Do rowversion/timestamp columns in SQL Server need a nonclustered index when using EF code-first?

Question

All of my tables have a rowversion column that EF uses for optimistic concurrency checking. Should I create a nonclustered index on this column for faster data retrieval? Each table also has a clustered primary key named Id. Whenever updating data, will EF/SQL try to locate a row based on Id first and then run a rowversion check?

Answer 1

None of your query plans would seek on that column. All writes are performed by filtering on the primary key columns which causes a seek on the index that provides the primary key. The rowversion index never helps.

To answer such questions empirically, compare the execution plans with and without the index in question.