Matrix optimizations

• Add to the benchmark suite: problems with mild and severe row restrictions, problems with mild and severe column restrictions.
• Add exceptions everywhere for when cols is None or rows is None.
• Add output array as an optional argument to dot/transpose_dot/sandwich. This will be useful for avoiding unnecessary allocations. It will solve https://github.com/Quantco/quantcore.matrix/issues/25 and will enable some improvements in speed, particularly for SplitMatrix.dot which is allocating a new output array for each submatrix. After the optimization, SplitMatrix will only need one output array instead of len(matrices). Relatedly (complete now): flatironinstitute/sparse_dot#9
• Cythonize row and column limiting for categorical sandwiches (currently just have rows)
• Think about reducing duplication across different implementations that differ based on whether rows or cols are restricted: https://github.com/Quantco/quantcore.matrix/issues/5 -- is there some inspiration that would come from thinking about the restrictions as sparse matrices?

Start with the benchmark suite and the dot, transpose_dot, and sandwich entry points mentioned in the issue. Trace SplitMatrix.dot and the categorical sandwich implementations, then review the linked issues for context on output-array allocation and row or column restrictions. Done requires covering the listed benchmark cases and completing the selected optimization work without leaving the related restriction behavior unresolved.