added optimized versions of static_vector operations, however preliminary tests on linux shows no speed difference