That's still better than matrix_add(c, matrix_mul(b, a)) though - you can't do A.mul(B).add(C) in C