Benchmark hybrid approach in Teds\unique_values, use if practical

[TysonAndre]

I've seen a hybrid approach in other implementations of methods to deduplicate values, since a hash table and sorting do worse than nested for loops in practice at very small sizes

• For small n=count($input), use a regular foreach over buckets in the zend_array to be returned and call zend_is_identical one by one
• For large n, use a hash table similar to the previous comment