C++ manipulation of HUGE vector using threads: can it be optimized further?

I have a very simple class that contains a single integer. The class has a single function that increments that integer by one. Basically:I have a very simple class that contains a sing