Yeah, I presume so. At least that's what Swiss Tables do. The paper is focused more on the asymptotics rather than the real-world hardware performance, so I can see why they chose not to handle such edge cases