The number of sparse * dense matrix multiplication seems completely excessive

[KristofferC]

I count ~240 of them only testing sparse * dense matrix multiplication. We cannot have this type of exhaustive testing here, in my opinion.