* Added kernels for the operations v1 += v2 @ alpha +- v3 @ beta and v1 -= v2...

* Added kernels for the operations v1 += v2 @ alpha +- v3 @ beta and v1 -= v2 @ alpha +- v3 @ beta, with @ denoting either product or divsion and alpha, beta being CPU or GPU scalars.
* Removed debug output in matrix-test
* Fixed issue with v = a1 v1 + a2 v2 + .. + a_k v_k for k>3
* Added is_any_matrix predicate
* Further improvements to vector, vector-range, and vector-slice interplay.