & \vdots \\ A_{n1}' & A_{n2}' & ... & A_{nn} \\ \end{bmatrix}$$
Where the blocks, $\{A_{ij}\}_{i, j\in \{1,...n\}}$, are square and of dimension $(k \times k)$. Denote the inverse of $A$ as $B$, so that $B^{-1}=A$ and partition $B$ as:
$$B=\begin{bmatrix}B_{11} & B_{21} & ... & B_{n1} \\ B_{21}' & B_{22} & ... & B_{n2} \\ \vdots & \vdots & \ddots & \vdots \\ B_{n1}' & B_{n2}' & ... & B_{nn} \\ \end{bmatrix}$$
Where the blocks, $\{B_{ij}\}_{i, j\in \{1,...n\}}$, are also square and of dimension $(k \times k)$. I want to find the sum of the diagonal blocks of $B$ in terms of the original blocks of $A$. That is, I want to find:
$$\sum_{i=1}^n B_{ii}$$
In terms of the $\{A_{ij}\}_{i,j \in \{1,...,n\}}$. Does anyone know of any nice formulae for this?