The * operator can be overloaded, and it is already done s so for the mxm, mxv and vxm combinations. The catch is, all these AxB functions have more arguments

The * operator can be overloaded, and it is already done s so for the mxm, mxv and vxm combinations. The catch is, all these AxB functions have more arguments than just their operands (mask, accumulator, semiring, descriptors). So you can `a_matrix * b_vector` and get "classical" matrix multiplication, but if you want accumulation, masking, a non-default semiring etc, you have to use the functional notation.