I don’t think that would be the case! On the contrary, there’s a considerable amount of added performance by running data separately. There are even tutorials about it: